From 8aff4aa80533b3665f8abb1ac3dfb28782d7ee3b Mon Sep 17 00:00:00 2001
From: randef1ned <46381867+randef1ned@users.noreply.github.com>
Date: Sat, 4 May 2024 19:49:44 +0000
Subject: [PATCH] =?UTF-8?q?Deploying=20to=20gh-pages=20from=20@=20randef1n?=
 =?UTF-8?q?ed/labyrinth@f0138d5667ab3437c8e57c5ac0b61d13954f8191=20?=
 =?UTF-8?q?=F0=9F=9A=80?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 404.html                                      |   4 +-
 CODE_OF_CONDUCT.html                          |   4 +-
 LICENSE.html                                  |   4 +-
 articles/index.html                           |   4 +-
 articles/labyrinth.html                       | 316 +++++++++++++++++-
 .../figure-html/unnamed-chunk-5-1.png         | Bin 0 -> 169809 bytes
 articles/preface.html                         |   4 +-
 articles/training.html                        | 153 ++++++---
 authors.html                                  |  16 +-
 index.html                                    |  16 +-
 news/index.html                               |   8 +-
 pkgdown.yml                                   |   2 +-
 reference/activation_rate.html                |   4 +-
 reference/alias2SymbolUsingNCBI.html          |   4 +-
 reference/assert_dgCMatrix.html               |   4 +-
 reference/disease_ids.html                    |   4 +-
 reference/disease_impact_score.html           |   4 +-
 reference/drug_annot.html                     |   4 +-
 reference/gene_disease.html                   |   4 +-
 reference/get_neighbors.html                  |   4 +-
 reference/gradient.html                       |   4 +-
 reference/graph.html                          |   4 +-
 reference/index.html                          |   9 +-
 reference/is.dgCMatrix.html                   |   4 +-
 reference/labyrinth-package.html              |   4 +-
 reference/load_data.html                      |   4 +-
 reference/mesh_annot.html                     |   4 +-
 reference/mesh_hierarchy.html                 |   4 +-
 reference/mesh_synonyms.html                  |   4 +-
 reference/ncbi_info.html                      |   4 +-
 reference/ppi.html                            |   4 +-
 reference/predict_drug.html                   |   8 +-
 reference/random_walk.html                    |   4 +-
 reference/sigmoid.html                        |   4 +-
 reference/speech.html                         |   4 +-
 reference/spread_gram.html                    |   4 +-
 reference/spread_gram_1.html                  |   4 +-
 reference/transfer_activation.html            |   4 +-
 reference/update_gene_symbol.html             |   4 +-
 search.json                                   |   2 +-
 sitemap.xml                                   |   3 -
 41 files changed, 503 insertions(+), 150 deletions(-)
 create mode 100644 articles/labyrinth_files/figure-html/unnamed-chunk-5-1.png

diff --git a/404.html b/404.html
index 82459ab..d040d18 100644
--- a/404.html
+++ b/404.html
@@ -24,7 +24,7 @@
     
     <a class="navbar-brand me-2" href="https://labyrinth.yinchun.su/index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -80,7 +80,7 @@
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/CODE_OF_CONDUCT.html b/CODE_OF_CONDUCT.html
index 7561ec9..94b3c24 100644
--- a/CODE_OF_CONDUCT.html
+++ b/CODE_OF_CONDUCT.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -127,7 +127,7 @@ <h2 id="attribution">Attribution<a class="anchor" aria-label="anchor" href="#att
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/LICENSE.html b/LICENSE.html
index 82a796e..7d06fab 100644
--- a/LICENSE.html
+++ b/LICENSE.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -154,7 +154,7 @@ <h3 id="how-to-apply-these-terms-to-your-new-programs">How to Apply These Terms
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/articles/index.html b/articles/index.html
index 5662275..fdcedd6 100644
--- a/articles/index.html
+++ b/articles/index.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -68,7 +68,7 @@ <h3>All vignettes</h3>
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/articles/labyrinth.html b/articles/labyrinth.html
index 7aa8e5b..806a11a 100644
--- a/articles/labyrinth.html
+++ b/articles/labyrinth.html
@@ -26,7 +26,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -86,9 +86,317 @@ <h4 data-toc-skip class="author">Yinchun Su</h4>
     
     
 <div class="section level2">
-<h2 id="introduction">Introduction<a class="anchor" aria-label="anchor" href="#introduction"></a>
+<h2 id="setup">Setup<a class="anchor" aria-label="anchor" href="#setup"></a>
 </h2>
-<p>ab</p>
+<p>The <code>labyrinth</code> framework was initially developed and
+tested on Fedora 39 with R 4.3, this vignette aims to demonstrate its
+ease of use on the Windows Subsystem for Linux (WSL) environment.
+Specifically, we will be presenting the results using <a href="https://github.com/WhitewaterFoundry/Fedora-Remix-for-WSL" class="external-link">Fedora
+Remix for WSL</a>.</p>
+<p>To get started, you will need to prepare the following
+dependencies.</p>
+<ol style="list-style-type: decimal">
+<li><p><strong>Fedora</strong> or <strong>Red Hat Enterprise
+Linux</strong> with or without WSL.</p></li>
+<li><p><strong>R 4.3</strong>.</p></li>
+<li><p><strong>Required libraries</strong>.</p></li>
+</ol>
+<div class="sourceCode" id="cb1"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="co"># pROC ggthemes </span></span>
+<span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://tidyverse.tidyverse.org" class="external-link">tidyverse</a></span>, <span class="va">igraph</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## -- <span style="font-weight: bold;">Attaching core tidyverse packages</span> ------------------------ tidyverse 2.0.0 --</span></span>
+<span><span class="co">## <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">dplyr    </span> 1.1.4     <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">readr    </span> 2.1.5</span></span>
+<span><span class="co">## <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">forcats  </span> 1.0.0     <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">stringr  </span> 1.5.1</span></span>
+<span><span class="co">## <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">ggplot2  </span> 3.5.1     <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">tibble   </span> 3.2.1</span></span>
+<span><span class="co">## <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">lubridate</span> 1.9.3     <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">tidyr    </span> 1.3.1</span></span>
+<span><span class="co">## <span style="color: #00BB00;">v</span> <span style="color: #0000BB;">purrr    </span> 1.0.2     </span></span>
+<span><span class="co">## -- <span style="font-weight: bold;">Conflicts</span> ------------------------------------------ tidyverse_conflicts() --</span></span>
+<span><span class="co">## <span style="color: #BB0000;">x</span> <span style="color: #0000BB;">dplyr</span>::<span style="color: #00BB00;">filter()</span> masks <span style="color: #0000BB;">stats</span>::filter()</span></span>
+<span><span class="co">## <span style="color: #BB0000;">x</span> <span style="color: #0000BB;">dplyr</span>::<span style="color: #00BB00;">lag()</span>    masks <span style="color: #0000BB;">stats</span>::lag()</span></span>
+<span><span class="co">## <span style="color: #00BBBB;">i</span> Use the conflicted package (<span style="color: #0000BB; font-style: italic;">&lt;http://conflicted.r-lib.org/&gt;</span>) to force all conflicts to become errors</span></span></code></pre>
+<div class="sourceCode" id="cb3"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://patchwork.data-imaginist.com" class="external-link">patchwork</a></span><span class="op">)</span></span>
+<span><span class="kw"><a href="https://rdrr.io/r/base/library.html" class="external-link">library</a></span><span class="op">(</span><span class="va"><a href="https://github.com/randef1ned/labyrinth" class="external-link">labyrinth</a></span><span class="op">)</span></span></code></pre></div>
+<p>With these installed and loaded, you will have a consistent
+environment for running the <code>labyrinth</code>.</p>
+</div>
+<div class="section level2">
+<h2 id="data-preparation">Data preparation<a class="anchor" aria-label="anchor" href="#data-preparation"></a>
+</h2>
+<div class="sourceCode" id="cb4"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">drug_disease_weight</span> <span class="op">&lt;-</span> <span class="fu"><a href="../reference/load_data.html">load_data</a></span><span class="op">(</span><span class="st">'drug_disease_weight'</span><span class="op">)</span></span></code></pre></div>
+<p>Load the pre-trained model.</p>
+<div class="sourceCode" id="cb5"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">model</span> <span class="op">&lt;-</span> <span class="fu"><a href="../reference/load_data.html">load_data</a></span><span class="op">(</span><span class="st">'model'</span><span class="op">)</span></span>
+<span><span class="fu"><a href="https://rdrr.io/r/base/dim.html" class="external-link">dim</a></span><span class="op">(</span><span class="va">model</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## [1] 7686 7686</span></span></code></pre>
+<p>The labyrinth model is trained by integrating knowledge from two
+major sources: text-based information from medical corpora and
+biological knowledge from functional interaction networks.</p>
+<p>For the text-based component, labyrinth: 1. Extracts drug information
+(nomenclature, targets, and indications) from databases like DrugBank,
+CTD, and ChEMBL. 2. Obtains clinical trial data from the Cochrane
+Library. 3. Mines co-occurrence patterns in published literature from
+the Web of Science corpus. 4. Preprocesses the text data from over 10
+million publications, including stop word removal and term vectorization
+using Skip-gram models. 5. Quantifies structured drug-disease
+relationships based on clinical trial phases, citation analysis, and
+network proximity between gene sets.</p>
+<p>For the biological component, labyrinth evaluates the network
+proximity between drug target modules and disease gene modules within a
+functional interactome network.</p>
+<p>The text-based and biological knowledge matrices are then integrated
+through probabilistic computations, simulating the process of storing
+relevant knowledge in long-term memory for decision-making.</p>
+</div>
+<div class="section level2">
+<h2 id="reproducibility-example">Reproducibility example<a class="anchor" aria-label="anchor" href="#reproducibility-example"></a>
+</h2>
+<div class="section level3">
+<h3 id="labyrinth-learns-mechanisms-that-mediate-specific-drug-response">Labyrinth learns mechanisms that mediate specific drug response<a class="anchor" aria-label="anchor" href="#labyrinth-learns-mechanisms-that-mediate-specific-drug-response"></a>
+</h3>
+<p>To validate our approach, we initially evaluated the predictive
+accuracy across various diseases. This involved assessing the Spearman
+correlations between the priority scores assigned by labyrinth and the
+established weights in clinical trials, alongside proximity metrics for
+each drug-disease pair. Labyrinth exhibited moderate to high
+correlations, with coefficients of 0.60 for clinical trials and 0.80 for
+proximity, respectively.</p>
+<div class="sourceCode" id="cb7"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">roc0</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span><span class="va">drug_disease_weight</span>, weight <span class="op">=</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/if_else.html" class="external-link">if_else</a></span><span class="op">(</span><span class="va">weight</span> <span class="op">&gt;</span> <span class="fl">1</span>, <span class="fl">1</span>, <span class="fl">0</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu">pROC</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pROC/man/roc.html" class="external-link">roc</a></span><span class="op">(</span><span class="va">weight</span>, <span class="va">score</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## Setting levels: control = 0, case = 1</span></span></code></pre>
+<pre><code><span><span class="co">## Setting direction: controls &lt; cases</span></span></code></pre>
+<div class="sourceCode" id="cb10"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">roc1</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span><span class="va">drug_disease_weight</span>, weight <span class="op">=</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/if_else.html" class="external-link">if_else</a></span><span class="op">(</span><span class="va">weight</span> <span class="op">&gt;</span> <span class="fl">1.8</span>, <span class="fl">1</span>, <span class="fl">0</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu">pROC</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pROC/man/roc.html" class="external-link">roc</a></span><span class="op">(</span><span class="va">weight</span>, <span class="va">score</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## Setting levels: control = 0, case = 1</span></span>
+<span><span class="co">## Setting direction: controls &lt; cases</span></span></code></pre>
+<div class="sourceCode" id="cb12"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">roc2</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span><span class="va">drug_disease_weight</span>, weight <span class="op">=</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/if_else.html" class="external-link">if_else</a></span><span class="op">(</span><span class="va">weight</span> <span class="op">&gt;</span> <span class="fl">2</span>, <span class="fl">1</span>, <span class="fl">0</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu">pROC</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pROC/man/roc.html" class="external-link">roc</a></span><span class="op">(</span><span class="va">weight</span>, <span class="va">score</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## Setting levels: control = 0, case = 1</span></span>
+<span><span class="co">## Setting direction: controls &lt; cases</span></span></code></pre>
+<div class="sourceCode" id="cb14"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">roc3</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span><span class="va">drug_disease_weight</span>, weight <span class="op">=</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/if_else.html" class="external-link">if_else</a></span><span class="op">(</span><span class="va">weight</span> <span class="op">&gt;</span> <span class="fl">3</span>, <span class="fl">1</span>, <span class="fl">0</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu">pROC</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pROC/man/roc.html" class="external-link">roc</a></span><span class="op">(</span><span class="va">weight</span>, <span class="va">score</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## Setting levels: control = 0, case = 1</span></span>
+<span><span class="co">## Setting direction: controls &lt; cases</span></span></code></pre>
+<div class="sourceCode" id="cb16"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">roc4</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span><span class="va">drug_disease_weight</span>, weight <span class="op">=</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/if_else.html" class="external-link">if_else</a></span><span class="op">(</span><span class="va">weight</span> <span class="op">&gt;</span> <span class="fl">4</span>, <span class="fl">1</span>, <span class="fl">0</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu">pROC</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pROC/man/roc.html" class="external-link">roc</a></span><span class="op">(</span><span class="va">weight</span>, <span class="va">score</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## Setting levels: control = 0, case = 1</span></span>
+<span><span class="co">## Setting direction: controls &lt; cases</span></span></code></pre>
+<div class="sourceCode" id="cb18"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">p1</span> <span class="op">&lt;-</span> <span class="fu">pROC</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pROC/man/ggroc.html" class="external-link">ggroc</a></span><span class="op">(</span><span class="fu"><a href="https://rdrr.io/r/base/list.html" class="external-link">list</a></span><span class="op">(</span></span>
+<span>  <span class="va">roc0</span>, <span class="va">roc1</span>, <span class="va">roc2</span>, <span class="va">roc3</span>, <span class="va">roc4</span></span>
+<span><span class="op">)</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/scale_manual.html" class="external-link">scale_color_manual</a></span><span class="op">(</span><span class="st">'Drug trials'</span>, labels <span class="op">=</span> <span class="fu"><a href="https://rdrr.io/r/base/c.html" class="external-link">c</a></span><span class="op">(</span></span>
+<span>    <span class="fu"><a href="https://rdrr.io/r/base/paste.html" class="external-link">paste</a></span><span class="op">(</span><span class="st">'Stage 0:'</span>, <span class="fu"><a href="https://rdrr.io/r/base/Round.html" class="external-link">round</a></span><span class="op">(</span><span class="va">roc0</span><span class="op">$</span><span class="va">auc</span>, <span class="fl">3</span><span class="op">)</span><span class="op">)</span>,</span>
+<span>    <span class="fu"><a href="https://rdrr.io/r/base/paste.html" class="external-link">paste</a></span><span class="op">(</span><span class="st">'Stage 1:'</span>, <span class="fu"><a href="https://rdrr.io/r/base/Round.html" class="external-link">round</a></span><span class="op">(</span><span class="va">roc1</span><span class="op">$</span><span class="va">auc</span>, <span class="fl">3</span><span class="op">)</span><span class="op">)</span>,</span>
+<span>    <span class="fu"><a href="https://rdrr.io/r/base/paste.html" class="external-link">paste</a></span><span class="op">(</span><span class="st">'Stage 2:'</span>, <span class="fu"><a href="https://rdrr.io/r/base/Round.html" class="external-link">round</a></span><span class="op">(</span><span class="va">roc2</span><span class="op">$</span><span class="va">auc</span>, <span class="fl">3</span><span class="op">)</span><span class="op">)</span>,</span>
+<span>    <span class="fu"><a href="https://rdrr.io/r/base/paste.html" class="external-link">paste</a></span><span class="op">(</span><span class="st">'Stage 3:'</span>, <span class="fu"><a href="https://rdrr.io/r/base/Round.html" class="external-link">round</a></span><span class="op">(</span><span class="va">roc3</span><span class="op">$</span><span class="va">auc</span>, <span class="fl">3</span><span class="op">)</span><span class="op">)</span>,</span>
+<span>    <span class="fu"><a href="https://rdrr.io/r/base/paste.html" class="external-link">paste</a></span><span class="op">(</span><span class="st">'Stage 4:'</span>, <span class="fu"><a href="https://rdrr.io/r/base/Round.html" class="external-link">round</a></span><span class="op">(</span><span class="va">roc4</span><span class="op">$</span><span class="va">auc</span>, <span class="fl">3</span><span class="op">)</span><span class="op">)</span></span>
+<span>  <span class="op">)</span>, values <span class="op">=</span> <span class="fu"><a href="https://rdrr.io/r/base/c.html" class="external-link">c</a></span><span class="op">(</span><span class="st">'#cf4e9c'</span>, <span class="st">'#835921'</span>, <span class="st">'#8d59a3'</span>, <span class="st">'#368db9'</span>, <span class="st">'#302c2d'</span><span class="op">)</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu">ggthemes</span><span class="fu">::</span><span class="fu"><a href="http://jrnold.github.io/ggthemes/reference/theme_clean.html" class="external-link">theme_clean</a></span><span class="op">(</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/theme.html" class="external-link">theme</a></span><span class="op">(</span>legend.position <span class="op">=</span> <span class="st">'left'</span><span class="op">)</span></span>
+<span><span class="fu"><a href="https://rdrr.io/r/base/rm.html" class="external-link">rm</a></span><span class="op">(</span>list <span class="op">=</span> <span class="fu"><a href="https://rdrr.io/r/base/c.html" class="external-link">c</a></span><span class="op">(</span><span class="st">'roc0'</span>, <span class="st">'roc1'</span>, <span class="st">'roc2'</span>, <span class="st">'roc3'</span>, <span class="st">'roc4'</span><span class="op">)</span><span class="op">)</span></span></code></pre></div>
+<p>Subsequently, we extended our analysis to encompass all human
+diseases, aiming to assess the predictive performance across five
+clinical trial phases, including pre-clinical, phases 1 to 3, and
+approved treatments. We evaluated the prediction performance by Receiver
+Operating Characteristic Area Under the Curve (ROC-AUC). As illustrated
+in Figure 3A, the ROC-AUC values for all stages surpassed 0.90,
+indicating a predictive success rate of over 90% in distinguishing
+between drugs classified for clinical trial or non-clinical trial.</p>
+<div class="sourceCode" id="cb19"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="co"># We perform ROC analysis in every single disease</span></span>
+<span><span class="va">disease_pred</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://rdrr.io/r/base/split.html" class="external-link">split</a></span><span class="op">(</span><span class="fu"><a href="https://dplyr.tidyverse.org/reference/select.html" class="external-link">select</a></span><span class="op">(</span><span class="va">drug_disease_weight</span>, <span class="va">weight</span>, <span class="va">score</span><span class="op">)</span>, <span class="va">drug_disease_weight</span><span class="op">$</span><span class="va">mesh_id</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://purrr.tidyverse.org/reference/map.html" class="external-link">map</a></span><span class="op">(</span><span class="op">~</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/arrange.html" class="external-link">arrange</a></span><span class="op">(</span><span class="va">.x</span>, <span class="fu"><a href="https://dplyr.tidyverse.org/reference/desc.html" class="external-link">desc</a></span><span class="op">(</span><span class="va">weight</span><span class="op">)</span><span class="op">)</span><span class="op">)</span></span>
+<span><span class="va">disease_pred</span> <span class="op">&lt;-</span> <span class="fu">pbapply</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pbapply/man/pbapply.html" class="external-link">pblapply</a></span><span class="op">(</span><span class="fu"><a href="https://rdrr.io/r/base/names.html" class="external-link">names</a></span><span class="op">(</span><span class="va">disease_pred</span><span class="op">)</span>, <span class="kw">function</span><span class="op">(</span><span class="va">n</span><span class="op">)</span> <span class="op">{</span></span>
+<span>  <span class="va">x</span> <span class="op">&lt;-</span> <span class="va">disease_pred</span><span class="op">[[</span><span class="va">n</span><span class="op">]</span><span class="op">]</span></span>
+<span>  <span class="kw">if</span> <span class="op">(</span><span class="fu"><a href="https://dplyr.tidyverse.org/reference/nth.html" class="external-link">nth</a></span><span class="op">(</span><span class="va">x</span><span class="op">$</span><span class="va">weight</span>, <span class="fl">2</span><span class="op">)</span> <span class="op">&gt;</span> <span class="fl">1.5</span><span class="op">)</span> <span class="op">{</span></span>
+<span>    <span class="va">ci</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://rdrr.io/r/base/message.html" class="external-link">suppressMessages</a></span><span class="op">(</span><span class="op">{</span></span>
+<span>      <span class="va">x</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>        <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span>weight <span class="op">=</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/if_else.html" class="external-link">if_else</a></span><span class="op">(</span><span class="va">weight</span> <span class="op">&gt;</span> <span class="fl">1.5</span>, <span class="fl">1</span>, <span class="fl">0</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>        <span class="fu">pROC</span><span class="fu">::</span><span class="fu"><a href="https://rdrr.io/pkg/pROC/man/roc.html" class="external-link">roc</a></span><span class="op">(</span><span class="va">weight</span>, <span class="va">score</span>, ci <span class="op">=</span> <span class="cn">TRUE</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>        <span class="op">{</span><span class="fu"><a href="https://rdrr.io/r/base/numeric.html" class="external-link">as.numeric</a></span><span class="op">(</span><span class="va">.</span><span class="op">$</span><span class="va">ci</span><span class="op">)</span><span class="op">}</span></span>
+<span>    <span class="op">}</span><span class="op">)</span></span>
+<span>    <span class="va">ret</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://rdrr.io/r/base/data.frame.html" class="external-link">data.frame</a></span><span class="op">(</span>mesh_id <span class="op">=</span> <span class="va">n</span>, ci_lower <span class="op">=</span> <span class="va">ci</span><span class="op">[</span><span class="fl">1</span><span class="op">]</span>, ci_upper <span class="op">=</span> <span class="va">ci</span><span class="op">[</span><span class="fl">3</span><span class="op">]</span>, auc <span class="op">=</span> <span class="va">ci</span><span class="op">[</span><span class="fl">2</span><span class="op">]</span><span class="op">)</span></span>
+<span>  <span class="op">}</span> <span class="kw">else</span> <span class="op">{</span></span>
+<span>    <span class="va">ret</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://rdrr.io/r/base/data.frame.html" class="external-link">data.frame</a></span><span class="op">(</span><span class="op">)</span></span>
+<span>  <span class="op">}</span></span>
+<span>  <span class="kw"><a href="https://rdrr.io/r/base/function.html" class="external-link">return</a></span><span class="op">(</span><span class="va">ret</span><span class="op">)</span></span>
+<span><span class="op">}</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/bind_rows.html" class="external-link">bind_rows</a></span><span class="op">(</span><span class="op">)</span></span>
+<span></span>
+<span><span class="co"># Next, we divide the ROC and average them based on MeSH structures </span></span>
+<span><span class="va">mesh_ids</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://rdrr.io/r/base/unique.html" class="external-link">unique</a></span><span class="op">(</span><span class="va">drug_disease_weight</span><span class="op">$</span><span class="va">mesh_id</span><span class="op">)</span></span>
+<span></span>
+<span><span class="fu"><a href="https://rdrr.io/r/utils/data.html" class="external-link">data</a></span><span class="op">(</span><span class="st">"mesh_annot"</span><span class="op">)</span></span>
+<span><span class="va">disease_cluster</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/distinct.html" class="external-link">distinct</a></span><span class="op">(</span><span class="va">mesh_annot</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span>parent_group <span class="op">=</span> <span class="fu"><a href="https://stringr.tidyverse.org/reference/case.html" class="external-link">str_to_sentence</a></span><span class="op">(</span><span class="va">group_name</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://dplyr.tidyverse.org/reference/select.html" class="external-link">select</a></span><span class="op">(</span><span class="op">!</span><span class="va">group_name</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://dplyr.tidyverse.org/reference/arrange.html" class="external-link">arrange</a></span><span class="op">(</span><span class="va">group_id</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate-joins.html" class="external-link">left_join</a></span><span class="op">(</span><span class="va">disease_pred</span>, by <span class="op">=</span> <span class="st">'mesh_id'</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://tidyr.tidyverse.org/reference/drop_na.html" class="external-link">drop_na</a></span><span class="op">(</span><span class="op">)</span></span>
+<span></span>
+<span><span class="va">disease_color</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/group_by.html" class="external-link">group_by</a></span><span class="op">(</span><span class="va">disease_cluster</span>, <span class="va">group_id</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://dplyr.tidyverse.org/reference/summarise.html" class="external-link">summarize</a></span><span class="op">(</span>color <span class="op">=</span> <span class="fu"><a href="https://rdrr.io/r/stats/median.html" class="external-link">median</a></span><span class="op">(</span><span class="va">auc</span><span class="op">)</span> <span class="op">&gt;</span> <span class="fl">0.911</span><span class="op">)</span></span>
+<span><span class="va">p2</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate.html" class="external-link">mutate</a></span><span class="op">(</span><span class="va">disease_cluster</span>, parent_group <span class="op">=</span> <span class="fu"><a href="https://forcats.tidyverse.org/reference/fct_reorder.html" class="external-link">fct_reorder</a></span><span class="op">(</span><span class="va">parent_group</span>, <span class="op">-</span><span class="va">group_id</span><span class="op">)</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span>  <span class="fu"><a href="https://dplyr.tidyverse.org/reference/mutate-joins.html" class="external-link">left_join</a></span><span class="op">(</span><span class="va">disease_color</span>, by <span class="op">=</span> <span class="st">'group_id'</span><span class="op">)</span> <span class="op"><a href="https://magrittr.tidyverse.org/reference/pipe.html" class="external-link">%&gt;%</a></span></span>
+<span><span class="fu"><a href="https://ggplot2.tidyverse.org/reference/ggplot.html" class="external-link">ggplot</a></span><span class="op">(</span><span class="fu"><a href="https://ggplot2.tidyverse.org/reference/aes.html" class="external-link">aes</a></span><span class="op">(</span>x <span class="op">=</span> <span class="va">parent_group</span>, y <span class="op">=</span> <span class="va">auc</span>, fill <span class="op">=</span> <span class="va">color</span><span class="op">)</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/geom_boxplot.html" class="external-link">geom_boxplot</a></span><span class="op">(</span>alpha <span class="op">=</span> <span class="fl">0.4</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/geom_abline.html" class="external-link">geom_hline</a></span><span class="op">(</span><span class="fu"><a href="https://ggplot2.tidyverse.org/reference/aes.html" class="external-link">aes</a></span><span class="op">(</span>yintercept <span class="op">=</span> <span class="fl">0.911</span><span class="op">)</span>, linetype <span class="op">=</span> <span class="st">'dashed'</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/labs.html" class="external-link">labs</a></span><span class="op">(</span>x <span class="op">=</span> <span class="st">'Disease categories'</span>,</span>
+<span>      y <span class="op">=</span> <span class="st">'Prediction AUCs'</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/coord_flip.html" class="external-link">coord_flip</a></span><span class="op">(</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/scale_manual.html" class="external-link">scale_fill_manual</a></span><span class="op">(</span>values <span class="op">=</span> <span class="fu"><a href="https://rdrr.io/r/base/c.html" class="external-link">c</a></span><span class="op">(</span><span class="st">'#9fcdc9'</span>, <span class="st">'#d4bfe0'</span><span class="op">)</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu">ggthemes</span><span class="fu">::</span><span class="fu"><a href="http://jrnold.github.io/ggthemes/reference/theme_clean.html" class="external-link">theme_clean</a></span><span class="op">(</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/theme.html" class="external-link">theme</a></span><span class="op">(</span>legend.position <span class="op">=</span> <span class="st">'none'</span>,</span>
+<span>        axis.text.y <span class="op">=</span> <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/element.html" class="external-link">element_text</a></span><span class="op">(</span><span class="st">'Arial Narrow'</span><span class="op">)</span><span class="op">)</span></span></code></pre></div>
+<p>Notably, labyrinth exhibited high predictive accuracy in determining
+drug usability for Stage 3 across all disease categories except for
+occupational and stomatognathic diseases (Figure 3B). Also,
+cardiovascular, endocrine system diseases, and neoplasms garnered the
+most significant benefits from labyrinth. Detailed ROC-AUC predictions
+for all diseases are provided in Additional File 1.</p>
+<div class="sourceCode" id="cb20"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="va">design</span> <span class="op">&lt;-</span> <span class="st">"</span></span>
+<span><span class="st">  11#222</span></span>
+<span><span class="st">  222222</span></span>
+<span><span class="st">  222222</span></span>
+<span><span class="st">"</span></span>
+<span><span class="va">p1</span> <span class="op">+</span> <span class="va">p2</span> <span class="op">+</span> <span class="fu"><a href="https://patchwork.data-imaginist.com/reference/plot_layout.html" class="external-link">plot_layout</a></span><span class="op">(</span>guides <span class="op">=</span> <span class="st">'collect'</span><span class="op">)</span> <span class="op">+</span></span>
+<span>  <span class="fu"><a href="https://patchwork.data-imaginist.com/reference/plot_annotation.html" class="external-link">plot_annotation</a></span><span class="op">(</span>tag_levels <span class="op">=</span> <span class="st">'A'</span><span class="op">)</span> <span class="op">&amp;</span></span>
+<span>  <span class="fu"><a href="https://ggplot2.tidyverse.org/reference/theme.html" class="external-link">theme</a></span><span class="op">(</span>legend.position <span class="op">=</span> <span class="st">'bottom'</span><span class="op">)</span></span></code></pre></div>
+<p><img src="labyrinth_files/figure-html/unnamed-chunk-5-1.png" width="700">
+### Reproducibility statement</p>
+<p>All other results in the article can be reproduced using standard R
+code.</p>
+</div>
+</div>
+<div class="section level2">
+<h2 id="session-info">Session info<a class="anchor" aria-label="anchor" href="#session-info"></a>
+</h2>
+<div class="sourceCode" id="cb21"><pre class="downlit sourceCode r">
+<code class="sourceCode R"><span><span class="fu">devtools</span><span class="fu">::</span><span class="fu"><a href="https://r-lib.github.io/sessioninfo/reference/session_info.html" class="external-link">session_info</a></span><span class="op">(</span><span class="op">)</span></span></code></pre></div>
+<pre><code><span><span class="co">## <span style="color: #00BBBB; font-weight: bold;">- Session info ---------------------------------------------------------------</span></span></span>
+<span><span class="co">##  <span style="color: #555555; font-style: italic;">setting </span> <span style="color: #555555; font-style: italic;">value</span></span></span>
+<span><span class="co">##  version  R version 4.4.0 (2024-04-24)</span></span>
+<span><span class="co">##  os       Fedora Linux 40 (Container Image)</span></span>
+<span><span class="co">##  system   x86_64, linux-gnu</span></span>
+<span><span class="co">##  ui       X11</span></span>
+<span><span class="co">##  language en</span></span>
+<span><span class="co">##  collate  C</span></span>
+<span><span class="co">##  ctype    C</span></span>
+<span><span class="co">##  tz       Etc/UTC</span></span>
+<span><span class="co">##  date     2024-05-04</span></span>
+<span><span class="co">##  pandoc   3.1.3 @ /usr/bin/ (via rmarkdown)</span></span>
+<span><span class="co">## </span></span>
+<span><span class="co">## <span style="color: #00BBBB; font-weight: bold;">- Packages -------------------------------------------------------------------</span></span></span>
+<span><span class="co">##  <span style="color: #555555; font-style: italic;">package          </span> <span style="color: #555555; font-style: italic;">*</span> <span style="color: #555555; font-style: italic;">version  </span> <span style="color: #555555; font-style: italic;">date (UTC)</span> <span style="color: #555555; font-style: italic;">lib</span> <span style="color: #555555; font-style: italic;">source</span></span></span>
+<span><span class="co">##  backports           1.4.1     <span style="color: #555555;">2021-12-13</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  bslib               0.7.0     <span style="color: #555555;">2024-03-29</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  cachem              1.0.8     <span style="color: #555555;">2023-05-01</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  checkmate           2.3.1     <span style="color: #555555;">2023-12-04</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  cli                 3.6.2     <span style="color: #555555;">2023-12-11</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  codetools           0.2-20    <span style="color: #555555;">2024-03-31</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  colorspace          2.1-0     <span style="color: #555555;">2023-01-23</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  desc                1.4.3     <span style="color: #555555;">2023-12-10</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  devtools            2.4.5     <span style="color: #555555;">2022-10-11</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  diffusr             0.2.3     <span style="color: #555555;">2024-05-04</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">Bioconductor</span></span></span>
+<span><span class="co">##  digest              0.6.35    <span style="color: #555555;">2024-03-11</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  dplyr             * 1.1.4     <span style="color: #555555;">2023-11-17</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  ellipsis            0.3.2     <span style="color: #555555;">2021-04-29</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  evaluate            0.23      <span style="color: #555555;">2023-11-01</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  fansi               1.0.6     <span style="color: #555555;">2023-12-08</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  farver              2.1.1     <span style="color: #555555;">2022-07-06</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  fastmap             1.1.1     <span style="color: #555555;">2023-02-24</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  fastmatch           1.1-4     <span style="color: #555555;">2023-08-18</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  forcats           * 1.0.0     <span style="color: #555555;">2023-01-29</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  fs                  1.6.4     <span style="color: #555555;">2024-04-25</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  generics            0.1.3     <span style="color: #555555;">2022-07-05</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  ggplot2           * 3.5.1     <span style="color: #555555;">2024-04-23</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  ggthemes            5.1.0     <span style="color: #555555;">2024-02-10</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  glue                1.7.0     <span style="color: #555555;">2024-01-09</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  gtable              0.3.5     <span style="color: #555555;">2024-04-22</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  highr               0.10      <span style="color: #555555;">2022-12-22</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  hms                 1.1.3     <span style="color: #555555;">2023-03-21</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  htmltools           0.5.8.1   <span style="color: #555555;">2024-04-04</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  htmlwidgets         1.6.4     <span style="color: #555555;">2023-12-06</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  httpuv              1.6.15    <span style="color: #555555;">2024-03-26</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  igraph              2.0.3     <span style="color: #555555;">2024-03-13</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  jquerylib           0.1.4     <span style="color: #555555;">2021-04-26</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  jsonlite            1.8.8     <span style="color: #555555;">2023-12-04</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  knitr               1.46      <span style="color: #555555;">2024-04-06</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  labeling            0.4.3     <span style="color: #555555;">2023-08-29</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  labyrinth         * 0.3.0     <span style="color: #555555;">2024-05-04</span> <span style="color: #555555;">[1]</span> <span style="color: #BB00BB; font-weight: bold;">local</span></span></span>
+<span><span class="co">##  later               1.3.2     <span style="color: #555555;">2023-12-06</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  lattice             0.22-6    <span style="color: #555555;">2024-03-20</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  lifecycle           1.0.4     <span style="color: #555555;">2023-11-07</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  lubridate         * 1.9.3     <span style="color: #555555;">2023-09-27</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  magrittr            2.0.3     <span style="color: #555555;">2022-03-30</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  Matrix              1.7-0     <span style="color: #555555;">2024-04-26</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  MatrixGenerics      1.16.0    <span style="color: #555555;">2024-04-30</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">Bioconductor 3.19 (R 4.4.0)</span></span></span>
+<span><span class="co">##  matrixStats         1.3.0     <span style="color: #555555;">2024-04-11</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  memoise             2.0.1     <span style="color: #555555;">2021-11-26</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  memuse              4.2-3     <span style="color: #555555;">2023-01-24</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  mime                0.12      <span style="color: #555555;">2021-09-28</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  miniUI              0.1.1.1   <span style="color: #555555;">2018-05-18</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  munsell             0.5.1     <span style="color: #555555;">2024-04-01</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  patchwork         * 1.2.0     <span style="color: #555555;">2024-01-08</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pbapply             1.7-2     <span style="color: #555555;">2023-06-27</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pillar              1.9.0     <span style="color: #555555;">2023-03-22</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pkgbuild            1.4.4     <span style="color: #555555;">2024-03-17</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pkgconfig           2.0.3     <span style="color: #555555;">2019-09-22</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pkgdown             2.0.9     <span style="color: #555555;">2024-04-18</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pkgload             1.3.4     <span style="color: #555555;">2024-01-16</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  plyr                1.8.9     <span style="color: #555555;">2023-10-02</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pROC                1.18.5    <span style="color: #555555;">2023-11-01</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  profvis             0.3.8     <span style="color: #555555;">2023-05-02</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  promises            1.3.0     <span style="color: #555555;">2024-04-05</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  pryr                0.1.6     <span style="color: #555555;">2023-01-17</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  purrr             * 1.0.2     <span style="color: #555555;">2023-08-10</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  R6                  2.5.1     <span style="color: #555555;">2021-08-19</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  ragg                1.3.0     <span style="color: #555555;">2024-03-13</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  Rcpp                1.0.12    <span style="color: #555555;">2024-01-09</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  RcppEigen           0.3.4.0.0 <span style="color: #555555;">2024-02-28</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  RcppProgress        0.4.2     <span style="color: #555555;">2020-02-06</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  readr             * 2.1.5     <span style="color: #555555;">2024-01-10</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  remotes             2.5.0     <span style="color: #555555;">2024-03-17</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  rlang               1.1.3     <span style="color: #555555;">2024-01-10</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  rmarkdown           2.26      <span style="color: #555555;">2024-03-05</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  rpca                0.2.3     <span style="color: #555555;">2015-07-31</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  sass                0.4.9     <span style="color: #555555;">2024-03-15</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  scales              1.3.0     <span style="color: #555555;">2023-11-28</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  sessioninfo         1.2.2     <span style="color: #555555;">2021-12-06</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  shiny               1.8.1.1   <span style="color: #555555;">2024-04-02</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  sparseMatrixStats   1.16.0    <span style="color: #555555;">2024-04-30</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">Bioconductor 3.19 (R 4.4.0)</span></span></span>
+<span><span class="co">##  stringi             1.8.3     <span style="color: #555555;">2023-12-11</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  stringr           * 1.5.1     <span style="color: #555555;">2023-11-14</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  systemfonts         1.0.6     <span style="color: #555555;">2024-03-07</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  textshaping         0.3.7     <span style="color: #555555;">2023-10-09</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  tibble            * 3.2.1     <span style="color: #555555;">2023-03-20</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  tidyr             * 1.3.1     <span style="color: #555555;">2024-01-24</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  tidyselect          1.2.1     <span style="color: #555555;">2024-03-11</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  tidyverse         * 2.0.0     <span style="color: #555555;">2023-02-22</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  timechange          0.3.0     <span style="color: #555555;">2024-01-18</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  tzdb                0.4.0     <span style="color: #555555;">2023-05-12</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  urlchecker          1.0.1     <span style="color: #555555;">2021-11-30</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  usethis             2.2.3     <span style="color: #555555;">2024-02-19</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  utf8                1.2.4     <span style="color: #555555;">2023-10-22</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  vctrs               0.6.5     <span style="color: #555555;">2023-12-01</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  withr               3.0.0     <span style="color: #555555;">2024-01-16</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  xfun                0.43      <span style="color: #555555;">2024-03-25</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  xtable              1.8-4     <span style="color: #555555;">2019-04-21</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">##  yaml                2.3.8     <span style="color: #555555;">2023-12-11</span> <span style="color: #555555;">[2]</span> <span style="color: #555555;">CRAN (R 4.4.0)</span></span></span>
+<span><span class="co">## </span></span>
+<span><span class="co">## <span style="color: #555555;"> [1] /tmp/Rtmp8Pmrh8/temp_libpathd601a4ba9c</span></span></span>
+<span><span class="co">## <span style="color: #555555;"> [2] /usr/local/lib/R/library</span></span></span>
+<span><span class="co">## <span style="color: #555555;"> [3] /usr/lib64/R/library</span></span></span>
+<span><span class="co">## <span style="color: #555555;"> [4] /usr/share/R/library</span></span></span>
+<span><span class="co">## </span></span>
+<span><span class="co">## <span style="color: #00BBBB; font-weight: bold;">------------------------------------------------------------------------------</span></span></span></code></pre>
 </div>
   </main><aside class="col-md-3"><nav id="toc"><h2>On this page</h2>
     </nav></aside>
@@ -97,7 +405,7 @@ <h2 id="introduction">Introduction<a class="anchor" aria-label="anchor" href="#i
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/articles/labyrinth_files/figure-html/unnamed-chunk-5-1.png b/articles/labyrinth_files/figure-html/unnamed-chunk-5-1.png
new file mode 100644
index 0000000000000000000000000000000000000000..cb5ec330cf1d55ce0dcd205db0e7c99b69070333
GIT binary patch
literal 169809
zcmd@6cRZK>`#z3eRw$e7*+BLt*@R@2kiGXvWbb54$j&NKviB%6Bq^ehy+SsXy?*Ca
zy<YG4=lA*j{{P-?-^Z=zv#007<8fX0b)Lt09LIS@-cVN}JWYEVK@h@g$_knYg5!lC
zSY~*q;Qtt5eldo>PMKd<Qh>kVwI}~35dOlytE_hqK^Vf(e=uU_HmwkZ8M&q)tL2rk
zGVbX`F|d!jv7YSs;RTf@g2f^eM-<GAf9euPRM2hGI@g;#YNR(PNN{AY+9Mb^H(4;r
zqE2J}%r(r(QQisu8i!5BeBt>{g?m6?YG%4s0{0^RWbdyKn-7=oO{8Ux$9eW<oF~PW
zMM(eq5*F)!O%xRI-xo(XzVW&LymC%O{Xh06731Nkp;D91&DqXEQQC9oa%fF$Y;a=~
zZw;A|{2g5qr;reZ(`Zex*0>6rS1gnCBlz9hHR7bCH;XrSc9c|9RE)x975==mN=loL
zUvHB)obugSyL$KTU2S4{1qH3ubLY>)mti1Y`-?9-tp3hR&OWo06r+4NQE@+iLB<pz
z5z)IgVrj#HHn^_#6=`W{m*G#JVwy|;&S`$%^aY2%7eW6`NI~B4|DvA~8%nUs!c_9)
zbaW_@maZ=RJ9qB5Gmw#zN<PJw4bm?bF;L`4lH*B{KNtU1-|+9m{5G{Gk@Q;llpY=w
zgkWQ1GheuX9nAdIU0FflchKvABXN3q`j5x#6sMK*G~w;3y@l+~C&CUSOiU(2&&dC*
z{mt`Q>*NizF{cV|mr@R`kG~23wa`0crwSJsRK9ktHA~W$&wt;o>drVV3k%Eca%uCD
zIM?6RwO3d=A|X_YWfb<N@Yq@!_}1WkFS7$aX!*I;*+fiSJZd?Z`DtAp^|fo)l$DkF
zIDC)x-Q5=Zahgw#ugJ;EFEi`^`y$6B|6tGQx1k6UDtA9Fj>NFRi`1aTvGuk4@>4C2
zWpRmf=X_ciF1uoHZ*R}7eOYvih>cZJP{3|F-gj_05bTwB@cXwLJ4KZK!^FQ^i8j|q
zQbFF;RX{yM47TL)-U7e-l9tP4tn{&NcsvYTk2Rs|bB&XM-`2ow_y|eu!ra{Sw{i=2
zTDm{eLMlwOCEj}QNpv!zo3~*DGs|b`lFQL-taM3kKK#C;D9p*X;o->TFkZgMzt=3t
znPByD7iUQ3G2$rMSso((^h8+x#to7i;*ZP>4GkF-VO1<8>kS`GH2Ov_cXf5S(-Qyp
zS|Y8!Gwt}_=jJdFM*rV8y?lHy5$lhsEXjK_(Y#AsQgCVuHnz7*mP!6zET1&L7Jp){
zJ0-gN-uy?>6s3pMEIg$~Zz%1nUk_TGKX<Mo$+)z${@z0SmX!bgS38Nn8|JIn*KTZQ
z*kRl2lf+1zN}kuxF18MTDJdzuGiR_6TOnF{NuR2onG5g*gMr7tm?+Mk#Xw*eeJ(X=
zeogdmq_p>a(x%4CEyC+0?u?1Si+VL%*de*@u!}AGtGy#!s)hs`qpz=zTEsE6ZuRps
z`MYjzMPKUv-MUNMkUO$Nj*W4KiicvzB=E>RK>!PvC`5}R6z(~aoGmUR<6%%xkp3fB
zO+4L7Yr?Uyv4=Hw)%up!*5^b-D5Z~nle@aQB8b`(q1LtWH<u-_{!XNJx=+ef<K1aY
zO-;?9etg^4Zj0uc?<OV~PmXt*HwSK;U?3vS6OKaayb#%(E*)rHyDIL!6w)hwLLV!+
zP5r#zXK#bx-<sUmS}Xl1>@XP7z20=dbAjwy{hYDys<4aqaD~<9MYn1)CQ{P)Vjkm$
zh?iMDr`H>P5s{ITFN@Rv_iShf*Orx$O#1zbsgv^g{o{1E2W<U`i3ubqBZEFe!n;Dp
z9HyCLd940koy%-kpN5~G1dq|{knDhce5Qb1mczrt8@jp?+J!pLth$qRD=e=Wg{%Et
z#u$Z0+MMZ5ra+1mBO@bHa`FeuL#4NMo35wui23|xN0`r@!+iGaSv46O+@owb5re6-
zv&nRmO_ukmdkZ}s{dw0nCw$j2kok{kSiZi#4O@c-)%4u|1}?J2(R#4i$rRn~?d|O~
z6UkOo^2T%B`1|+oE#u>jVL7n53g};HMP0r(&&<Y#+nvmfg?voq$C~Yo4Udk#ZWInD
z6MYftHqzv<h=>e2IvY8xjMZ0;%&@>m@7#K+F03zoZ(cRCL+<Zy#>h7=wjOj)nM8EQ
zO6{|U5i!_nH(GQ)n%<ahLxMujP?^uPN1j#xs0eu`1rrnV;UF(e`|;7iklugx^~OLn
zcBy{xJDY5O))MD$Wf+-PJZlQpU}C2h7H|<|6_vMn*Hb>fy7iJaIw>L|zhdnKvm!q~
zzsDTz30^txo*5JL-TnTa6wUw=)ZE;xhOcj8ViFP_9=&Wujz+buYdSi~eXpCF1ONAp
zD{mxc8m=DgOxAiwgu;|Wbes5ZosEi)E)+%&?a~IkMOW;(*q`^mF4aTc05MT9;p0PE
z;Sci`zadx<5fRPo!;lKO|9tH3LJvdkwLu99i6X{Ff6q-s{yF!osqreiggObY4O0ja
zBYp%|u3Uk+mp3%2yK7e60W+_q7tG8oEPOWGf14)w!Gn(J2*h;uT})h;r8_O`So9_A
z=Pus$MExbC?T|Vt2p=j;4sS#7b*pSiktx_4b>6>NgM)($E~>)-qTH4T2~UT<wM{oJ
zRl0MB=g#<R?K_z}lYvZa;l!nRcz;7eK7ZjC?Dh5LlShY#en^n!D+OI$-MrX=oU3Qh
zfKSC^h^TbV7&UmgRClPrO=dy1(Kj0G?mlBsd#B@zhO|*G|DSs!?R$``9&W4q+U-@H
z1l$^jbb+|%nxV@c{En~>MKUa%NSCF6m#?qI=clr%S3G$j!r%Y;Rg2a0-<{|W*bz@W
zz$MmM-aOun4U{1Xv6p0p<E{tk>(B_sjJ~!MGc_t;JN@UC)P@Zq;os{fR>%MGllDuC
z4b54%ttD-GNI}Twml{mJfAr;+$p1a+=g7{UJv-ga+l-9}IgL>8Htm`lg~J%_R5vuM
zulR0rP53NxBP}^sqf@6Y{@rXoc~76p9KI3e;(7?E(f+fT!oL6pckusx_5br;I{3Z*
z1>uMYMrJtKZy{bTi9h669Ij-`^4}hQO_XbF8l3WP%S8OZfT;A~<e>SaaMMFTQ1F@N
zD+Fn3YJ&YR;Nh63*@f+ViS>WJGJnL%%8G@Tml&B^T{SAkH{NIZU+*&B;MK?oiX&oh
zEFcd&5^vh`%k0{=G4aD_eEYV2YiH8$%!#BgC9=^Iyk|SJa63<T*1~+_)9Wp5aYc*v
z|4ivkhW8&ohBP-zqtn7=^twv3Q_O!~h*cx&A)V+Q=F699dU|>W7;Rv$1?<gH!2xU=
zA2%A4^4)eU(yu;!@#4iemCNB*ZpzOMRb2`msj7LUZYC>RKicT;_wdnNz8hHzll?Fm
z#;0^`7R{&mWo0T`besX0f58sv0ER5d(>~j$1NNNS&45h&_kXuSe3@EW(qG-A8LM-_
zf?cg-XqeIGQLJ+YAhN|O8EL%xe-`eXebD2CglAtC%dEO7FI>DhpuBE$&lnQe;m>Z~
zmB#A|TAlx0%A1mTUluts5|vAbu#hLB&c@w0kBT9f+YS_HSJQWodBYcbA3llsmlU&7
zh3v>)d#pyieJe*m&4;+w(*pq9-Dy7I?ol8{;0VjV@mPhL=@dbb+%uVgfTq6kM+{fO
z?qlJ~z0%47jJ>-(>db-{d@Vhg*>s^N6Lu(WqRPen--{0*GDoleUKu7wKELsF1~~Y`
zP5h}FfpMrD3+od?B1Vab&!2AqUONX>0A%eXDUAmqA=nQYC21KM88>!zIv>&qs`jKr
z{E+?cX1po+Y`{J~Jw5tmb^tAW#&)oX23xk}g<AUON*fLSw2^9u+rO6v&L9wS*$8L^
zuwC~~82e$cGG1O{j~FGR0I;^dduI;8D^<|u(Y164_Z{!!&%5mXv_g?;dn7p^k~Szq
z`rw<~)+dY{@4b&>iBrB{UbGY23dhQkfKh__rFy0uKrHnviAW5rQ;_y@G#Bs@7Z(@X
z@z=sN&XYt?W7t;fqZ^0eibsfkamnDahQe18mxxn;4{wgo1r`>Jw6wH3K+7Xlc36$Z
zvDXq<rvc}_V+uUHUslGankjx(Lvo7(-hGh7dHbA%gkcdA9bI&qppE&)RO`2D2P&9n
zAzH#0s%gQHWXunDH}QywP9cZCzi51W?am345(XcXV@+VWco7HgqaD%`348+}v!_r*
z>DJz10Dy4QZ0w~*_PYwJZa%-=y9!sYwzjrnT)leLYd+0}57MQ?&e%QMFE3)$V1!>@
zs6EKeW&*f?O-xLTv;d^!cl^e;GF*wv+xS~El>cadxs)}b4TEp#s8FXcJ|lxbR8*8b
zNu{zhM!#55QSlTa!;_K=Bem|vlF`?HA{L*Ol0vnzy80=Zmy1gQ4j<A2m(R+|q8wFL
zSEo$z_xI-}@<dl0HYgHQSy|a}ZNEyNyT895K{Q`c1Ox<7T6}o_o`ie@HVxoCA%iA3
zV8g@1D5%G9c6N?-8hhiZaVXHz*@^4n;bGmqu&@wa+L(|)J~lr7DS2qoO<7It?9kEC
zQ3tOW$K#}=(8rI-kg}yL%~x6wGO+CzvsBd8!vG>S`tERW@!|#RG9rdXM#YQ{Z{SK0
z7@!#CGf;o|C9*VBi=)~;=Y4Q6CItmW#Il{8ofynu=$tl3lJ3MtV$em7Bo%s}cTapd
zczHD(+Xe=z2Rr3N{~I^+vxXtfrV=En<n2y9Al}#-GMW9o_{Dst@?10(Z<yS(7D#}+
ztOF}6D`j`a#T`d$2pxt>I+g|sF%T39HZ(SNeCJU~(9QPyi4T#Di&e0A^Y~z0{@D%F
z(HbXn*G{PKqz^Z+8X6kPyf)1h6ctC@vVxhX-~^<KI%lPh;RHRC56^+kQRlY!co=&f
z_E4GovQDK7737-zt!3C?HK`8;6LAnFC8fg0E|HOu<^t6K?kE#nE?>SJ5fg)hK$0{_
zni6gg=?}@y!*;f}=UAI)JcHm75LBfa-g>2FPUl)bR)D<@P$@1U0Sockoet+__4@Ve
zSBD57?jn_qcb%Qljqd?sa<k3~2nfWbrs5%Zcz9(Ttp?j8`v(UG4=+KyUR+#6f(DBX
zM%^^v(t7S17cX3RnxBt{e9pU06}_Y7;Zd`DwX-Mlip$#A<!GlwHAg@#2ANZzdV7^*
zW$$-jm+BOr9fgdfbJ44^q9QIcfxCq1iU&dGBZf*29*1h%KFnJ$R54W6<sH^u0bamB
z0N5;ZwRW@6{!4R9X|fMa(G=5>*@tBXoF&D^xZ<(W0&}TGW)E=`Vq=Hrq(5#VD;ZQV
zk04<_xquA?(|vA(8fwA=(??Kr!+w4Dh{2)1<mF4YEWe-E;P|CI#Bf}cxJZUi&%jXb
zyK^U@+`BjPN;sqo{j*G@1B4p9yt|@=L4jWoWh+(rL5l6EVtJoNMvkP@r$ci-REENt
zq=G81thQFEP^XwHKwfLW=P%ySx(D^(!$HEvg%9Z$ii=>ECMG6M&2&VgYu|MYTkISJ
z3ZF*fmDXX1ehIE{qNoL}9}aBN`fzPS!83LQg{&A6lhnh828rCx=f#(M^(w8+jt}=X
zHa1!+d!?dNOaf?*4t8n_Y#bmLbgM-b4%kC!N5Q^2zA`swRUU3PT*1zu*h0Iuwr1|E
zc)bzhra0GwXE6t*8dnMCZqCy_$<^ZM5Lbtz-`#B~P~K3zcjor(+vP=hj4#+5v%*A*
zKB;+mc|{}~>yf7<BqXRg0mYE(;6O-4Mdk7h?|P%I%cP{$CwI9;N$ncXri)hVuap?Q
zv1MgtQ3~Oq58Gtv%2s*J4v&gSXT^F<1n*#qfjUQ0l#{NnAMEH{2U>dihbY$6Z4!_d
z`sP!b()hB-yV2n^d4v0fD_7{yfZUUD`9Xd@J7DMn&CbZ^=(c;klHsYT)VZt9O?#}#
zye7uewZ%!r|H_fA6tk6Hs5qzIhMqZ_4BLNZ_E$3Ft=h0@m(7Yc_Czb?ca&6A21%Be
z9M;C_H-7(a$7S*h@;}%%n{4t&g5u9BnL`q`CQo#3J`$M-IB-f>g|M=*wS}WXu9qq9
zUco_W@PRgw!Sfrfv9YlT+gYcWS9V`s*p5`a8kl3Cr+2*5#`ow|HN-lnCaTd|XXjhq
zP+x4i*M=GL)b2H4VPQFOtgoy%-lGE|uKHC&ZEdaC_t0m>8q!BqZ?Q`i&dF9<e{!^M
zgCcCaV44y_4}2qpQmF~48K0pNBYua$B9#t3BPbh-Xe*w-D0g;a*!|Tj-3J)F5=++L
zJeqsN-DFf%3En-Vc>o1BGY7|Mqc<KMkiw%#&ns#<NdC=<1(O7gd-FY+XJkn@4qe>b
zD13)$9#?9c{ZP~OFn>`&kYJ<zB0L?o2c2`n{f{42FR-$jw}lbf4p&_CSozl4Y!Vrf
zI#p~0;S6?S+xb`tOkQ5zhJ&>?+Ts%-_~cJ@3O6C3T0r4QE#bwd7(*Khamx(g{I^OQ
z;((*AA^XDrtXaBQ%|VpUeBO%}GxML)Wzz6-ON`!FYpy{s@j6(ojbYUn=;Y)73T54g
z@xa4f7ocSH1}aOrl2k(L>cpUwQ@%iked-iiKcbY;$?=iPR0{?mg7in%6uy4rYc+`0
zDbl?UpErl=*SXG}7WY`uU9_;Ww47?WkA-5N!qO)DDh-Y#^VN}RBxq;Cw^V=WZ=rDW
zh%HHl-}W=}%49Q>ef5ZKHsmH4?ztWkCK3`PPc!oUd*y_3e^JM8L6HCTCC@`%kbL|0
zEs7rVwK#J9>+9>q?oJ`bK8qLA-~Gf6YHwGlviowX`Di=6%R4MI6eBM-ft*zGJxu4-
z(#CDtWG=m6C~e*X;QF4D#ZAz5MfHi$t1*)}r;)1rwY8!s4ubFJ7=judUO8cAa7NfK
zU26T_mC!L!ES^aOC6%nE=3_C|VpzWh?_YO;k}EQ(RWAyJ|KKvHA%L$BgLHo4!NtQP
zGP(v;%MC595Wr#4s|8vd59u#!KJ3<kz!{>-+6HhVChRd}?GAu@eYMV$92)xS3SU3-
z!?Xe<vhK}t+FE%zZ`kN_7WvpiNzKSegaDW*as9W^?5X#+SEFwmd*j5%$KSnoPu|v+
zT|!c_+;LdoAvHfHKwXmV9$*&WbiGCQ8f<~tcqK+zWo0~MDw?+mQSsf!Y)R#};Inx5
z@boAlAz=qh-^Oa~<lFIa1|UH0y1REk9xb!&J123t(I%1|4XGQl=jK9o;M<N4oJOB5
zIS7x_3k%^8NFB%O%E^EY%y)?gf=cUhd3m|aqND#=2$dIah<j)*4;B|{l}bQT=B}l3
z9qyVvefo4<a`Gt+4UPA6^Uk#p$2aNpAbQo+)unQmep!r5OS@D;_s&m%jg3voPqnL~
z<7x)p({uLK`WbpUJSowF%})~2((p6YaT9x(9a6tPJqK`)ot+&~VWTe-VohSkQ`XR+
z8ft88<d&LGOi5wuc~Pjv@l1;Y(2}jt%mq9^Mj1|DGKz}0HVu3L;C&k%P3Jbz)C`8v
zJpC>0H25+3Nmdq50RIfc{LS?Jik(D=)g7JPA0}>*MPdW(XWgAkwv(umN5+<wU%HbR
z8%rGZD>}~Y{IxL;woRUr4z`G8+8D~<<MmWaR~k!!U+d0|%rXW(nT)3iTFzL6rE$j|
ztX*=Dl1CHqk3bQpkr2qlW_v&9#N3xC<mBY|90s`>pgK~VgMf8+{v%D1ZaF59Rh@7k
zP?(GU_$gf!SLI|-ANU7KF5MvD+)O*7&YmTTF7w&4N4XJ*+U;;e(aM5Lw+wl3im>2!
zy=p2S1|VdVwFo#^6@@uBh5BuGqm8()NUu_bDJ3Gme9EMo@GHQZpne7YtA76e7|71r
zc%f_K&xcIXj5IVf*4@wZ@}@r|aeDpkQ{rZ|(Y$rb<zU+hk%3Ul#p)$YOITu9kD16i
zTXbz>qRHPtb*kEFbkJOYM_Ss%u??_ofu*6NV^ON=>R@sDutQA0OT6<~9Y&(D@AXOh
z<t-n6Rs^b;%G4=9tR=fXr*LreR5u{YL_5Js0XAy{rse0aUl<4!1y!k2pSrq8+3rE=
zsp4q$1-jwKmJcFRYTU@x6ae`XcK>-Xv2Tuh2kV#QojvYBb_|7;M`V<hBRW!HFFH28
zEHP>32}sktwevOf`t7^?O0nMz83ddsT0R|_VCLbEEzKX_uqh;K5WO=-LOw9)7l!IT
z!a_hwvTAtTf$cwL{7NA0<8w#Y>YRQ8TIb?Tzvkbh7Pe395QEwPB@NRuqpcvQK+#5o
zElZN{hEu1gCF(4v>IWh^5rVd`w3Cx8ZYWTW_P6Al1CF?|&X5k@@ozs3MQ8ewt?(1q
z+VznJ9AD41f*0>qUkxNNtyS^Q!YU=c_#I<ef}1=2Nv|fw$!eMO=3G)@qUl7xw%GoW
z$1ZkFuwtQn$^80Bl?rpYc@aHE*vX8F={Po780}+$6CH)yns_X{3~|aBvd#mWC`0n2
zwH^Oc1a0=)LqDw{F4e1{q3kmm>eY|%Cr*6;@Xel3#vRKdYf!^xWwo5N(5J+7Hk_B8
zlbf3mibt8n*=Z=gbv`B9vpeY4C;%>*D?ZXd<08{u&HU&kz^~<m*ChoB0mrpF*nBcQ
z0bxR<nk8PDYkR688_t++Xa90k|1&ItIc`E+)2WpeuBM*B-&I_`PTX7UA`t7yeK!W%
zCAGIlMqD*$!Zg^|_2(!ko|&v0&-m;Y(<vU{AX}_?`m(OK*JcRuFspFV?%turOsleQ
z7JuTsr?1@uvBZyfSbS9AZ!XgQ2&~DHgp$9NQ`3xW3$`;B&<vbUfAjA|QL|p-yy7N{
z4ihpxhvxxoq|FbUrAXd`&Ghuo#YIqC<N#99NxEtmJUC)ON8K~2Us6&cQrq`T9yk5N
z!Tx?!o4Kw$N74xWEmKowOf0Nop7rnFE$`8VMN${d&CTIJeotQlqOaq{&zgx_=v-vi
z=9w{xD6Ew)dbz9L!F^>Mq-r1Tag6b;&X?^wQH^B5EhERqmw(~=dvSpGt%;Xz**9bY
zjGYEL9vcAwKAd>Myv*M^Dg%d(5{Y~KSko-6!s4z@H6!UTtKfO82m8b%zVvirI|rg8
z4iXr$n8k*{FD`Su7p5}h{W7RmgghN`V2Q7jWxE_YMCo{6=`M4NapgY*j#82>a3O+3
zA`@ld2e2y&0E8exKu(+OZcMWt;K;Us&UqTK1^kq>3>*-|hRv@*_vh9e23hH2t&3!P
ziFnV8u+kbadE@)PP%GY|{zuH^Z<+?z-d`Flrn3~FY1*AR!(~|a*>Wtt`FQ^fFzP?X
z--yDtDKL!vA)B0-7{a1K99CJFKD-%x>(*n=F}5wI$ZZw|{+$G2BuI9*Ows%=XLv3L
z_Y8T%Rt7Vk*RM~aE}Jt~9fvD${)EoX)0xtNQc}B9SX}y5cp%?VEd%mFLmaZvipxaJ
zP+GAS!FYk@vbUJqPYIrP?bFifgAt==e!AX_OL6#8{#<QTj#28=afhZ`((^%?s(+;D
z;As6lHk}efJx(X1<^Ur@Msc_NzV>x;R#lQ;H?)LRHwN!PG;Mn#9LSa`b#i!*ORqv<
zOcaP2SUR)TV4QEwfznV@MFUVR0$~cmEj7P6Hd<UISm@L+-n7UaTIP57WPRAM%E_Ks
z5~Oun+vo9I6j|inQ81!6^pbh`D{<!I=)N^`Mx=$mUG5)f$f12Pv>cH$3m$^MUI3uu
z?DMz*t2*u+Uv|j<31Ddfc=M+>*a2zw0sWNge>xNs(tO*;G{N$KBOezXIY`KGyV!y@
zJt5CC6tp1T2vbfn9nxH%|8!qno9oW4BnsUkHm4)J?~>sx`5<g{x`|6-BejwY8#CF9
zamS)xY?}+;yr#`z`v;jBRj6#lE^A9rJiE_~_g?wB;PBei6hb9hdwTie#OsSZzCOP4
z9P!M&Lp5v`T>`9L%b$S*;sf=p(hAhIxU8&44<8a><KU#N+y;sQ2%*6uy$2B8-_N;1
znWd}x8J#hC82RSJoAsHER(YMI6p=t`flQ(jMsEI=5^@T0B(Krs)1v>EI#v9pAjTaM
zvVDs@3-tbS_JN!*^4gnAfy^k_-q|V7A-qZg+7cdcM*tx4;qBrCv2MZtp;+hMsVT&a
zMg$(6!3vjfD9-?8g|nAXaCl;?FDq4iVf^%RjIRs!@$MEiV%i#occHy=_E%na0k!aW
zEF7hlfq}CSK&-pr?uG37+vhEAqj&E*v&*zQeXzcH&&@5=O5f9uG#2j@OY5XO-+N|3
zo*VOvXEDz1Y$wSOwP1+1%Z3&sh=~)QaRwdXj{1FVX&Q8<3q0M#8GvddBO@z=lDuK#
z?U}jcFSbIT?W?O9voGrxQ_{w~C~0Fpm&1XQIkXpLN$xIvzt$jS=Nf8tLB*4RI1M}R
zwVNulAj3D|h_F-5SfAyCqkKj1b|<Oj?&aRm5%57TW1o+5d1h^4vOEjfqY48+*CTs-
zd%!*dq6|$)xK&(*vT8t@#3*9HUV^gY8T?8xGdgmWyt!wD5%yStvm`fj$&a#gy=>kQ
zR9b!SRc~t(XJSwI+2Lfrg2OC>>HOm^zRz>&0h-NkHh8pOdgYeM-jCB4P)MU8U04<6
zrcnxVb+pz3Xi|5EM?fVH6;4wv8Axaa<^gc|IoDn1xy~3BeUc^L0ZHxn`1r@ySL6Vp
z9iEQ^d1lz;M+3yr`|fUeSy_yViVEPpr|0KS!Lsu|0`mCn=;#~88RiN~>6?-HeeXYY
z$qqfr<WODv!01is!TagK%mCiQ*K<KVd0pe@J}Qw)1!Xr<ell8lO41`2OYv}N2WMV!
zoFZ_ffuGt{ROyVg>KeEeNG0==xeZPQ1_r98i+p-vM*M%1)G^jsyvtLq5AX;HLvwS_
zgA;(%sE5xI@H!w33XpC;L+W~ODNqH(4-~cK_JjL5Ip<L1SFB$h4B8SvqB1DVT`g3U
zm0N)4o}Qk@fD57`B&?F66|35l5U3M#Kz>De5(vY9{mQLB(LNQLK7dr;Iyy=ROhiOf
zRDsr&uXPtLTwoR#r?u(Lx+*7!33x0%J3DrF*DC|W45+l~V53#n8r--M^y7yu=(c3H
zOAK3KQBc_x*aoz^fl46)_=gxeQ551TC@IZuuZ|8dDp~Y#3Q@=(Z%Tw6&U%Y+IJ{ig
zdJrM$6`4P1X>Dn`dry9PZTjYF=b&9RltnzA*$q3=%!G@Cp*k((LzUAyhs|ly(Hj_;
z!bbe_Z<q=sS<$`eQ&v}BZwB}Z#C2*g-SE52C`7x%8z@JN$N&XuZehW1{gE17{p-nG
z{2VA*j+S-;X#_wmbaW-K@td|FTqaHQz*RodRnP64g4>1F=zw%W3cL%j6BeKwhy@&q
zd903Tn@QML)1tIByxj*Fyvli!5t13FVk~0}WE+d&iW9X*6jW5<@XN9<zpz}mfDOT3
zx7z-!$NEIfw{N-t{+_yJOhSD`z#vWm)h!*P&kvU8&!3yZRc!m8<3FOm91OKG#6K2(
ze$tb}1?lOPm4^#izCABI@;Bdq`0y6Ub`mNoTv%l<Z|}$hY*`uLu+7cQnd6l=z9(ov
zF_))5GO)ZQ<}wrFwes{Sb9;NcOrB=z<RlYt&8Xs}{_+9JaIdTY^M2-))@fj~>fXG$
zeVsn0t9|jyl6VL%CisoCW>23vO)_%8@6uS)WS0{DC`7343Uj<^d$wKN*wD00C^tRZ
zA91Z<AdW?z1SX3{(Asq6TRG5trDcwej$GC~8Ii3Ia4jI@-rz|3Y%7#1V09WiAyH66
z1AeDQ?(YA#Lze?pnzXbiApD~hHmc1-I%-EbEGVNQI;C7@ut#c~qJdxmsaLLWU<OE&
z&f6ybp_2gzgs}a&MPH1+@tg%T7xC!PJ8}&Pj70TwqHaHKpeu`_fKG-?5hq&UgYGW&
zU-U}!nECd`lQ)czE=<yQdnzx*m@Jl#i7B?KN~pb~g9YwPPfyS5$Fp!g`>*V~2L~u$
z1f)J%-OCe8hImeRE2(ab^b-ja!N}U6sG3?{c1Rq<K_V3}nV6V>QvpGYg_ZS`q|fjC
zWRHZ$kK+;(v5R%0r>0DSAvy=79h8}XQV|5i081cPfFMP~6-0C(+#Wy@dgi%4F+heJ
zUkrL%4sAkmav0kIaeiG;`^O^dB<=lgsSuEmmZyco8|(rB66&F=pQrV=YX5j5h$X<0
zz}tLCrXjJ3je;)KY^lb#75~H56GL5&1Z8G208t$d$G{Be9Nk7AVPpyf`f#8HYMe%=
zrZ#rUOo8(RBAjVzZZ5PVn(DlOK>v^3n>TOHQ0D=cjMDrt<K?b%AApLKJ~<QwI1F4r
z0`aP0Z$2HRbNau0ff}6!4-?(i-oNG}uiZwd#ogFnd#*nMhep~3V_<}sS&JkkCHc+U
z@qqj{|26+<<JT`NWOskx9JJ-#`E+NHDVbBco$kCW0!?W~`eZM*%cEMKn}>>$at0_&
z;1w_V`T223P^vIZ#>U3tB2ZH10-c8nI>7OPWe-8)WyQBqq4Vdlz-}O8SGobAS>JIM
zK0f1mu)S(#V`Ed}r>RK>+<s8M!`k|~bB*7%Yxr=kg8C1hfh0$15e2jvBoBm{kMHx3
zElv$}b=2nq;-Aa(TO0~X$|sVZKw}<)xDnSq2Y>MJ@ku9@&elU-0-_oh+4%WW4n`NV
z4QcYcpddMD&qxqJLf8PP5CnS;C~HtOQBl*(j0K|a^z7^dlYl)OG|cwpD&4$sgL%S4
zB~R0@Sw|taj~T>i_QsmQ-<UmJUH8>rVj*>}U+eodURO~;5e;&KokIEgbwLAT66{Sz
zY;X}F(fn36HqoWE>FIQ!7I6<=cKbmB+2K^Ag86{g2cX~pJRk`0*3peSpbt?eWM*b+
z^Mej9OM(s9%M_3prAc#QBE|C;FEI0lUV&&vIhv2Jx<Pic`ip&aU6R)sUIHd#Ga7l!
znim9h$vCBQ#eE`Yn}Ufv5oCJe7Rl@RRoa;QXsyGj3aRSHNHyn#|Bew#V5qCVm{mU=
z90bwT^!xjG)@zr*9|nxo%$xP5HjtA*t|wt&h`}dkO-<YaN;*M9DiSDf$B`=h$NlOQ
z5ZZq9=SAPt0~3$LuaAP?YMq&&D(_CNw*CaAHXRvt7{A~DC>W3++t^)P;k3Dyc7vh9
zscNFi?iSKw5i4n3H6ACSRX9)v!d`u51<)RvYhoQ%c0#n4Ha6+MDY#>G3a@&2h#<PE
zth8WLi|i71cxSDHiby~h6o{ajfv~{K{Cwpi3$_01pux?}xh`v7gM9qHs|y2(#<xst
zngGUxor6OzAV3N<mF`9=!snWoe)~vPR940V0)~@m;Fw!i_ibhQHe_1>Fy*3+5Hl<|
zuRpsXliMd#tOMpp)KD0Bv_*BizswuWFGK;9*^KK|i{O+zox%?B!6wl=lTAj>oxIyS
zJ5&ij<`Y3T*dus#30<x{_tkIs$W!)T-_l<HrZV-u(8O&fYI}~kwrym3yg)K*{CYsx
z10CWa^Wv_ibAAX?MwOUaf23LcH^f*%t|>EXJB14AU|)HcarvI=2WqlwRJ_LRz@(~f
zn}BQa(xpqEm$!hn!wovkbP{m5*$I@C8BAE~r%!|kIN@-W^IrTa0IrnLcY8(l=FNDq
zuX1`|{sF@riXtK-TY(w7dwl4<GSN8td<Cd1zQAKY*oRZYHrXBEGI<Clq1)fcOLG)v
zjwHXYY1og95D7a%ZYPUuFHopv+0>nAG#;+W)4T+M73TN3rd@rPnk)%(l1c}5l`#H|
zoF}LkV_3CVXSQ!G_Rx7;RdcB2v&9e}<=W=G@B+z_pNt`V7d%XSm-!lebdMGPq`Q1^
z2uoP7dsCs*obZ|@Ou$Z@jz0osj#|<<P&!HF;&xQw4^mJ0oXWa7UDXYpuNM8ED@$^+
z>_A%wXmOUQ(-|araH!}(WM<&JT_VJB1&lqA+GZXQOSLrlACx&yN+GCfSMD;an2^;y
z?_=@A!r_XQT80=Y)~Qq?x|EcZaG(J(^W=@H9EN%dB3~0&naQt@K(^m3|M~Or($24q
z4U$nl(Xe}ZR^d@)FT#tnv$Iq0iaz9aP#>OVFH(DP>eQ*IooB+0A)smYl^9(qzkSEh
zkY0#ZEwqgoG&mH#1C@?C=89hH06jY6`?SIM1;}rZ%HMzb6bk%kMn=XIly(q0bB9pf
z8CAPoU4ubYPheF;D>~Hn5E>e49%~}+e@}=uMoc3{m6d$5ZT|cGeAK!b1e(GyTRy)>
zfQmB2-A}_T2pPOwv58P+txUb^3CvLGtLmNAAPg+NyHhftUqs7$J^k(P;Kh3xR8#sK
zSxBk!LQq*1uC+lN#S2E@@E7)=TK<g94R^Ke+cdD(<JGWb`Bc`TGS(0Hc;1J#zr4(h
z{A7DK!*rq+ZfNpm(w_G5oAp5QGm*heo@l+V^#(2BQDg_QOdCWWhpOJVxHRxrNgeN6
zRCe-TP;<R^uY26tOQO#4+~Tp)2SRWYseMQm1}hq*_ZhJ2{TM3Ere1*PrApQSRGD+b
z<|(v{0yH@b?}C95ALOC?U=Tx%5!Suk`r3AE;5Y%?&pP_0zrP*3RPGr`CsNa%TyY90
zNs%N^Fz}{V@x&JM*F%g?S`mQ!B2sJSK<t2vi(4K8p!GZ$5FxbzNfX+Yf(42b2t)7#
z4r=^#cLcCuO+G>_`J3d4kX<kcLV%nB`y0@UyoJ;Q1UOpLspJ70_u#<;ET}0zgH{$2
z8u|>-l}P42qf1ws?s#7yi&WdeRWXsi-nDXogM;%d@Z?xz;FUAi_^s+t8sc*mDvCbh
z#x%cp#~OWyEVE}!t|NLKOl-V|q6`%WkKg!Q7*fvA#Flw*^m&R4F}?Hk6Rll1nhV}P
z4rdCWMzv|sSonN?+vfU1%?!5#P86Sd_glAa-30e1>}{9+ER>Z3jLB*h%OrhA;Qo!P
zjlmk3;yBLE>#n{{kg2AD+0bq9;y*=5Ygx5^98#{5_ZBJ?RQiL`l!cFvgq%&ieRp#f
zRjcU5?jridgG-zGoSFu!H;!Q3!C>&50U8`QiBsUOocjJ9vIuHE1(pwdK{;SsnrZhR
zX#Mhq9AJ~YiwpnE%nWL*1mX{1JTpK=2teYQ>SF=3?+DRq_-ZsIcvaM@J7g`veMJ_p
zeM;))ajkvS?PM%6d^x|E=JDujpW0^2>o>Xu8Jc9H@7=I4iRn0vG-h_@oTU76s3Jsb
zopaZ9N{wTHJU0jPjp8W8G#FtB=-TqdI&>Eew|pcSy}y&X{rHT5OUxMZKjNAn^dKn4
z5-!Zo&wz@DG9NLFnZ57d%RqgnU*l+}wGI9vlt=518Q4gzD=jVUY66#UkF6w>AK()N
z6K(#@KadjqKTyuy_E8RdZ5zHs1@?6KKOl9xoa1Q=v3d09(Z=ScEF76WFjf_Ye<;%d
zp6S`pvYE4UEa*hCpaGJX;0Gf>7`w=c0q*4|y54qyWZ(k0>fYx>S^H=|)C*Z7_BI=j
zejzV1R{tbwz2~UY@{QTI5YyXVI-eJ)1O?G~2byl%q?j$(jh_GO{}o18w6tlT1Ij<O
z?N4cfxTp~jG``Kb6ce;X;N_R;#+?ay5c+Or!b-4kaS?zAOv%#n!gtrpS{!E;AJJc8
zTn@-pjCqT4pa9JPzN2^q;j{XX=xQNMaFK+DCK6TV;FN%$43-kLm&JUjaU3RtrLMH;
zeFr)?E77I6Cr_fP9foq8Cz}QydleRY<R80uc)Ww~6&Dv5^!)h+$bGULNpB&tLiwm<
zYHC`oR-%)v#c>tLLzIs~YXd>STlC%9r9*ANYSZ^t0|^XbvQ<pGtmT<M6Ll`hC3t8p
z&pS@+NOHs1yKCHti9uFV^~s`(O2V7crV4U3+^zibzc$9Dgp1dc6Vs<)|MMDq&z(i#
zH7p8>x=?pDC}W^4zO{8N1J4^!6okMqpiK^_R|PHjq1270<bDfnDvoX|EiE#ZR#vSb
zCPGou0&e?wNI`(;wRe@@K!i)^LzBh&MB^spWfYhEvz_qU^<O|C8VtA-)JC*I2YKo#
zbV{)E@VtYf0)TfokOey3)wkFws{4rsG+<6a%o-3Ucu);oMfd6_buFe%&Qs?q*bB>a
z4U@fuL)*Iu{e%5B`s;-bU(eA$@;`XF`>XA-ZDrSZ{Y}xdF_b0F!Bv4Av;hm1^2QmK
z3<wT6up7B7eZ3^)JVBo>>iiC3rrL)R<3?H(eSnUto+%y*I2{epbK4VJ;2wcWn-8pH
z8CN`spn;(c0_rAECTTHSFvj-7y<hQ}nK5tzXX_WT<iWA72tJydxr)=ZlL2kT2DJzR
z{4h$aKV-}D!d&Rh?t~hZAF|?DgLkDqxoh<R5vg`_WbSb!ky`pQ`7hSWO%?&Jq6HG$
z;EjzI(r>I=d0y7GN%Nr0_4j0~?a|JShaLN2-`9NO%Y~^gqQuzEag-$j*ML<GG^t?2
zuvKfWUmz=7q9gdDKoM3J6uo(K1-J;SnvbANm6>;(b}`~fu>eY1x6JhZwP66z2~Ph6
zs*g2*Hr~1Vj-XI&cXu{ctwsaz$(NTJwe<|=D#jbXTN*aIc9>sw!jTu9UzkfP&Q4Dt
zjM-m*EjnW!?^A2dY8VK5_jLuzS3>$OdORsnT}BRxtM%9ak?32Ut&-9+Q&~)$Mkhb_
z&eZuG%Y_~|zw3zXE~MfScyi-$0Wq)+mVHUz>0R_lRFyn#S`hzcrx>~zCnLl1d-q^7
zdZ)HM?dxBYI;?O30Nsd>`^TkMbho;-T^e=s8cRcn^fOA!q<sI-Z!kjxOi=f|1+3|i
zR}5S5JZk4fu>{Z^(FdT(+FuOA#Xz}^XXY_P-NYs#KAqt+B8>{>S}*XJW|+<c%1(+L
zb>k3z;<1FeQ^Rd9{I|~y6NI8ujmFMbubuND*#Z;{JVH@YO~A>KC+eJKKXN=xKv0zA
z!6bx&?Rp=capLi8+^5YRZ*AXj3F2%V`PSnxwOqTa^u|nniB%bm@8<+IdjpRvOXPvE
z)kk3|4Jq<l-y8dh^CYk5N%ZQ348PQU<Sb(PZu?}=HRRu<Dp<C_uSU9F7ox73$NOrq
zkoe{#d(}pU?J_5;=jA3Ibeq6NIU)y63gFdOcWR)J$Wk;i#8`-iZVZ8v*#}Nba7vHx
z!O8(Eu;6Sl*E{q<ZLMi<tT`@i3K#h-A6`Ft2lKHj(Z3!A1@}AP9ziL?Wj<h{*m$4?
zff|fn>a+Y%;=Y}33ttw0Xph>+r~8tW5ZR#p^fcCkiLA@yjMwxnlS@Q59oxSDyED?8
zlC9XheOPNBS5=>iF)cALJy|*UReI*eS^Khpe$SIpc4Y@qnLqObly_B=|69^2rW<Td
zCeig#8DsQ_OfAniWEf&ZxmQ{Kcd3%v2hYW?XIY#^26QE)4|f@a^wnY*noD4Czk?1I
zH8tzI#w`Nr!?ies)z>iJf7mYS_<<r2+AN4cGu1=JSP04$LM(e#=2&}Y{4RKK#elAf
zx%G|YW0oWxxYy)>6?oqNWD(oc7E<r+1Ab0&&;P@q-KaAvf_iSj2oetHNqdw(GT+75
zzU9l0zN5a^L~ZZ$IH|fJG_4%78`Dcwyc-$c36vK3fXyp?6v=2R-WYT;TK~z?_iIUJ
zoi_&#r&6FpMX7xFH~lEI`61t|f6)z?H27oPxzpX<-M05@;aQ(;f$)7j1Mr{<J$Z6;
zzMgd?{{uhB@AqLX=wdGE8934957_4(<cS#L!Zc2PdytY(x_|tZ;gB{vYEX*Dy8CM8
z^^mngmRzR=UZW7>mz^(}(uv*~hv10a(~Fhe%^$Hq30HeZiqkld_JX1GNPcqyULyp-
z#SPbg^pG|InwJ>l!wAgQjUe_?!;y+nEmsWBGYAY>>n{~s%oj4y3H%k@w;xv8jEOz<
zh5tp@IGNvRnON9bI=&y#-5k>g7Ohw)|DS?>0C<)4AaHjY7l_c--d<uTb8lz1D-)W)
z3IHcw_?qZJM8pzF^UQgBg`X-ycqVj&tMR4G&)a3;D#O_a?yLO_oF{hTT=Mu*p^SQ|
z+tceo&y&@|&?&s{I~_`eAV6{p53}K6f;ak6xkaZO33jeO+H3@4bQ90)^GH&`0gMB9
zkH$&k{};JFR|7p|>S!$pg*mEF0ION})x-P$GLg?ESBk71K+hKXyC9~6P?6Qw1eB5)
z7^|R~`tkW0R@2e8LBg^D>2>$c7`kvs!kCE9&YF{JW>i#EnQ1F#mHohJxD2>--a_6g
zzH<@)=1P!hLrY6}!BPK+kWM7gw`jQ127I+Lr9*U1wcGppzMmW~OB)y)IE<FA9LXvC
zW(FZBKYq7&%a^)tjVk~A<SVVC$&+0y2a|ELibgypnOLrm5n^YC2Gt5#G-17SaOL3)
ze13k79PCY<mL`8Cfz6f5-X{;U?eC~+$0%APYCVjOmIsqWx&ML3?%tkk;ea{l?LWSi
z=XTYhgZ#g(?1CykNBTXy!4nk&?C>mXBfQh6-~Qv`VIxo{RbE^>Py-BxDwW{s2YAhP
zC-(oJVSad4&!85_$R87pZ|Wp0-;*d`2;2#{Q8)nia9-+gLmte(xlpOp`D}5ZJUvX3
zf#XR42xRDNvSs^0gAXQfma172G*FCl)vg?srd1COjW}3$?+yPP(H{-aUZaA^Kodnw
zOH!g?!_J)N9II!p$Ewi5y^Z%~sy`%&P1GkGG;8s|z;bJW;lZt-+r=4Y(8y~6P6$3|
zUv%}ops1t-&Vd}1OH6n%j{+_rl~G0?KqD%9g2F`#^DUslb@IRL@^Et~DGlag@KREX
z-Z4z-`7#9#Y~311nlNIfNASL3y(<pXZ=hjzLiIyRchKDBQyg;;KqunZB0A~!0DO~Z
zX*shE0>sCR%j95eqwFmONeBff(B(u5O|aAa#l=^kF|^2}nGr#q++Y$To~f&aa_Ajo
zPEw%W00~%t0rY)O&pWiw6fj_JS=r|o4)u0lSPkp$VgSRe2z(ZBqJSd^{lG^AU9;u7
zlTVf%x`iv}4vYMqoa~w_cFrK*7Z%Q+Ki~G9jg#{_*Lgflpa!n-{`g7#fi|GXO!=+B
zC|KL6x*e+Z3rE=L5`-Jyb*rDU`(#7X`z9c>w%7?TZ=vMNVi8BIz?{2yruL9d>Cos`
zMXRhk?4x1ATR5_1dkuxI8hWFOYE#q=3gArfkvbOM*8Xuz_2kR`^5!H6f7Ug}vE6~P
z*Y1C?^d+UlB;gZ{viJiUv{6?*a9b!=8f)^WyMO=wOcS8FnZ=jcVVRjmc}&w(*ceKW
z=v#s4nwp+wxx@57eD*X*37AO3ubwN=p!yD~NS1^y1OY8_7lau;@1MZ*TDFXizEN2?
z_#X@WOuNhf&9lF#>lTN8DYO*`Af*|ILy8BKyv76&k)XAV5vnZ&t)#%ONCdF94RTGn
z&(_klLYm9}@dCKcmZ%s#$PPTFfgzy1i|tN;tWi5K5xrOt+9aGbztG9PoEs}82UQzC
zC}ZU*ccFts>R?6iaM|QUz+h+uR?geNH=w4WX0XO&_@Kk|L*gaq<0u?<NQU+YD5Ovs
z2ewD8?l$0I1a;s8=hN-zjqI)mo^Z0Z0_Gq9Q()!XEltgKUIlP{_LW;Gf%APe^5mOE
zY%7o(j(gDTGzG{T34$U6;(Z#o38+Az03is}Xkg;xdIyG@i_mq6AdCUKW}pl_;m(ea
zzYnJsLBO8>x!6E7+9^ZIUthIh8${5t&7SVp{s$tk6EP4B3=CUn=O6wFjopEVo8fi9
zQt>pgb@wlC!GWPlA<SOZd{s5}(+WS>G?7}!Pci+L$YrCObXMg3p*T1*DXcs^JgCv1
z`qnV7o_@7GPW_nQX?{sbBfrMRh6b)<^>UFkBOOo<pwaPu|4?s7$6@<VIE^zY=CG`a
z>+4`$_`Nzh^jsZG{2eUGKHDpHo0SK^@YcC6QzE)sG@7^2e2l@V<*?a*hBrBVY<zQ^
zC%VfBsq;QA=@L0xgK<Of4~o{h8m6@0{Yc&bY>ktFCk(ktu_gWG7ayP4toO9#6dX$j
zHQR*IB%9$5KfbCrrYY!&ExdD=<;&fxpM;LA?dnOJDOB%WYc|ufdlHm$&e~o!RiCHc
zDs}PpRVz{#SsbcuD@yBt!1rTy9|`X9MhE#B_X_s5NkXfQ)IyI{VsD0T3G(8_3)IU3
z4m;h0<KqKozwKcgBnS#gpgyL?JtqfL=wSgt`g%9Tv=#<I6*bxeWn=Z&owlZ?29<t5
zl6|IAc)8PP7;b-K#G&+=TDr)$CjToA*+3iu1^F(PsdS(=KAsfhq;lxC&2$->|Mo!y
zoWmeI6#oRG@9DYeHtoMMJ|Bdv;^%5IGC{4g^%*8jv-9FoCF|VT#}2NIB>f~H{oH`b
zs;Cfvdx1ZR$;kq9`^MlHQ1wq!wJr+)Fjv^rBvscVcBka4eRa>TL9G=3XEY5VLKrj}
zQ5#jd2lDgMg-xORi~c%#v00L3-phluJKDi~NXuiR#IcJZByxDKxPhTUWkL~=WWCeV
zaO5yJtb+#zp^JeOe7<yabf|?lflcEbFfe+8*us^+ft1T}Z-|K@g!T;whEAA?ps4WJ
z#r{$gX?iI?YT(w&Dk^3uJguO-R056(2*>KjS*RtJld}y1Z@Jr|X0ZjYNz+4MdqCXf
zzs~WsqdX}sH5HVB^U(7OwniDC4AGu!_=;so#q7>lCM*PGvZ`&GCws5VejNT}5Y66n
z@pj(Z?%nZWdCp>MT?DN_^Vf;vl`ZcomZ;>JTUVaO#*FANf8eieEY$+4^uod|xEaCG
zBwO;+q2ki=1p$j+wQdBb#3`28w39X-B-(td#BUnimxN8;RlnNZ{Ia6v>@E(n5!fdh
zw6uPX1)VMydNs(jU@Hw)WqnXy&L42RXCnq-q~U6EH1HKIgM*Z2W@e~I9cpUC6k5n4
z9zG07OQX5hT~uA|UVQN7<@4v*`O4JGIeX_{S5;By=;&n3l&nAuq`vHW0rkV!^fH2(
zCvy0<j?O6r+J?x0?szV$n5Oz-E8N%VQSl0Z4G<YomjZ%72twUhAZ5P%)oMG?cfm)f
z+fa*@1P>k!vO7#2*26pL@{5KkWhdGB#IH2v129dSAe1^JN&Ebcs81Rh8KH#=TF{GA
zk|UW2C4O6dlH8XEWA@UnGePXQ>)|nP$aK~u$5v~{DkKc7>CovUwqs^({myp`v}5al
z`F<tfS-f_pmzTQ>!H`01T~E)m1zHs+tA8h6s2shx8(w}YO~JJ$SPu2;LR!WJ;oS$f
z_ev0de}GHO@E8DuSy*&vFE3sS^L~(ac6Yt&-Mc|R<m!0q`*56&J`Jwgk!AvmuF5%i
zd)EcL$vI*!F0S7^v!)gnZ4EX_+LE<lD8Pisn6W%WrJr*@PO%6UG@ofQbmqMddu(T!
zNqCdRfKl@_Cu_Yg%ci*c$}Sxh@%ZP1{e{njhm1+>_7;{m_>$YLWo)^;OJ*;i>+Bs*
zNGpVgq`F_RFgHif>BQ=f*Pl$-DP~~yI^j5gCp;2zM=n5Yu#Eh9K1BRenY;^*YSjAX
zG6{ztdwYlKrA0-l@)N7RlX+U9loE&|rgnCb#8SIK?#t!YHKUI&UQbnsEJ-QBc7lcF
zVrE7_WfAk+<%84#K1G0F_W@+y_RZjVBCH@Bx%Y(6Enx3adhc#sL#^Mzf^VZ1D`m79
zIG4Bj4~?}RrC)bdWG0berx4&RSr4D@s3k7JKi|Z^xjO3|o8H5{lPb1#^AhPt_p$0%
zCVQ!IdC`2fUV#L(6HJ#c4;StGEiUJlEX`gvy0vwm3}T38l<V2=eP^C2@19;NqPt;b
zY~2iu%;-t=5Q6B5x^8rqoY`F%q7!5MRXZ^hnc7^vUv;-={DCkYdQc*2bc*#`-@e7f
z$H({Dn8IL`@{523Ik76?{WGMRJE03T+(SR%eHp~l(-Unf0j;eQLQFjLra?RNN9=}x
zqt`m=;3Vh>#yOK(tyRBmlw>_P2xW`gnlv;Tx0~~GadV@g7zk4q%II>BReexq&atwF
z4cpk+;Xk{Pe=kN?tB5>3{+aJy`3QaC=I_N@sJB5vf*vjQ#*K{<s_BJmy&alJ@>TOR
zY5aKWGE4*o1$$=a1^>k1U&aZ`p(Juo=u|ey`tip746!G1=){JYC^y6r8SRp{S#$d-
zZu#o?Z>L7m4}KE;i4P%k_ho`W!!ZM0Ud&^K`o`k4!)se>N*({FiSspXF<=AP-QAsn
zdoFT)cH_Oov@tVD3$!o*4GA8lqR}7>Q;^xA(lfP=ab11GVPRo`MrZJ22&6r%=9QLa
z0w)+CY5EX3mW4X^G;oH1I5^*%-FpF8ZI&<snqcVpEe^?rW;np(d^SDwnX3cyx-oCx
z+SD>$5f`^#o&6b9Q*eO{|M2JtN=G7SNT}S_4k&1pGVCc;zAz(kM{0I<f8yum_3C<K
zJeKG6PqBGfVjiB@yy4##HerudEdL;_&&MRaGWl{#{AfY1==ks+(vs3zuD{x~dA+*S
z)ldolZC|I&{5`t!=vGj^CkQSk;9t$)$uQ7~GgRMHpa^gR14$QgGDk_wD0o^#gmx($
zC6Y)5Fdmq3H^%`x3BFxg;2=VkF9hFUZwII!|M+n*L_L8m@RpUqazlf$3`f$>C4&_B
z=rf3`a3ND`tq+z0e@wF^w6q^M?-)Idi2<933N(<KI^^M=2@ea)DJnvlZ|i*?US1{W
zRRQ=Gj&p|cM@)Ke`IFI?oYHm5g0nTQm`0F}1O&)XWzTEE4WF1;$=H|?AYH0RbaeFP
zMq<NW-XS!)X740Boc!;}dSBmsEZ3Jcpg-qua$usGCPD~tTA$ah!!f$_g)N)sb&^vT
z9~by{mC*&!mU5r*=dt(`bnG4p^{RO7Zm>XNEiAqQ79d@H{Tc8I+=p&P^Z_T}!Lu(G
z<^cWrU)4fMN%>=-kP05HfRg_(L}mej$FHt{((JbM6(7L16j&wt0AW_4*6R%bIh$(+
zRM1TQG4l!)j4PzNS{RH)XjB3P%M5B?+sP)$^*VR&opDbn{?Mk@c{3we1+*gvEDVI4
zoSaZ7w?dwSs`V-$?n<Hx9=}He64bnq2Encb59!F$Y&hB;MLS0zWrGqy0eT(kCId@8
z5`;R#VZu-KK0^`3=AeNydfB@!F0#E@l4f`AJOZs@-v3MZ^4^>{x)Xm*xUM$W44q*K
z5yDNJy7)Z)efOufK<gh_SFhMAZxP#g`H2$Gscb*sRUsil@Th;jIv$$3ORz{2b+LS)
zlvD9fG_vo6NM!Wd4TD3{O^pcv&jmI%3+RXdOCS$eB!WSTpFLP_9xAv3-aC3|S%sFR
z7U1{ep?P`n%S$Ntu%Lcbl`PkN2zOUxSWkwa&$a<q3%IL7x3pZP{Fi^iV`-*gM1a@1
zfou;4YYHl+eq1qXa#_A2-JQaVVj%}bSMZ7fjO=@%Mu&3qpcSI{5gc|-e0NGZ#l^%-
zVLH%F0HCquqwRQw)tR?#ZD_+b+RINXd>YzxQOW^2@b3QnehtdJ^e>mcqP9h#?$4lO
zyLhn`(mI<)mgS&U3DMB6Zuc?2cbyB3b0dLHB~)^Fc0|}q1<`yi6d(<O>kS^&kOpm*
zBnWyRC^f7#_CkA0`kq2955tx6GpgJhL$562sOfI|8tzYek}_@L7OE1`sFXh@t9?Fv
z0FN~)yE`ohJY%j&b3n#%%Xv8Rv~;0ztv+Si_ZHuj-zpsFby~d*u7%}G8|`KA%p7YY
zGw}Zv3kTraEdfzOLU*vDK50De+t%0mN_OU3(>=O7))EHNe#Z@=Wi=J25DD*n>X#=H
zdD_kLZ&IcQM&3-!T$4ydtm?_YHv`%7ZsS&4C(sb6E5mX8^;vL3s>h7*75vJ_V`-3N
zXCL6W61XPxFeK~*>$Y(HnJGP2PpXTa_-Jf0#wJWN&HG*w&?`7xpm&f$B4<A{wkJ(j
zw5}9;FEu;xibtw<>PxAvq0)1E`vVIk!_lF~NzBjhANgH8lbh^lndEeBH}jUOsCmw|
zJe>;3_6qL4Yc>iK|Kx)2p-O^Y_;#k<3!j3Q`WSTdJc)7V7mDaE?@IpU^>Xtt8>qj#
z@mj~=iMywr-Kqb4>iHh6pfv#$&0jv7nS#d$o_T|bp#4?q>V%bA3fjCQ3>@e|z}PyR
ziPLedl_%cD-N>3#LJ*CLpM_3^6b=~qq_0;MB^6Zu-A72=3$&3AT7(H1P8AgurDmIW
z&m}36eu<7i6FrioA2U%H80_aMR_NECcq#SD4wkjS5E;R1b{9r_3bX0j6B$cWNxJ{s
zU`X7885G^1ft^BN-9_ERc)8FFiTEhd_o=L5w!FQ1CM{cgj-&W!F2W)}M-D@VK$uZ3
z^d+Cpbu)7dqu86||NW}F3)9nmpIL;?z=*zIUf%04dtb}l&mq0Fpf|V;Ii+Y-(+@{D
z>xYwP-x{Yo^~Jvvi>@36JPJ=qRaI3!{H6sw-T=A>73IVGY;OXfU-)4rd+EDNch!fJ
z`a@|vY)quZ1h-`B<HU*Gz|C2FRns00PI*$qw5un5i&F5vU-_g4ls!}cbNIn$&n4Zw
zzd$DJ`RBf9&z3W>b6O`c9L-`UjNv>LtT%{Cf+rf4r)0|~{#oij|9bukO_lD?4fDd*
zQ<__!HxQhACgTtouZq#p)T?KOk4tXm;Lz>3?3@3DR)A|b=_c|2%*6A{n{;AGTW5nF
zo6}V;llw}<m3mg&L`1q}Tt$OU3k&n+pV+QDjH}<1!8ev$idf%2ALnd<eYtbLhlj^@
zacSTAzWy6^d>!P5#OUW9nm-d9x3B`X=@xKkSAXa*=Z8mA+=sSTV(78Fe}mm}l^7rY
z|3lcDKx4gkZNPUEm1v$+ilkDBM3X3$5*39IGK7$jlrf5OgjA9#Au<oiEMumV43Q~>
zD2hzUkny`V=e*DJuJ2vnx7PEVbxujQ|FD01U(?P3t0}9X5Rw?n2~?{U!h-KjSz=Mq
zlQuF&TF&vAKMpK(^+wZAYygDR0jo209Uj)jzfQVm)md#;@l<hpIj`Un=g#2q&2Ahu
zX|8Q|M91fa&vJ9V^swhbZl{;M<HES4jXZI>J-jAQCA;~YUVeHrb){TyD@P-K;e#d2
zVN)3)C}^miK%5i{w|bmJ1RRs|n+On{q@silP|IUa(%w$r%>EBv@4j98``hH{ZF*OG
zl#ZN>G%GIVG<#4g$T{|50WQp;8yZ_>h8^SwJ0fE{qO1Vr<^8P7=N;`DHB}z?x;I#U
z`aocJgo*??H4i>P`u6wVe`n)uGeK}0Nzc)#s}E-3N?{<o6y&e*Nv?%^^2Jcp1oDc4
zLql6}Sisw9C$q%?BQPDWP`U3PGQ%MI1?YAPpc@Rl5d5iYXv}~}2;JTxuJY%bPR{+a
za+5i(&PU;i;yTfXi)dIXEtHBrrBwfNm?^;fpDUwK<?sOC8yaU)+~XwC0+37)9L3x*
z?9d6o7qDiFa<|XPXRZjgu<U*(YFXA%1Ax<H@TU^gD=VQck+87X$2i3GD7(CTF+^(j
zfVAUXVc%cLBgNxR&MeUBOvql7!H|g_JWysX0J(-VGZyGNp&c~%E;COr;kyBkxc=5Z
z6X&+rj@*$i*qmx5(q{DRfbrc=e1&FVkL<(yTRL{1o_d$hPP(|bP)61)hVa#fzn?fd
zIN)!`HHquk=5dmcJ{)vN6Ujp?3LHJeOyev&8EI3(o{iQQir-x~x*Nj#71|ICp#hus
ztGKw^@$rIaY~3iRgl-Uy+l7onpy|<=cm>OU+Y4HHq0`YjXsM$IcdDd7!F(rW@8G0T
zYC-53GZO#|YoU5T6I`*s8#)oVKK7dTZU{|U;Kv-(>Mdlj0zk%{mKII2rgA-+3s|!D
zmBuWYN4=_<*87NDOfxe;w#_xsmB1dMz`qTX1MYX6MZb<;``M;M?aN~-g<-OQXprkh
zni~8dgvz74vAAd})a8H1*X$*_Qc8oViaal?tUMpe5o*j}guW9QxsBtXr2`a*Eg&eA
z+x?A$&W88G&`bgzf~lo;W;lS>C;2V_*?ZDFQ#pwKqOLDqytu~(*uojbyRsrd{(@qC
z*MaNJW@GtB_9)hWOJQahcYO+CdqF4hT+Xr!=&QgB#hJEel_`EmJRJcO5E_#>w04vQ
zS5dJ9Np)Kx7wR4E)2iMd$>?pGoLC9qr4aoi8=he3Wx3GX0fdjTq^Ja-!}t-uKwJb_
z%s^*N5I!LkT~M&HAoyrLgZ>4|lzF!g1uSn5lMTZCCarw501pq3c%V!(n9A05{-XC<
zFcj_uVTI(cK(n7DU;#i3F}bO}l;91U_S>E2FrR}$TC{4_UzoVxLU{?5FS0c_pjv`e
zd;w)j6&20L^IZ@u>F(tiB!3(cO6OzM6jQI!AKb9SiJyFIM~{Awx-d}Jvg?*;kGBfn
zaSR`QK*ZlRHgX_7ArGwsKcpQF4i1%JMxx{WdhvX!FoWR1xdY$^7Bt28v(eKM69a7s
zJx_n~X9@hotBJ}%VrKxuoE|5=ZSb_=jOE|6=Qv(2%r$iZagTlWJ$q;u=IXmB(PZiU
zjz3b~Z7xqQZwOxTlw0lIuZ4rjy>D7>n2L&Oww#XMwY>Ywc$ICXr1jK|`}$|8(^2Z~
zOP?QfhKSb7+xth<5K=c<psgJK9^^(u$6rUeL@a;L{hVe-Qht7n3?Ww$5VK;3vI8X3
zM1w$y1}!!Dd`THd1rDt&!wz%wTUd=C&c2N@Ci944)9Sj%7@-Q*@EU&q7`l)$tT1d*
zcFq3QkHuHV5jrXW7N9zlSWaL>c$>dY-NsuGU~w*S>P4z%3q1Qe>i`-$Lx#VraJ9Yx
z6a`h!nMGSA&zw09%{gL*NRW?Lf^H`$93&74a)_8SFup*coQq$C7PWHUfj3C?+oF<o
zeI7}40Y>5pM!3LjU||73&7fY3tV&&rF*sJB8>tQOsZd}d(}GnD#uf~9u(y&wptskU
zN1*VfQD(cBfo~d#+d=A-^Tvvdp6B!2xNRE?&^O#JX5J<302+>Ay4x@3lNTh}5u~ZN
z0>%<VIx^$ws(vFU0nId?Mn*Qc-R5g-`T6;9W6ffCd3hNt`)t^@&H2wKTurV=VRa+$
zhEBCcNY2B{^}22Zu0`bO9Li6DAf2@U2iiuS8Ae)KLH+Uk5YpqfU`9KJdJSOO?tYmp
z3kwV7G6~XP3<B!~tOoqT!a5IZ0iy_7)gjGEGcKi$jN)EHYEWd=)uq*qz`aW|Lwr|D
z%S-?e;2*vp40-kH;p${O!qTTs4OAb?PmLD2dPE?wW&C4<2_`;0;|4&<luMLzL#R?;
z2$FqoLn$dKwfALm&+le!)?qGvX8yEwmGCw0;%$vJzkXe7%UCpbHn81@q!OU!Yemc-
z{!x7m9bJlWyfly2t&6VkJ`Uc#udM|C40*pAYvjI_J#2TSI`tv8rn@!yiDT4jjtol^
zqeF#$`Y(sq`lU1F)_qQ7<C>a(Ia%|e=Oq6YV$m9B_aNdHhv;p*pS#|9;gQ||6a;li
zw+gSP5C9BleU9@eZ2<NIr7<8>7G#hLWsY=ggI<6l{&3~AmuN7ke$X_Ax?07P%L|0Z
z9<g;Tt*uH1?-AX#3<DK~llBb%{=$9}PtL88ADa=ewQBXjQPV+)5FmWdfwt2;QGQ)!
z>!)-JL3fvg#6+5XUjzjabG&!oK7^~TTSo#h%8-Jbq%JrIu0CoFP*(Wk@8eW+1%1T?
zhjh5`jhNPU!2(oGDAN2SyI*iE9)MAW3z|hYj)QmPc)yu-`_~AqTe>vYvN@0sfmkr_
z+(kp+{%q@(Er(=f7Z(>7D?06f#Q=N)Pa~dm5a(p5&`<OcVo@Q}n)FnN6w}glJdOC%
zbI4U<da_r)!{C9y6ACEMc>ET%3SV$7_1TZyzyJR0+?^rRUg;NU820yzvGo7?C4&PC
z7Yos`0`1k6l}{vl0pWW2`cf5yFgj{>4**c_)j#nPF}a(Zc`l;y)DI*6mnY59)>l<k
zR2Yo)HowbRvbH<6WiS|~rlz;%sj;<7gv;%+LD{l@SA)1msB}#p1PZI!49JsivAVjN
zDK;EvKm81h?2!c)jhCI6#Pu(K3XGKbtZ6$d>5FgdzIAVg;+qx;I32V)|KMVvgn&wY
z=kvpdr-mJ$2SxD=Gg)2_LLcB-e$UM-1R<VMJ@kP1Y^%g>rB_SyW1y!r8MG3Ij&bEx
zRxTSI>e1D4&>KSoGYo}*xvN%3WjM~XC^K_pMKQ~9*A5qiF04txP>nFLe4j_*-jXX9
zEt1*X9>CLTfv6j$dgQe&324oLymzC_qwOC<p#y;_#%sS>gu-hX!GaFs)nZKoS8HD#
zZr`qW>Y9Z6_ZG|UHNSk?PQncltzddmf8Okwfa3a<jeJCc{`@*|UNe2yqZ6~y;U%6H
z=KQ+aCr4^7v+d_wcTz|BvQO(T2xOScanF+bCi7f<v)(pa_pmeF_A?tYo<_^Z<oPq~
z9Dmew@L!wi%_<h{PIl-RoeboWt;)2qinh^=TZMZcl~hzzL>eK4H_Am$fSB>Fe*y>*
zOsxwFXKih*dfs1)y+kZ_F|n<kd*Q-`##YMtk$H5~5CS2FI@olaEzr$|%UX{ZJH3H&
zlwk)lXjgJc4@xo)dxAn@lYV~@<Y2XlUmsaoxqElylI8I^G6F2=yfHn2fqnfpWoS#`
z?|uaN=|8cN<ikLZZ84<kK4H**3-5EJvXh2fLz>O^51aulH%l8EpL<++dh#W|fWTo;
zU<NudNE6J=7TC2*Q^!qNQSp05rs0(<UP-K<&px={NwL9UZFLbR&#hlBwPh6-4r(pl
zSgS}k{Gkrh)CaNxK??nIo@SPbmsq&J|HhB+QfgQ=f(nlqI(3}^{o1;^U6M@^7oYyn
z1nuEs|6Uq4mcREWvx7mdv69C54`b>_$H3L~^Y7kKVmr*eBu*!y)7{tiy^HVr;%}hu
zK%Dm~hXnrm^F`2VB}09KMy%eskIbHfe$SqLKrUEy&$Ga_E~ubS&n$Azf981^LiX0(
zOF86kc`n#dDC@r^N9wN`$98RM>wd#vn4gs?j*8ykrm=G@OKAW7b5`Xn3F{+d?r0a6
z<vxuH)6ScoL<9A74s?tjQSEn3%@i9G5<Dc^=eUo7^EA9ptW<j!9SL!l-Yvn;#syJv
zUteGGkqmH`Zs5r>q$Iw#bac#z9f2g~kh4OLLXl1(r~Qh?4wvo`wMoe{MDjeES%^C{
z08z+<%Mr5*EsEm2J9W=yiME#W?;d@)uq4VzwdUTT4b9izWKP}e=OPv}fj%~*^!=Ps
z{$JeHK=U5H_6ZBMkECIO>_w!&AGY!7T!16dzX&v9)eDy^DvqhfU8li-tcDE@Dm>me
zYl=t4(@opi;0aaKmbb97>iit)42NLoQ0nW5zBRupq;|@$F_p&ynLhf#daK+z^t334
zt)NznZ{Dm~eGUq^qg;zoz!Mt{R4+c3FDx&AWYLt0N1<Y0HzaEP9mxWFEq-yg&yK~<
zVrC<kze!`;3TKM}Xff_lu2+FT)ym8oIh(*u>~jUwo0AIz_{?Wp{1*RPOZ1)ykKDFY
z7KZ8@Gr?K~wWPtN8-e~de@dQ|(iH62qCfqH$`rkS-v<?1Ao?mxU3tAFe&}@f51l{F
zp<hRa3)A|P7|)PXfiiy~u+U%0;2Jmr1duLABfMzYvOI(BGjB7%R3FpS<bo`m5@j&g
zRR0vp?E8@`)D0<c3)<^TI>8<u9vVOnh4iI7)7a0P($h)MtLBTIcciubKmqzi<Z@48
zmda6}o{AS3iP{9+j@1*bOEy|a-B6b8daCkW2wh9DofHK>jlk6i(*hxZVk7`#z(e*Q
z>BFkilJM~%4!Q%cUe9bn&XReUe9N`U#BP2NY?(`>QtgmMqS;Qh)Z*}xFt=MehMpUq
z{n>ZD-j!Wq8nj|}2S6|W@~j3j&MZ68eWUyR0+ojHC^{9S<CI=8c&-*>z$zyfW$*K-
z@5&21%ty+~>=-9z&zym9w!ZE+JzNJ5u0w-=R8#XsquK2N3-qbpX68`oF)<*tIPTi=
z(eX?J?7Pei5+UK#VP;^=43#kaadu^EIBQ9CCTdY-T__|LOI+nmR%f;4Xs8qn|8E+}
zMo-pkhek|$W?-yd46=`IW2%EpW)0-W$3SAkT|<2<3>feU!S2L{0sGn0&W&}TCCF?*
zzl5_10fVHdz*n1e)d`8*8e4@f9ip%UuobiKR}p950fg<v5&xaGxht)1-4(G6EFBj$
ztji8PwWjt{ee3u2o1ZU_xE&kPU649{-niRtFjD`T=f1Ezf8{DmPgew!bG#nrK=p6m
zXMXZy<D$9z2uMQ<w+KC8lnr5ND5dc`mdzJ{2n|m9%bP0kBq6x}9>TC0416&1b7Bm8
z-7)X3I4CJWn|_LPWn@E7Ng=3sd=z#$pRpe4b$SC|E?(0ditGZL4av*v>Tgk9`@)?2
zD(@ZTx{dA&4x+$;L*Y>J*cY0v&LY}LqGZG)?dsu}kyrbsucjz~4l=rs@(@?!RZ`MD
zSVQnIx}j0gLS*ayn#-FS;6<d&IU1DcIR%WNc;P}Wrf~??iBPoIX?pmC=@Fo25#JYQ
zvTS<960mnTx;itUuhqD(Nc-m>lbwTODJ!$qixouWFB8z)T(tb#-qF@c;~TMqksZGl
zc<A<2R4kafhNmtfai5IWr1^m{!yYl|i32N<b%o{_;x$Lfuix6+JFU&nfK<ZaO2em4
zk|tVvWFwgLG#;ev0|s@*=zq``zIvR8Pd|)^SA?!z#INbt<VC#GCdP{H0_`*)?>;bj
zg;r`#_wX*UBGnt%eiP)(MtfIS*H@Tv=pT4qH~)XgsQ?8r-(_c!O9@i#@8&p_!ZI@#
zpRd{o#$4(QjS9RY09urSGG!OmT({AO$BNSMsT+cCl_Urdu9Qcf-+6jr2iHBxmuJZ(
zNe(a!^iBvDgnf7s$dSv-I(?T^)DqY_-?uC{(4Z+$=4VaD+T6#z{0_H%IhERI=U?;t
zR%JElzwu=E5-M`L55m)MCpC3n|1M!+3E;={Bj7n|g&$#etppN)`QamkNhcQv_B0Bm
z^<3eig@Ps;4FSWt-Sk_S_8~ibm6%9Lb&HoRqm;54%j3oNAg~N`V>W1>f3x7eeszGl
z5HQ!w<qL&p+BG+;pI*o^pN&oDDPoz`?W%9ZP;I>G3b`LACP4;f=MU`3FMC8PZ-^uc
zL1s^B`V@YEkRM7cL*u(~>(+UAw4(<V(ZKIVBYPiU@fv*cJMa_2A~zG!ZMdx)`1sfu
z6qp7q&a);pw$ot50+Zem=p<-i%e>FGm3D%?#&HD$Vs_{4hYq_5W}baW;nvnn--ApM
z4a~};cS5Ce?oO({^@`IP;P4-;Sy!FZlybgz<e8}Q&V7T29}M4Tj+DIy!U7C!_}MMX
zFmmdsp^%W47w8OunHdYuM|P0g1HCj^$Vlc$j}P=C;L8&^T99{s7ph8%Dg?%84@8RO
z#CUJwJ4&Fm1Wb|(|Bo>;iio`hBLTd_4s!@28$&3x%x_UU<KU##(Fqat3BC#~9}F>}
z@oY7CSeGOd=Ey+GrE0shUW}pq&8g=UI<x;bPvro!a*kH&1OLud$q;Cx`6MDAdCo94
ziCbFkM_=J(Fo>TbeWdN9n7{BmPFLs3hP6+tt6wD~-8*zb@9z;VnKdJxD%8FZ!hvcE
z9@FOYaS5L&_=g0O4dJAi4*-jG-Q&Zb#a!%ueHV`Y#WMLOuOUg^?=PpH7u*|^pI1hh
zwl?-CYZJM0%cd`XrDBRP=QBUQa-_|+4>GR?QE+w*E-_-42~Gh&*Q+uTStcaiX|KZP
z2r&?u+<>^JoUXf4+z!v!;SGy+*A4r;K>N~G4a#eZ^Rg&M_Q>PZbBDLD$W3^lb7JN4
z6*%YOiX6CP{kC5`8^e3tL}Rf`pK%$$m+G{V0vMhQNMxOK_49p%H{f(yRb?UPuv_3@
z*4`+OhVg;qjqt5TnE0kWPL==niLd~Oa0HMAwqz9ZW}!DiEr+b>k4nqs`;xhkUeB;A
z9k_20A>hx<=YAS*=Qe0DFm-1n#`5kwe)wS0nFPGsyAAXizYLufwY8coozGuO?)drn
z+E`h$07vyf^P`ee^Oo9CP|%8TnsX?t1^Skm%WbchVF+dPqN?pNc5tnM-W`a~oP`VB
z?y-@WVm5{^(AvV+R|!cN>^kqZBt2Nk>fogeg-??^RL@2VYu7rKYh^M54$ca|RV|Ua
zYTQ*HsbE77I(RH2th%7T4#b5chJ;-3&{^0*12Ci)J6<5aN(wx3<vO_wK~H}{hli{T
zqa%(-=5n5*=Mj*4^#_fImg4i&Dz3a+cJk7CmuXff_CkC_X`_6z2qPfZtK4kaG|^-s
zt90SQMuf@X(V;{Ls_BSav4+P8LI$Jnz3iKy&migqr$Mf-QRb%)U+9Cn&fH`!84smk
z+>d%AleV9j#|+NLyc&Env^Lw|>Cv`tqsrRzxhoF`Pan;xfYOeF76!1(zW)givQy#$
zk8Hs2v9CMX`&g&7)|=Zp%fh*FT+DF%@p-G!npZ}KABr1UeKM8G7EgVP>3>wBNx+vk
z-(8k5S~}5xW`H^Jwx1N++eqKe_z|nb^}!+D8J&WcjSY`XN&krWIw6>V6B_t@OknP1
z8FF9R+aCds77StBZkCfC?1f1ao}X4E_2)<C7PkJ%oq6cW%yT9KS@RdWKaWoIHSG*9
z9p5x}76Z!>Qct(?{O7|2@5gJjQv}pvx8H&IB8g$M+|YBQru#*CzIy5pdA73k$LAHN
z-;<p>8>VjV*!!Y7cAd7eJ!LW~ihtc1vyQ%I#mxNIrpK0@Lz1W2zFxD)NSu5TG<&DT
zmw{D+=OVXXDmYqND3<A4Sa)kYL)4Ca*y5vRnDIj1OJW9sys2z2rhmwtF90@b0w5j6
zE)+V1;lZb%7s8PN@++AkC))&(v%h<q)&qWET_Je>)DAV(->w=6j6fnw+mq{%r2NB_
zF$Kyf8pT-x2}VK(RDQNG<4eS!XO%DAtp0b{qq4*@hxLP+H{}u({xAn}_Fk=tX&g~q
zK4P0<*>`s8!=JJJC5~cUJDxk^de?zv-}EG-Y&MrYzYW<}jxZNb_tCsOKjm-XS=LTA
zzD(jyEYjCjGV*vr%$6OTVks>ikp>$UOqd-HqcG9{Pil1tSy2PT&n_9AjND_?n1}B6
z(bK2Z>H_`cl`U1rnpOe2L}wf-JX{&gnyONIa<5qIt$CcmJeg{5*m_QR-yaH*&UY_#
zTeJV!$`zZ;-?4-jMykf$IJQ?Vb)n+7i?+$D@s=~WO&WIY^ZT}GZm!3HdCK1~zvwUC
zG5xt@!>jbkwg;jFc=0hky{%}|3A$gpbQS}$!u9xf_Wg2arFt@kI3BDQn`qE27naL%
zScJfz{&th!HAnW2baMWE@;y6!f*s|D)b>se{b2X|<{onT?nVDbCbt?PgO!YvapJlb
zS)Bdm5+|mrkB>7{)En@AHrWiVBeds<6c%JLGv_umtR=BHEuMfBD6_Ez6Y9=OGhd~r
zuLVXMn2R~u((>)ckMW;=OW^?RGw++YySbz>R;=K4l53h<-Y-KjP2{3fah(61aMb8p
z@@x&6)=!3KHa03cEjEhEYvv5}8v3R*d<IYB)CDN-0@}y0t5LH$)jmOPB1SK^`Y;$F
zXZ*1w!ID9K7*^0z=o702D07KKH~RYXF$+UvEEoA1s+d5)l{m#E5Vr~9qWSyxzu=2u
zET7i|WQKYo==dcd^@Cx0(<bsypgRWZK)o6wCU7qKgllb4<BNNt>3thhWDcGZFk1A5
zxbxJsLW=)mZB4*od-Y3~W<$n9N`DZUVV(k)2|ERcZaK;_Z{BS*=#bAA;RX)={A7<_
zf_xvCX0Y8Oq105y1A)QNE#mXd#Y}jmG2MIlO|$Wt+4y+Gq5w9^{`6@zgh0nNGy<#p
zpju~UTF9{g%@Spqla24wLQm6o$x}>=JE`k{Y^iNg-G4*VaLs_fOaH~4ZP$N39UAJ~
zzVq-ZgHexwd5M2LoT?DJkE2}%;8f1L2h#<RU=2;pnNxHVws|~41|!&kY%mDWk%q5_
z8y>*1IQJ5gXj6bTjS@seqEEck0g@TGCi$-c={aCb1qWdVM=;P1@<$v-yQ~bz7StjI
z>$SDDN$--!q6J{K#Q6ZGC9%n~y(TlskPoF0sSoY@uDQxbD`+8ia__aTa{*98-rozN
zsRgqEQYW+zI-Cxd<AQ__%C`OF*hP#F{U=h1u-0G!FzSM0xfo}G%M=Mo8e`xbd75#h
z6W62flD!30YZk$3$T0>d#Da!`f+x@rTgI*g6{nVA`2{Mue_Bl%KK!ICzIW`w`$2Ah
zj3Jamg)vtPl3&7WATb4%%3w^6_t?fETbV)49qCRHhXvY2v&6@?$K|Uf-=E(lp|+xi
zOCaStDv0XPT5W$NdRNT5&mB753K9lg(=}Xliu}Vug91eBUoTx?cH_o9P)iIfUJ#{)
zA`Aw`rabfrUb9;O)ZfM3A$FgFWH2YsXJIMNvbCk;%kc2<s2d+Yel$UP3g!Bu@gr^n
zGDjY1FFWYe=v+@h<Sv*W@V{?Qru`QML~$~RM%)PHExOY}46rh2;0@EQHG$GzG{Ish
z^7bH0`!LV~5a8r{#}jecFBTB}7u(6v^$-^t4|S`;?@r(Y?al*!(s5;h$6W`9<^tC>
z1buLb-EzTeX)tiU3e@!-Fk+<Cp&td`l|B~6P)e@yiDZJ+lvY+z1~OunV3A)+6&mHD
z55f$6w>keZad^Y>L8kVB9->?g(h<Px!eCHv7ubhIixy?)oyS;5U=E(<<0$_<rtcDs
zGRI8~={4F$uhA+aw$;UTD;7R~_hYYo>dDBcW&TH8H5I(wkG)a;w`TgnQmug;hT0^<
zHMl%Xk4_L(i2O4q5?eYtI#yqiwGsXEbty7SP#B!O`wVn+?&0@xN=w=BSV-XA==1{g
zp9|23p>AN{jBpDKXIfCa;uM<6*t>tfCK!DB{wO_|bkKD%LFJ3N%T5m#?=LXr1C!xG
zVVoT<r@T2XMk$^N2Tbvmixks32eDNj@F@6o9Y1zVK=1`1Y06s1){HUSLG)eBcFJyn
zuTgJa;LR=`M1Ier6)OrlVk7X}5vNPoRSNocQhk!|6><ojlG8C)kY-Jo8JyEzok1v(
zl+Hyt3Hn;{kyZ4@BHgXX+JwP?=ZInkNEwLfPy;Iwp&g+1K?5b50gSdb#bhI{;$2dx
z(weYXx)F?LqwOC<_;4cSfiy}i+E!-C$IJT`#i0)fBnnGlU4Nvc#4I~Z3P36@8rm>u
zSxr&V^Mfiyz$>0UeOgs*`}_N1lw4*8Xf|;e6Ill-BDSraYA4ZtBP56h@O=Az8KSN{
z^*cR1JZO)|p?s?*05STVL7HueI@PIg&R;P!jD*YGmy+iKH2tor-aII02Drb?IiASb
zn#ZCGIYhGYlFVR0M+eL~GAUC$FdIq*?=B=p<TvZy0F4b+a0UZ8@kcSe+THKOY<pUF
zz0WrZ0H%<+L>gWcWP5pbo}9IszNqfNR;-)>iOR?Zv(uw{#!$p&<jJ4cg^(D4BdV%C
zzhzmmmF&Q?TbQl(7s|=VWV_TiG&~csnuK=p7?`)niT;!jygqI}f9ft}OSQgNxe^`i
z$~%MePaQg&$NcQU_%5%EMz1B3%I`ietoFN8CAG7PSAr&kMHM12n0$#&ZiLT!HZ3NC
zS`si8gO(mH!kmj+a=soSid+DH0#JEIiwPSV8VZ0ghs1-6m-qgF<7*w)i>{u$6K6W%
zW#H{0BFAJ~rsDPjoSxTt1a{lzDX8`&hQvMf^d#RWEiTZwlsD0zpzq?rx_XyGVzi`!
z8076+BTH6Au#0#d^gimI64r20n>aX>kxBSf^$yPF+p=Y@j*gCUYZXr4LqL>p3WO}T
zB^Qw5oS8ESnYW$1RqQY3?;$+~RqE~g+{XVniv}p{oDMBCS=vj|RmwJbX2JoJc*=%~
z7APzIP7s%<TqY+i^*UZ~$rc;~J>foF!5td0(eu5x=x^1{@_HTit%<72%0+9nwk&qM
z&&>YIH9RcruYCnFGgmL4Cw9t4wUfW|EN%{Wr`cytS$QQTup88%j-badyPAsHsqvtk
z`{s;hT+<WM(qF5diXyuPY#F0)=mG>7QwA5V8fIuff9HI=dF18UAFh&3mR(il);Hu7
z{@iqmM0QZy(3AJ6(blJ`j<-D9MmPEG&i2^2q?N}y)XpUHs!9p|-N#~ZfO~{Jub|3y
z&9rLkFE~WV5(h^cV&<}`TGkt7eL>365mZA;ut472%R26!X-ksbe_N8GF1coU-;vjq
zkRNZY%gdXMIUeT>1?8Z3pm*$NlhN?xCu_|r?n~E8?ksc8CvM;qmN(2XxDX#K{w-L>
zus}UbjMYXm45qjMO`E0&P!C9sj7r{F5_M8luw+4rT_vUWK(4EyrDYGaO#Ro_cdPaq
zZ!on6Y5QMnTjwz3z{EeGEnPJKZhGm<z)XV*F;5^H!TV^!;hbCb(@36Yus1^Xz^1vn
zhGmK)g8xWXPxvi%&ZoRZ5`=@CXoVVtB1DlCm*4O>+F9r6xo;OLw|(SJJ1HflG*2z&
zy_np<*LO>y+WvymYbv#oEL~s_Xo+58;V&n7cg5<PR?zOJIL8QUh28J5s4Q-Kos`sO
zwwAY~x6<r1Aa$Ts3zGgfWrX>MnL$c>0NlmbW_a?c(n0q-X`}l)^tMHvD(oHg#l<c+
zCspby+r<!kSS@`+u6PXTEW$DZ?(sNLESLWA(CfHUC=K(+nQFst_lRq=S+!=3X-Cj6
zPi_^R+Wq_X3HUEpdCD7}zV=v^P{crAZ?C$J_ciZLSM>DW{nohbG|v?=oBZ+HS$CYX
zyZ?q~cMXfDQV1ZLXHUO`Mg4lcz#}#sjg=<ADdmAkzVfL^wubQ0eTO{?1OMD{d4GiI
z<=L$`ExZM)n9&utJ-e2A6>7<)=c>wT@q~4b?FmlX#D2Xh`SokfMzNTMT44Bhn4U6O
z?TD<Y{sVd61cNg%ET=CLJ!tOz=hI8S!Q`)L-*3ID%WBFV9r&~GS)Eyw`|#Aba-u&-
zA#yQ0`+ZXNfre2Y#ruvenfH+ksE!yN^bZxjW94Ir)2WH$E%9+xS+Z6ICWj4Z8XOm{
zFEEJBDB!-pwr>9={(VK86;8c4+~niF%R8#TmcK@4LE=kE?#quqTVIZuk(s>wJkPTE
zw5m!AA^#}=fN04OHZl-43GQC{*X4JQHvvewn=?54OzMcuBo^Nl_o<AwSA>Wnr>T0M
zugK=jj*`76xtHKLXLqqTQf1|n{8NPLv~TBRK+YZ>C`IfIW}Mf#bSdx6LYs@fzQZW>
zc2qFx*OIv^;A=_^jDlL~S`S>cx%iV@3`jH*Hn%prIH=dVdKPcYg_wDzJ}sjY;1V)w
z)L{haxU{3b-Q<+FudnGyOGu{4Xx-h15?98ZmMLf|Y^WUXN>hGWP;k>W`=Xd?KK@@-
zckvE?Tl=QC)!)T*We*;co}LG1LHh#S47=UZ>nw3<&0Y{S5c=oH5QIGYhx|ot*frz4
z&cBuL2PD7+F!#%sFRON4nrp!0rz5OpHk=0wA*^lpFuWL4LAPFf#P9jgDyY%6ojO1E
zuM=s&AhIIri{hU^C3bvyPg)i`hF+*?AhLow21TO~^~f2-#)7qb7s0E8cHCZO%h8i3
z@0SWwK;olEGkQflRYnweJzgawd~Fb`Aw2Wrw$EuA*{=du@n6~Ea_R@$j}d~XzUGgc
zuK8xmp`LW`ikYqH>s!D0o-BW7(&i|&w|~~CaKNsy(S$E&ZiLimB}^eN!TaVdSa1|A
zC~BHn_xW%+0<GYdi|`BAzj9>(1_p}L<#CT21W&OMdJ}9yYPIgXZnhiw=01r@Fs#}a
z_B+XMSVRa3sU@j~m;wJ(3tIikOyFMKH4~HD$b7nfJ)wGk2K2euz3I@n`XZEXFp6XE
zvSam;po)*IlhKc{5Cw6vgmudEEt~(Pr)QyD1C!{#<dUBJf5;^fI>zuOkfUiLHA!-=
z-m|uQ+Y3V03DhQRWV*pRyy*4HbpX`dpI&dDN919ef6)_fukV>%0e%;^WpsrZJO5FV
zVV>MN4L{9hU=`2|P^Ac>fTre=q<KV@J^|Qx<H%%&EFmv$4<5-2LSIt#t9TI?sE$6r
z*yWhF^gX0t%wmA7>EzZaSx`~*PBKy16_~PLr+CMG&ofahE4ClP7zGS6mwPvGBkwd;
zp?h_x;MjAzl}t&{A%%z_MUA{MhYlS|OjblcO~E}dg-VY>Q$V`C1%sdeLjf)M`oAe4
zhyU1<8iN2A;J;fycOmg;Ms~l-wXQYGYZ#sre-Z${mMqtPeZqY(hoh@Lb7qtI5#7(7
zp_8XtZgBjxpE%qVn5Xnaef7CaZsDNkT+>ik$s<QcX6T+|F-G8Y%EmDWPc&$5^3eC8
zD{lsdO6#(+*S11O#-LdaoU8Y>|1ZN(`uf?yZdgKulh_p$bPg0;5INn(NR39DvYr`U
zklJEnUF04$(YPB>zXE>71GF398%(1tn3TYLfse1bd_zaoB`}F3zZj0cK#SuDN+QiE
zbdUSLS^&_I=v`+rGGQ3hcUy;}m9~#Y0o;dK^*dIl7MGL&2*^tabUVtGgQ3$HC=P^p
z0T6|LZ%7qH=%WlGNpVXVhyMR)B|MfjnSXwlg`R?CCI%u5pWWZyDOTTTXk6dilpL#&
zVYTO)VtwOhRg+%FxQDHa>~szMGS=LB^*rdGt6dJ2?IYMkLD{c|-Uba4!ler#vL|xp
zXMM6;Vgig_vtX3Gi|S6Rx1uhLWkJ41vJ3D``Os*RJ_M%flv14@eL@9Lv4A8{zB%pN
zLY0AIogQqGCs1JO)S!!6PO>!0UB`M1^$|=g>Fb5%J+;DE|4#?g98mX6Zf^e<HFK$Y
zD7qBzg~`ua0a|AYLi{LNR>HL)_CtoZzyMHd;(u`pH>1qbswNEVv}jMk%DVKwF5!dO
z9qbeDm;aCHD77?mE6>h*;AgZjjULL6Rx%s+KgAd>Ys~5^`yBf4_|8in7t^mP>5daY
z<9PIK$rd^7xT4F0e4IO<V>|CmjO}2K1Tp=D4iwI$8K5@2s)Hvn^Rj9IV*NEPU2*_-
zt&ZzzR>4zqYfb$=R7BKhSH75#?Ks_(&et|IT{LjLD9(-<%!=f+jO>T98_Pe}45bcE
zwy(Esu8nm}C<>iu7rXqUTfu0&d*s$Otl^o>Anl`tx~{1)$a;z`1a^>rpeOxYkg!XK
z2(NgDH+#02_a~XpV~+Ayn_k>49a$5ixB2>(RDg{r)IN7Yo_>gwOh-{5p*QGUrEfR8
z8vwhBL#y4%FP_WZx-Zt*v3;?)&-0Ca^Y$+B|L{SB!ALTytNg(EQ2#^LlHhkgmYdA}
z<t4MF?3I5l`%o^2-+;%pL3drLMQX8qi>F2+@s#*iud*>+C<<;_K7S4xMC_6T3bKZO
zKYl;BFKBgo@H<)UrV6F2=libL6h-ir^t!P3@Ex4%m}L0eQ24FU_bHRqTs=MtLtnm4
zApjOHa)@BG?89Ud1dSW^qO;;Z*X%zE)AUth?2#99k1gmh?EJg0&ppLPh5PK5bK(}g
z7Zg6-_t+-zdq5*<*JmfyfZ6zwZXY_~q&od+*=HCS7y$3#M3yKs+uB2zJD}&a1EIDG
zXAA2(4o(M=^T6>b)`xxnf{Hw%D7M-QuhlR-0M}ez*rTXjFQvo#Yj=(Lu1eBVwPmsY
z$7eK*<#+$XuPmdV{av6CJ{FhTAFo844V-F`rT3CsuX99-3-`xAatImyk3Z>Qh?`FD
zW1b)w%A=uy5VSA!Xg<f76uoR%?UQT9v1SVWrGK0_NA6T)q}EuG*RK%h(3HVd9CNdN
z!&3+spmD6)@6w%r7zaCd*o;HbIvzVF_QDLk^IWs<4=Y9Omx}M;o;t*`r)Pxa$rRv8
z{|J60B`tjvqejLMC8<xDA>a~Vb2@sqH3=~CmYhgt0OmdpYKxBt41W<+_yo#Wsb1dw
z4I|*5OJc@U{hyZCPW1?1|ImFeKk$6;N|%b6q-dmy=9S{HDBkl1>oz+Mj-*e0(ezhd
z%p4!h`SH@ze0``K@8Ya>=ICEfIl8}>hc0An7<p4d8_pODtTryH78Xt)*}}U$@6qt=
z#yFupV18U13+GMRTi0_bC{+%2MRo18zZsXLU3>4^E8WhDpVN2z?D}a{Nsy`k32t#j
zU-H|GT&IX1&A8#z!cZy18dk?%bc29y?p?E+TDpkm0d??zXg_G4OY5RpSl3-p9q8%m
zas?Q}DLV1V;pvFw&Uu_Aj!oG!%wJmFY%F|NRCJxk{IiG%zp(i(MzgxLNX5Bx$W;kC
zo|~VzDOQw2M;k=vtdZZFUDf{N=PGyh+^|SwF&ItU`nk4H4wj_{>1W**PXU|{+h7tu
z94PbM8Y(JSnOTSgCS3pJS<rqnm&eB}n}5tN*{Z+o6s=rFOK8@3!%*n@6|<@=FqtZR
zx=Ed#8MqS_*XhaCKO^;ZOV;l-kv|>BM-w_W<!=@gd!I9*qbqrZi94ji(&h2=BcP51
zO$hL!Jdp(pum(q_GYX$Nw5zP^co-x@q3z0g1??<$jZ)$f5Gy@|YZY-uEvUvR#$RTV
zQVxv-+wr9pyFV;p4VINk@yV}yo^_dg3rgB|-JE5gomUS3+LUOar!?(C{FM&VoS!y1
zVgTa7#5)yA;l=+YH9z=l_;sEMyystj{P@$*vSGjxf9vn}IFpbfgViJt{rwk$OD2+6
zNht?R0OSOe5W&7*T3p{Xv&CQoRIz()Cq(#={{fL&V<8Z%z=VV+A0s;q)gfbNa?{)X
z>%MkWm!$h<gQ`Ebx@S};ElL{TSULTwQG_-Y6`wtuFgOlGjufc?hShHU(f-+xx>P@L
z@Q4Efs_$Ys?Aw9@CaCo(=o;X%k{R9=evP-YMa+8{;5y(mryZ++b^^|Zds!kZ78qaD
zYYe4_r?bgnE;^1)-AJ>Jf71@B+be7{1ZgA=G6c;;El&S!r4Wk&YT#hlp+iHY`Q?&_
zqkt%}hL9OPTpGAR#3{E{E~i5uNV{=o^^WL~&8MH)p*g|yKk(evCQg2jPE2d{KN3r<
z7GMKYB05>tXj6F1B-(~C8Rz=?`okAoGnU8MaY`Js-&8%DMWBcE>1FA-&)@Uc1T9&*
zTu9cXO}{cPFMWP2d%4QSr+JE81-1DzUxY2&%3Coyv%c15a;~7pC+(Y0`9zz_9Ltz*
zmK9s*45s`PHEr7$n|=q+$~DnFX0W%>#tB?9V1c>7Y_AVSb|%g)?3=l<z+a&KjzLqw
zZ=+(-d52GZ?u0mxOH`BtNkOE_LYIakfVxVU1j*ElUyokL%jqgHuW`jwltrAbFVOWt
zcf-J@NEVEa1j0*Hwf@4+03a-({)nKwe8mb)hJAg4huD*|dM^vIQX6h?7hZH`-QG0x
z>Qf8XR87iTCH?l~_Ui5H1_2q;hA$#cu<Mf&j!F4}M)u=nPyCetQMD!~#^Ec!eKtG~
zM=a!PL9%Z0@E-vI1cHQZ+3X~5gM@y&O3EVitQz${BI=7Omt;bShin8es0Gzl1ATot
z2=M;`Iu5I2sTKD^<N`*7PhI0<V;_)tGdP6t{c^EB8!Iben6Pck2bw{OLdxaEq7Q9;
zHmp{lnD7HPPj@x|_Fe`{2yH(c8mn1(!KTo~M>$#LfiJ<5(%!q~kQD{m8(N`G8`sft
z?KZ#KahBHu*moh+Ab66p+xrEK8buIJE2&$b7Q3mY@2%O`YE!<HK||@uf}ie-D%0E5
zvTm(I5km7vIzr$#FiOCG%7PDv#2U(ahUzc$8tYf@<G4*v5V!y{1ULz1SZSH4(87R2
z&oCTE64*x5ZsUro2K~^7Q7x#L7`XA(wZar|GLR!)<96tSaKue+ZO`vEw>~DU7n+F9
zR<he#63{+t>GXTMV=IDvNMt257>9dBTe|?W<DCOCdidC}ITWY|D&_HKCI;q)=9ZQ_
z*cC*IQ#?zA&66^6_*L6FL(iQT?_r(>xmZ8|_Fov**SMIWgT%_1*-&DmrKY?$T8EOC
z*Nn?WDIAa;QFtvjgFedG*mrSqWF;nu1B`5RswPmC-?&o%Hj0)45Zi}g7IQmnM1^FN
zgG!paChC>i#kHtnGyTg7p`?wf0<u3J`sJTP>s5c}hje<)Fc!Va7{<#_p-&|3i~R;e
z=9lJXJ9Pg|>XT;2kadGM2K_tvlK`w!A_@UA2M*LX8~m-EzDjVyW#3AbLVZ)$$u!g(
zBw$1j<XZ2PIJHZL&(rkeuW&iu!MDscUtfy{s@K=}=|xMKnF-ZD@0@9ws8t~}D~%7a
zDM{_u8jAPnm$t^(TwP<;0N@!aTycCIUXieJ$Ft!M_#b8tp!bADczC=ton+m2RjbVK
zNP%3}V8Mf}G43qh{lRj)`&`58Qg)w@tYWt@u8&QtDv-0$_ph%`m-!(4+Put(!Wxmq
zOV(8Id$omNQC*1lCYTCi3#|mTAB{M$Et#*a1a@9c*udD!z99rHQFLyiw%h#suKKZd
zgUfBETR-KG5SG@ay&-CLJ#4X#C%E5|9}u1q28^~OV2i_60rJ*dNV)>=Z~-cB9766~
zxpiQ}Au$k$WkOroZ6uBz%W0g5t<VNm4A3tZg&+=T8Lr&%e6*Jb4>yBhExKN}?D<`g
zBmj+blD_o+86L`-bfri*Yy>sXxu)uGGq@TT34bv&x;|LF@6G}425A8^Oos>sdEK}d
zCLCz59X1|+a41~v_0qFMCL&fEDNx_LnjWSVF31jRRRbIhT{(S#8&;spX@MRAuUMBZ
z+lQSHOh_%lRt0U{5lbvkfPg3mQ$*0s3o)&CJBJ0^R8|Qp_?nK@tN@t}RIcNo?zCfT
z3GsFi8)g{3FDjavvEgEK%*x7Y9Aw}>3wd(D#l4?silQ|yR^_=c0b>aQ(<r4$jgI0S
z2ay$#Y=a8uTY3HFO~=sC(8fV@HMDMCzwsPGWHk9h$a@z2HCFNh6qR72;~NK&>~%tV
z%wUEY41VdqE2Nl4@dvb*-UJix`10kZH?b8JXQNrfDCuka_Sp=8xkSf7u0TWw;NgV9
zSeY%mOS{k#5XF%pAmDuEMF`uQ%ugdOb?;8{-Q-(SuXbK~<nbe*^9$Ec*qpzP9YwA_
zbGLX-s$T7n{$Zrm`wBzDwl$r~hF?L%GxQtJ;8N%qXQ3+SOx`oum&|Ij1H~D2&UwZx
zrd4C3PceXl0An=&=BQBFl2b1ZshBb3KG-(Xg5K$)`Gdk%4Lm0}shLH`f1g8wZff~B
za>AE^(@=yFc(m{_1qTJRvBpQoe+`#aEHj<&-MvkuwZOJ9N?>}@dhs4d1W2ummqip*
zRZDIyrhp=d&~N+f^us8Vjh&bYqOmNqXJT6ag3DivP!ITC)RA_AU1Z?APqeaSfE~Y_
z51V5jA~p0yboBM_gM9n~OpQsGyNd4{gMlHl>d#9r?-_cRS3rev1lxg7HKWa<u&5J}
z<=bfxU%WW5{LJUQG@s);BMO9eF&&ByR!~<DXk4`*yndz2-foyl@jG9zIY!(4jZ1U0
zIuMdo9;>!)+ja!AJ@AN^Rbrm4Uw5gwthOz9#uB5IxUW!*Gfa{LfT!@63+rKjDNXry
z$2(dw#p>Xeuy=#`#Uq>Q=;C^9Q^h%7Wnq>NZW!$&ZQ|zT<z1l0`QbSBl)xd%q#vVD
zov-pT{`G5p<Mli8m9`%Wc~IE-REJ@A(qRVU@m0TjO$o^tHqeNQ6MrJ)Z1``U1GXgO
z!STbu?iMUSd4x!MjH8zmPn|e%2r@^g8_h*HXgL?$h*(>of!sAOQJeh`Er*-(r~~-D
zeyjaE7IqQ}UQ~24x?SJY*!aURnIoZd0a^+&eikLz10F{VljTsZ`U?JtScz_&3E<VX
zKJoC-lNJwO!z^g94FOZ9dJ%`3TU;Dfev_xVi$!AXp8oNova&M<w4@9b=ALXEf4mJ6
z789Ne?(TQjxSYP~z=X@{St<qkrVnhzWlUv3gRhn@DEf8{5oz{<&sT#V1#(hVFA*s7
zDqicZcwqHi|1_^%Gs99(yxp|ITe?6k3T|K~-9Cl$=i>$a{#c6wEd`^ryC_nG1%=9Y
zXX~}A2J82@722QR-c{=;&C|O3V{&xhY$Fz0;l%i!k*|(!X&5G6dYZ3ee8{mOOD|Tc
z<v{>Ua+c_J@9<W$*e6NHOuXwNl|l-|qC*ER8eKh?S@*eE_Ma!+5a~47h5XT+_Re7N
zpa&0-+&xHBnaRbEW}HC>9V|ry-RF4Rb0EdSJnV=87&}OvAfUk`owZ`cb8J(?)HVk~
z)f>Jej~5oV+iDrih6;dd!v+S<bp`{UaUCjhAto&|z7hGA?lLj45`6B#kr5FUR-S=4
z#b^;1{2dz*ErdW}+&2K0ZP%S&$E9%WF=a`Mjy81v{l56L)=8yG9e6+lSF3sk_G$hc
zM!bsg>d2CP6P<Nv<8Sm-PdxDQ^kjxDD4;FT3aa~oXXRO|KR%`PdO>wJrw#kN^w(%<
z{rdH*!eVitsAgR4chT#rzdsbh^-D>ahD#SM3g|b%HW>(@nsHr7vg`W_8!{b`8c+`t
z=bUxeyPACQ(TQM_2JC!@ka$<3`w0jo$my5;`<5<U3N1`f&S=}?U|}I4p)@y?uV%Df
zSB+IlkS*|0-O2swA?K*XMIpD?`k9Ze#3{+h#^qU4RReVhYU25r%0FY6M$j5QLN*aa
zY|=If$VJH4A2eSVKVA^_I(3E)F5l3mKd<I(<@~$it@@rmu{NwNz4zjk-6Z2&&lM+*
zHY-2$)iaXjyr#2;uTb^P!kG0*-h-_B(4t<5R@l>#;@jCE+?JBcw^ruOLcY3Iua*}t
z3!lUt8wXhQusXn?z|F~N9z`oa6~kaqJ6u^TB7SO@OZeijPDjb2U*;m@g8&PK3Z9KQ
z13|RNNoKpeqgtiOMJS{@I3_Pp6wM+1HZ%>`o<=_qv>tKsk61MxU|5uE?+2Fttknlp
zB2*ePF*2dw1M8HOX>zXBvZ2_8|A=5vs^-<Lx=RXK|M%-RndX(|n|ABO=r?Tt!5Io;
zQI&6}deM16)9;$2RK)A{`%ondj?UaxA)RBQvpC`236vR7@Q@V;_n-U=Fdaw$nR8#$
z`EVD=Xbigo(<q*eT`7u(baZ$qxDr26M@Gba1_*m|XgU}HsEf8Gnlb!ifgww5H^25g
z{_=;Rub0=~w_bIX2RRn#`=nv@;zo$x!~O(RLH=Ujr;>HcFN13w;*r%+W@Dgmlgv|b
z4wu%vOzP;+S4lU2G(Lpog$t^GpFDk9r_ck@AQwiXMKME@I7-G;-Vh!c%le?Ujst>8
z7~*%w-0ygd4h3e{Swiu`n8XZTx#8W6e|HqjkbV3I1GlZXAoj=VVGyLtT4}xL>+9w(
zIId;o<YqGv{2)A>6Tqp(i_pehpDrHj@DCMO4ul_@0~ya6p|N&&nB?;3%5%I+G#)Uy
z71HXG(m|NP!otERx&-|>#vg~3YUVSrBM6{4VJHA5?+R!YG-F6+XK}0Av+);BQ%It}
zUU%lj6FV3Cv=DZyyOLAOIQq!DVwX3_!UiB2J}?Gt4Gv5p{m5!vG?+xf0PM)+Xx|o_
z{<iOOieOuccbV-_Pg5b{KZKrJjSUNAPK+49@A|~Fskr?+GGL+yF5vJOwFwbbDR4%U
z*WRaFoZ7sfx~`>f3F?F~kAOYIeu1Jm!o36l5H^H(sn!}G|GWh9T+SU%62>gys-3~2
z#mqZ+9(}~xwjHw=h@w_&?-$`%i2MMMWV2~!3BCzUH0YJ{(CRq>y+^jLr=QY?jwe3t
zTz`CYj;YO=xbnqr>g1a%z~OGKsbhiHg|t~&W51jg9s_~;0w$GMncF_j)o?`tN=cVE
zz~NRHbLpa$uqJ|Kv9?gC4q~DfXv}a3rpcNJuI%-Gw<U2MK<kg9U~mP%!Xb0yc;}nb
zQ0mgC2TO?Q$yZ3Yz@fH)J|A3-4j2XZyszUbSbG!GC>$e;SFa8ik}}4B)F$Y;=iGg+
zk0Xir)4}e~k!E`o&{O7grMj#X^)IuS8b%*fwXL0W;D-W?v>CU?M`(E(g5c2+!YN2A
zc>Vp?gU6`4^i6)Q-l-?ehMdL4f;P!wg6nUjtQd;f2Hu;tjG~yloe2DxUsJP^^ur`4
z#lj+dS}kHJu?8@6a>9y{P5OJDt4<2%4H}EaJD;+!PbzG<H*S+t^W^6=BV)MKRLZie
ze%KR>ZBa5H_l(=*5)m002WW?KDOLA=%FUgH;Q(@f_Ihj`BGmw5f5WPeK1LyWZHV<~
z=(t!JS1Me#=?Lp}eSL{h1nL4rU(~Uz7nH1vh~RZCtU>Ywnv7x}A?Je%AQiWs3whOk
zhlRt9ww`I?^3irUH2-_oOWDNz&@49THy5FNFD$^K2)0pkAZe5sjJ!<b*9MI&gqH*6
zJRG0z1~1CW-kvgcC8-p$)X>AquM;KQKD^lhH%I`fI{!Mk5DH6{Et<(Nq3kxl$<0|d
zKg}`l?)9#^<@_?oox&$IugIDdth{JFgE0)oXpQw8QwV<PC8E&~(`T~74AV|4Rk4ZT
zjbJ=yRgMX8KxYza4)E{O{=Ds*k}U%m>YZP8UY$aNRoj%BcM02)G;;5mJeVPr{CLer
zM{s~hCM>p?v=+U%k%v!k9U`c58dv<ii3dW9Ah$PzGw-O3AYxY7GO8xixV%~5r6V*s
z<%M0Xty*%(m^%mDUt*wKtw0W%=-MVd1twG&-F4@bdH3w`w7G&KmJE$}ES->GGY>jg
z5;~wMqIIrNJI@dj+@FOpAM_P7`YU>yBO)TO&IoZwe?h;m_xoFS_XEHT7*t-^B4ZzE
zfcVoCo^HP>vSW#JzBwK@EKr(<r)>e^U`fl2??5Rjgwqg@i8mV}-`bvqh@Q-z94niA
z#bP`MiVd_R#0fzmMeTeGCREUW*az;H^Z0BcMF>L&z_3I7=C#I{%vzt8l$A(q0pLO>
zV`ojQBY3T=jp+z6{rmI_fsjdpo}g`ei(v;XKBb^ga4s5b)Z}WbXVcHZ&#A&?Ft+rC
zGn%8&IrHaPD3mW~T=AA$9Cx(S@Z1}nkT^i0NCG;Ij-SZhE5^@6INVgf2x9Ctx`1)p
zPb|^EQXsj>ShAGg(0lz6-Z-C+0#yoDlO>DK8T>9^p=GHsdf8bZQDVKJ01qswAbqN#
z;lOh`hGz(1gXVhLbpSI9#f?ENN--~(BnyDr3nc;O6Q(p9G7_TGrbV@r<3j-2LOc=@
z<KvHD?Jn&kL@l3%G@Ak{LV!qiZyAIwd<NJu$m0u0Zwqvn3vd!NmB?4%E%g_3D=yAD
z8IC9w4G=_>*MUq7q3-z&n6uFGFMv~>);mM?oQqd*Z&{-@f8yBZ&!1Ui!exCkYuT+D
za2ny5QG9b_hfMrnA<~3s>Fu69$xH*bl!*hWAI19M^8x52UGiz?!zSx?suThkg+;G!
zZ7Wj$o*(DwU}|*CGN|FhyUzj)NkY}oMokq+AeI&sK$=Z^z4Q;pf}6yA3SI+a#cI{#
z`uX>l4^xgJ@F(D6!Lo59BO@))2$5QWWoNqj??JkPkd?4hHr&_#$(&R-Ql`ZE&Ev?=
zEZyO;jJ>xDTJ%RbW<FtXIaK!fPmVFux?F&=0JlR1v~Pn)&@-Y82Ly&-e^oWx9R0*J
zCB6zwA6`B_(8{d!xBeyRdp|mL1OzFJyMfM)r+`wq`Kbuuq=;tEa!9xdiH<n#DkGrh
z{m$I-aQpJnu$Y*F)N`SU95=ha-xuAPS<7vrpdi;=<ZnAZxnS<xY?PkQ+@pq40AL&e
zoBoRxu_r9#E&m{{X8B`N?;mSx+<*GnYuQz3U3CzWSpkgh+MjwsNy!ODK!&>1#@%wL
zemK}?1<J5p-pKN=7T_<u1-tbxo;s`=V|!PTp~&G;%@sGznTV$t#9ZQz->~~!Rx+pZ
z0a^VklPW!STn-QKrPyF6W%te@J>hHq6K2eBK+K$cIa9xgD`{}l$m*EwCgs1v!?!|j
zj9|2T&<N{^gvxBG&7Xnd8{YQ})gIGGAF}0I(BkhOT8DQLt=L$<|Kx7P4PyY7kqY$j
z&ojXLXl8EvyqGX-Idb&iWS{q^H~!tHuKs(pb4KgoIsJ}@)V^Do7@^mIH<5h&S6<%H
zEz25el(-5I<Gy1Hm{52F$b?Ex7NW+`_7Wun5QCJ5i6<E6{O2)f{|faRtriH-!b|hy
za~3B^(zkewvE11iAB+6Z7?GenCms~((vVF=Gc|QoR#q;rbFM(+Kkio;$~mo~`nzA9
zvQaTMqHLI;hpc9E#M7;3lrgXZOFJ5Wj$31~a>|V^jN>9A&(5vu8-e9|kH)1g7VSc1
zK)P=s!FmPU)Ed>9T2Ti1a#7U^;TuT^iKzed2_YGW`8;!LYR((@12UwvI%sfaK_qT3
zuz*8k6(!dK&OogW=z9j=9<6mCKdzx}QMyXRz{B|?D{ZIa&AwVJU$W$F<NW$YJj*W-
z4TVyn5wzkdv_LZvI~(VGNI0}k@)MwK<093?t+IEL?>gmcUTMl)hx&XBvdn31W5=m5
zA8&6?m@xoy_5S`{X(@`xGZdI-qu5^g7H9k`R82WKxq#QNmxM`j*F>{4@KkPH%9+nI
zIP`6lBT4+wG;x}sihX)F3>-gPL0{Ea$4L*Q;gU_}7n}}luyc1`8kz)?*(p)h&k*0B
zNrMFJO2xQv>V83U`7rs{fzidg{@X<70j4lPk5^LG!12&@R|rt#S8+W;h`_ZeCzJt7
z9~=U9E&Ogal~^r6pstN)tVlc-;=N_LOsBcEg`D!Xqtu*FzZ<`<q><?L@nigH@Na`d
zDAI7r2wD0k)cOc>I0`!?uqVn@On>r!{!E2V`2w|yB=25EVjFMvZmTKV@zn1Hf2qD&
zKQr*}gm;5CcXG~{9O}9eUFdfpjUtvAudaMn-uos^`0}Lp!u2intZUY+)3rKAAFa=F
zfkQ{ma)#ZBpA0*l-_!hR78V?|*iH`F;+Z=P?KZiLffwZCl>_T}p+}g`Up#<p7cSKE
zxay)JP`17b-=11wyeJT(-IL>;<5pw6M<Gd0%urs6!&YjbD#Pj~S=iBa8I;5F)+JTQ
zr>d@BcYR)C`p+wVZY+1j2XEG#_WMvCZE8PwQC_LMx%N5p@9r~oyLph>p;;rUnAxKd
zSulP5>h6h6jxB}Mh%z}4;tT&3MFju#*IYF6D@l){9G&5l3!0Wz2V$+nZVPGjyU2rj
zSDz#&-6cEqnKpU87vkd*>4_DvxD_VH>(|riun6Xj=bYc%V&LzIUWYa(VJlG7u!snX
z2`s6D@`2VNyP}<##UN`2#W#bf!xy~8tMb5hqTe5zO|~Hr9Y&<OYPR#C3&qB0Lr%a@
z_l@90wR7xJ!3U9ULuL{{?Ud}q;l>OY7q3VlUWDd;5O!Pw>c3o8_YU^WfQ92g;*Gss
z<OLKwvgNgY*CGfJjuGJs7uaq9ZSx`9a%~5+M==r<)5f5rE;LmdSh5nm47R`o7gWZn
z9ByO8Ql=xt=cg%rRFo%#mLEG$jmV6Jw#hCz`Q|o>s0)C45hzh(=^q;tbJf$ydC*d&
zxz{Ht88N|E|AD<xC9y`Fk*Lgy|3T`Bp&yBy8-W<%RRZH!o||~?c>3_QtZKHP&p}ew
zU%fRrucp=U<DJ<~tu@IBuB;n~f{!Sh=O$FXyxX=3z*vPzJ7YOSau1*OQ@$RiA+*8a
z64#Q20r*3HwTkfyUXsB+QgR{v0ICOcgNTf*s;c_g4?Y)CKL(U%ix(}j^YU61ZG){k
z^xN>)lfwhXYf9JFOT+=Uo=j4>?v%3kbEcpRmRCak7A9xAV~Ysl55TttzIZVc#NnMo
z8^Ug0&}ylAiGl;Ky*MrpT72jz-2k^6d>&D0n)1!yuR(s9@~G&|?zO-HZ=q9RTIUJM
z6IyS?(=Y*%B6%DL$JmTPcAnGEYPk!Ieh#W=MU48)w8Mc$O%^K0k?kv5J<hA;V5Huc
zzuEoWE9*bk_8vGufqV!={a|607ISOH1_Y{LAEiCQbtniEmN6t&85mANk3v>1q8gDL
zcnh&?e&fdL_7wb{W+O#LfeP}eq}UOqfFFk>0T>yY5hy}P0LDSs2*^bjesmru#3N-(
z-1SYSD&7tWpp3x9=sSOk4=peSWeH>D*bEF1lzRxO4hlC018sAF2Nw@+AcSuhw_V=}
zNZsfSY)vU|-n0U6$GFL0K*fltPq%R!=;ipsx6W7rB{pU(>I@+WAgG!&b-?CMYGvWD
z#3iE$$gC_egcc^OkRL@x8!1%e<mC@<S%%>1nRpaFoXD<<_KlF^MY`vG8ny%k1W0H|
z>sSn3os#)dRoEbdjl!l*bnBF_LG6&<#}KHTo_vgZux1Y%{vOb1AL}M@P6r|y6*k2n
zH!1iC|A91sE2$eAy26*U5zY@BPwefffpiZG@`58*bYr>^K*T!bz5;N|gYHNGO`E->
z8f-a$b%=4oRl-mU-YFM&<MpKIdZUvk=M#8}Ckd-s=b%JU^>lM{lPD%T6K&=KR~CTc
zd*bwI9L$0wesCRMEqL``f;nL!U=^Y<@ysD16Nc5=cDw<U2!oOam-23@Q)c4Nh2&n`
z77=u?UhJrVE=N>OK|vfzb^QDfPl@s)ErFGRMF(~m6J^~0rfs0ms?P%U9j8q7V<ql^
zZt1hzsDO~?vnJWL``zpXVofO=0NjRavL^1*W?Sm&!PAj@kdgggG%h+PvO;QWZ^ElU
zw46FjY1r{tQ9A^{+ZY;5{9qDVEBN&vI|6VfQzwIg{Wtpvr-m0-<L`%@0|ma=Z3fL1
zg{p(nz(>N=myh0DD9P|dx(hfKQmi4DaHE6^L%)cJ<d_gu10lCpjM^9XqB{=H2KwBw
zvHmieCU@5nKZ)JdK?#d$TK9vpK8-|IO!RX>|IGF;I}VW(iCO63#^pEs-M8k=!5XjY
zfBr))(-L3QYbXbC#}+o)PTC;NzUGY~Mp0;cv9#v_To`u>TRGwT#7vAcq9Seh!g;8v
zV1zVoFI_j7NLiUQXi;gyaUcQz<<sp-cknWdfd{>-xE<Nd*}#}+UeSaN9-ZnM^{**k
zD)umfM%vTPsGfc<;shAc;2__l6SYl%3t@;{fQUFnr5`(4w_v*mL4YKOCr*Qd0cy0C
z+Zu4RI-)|7Gmdf{Mijn`d+~M1fNKkZVt;Bvk-7Y8*-8Am6<btvGz>IB$r<8-hd^L5
z><U{0AuVRo0vM;Y^VPn+d(lve8;t^FS+;`GGmx0D7%v7gfILf+;E;uYg?8E^>Rer0
zdk%nad;i-IoqT%3j1w}kD4<GSRn-+9fY*~~BI2uc`_s@+%gD>mVIcchP*EHrADS-#
z$lyXVX^EF~7LIkrLGW<GuBOHjA74_^DaR(Ltwo_LNmy8zsnc38R_HvP1_oqF#)T-Y
zhcTiVX;S<`RkoqY2zS#RL9Y>QGC}O#!KWS``D)g?Z%CV1EY#zplDEJ#b3%*>FUfQM
zcv(<~LL57{Y{|xyp@DZF`$Y6ahG^`-+YWBk0l`Dg1rU?&02scE*W>=FV)hnjsTp8%
zF4@+6i!T!e4fGaX+uA%rHUAfNeOAjCq<|*WGK>I^6Q5E|H~Tisv4fLH>syfrx_Y0+
z**o=@SV>=d1ZEV!RGVrVT9Cx*<Vo?;sHgFg`*|i{lx?DHBG>~Hm!zAyd6czpgs%<O
zX~=CH4)xa56e0D5Jhbqcu<nn6lZL{3u+&ezOb-tN3CeEBW)yH_#nqd-N_wLBgQb3S
zs(Yp#-uLPlRcD}cfmo*5nD-emphOYis1)}8`s0Ta<!B<~|3TxeDt#e(eTWL=zK@g`
zlH*WjD?QWZ=6mTp#C|3u_skNS_mG6f(0tSQgf*ZuR|%k)vD%qJE_2Qk$ji^qZ_fAZ
z1GyxQ{9GcH`S;}}96R;PL+Z{DDl-6VhW<%|)hBz=YP`0ADdyVA-d$ZYvXM2TerFE`
z2cQ{;O4?5$X-X$fu#!6|mD?O7XDxRi2sy824=7!b;tB5rC{{=ZYKK;;;`6~IPLa4U
zSD~S@f0qI@3dNHl<Yde@ud%K@3Dt%W{0x5l1&j}9X6@%S0%QbBy;^Ln*FMK4nY&G#
zi2<$MYU$oxl7=xzdZ=giFWbYL;|r1$Cdi<XkJ>n2%t22Jagb@}Y&fWS1eC2kM-)tP
z@odKF9?p`xCwD*8y?0jW6E*%PoHay;1qC;)f~sORWBY|9MBE(czx(W}Jhwx_`9$42
zCOO>OJ0dmS{}e>&!%=a`U3vlX0$tV}Sb?&uEOdBahyFgx-<SPN@Pe|tcuNNF@?yDB
zoW$zt83!XTs;S<==f1uymBo)^q-x@D6W4)T*19NdIE*5M<1bvU1#SOuq1Z&A$A{Gr
z`Hfm>xixupr4|HB0S&<TLYr-(vuZG+A)D|_hEpU3EyCpis8U&Qj+tM5+J6}K=r~)7
z$-pIm&W=1W2y1#_#(~if4~;RVjX{B!)_YP-VUpseUH+Y-lB39PM=B+2>62q6!-q<|
zxV~CQ!1*qi`y_OdcQX#B&1hwzEu*!mpB}S-a=Bu04;c=6hPlNO{JRg_!MrSAydzA-
zNI?7XVKd#}>Vo3en)(IRLbQ?_rhSMD*aaJYZ}RT+&B)y}L-ytcoq}cYhDTbp!ABj&
z-WjaSq(ccdY0CIpw@%eaI_-fOAWsIhn7_L}uL(<aZ30>k6_UjImHo;;($-zrvh4NE
zgx!O~uAiQ#j(?JoPU9Ec2$fpGbI?Y)IPE$<Cqc5!xc<}|PK4N-Z|eR}%KY8co&Sqv
zH}uDo1eqKT_Fi+5paR#x%r!78+EnXi+x<VJy?I>E``Z8e%{*mXD-@C;%MhV5C1j?I
zg~}Y!EJUIVQAjdJNisGFQK3Rau?#6q3Pp%Al$0UqJg>O-eV_ZB-}&e4$NppQ^=Q@i
z`}w?w>v~O>5)q*0zW11w$JCUNHI7cpm%sed7R3+Y(%YWxY|NslaNBxMkH=-FK6^>+
zejlb5SNj<N$G+G5@w{_eq8jW|7FUN{nG-l@?#PP=l&}u}1)+*DQj|-pR>eFY<%oU>
zwi_MIi=v{i)iHgjRn?pw*61Hey7C&8r>D=Ak9C{KZj#LEgoK28`<D_Q-*@1>sIag`
zQiQvd9((S6Bs#-j*zCD0EGOP7b<vM$cGhyc<Tnb-5j)D`Y*r@c&()=imul+UOmNe#
zeDT?JKx>StOFX;#+5mJO=>w`JNz{FpzZxgyyO_YKYej-lHsh-Bts%P8cQiN5QRDpQ
zK}mGHO%UU(6cXWd)xJm1>ez{AuJuVbsuda*w!N&fr=Qsp<&10>C~)Z<2;WiYUTTJ0
zy#t+O`jrhc-S95Uo)<p|0#%*~c{FJ8&Zo{xpSm|><B=}SR}68f*%ycURCq2-+j6DB
zg8N#@P2ObApHt;eHnsnaYt-@Vb~)Qc2p4eY@NeS%wx)go)dIs)Ke@xoQM>u54)FXD
zgMcD-eecRMDVypxUfJ6)z|SxFBhzy>U7fOE>#SaHo*uZ7-tx`AM6j;t7qr062<OP*
ze<x}~D*1d&<Ub)5+=uV0EuaL6I{_MZo#>0~D}PJtxh$62k4{fkpX5efa@+mq|87uc
zjQW6!W9rQ$j~0&N7DiT|oU<Cw-7D=UHJl40zVdx`YKYs4D^sQfpc{&9<{1<e^l<rb
z;~P_a%3^o4pZ@32O4zK@q|;{oc`ChTDFo>y=Iqn8%HAzowMsoXu{8-LX5sd~N4u8(
z+w<I`TDF^at@7q>Bkf-R>m>_|$`Gd1s<)@-eY*f8Oo1u=8@G2l<d)e|Rm~AaXTN^^
zL<4f?^>P^-MD=C9xa;ME|C@R?3LKpAw(4l6Q{U38+xrTazAsg~D3#j+E3JMF7h7tV
z55DryBx?(QmjU+N;WB`M6htMVkblRlzF)G1>YXs)tjfg0wJl|zO|?Cxp7ra>`@K<p
zB-aiN)fA?S629#-q$9!=@HxhJh!X60tL2-c#?}-PEy&u`NNI}#`6(k{IgSyS-KOhs
zs{Z8`I5S!OcXWj%;Q|pYdvcWR<La-i5MuP|2M5M(@7wnmRqVT~i}xIIUQu3T8kjOK
z-lYWqIlz_A_P9_T&F9=w;+iKpmSG~Pmpf+f`tgCPx3)qqUqNXU`Sg-~U5AAn`o0)k
z0PN3p=;;~Z@)3?UoZH7TrciZn@B8=1&9Rpx4J;l9Bj#DOUg0-s_Uyx%D|+nzcfQsV
zm@_^KvROjbiIHo+;?4aLT+PC2Q86Y0De8BaykyRt^z=;Qi4$w@XgA=usS@8Zrb@Rx
z6}J(9!yt*Wi))tUu$x*Q(p6fe<yYEqaEf|-f49z?TgJaKTllY|+C@!Zj;`BAb>e0I
zEr6wXdOp?9_M7Y4P&xxZ>ys+JyPF%D<BxAHEij%T?qahg0o*?ErAcav4TSc*eHH|W
zETh;&*8JcOMdxiU3exjm)!W~GX;YtVexW1ZzJGtFeD;UB(U(6n-2owz)YQUb!4)`^
z6uqE~mFsU~o_nRSZPyB_?Th^vs8zbQtyxv?NGE_-Wt?NQO&_8PYRfFOB}JKC8F?V=
z?z1>QtvNTSI3g}h#tdDLv5n$<_DNo&-+>QSD?NRGJ7h2dD@SyRz*;~~;QhHqeXX7@
zdzT9|2`jXsdiK}xIYHu81`f76##O7-v}scb+C&}AIGLq^1$&~jR=<4dH2-UdcG*|!
z|J`!Dt=D6RXrIa5H!tgWBDE&U_4CtK4eCh4F{Qu7N_sWyYg?v;jso#@XDm9XL2U{r
zmS~1~$oBC+GbUvd?9mi0qPPZ!w3?#1R}dw|21aZ|=eJPJeB^TIP;m6#ohSrR=P$fx
z*2aS^Mf3qex%woBtn96@kp<O;YFwoM0ogqL#lGID*;)6bd@yP4o1BOR6n<wj72!qS
zZe6>=St>VZ(j?vDTtAjHaD>c{s5Cb<G0A7Za9VC3o1!D1zpulJY11Vbk}T3bSFcI?
z!GF<I!}WV$eq@|-cG(n~3O?ow;G=Kf3VfA1Co5-NJ5+Z6<a*1B@j2=gczC(zr=1vo
zP<ufwiLxO#1t2Np!BevlU%Ur7H(KEB*U_%D<vx`G@n4#x`cAWYbx0$>yZMkh5ICmX
zV>IJqC<qEO9LWh7S_Gd7+%o_L85LR+P|5f9<v@PYGKuxFLx;1`S{IMrllEQ0eFUW9
zpUaqro5bdkqku*ui4QA_n8C;4w|%Duc_tBTT>5<NCo#CtEZ<}YcnID2+mxkH2=d4k
z^`_lDMi>Hq)>WVYvpc@?9^^cClpwMQo`j+YI-~KWP`+?c=RfP#ENW>B^4KGp3J>%s
zS_T;^s>qk=Yibf@P%{ut$PpH}a8^{T95ojzL&)pCxjN6y3#8&EorG-c+)I6EX@#i)
zcvMn|lb)Yw4YkVkKZ37|#Z>Ysk?qP7-mtsC^N)#Pd?gztyKTIYkFDXIJiMJp+G>U@
ztwA|Wu`IV_-)=_HDjcSW;t@-z+DoePu3gPJRO2#Ub^&VoLd0R`Aw$#%W7|k4{%xHM
z+v_X2pPakslv59_TLE`KFR^RyUJsV_h5jui-x7E@k@_1K$!K_OiS5uEu+}g2k<0pT
zFX|JBA56YJa*xJ<bz2Eikx$Rhh-U`;b5L?$my-F44RLX6J!XzP_W`_8rm>`=-+LOr
zX0OlqfBv#bpXTtsc=3G!`6hx?!4%_1uB$-bxRBMyJTY>a18%h8-q`?Ip+aqjj(1B$
zgceL{O6J#$CFlSKGQe9@8)#&xs>O^u*4*4b>6i!UIz$>7f|e}fa^NK{Ai01vsq0B7
z90$y<)H@~1ow+leTRA`5G`t<rQ;o^8GGuQC#2qs>&*IZcICCxCOT%>EjEoHO?NeEq
z?a)5*SIJ$?ZM&O)UiU3!{Vxw`2#GLbXFmRMp`@3{5*RELQvMAL{K=0!iYc>ZHDF*j
zH`NPPjaEAX8}$RY5;H#9bpdnZ_Cu&$!0bfp0|xwKN_~9Ac~&Z|?B5+k@k4L0G+r;5
zn>5C~7!g&J&Cix;9j-FA@cm2AH!3Q|KaNbBGe`N5Nm%(e-q+M=pJ!ToBFRfgKeM2?
zXU6kt6h-6<{X1sN(^XI7RyRYPoy195dZWI&&QkSFn!ek2SuNBkvHK021I#r6-&xe!
zx-Gg461ISG&t+tCprZd4*8uHC&i#c@NNESAUN}=KqB%u9+qC5pVMG?Hh(=U>8QI$a
z01By`-AeTo8>26)1DmI%hkfbw*OPrFJu>p#W|&%Qbr|nouLia6l76k_s0JDz%+U5K
zWJ0;ii!H4c3OMre7(QcjWI{5LXvmN_6>svR?v&a7rX<xWSBLclIT5I>fN?HKNJEtr
zfRwbNb#4A#9u*bS22Y31KC->oUlvi55h?s~z*usz-oyL|<GfD$QzX!3iRb>&1$lKE
zXuL9GtXFe&bziJ7`^)^SPH6^>>J%)&dAqE1ZPNdd$A<@=w1N41--`t08(5i0S(>OD
zv_sk~4j*J0lJ6PiGVNGI65mQW8l#rVCmCn+jisriAP@&R|61?l68)o<`gi&t4|X<}
zKMh!0S0SKup*<cDB}FZzIoA;brNvXR3S=uZ33!6l-uI6=6T5fsUL*P7!*cB^b41Fn
zlZDg8Ez0Yl2;5@ds&!mr<x){%UGr*hlWkE;0q!;^8n1*{P!vbO`X8u@GjusJWUD9V
zfRtp+VT%^2y(%5I);v9eWbwrYoIG3RxZl5~yHLljAPBuw0g!q)tKZqLC#{U3IC6IB
zc>Lkqj?Hv-j!vqz^ug$eUqj-KoAdxO_P;vW%_ggk!(=|2^+l5&5B_rWaxgi)g}p>i
zsw}@!@}nprk`X6I<sg)l$Up$$+6rlkBbs;r%?`dY`^+0(hza?@^T}Yu>=K1SVw=Zx
zv8ga9CKO;3$!*aYKO84fYT6$#F^ed5w5abl!<%GR^O5+iN_iWW&*Rs$8Ro9<qTfLV
zlPV;G5Jnv`Hz8&olUq2}v$wbS<FUPy2M`TUs^6MZIOoiSD^{CkE=n~xo!ZK9tASvx
z-hHlZ1+S904$hof3jBK;><dVDe%dST<$q}bF!UpQidpzOP!*BT?&<ZHY}LB;AP0Ki
z@J&IJh-9tX2j1qtO+VvyddVr>;$OM7SfhFq!yXxxm}ftUHt)rv@aL0lr47TAqXzaT
zp>wFiUhu7n)(hqgeQ!S**3&!-^lKjRZtW4NjG-iT(LXXHVEgt)qC5vs1YxoWzcv2X
zG}ag5soI<$UTBeRy%NWBRHlMfNVBY9JneQ;)JRt>YUjAQ`NBJ17pu3-{8`&S+81E6
zj&bUdT`z~t*ESsXBg(9@!`ES!9cSJNi1y7K1A%I_|DH}`$HrRs&9zi04jOau&>L^(
zY`nDUA3kEyR1Po1093mqvlXaUfAe~{iak8x<b~S>o&Cjo1g5CL%>V`9_+i<=-S=Ma
zx>r5PBx=iOZ<cvfZ+SZ5VGo<k-Pl(834Z6Di<ZBBes^I^RLACzvU;bjV@|_JhPG_9
z`V_N(-wk027$9*cXuavzIc2diiC~{;<QxzD=lIuo$U0!1O=ixdiBEj^a8jWQkc`A$
z;sIm9;B#g)k=ZAA@<EU)O=bIT9IgOGIH+5)b`3y49N2`|a00#UDH4OBnW+SU6H@Nq
zcQrIub4%-JKjX-#Q>Bg@@5E;;{JpGiwK*66xsfO^`L=W{Z6n$@wbtmdKkRm^2r7rs
zP-`4xH+?8eDX%!5<zU3b%2Yfo>FjVrq90G3*sOI91}^o=iY3DK(hb%7%7<B2Sz#}E
z{g5mi&g5wTt&BZ4_iFyTcXtnzx}rHH`@)+ja*m=KRMs@R!qc#<Lw1iNcutbRDEUb#
zY;wq`o<UGEDxQ_^(&}=+y*|8^@q{uy8@x+6TYiA2!Dosux!IT?Ombt}il2p#z%7Ab
zq81NBkg4U%H=&wfRj8FtXrA^fwW~pz)!(K|HC@)MQ$4$RS4A;6vJ|<%w@;EKiIV=i
zO%){fiE{hw^GbB}WZw2Is`xH^H==USS-r-<*h%eZ7%{^(Y7nAJ@D}*rWmP|%g<u!0
zIYM-Zpk2FzuWo(7%4Q~*NaZRb$;j*oN6QBj-=&e!j?7hv4m>e^J4e5H=1hJLEGBhb
zrVG--|0lt?!ob}v_Z()aKU`-n+pa19Tcn|dvxESVWD124XMxGqZ_5u3S=}46?-Znx
zRQ4NF3KOv*Lvq)(*<&HK8uclEr^VlY-~K-P&i^*OMkUp<u_@UuoNSMWg=q&Z7eQJr
z#PlGEfOK56KH^;pwD)OSe|Dwg`)nI)dPxA*#Qi;I3cEs{%y2ZqgrLDpCV(7EMwukc
z^EIW0H#Q!PjLf~dYPC<w=2!_K1wuu8Mse>+^eW@&NYLX#aU!{diVjP&DoV0}(7P0%
zw<~6{fl$o6gwv9X5oTqB|ErJCUJ}zT!Iy;1@PR-isq+l^OVnLd$W|ro_(DR|OH3CJ
zQ0gZR2#nz(Nv0OPjipa?_MnTRzk^g3`2oERVEQsV?POl7vsSGyWO?qY*h(mZUiil*
zb05_*1WlQM6_8F&v4LPW5&bAO5WFEi+cBReXZdwP0;6@0W?WKwg?0k^SP0cu@6vZD
z4G{_{hQqYVL}kViOQ|yJy8F=FNi}Z!d_+=hm=j0=hipkoVpdiGFr1V#Iu&kMxqt{n
zUNGEv+A_*T7#r#MCA<+0sE~1z$rcxQ?Z;X6bY~6!KYrxcSXB@JoM|#nK^&vpx`CH0
zv#gd@;=9tI5#*(M5Z{v|V_{(>6Nt_onEbM^&<m9oJ;}|KlxED7PMsCC?_P1RNe^tf
z;s?Z!dK~spf;~jI!bOD<RK|$IF^i*t)u7B4H#$6U>ZdDHYD~Rj;i^b!5)aP(fC0@~
zZW|@0-WFL|n~D!2+F+wgSgC|zprb1QV-X87urPu_&w4(!&sGgsWf$$q3YTD1h>GDH
zKiSi?fcJ?e0^-`lis$(e7*{8Ko#(x~cSoQw=7g5X4}r<2OnNxI{`YTL2AoY2{?#Uq
zqxJok=By8vnoLnasZSu~!dzmW;dC^(i|Ku(ToGF?ncB1nazAPhsIHu3zJJ0N67uBt
z0U8NFG{mxs+E{W^Id5LPetpuZ7e^(xIqhgIz6TC`48OHF)qUL9J{_UnKtG`3JdPi?
ztUa%VVa@>mmg94Jg>F&_?6_H9os)+Sbto?8`S^a{7x2~-Y5#qGaonfr$s93C*+e;w
z_TK#xV+@6j@Qf*N4S*zplu&WamogK<Dr2X@q96RBgE<ShS%B2|WMLUh<i3g!OcG{L
ze<7vRlh$1f%fxYjUI<AJsm%23D<hKCTDGhO&B$(!93RM8`X{l?;Y$TfH7inlSJ5#(
z+qZ9`+r^4KF!}Sj^;O<iLt@G4X*YP~R^u6M%Un;KIweYBaYW-HVQb^q7I}(vq*TaK
zj~p>@YlMYp_waL*|C{e+6GIwF9HY#!0~8E+6|McI&GF-s0`AF%o*lPJnHFLs?5c3#
z$n&@>ZiZa*`EWhbGW;Y%cqK%e3!E>!Xm1=x4u6iE$BpDg5#h=elU=q1ViN}|!ePc8
zhCeq8=Y(+i{Hh`Ah=p}x_tUxxZg|oeWZ(a3Rt10cdDZ)I5D$@-BKy_{Zn_ejED^7I
zD}cyXetg<Cr2F@y#SThnw?<QO^5JD%eZ(kToNV}W$ruH-Yl;U~{uL2w!-~tF<&;>G
zf4|k|pT|fud9WwbY>^Mp`$eQ!5y`yV$2~*}YORU^1D>q5bg~Gg1Xw?F$^O5oLA}sU
zOh)A$5XaO^+`^>g=)@cvv_ezlhzj1^c_MV^50?>5;A4KCN5lcXc*!VtBG-wK$VSl<
zH-FFnrdZQLY9eKkN)bhNL(@#DBiyLy6gYUbK>^1&s{glhc>KlROlf5Q5o&6URZCnY
z=9vXtTUffzs1U(3vhk?_btn`PXu;<H^{M)oZ^Ui$Amq^rcor}$*nNJ8$W!8Yp06ov
ze>HN&P<P9nr63m)Nw>8{=j^IdWC)dc)k#2+yGiMYoMo4Or7&l`<-8QMB?UCJ=Dt}+
z%6n1LaNNtQ;g7kdII`B)wvZutj220?*$b2b{0pqJDW$Lg8#cLjsB2g(N<k^agtp~9
zPvMRNM@JOCYAu0N9GAq2z-w7&%yR$nr8|w-lSMsc*wlsVVy?sVhg1O7_C{o0GmP{2
z^)a^)A&ddY!kRhN%&cMLtTWLE{W*g<OE>!0eax@Gp`c8ZL0}=CrPh$Jd<YkQIV_Rx
z&fOZ#8%Vum5G18^jLsH5S0!cslNWqoSv{hm^D#>V>a$zZ5;+z?RbZ($+I|rMq{q9*
zE)4TAYv8-U(ed1+p*Bf-XZNme^V~vN%ZMPHCwp3R_zCS_`47<Ir=y&VX_fAq{5P*f
zYci$(m0NuGPq+1>%_E$;E#MMSK9ZYb`7b<GS*<lxiHy)nZIe=d7aSD5qJsFS8}qf^
z(mwq4*I(yX4?Ms`Y@E3namMUckPE@Ea3n-DZto0f7}<5w<#;b0A*S@gCullK6DmSX
z>wWD0L~bOc+zo3ese2)ayKteAaQs(9pFOmfBSws%$v!$`(`pb^^GbhKxrnmJ@zN_@
z4wiFNQ$v$Zy;Um@N&(yQ#ke^zD3FcY2+YU)Vy$wtO&b~DBU3K{6nz$@agg}iRq{+b
zlMzYBBo<pmK0<c|=9L7^Z6`I(B?C~MP2G9rwL3y$Ro}{Q-xRuCn-xvGgZ5pjBBX6;
zuF3~3)4lGMs;%9@C6+O$1G|2sf0s5{H3!tWLK-NUignfgU>&u+`yQW;jrBK9%JBQP
zdt4Vf4!c20j5Nr65>b_d!h+s5!X108yh{_4Tbi$J!q1%93_S8SXuNBNChNk~^yKa~
zODGyvsc8;M=`zu|J<&P~bHg)id^QL7rbK&xe^IDOm}#D>ZsCPh`rycR>e0yom>Qw&
zxlxwEFuPWEF%GXd_;OmUvmXz>oM2k)JYW>DiuTSvp|J=1-X7Iw{{$yS9#3++5*uri
zp?9v2=9{4#@2wrT0^Gy;;>TfHzZI1k3@XKVR0nBTy>qH`#MW`N+~+Nmn@U9sgm{B-
z*-<&#caK8kV-Yee{?7jZ60p7hZ{4G1tE^6)szmIQUFQ>{wzw>~TW>#1u(ZkepFHzn
ztKAtfE8k56S(F5CvDBhV5f*b&|L2z-h7rl3peSz=kngi|rw|du0tIwYv7Fm%%(ExE
zcf|JpO0aG14RkP)5AT<`!q~aDPx}=_7}L3lK8!Pc7ut1@LV8WLXRAm4s40tCarl0E
zDlO*XfXu6Kdh@L)Gjj1B>vEsgbTQbW5QwhQT<b>re51>L2EtD`xoOHj=@zTwR;@J%
zI=5QC?E5+{iK~?_OQlKVzxDe3Lkh<6-h+3?6cjt^y>L$(O5U+gT%U1|k6oOQb0&UI
zOkAfPfK!*)PSQm2A>Z3C+PtfJ$)|?}$9m*eZ(MKX+4jJKyb(`s*ELz5r?@-xt82?G
zI{mYPa`#$w@;nwg&is&nMRvg+tL+EO4{mqQ*j|)+C#~b`>=Dg(86ViVATv+BZ?`0+
zH%)4JEW0|d&)hx-ucWxmD?Ge$!Pyl-#hS0a>(GS&N@`H>VXe#>_a?-u^D|g6^4+YX
zf^u@qBU)E+l1pql=*9xdU}``aR{G{)VB_9?_wHY|T{=d+sB3du8=IArO0W(l<m9LV
z9ufi83(4`u*h5M5T4}C5wE{jc<}#o3D{lBRv#%gCu0^WFX!zkAC$|H$Rt{3{p>@b*
z&)r{hy-x(^#tv#X$zxm9Y3tnebLOq!<dZZR`e7vnWw5Ane=!B1K==-8J6=~{^7;d1
zS2%c#0`!avN+~f&q*f0Ju$O^4NbU8UOoB62jJwQ_K&(Yp0Y4ytPb@G#1`P;M8YJf;
zs~dVIn!{<IDTzzU%eRDwhmWa7L8Gm!%UxQxY)J9@_xJTv3QCWf^f;7fNrQW3_F0v4
zVDBrQnj~M5kr+hvBbhmT_;C8%taS!KZI2^)n%(=--8xjMe~ZtUm5EYyz&F<!RS%g^
z+onyMgzW5tZI8R3VIVsPE}c5m_rWpS>X{nCNFO%cRxhJC1vyc1Uf16(P;|km4_~iB
z%8$3hlAG}8(fPX-(?a)|4Mzf{>v|*P8q4X8;YQ49nJty0qoeQVt*223W0uN{HgD<q
zp<aopF2>fD3l`jHm-0i`uVDT67Q=4Mt6%i&_LQK=Z4IaYJ4R#yR@a{xZIWY@g2?3b
z{iLKZ1r7&Kht8eOPAR@Ja?bbmrf#-#ta@M4;efB0IEHzW8i9UA>a9UQ+`Q<cPSll3
zg`!gEE?T406^DyF)=&x}Y5HtbuxwyEA!q~*MLWl1qg0;$a76jW%kDRiU4x@My8G2a
zH{}NnUM;^O;TDuTo>U)Dm~FAo<4YXD6(-@Dnw5*<s_0*2t_Alv=z?J{mCjwdoVdNv
z`fgcjyRa^nRk1;_X))C@HgRQkvRJY>&dK3O0@}FKjv>JVed_cHusm1X*Zn}*Ahrcu
z*SLaGe!hPmJDc0s+Q-A*g2CSu6UWe{aP-<Vge4n7$K7xITpm}ncXyT&eGE}u8-jx7
zd9*f18Yg~SE-e<JVaMi6dmXJ{WCGQNOgu3)HN}C$T)xJ859{dYYz_|ApljqhkxKO2
zw{O~wYf){4b$xbuHuUp-z$t2#hKL5>F?3y<VD5T>B@hGbaMHYPdBra3S=4e7H^BWi
z<xC9GU-zv!0T1$Fe0?YHO+Iq|^oNC0M;4yx4#jGoH$?T=gn<`>wKS@AYS*sKC)!xM
zY5y5jaYN1JShHO?bIZMB${Mc0`v$mKaBHtz1Rx>`)k%|fX6C*U)NMT}EmGc5ou{ju
z9q}x0U2i{c07;$8_T;#uLzv}Tj)Z{^M-ARM>s<9vTN|#`;a3}bg~i5CuB6MjQCJrJ
zuw&2e-KSPwiHT8R$#O;JFB>SCUUWpeXb7m!3wX8n@27jdy-<ZoNTJ~F{zh1|k^&j)
z(#TonKi0jSV0x(>S*>bH_rDtXSXS%kRXP&z?5a-hD)Ad2t~=BQOtL?f|F~T9Nr$y_
zc9vc|GTDuaxumSjgE=A+hZ7W((yk{<m-0`pxpqyx%(gPb^eCtu@?pt@;I7HAnv#i1
z-3w!Q0zox=V(C1H%d@Jxgtih14<v~kP=GdKxS$=_J?_@NMN-Gf<appPF)pKHroyIS
zuYW{;%_ihZc$*03G1S}}3>G6=DqucBz#hL1RD?`yz~5Kr_ewGi<6rw1&*A6rn$y%s
z5hH`->1J?n96feyQ}<k{Xrbb`jc9P9To&%!wG+|6iK>NDwF8<P4(`#^%cF*ZB+cN|
zr)BagHz0CnD))SJH#WX2uFU3PH6>CmYurJer$j@;99s^3h$aSMdexU!lMWBu$R6r<
z>5!_bs$^i$uTAgScSEe%Rl$+K>oOby$2|LxP#)^AdnU1lhG5Cacg{sRLQ`xw(63&s
zT^`VK8wLbmtz;<SP136}to(=C_R?}QSlP9C14f8!AUD<UxTh^tUgn1<?}zy&TVD0u
zzkj{oL))c`zlrp{Fmh14os*KknWdb*Mn=v;6lVf7a+&~i|N2=rnO0w7{WyZ4hYd|l
z|3+9*GrE?S3QWM;w_TR$<i^c17CwNKz-^S^BDny?5-fr`iqJ$KWPY3g8hlaOijAD)
z$lRrh0ENK@bd$Xd0D2qg{*4#QK!msP^CEg<qlqpIm{!I_poHYrA75YnOZmA?wGJzd
zfs(k&TGRFsydYD5L~_DPkcq4q3vY&>yS5!!>`lIb^m0ICKq2(E3}*C#=|!48#kOrN
z!@>P~qz=#NgDa=OyI0TGl^J74nvZTr0(=Sc`O?0nIFa*_k0DbP0#Y?)3%O7{_=yL4
z-B=RjR*w?^!4R&z4x}F_J3MYPa^y%^1LEXFcOcbabm4Ildw3~z8>{5=JG8a6huhn0
zO-~B`Ww<P?%fr(_?%UDL)>g<&8%!J%A|^L$){HesC8uFvaO8>)YAqQBLG0Z8y}Iq3
zpB%aOx)aW0Nlyht#QFNu$SwGJ#LEaL1Al>_0Ap?d0Eh&Wn*!sBPr>u%F_kH+)J^+u
z_iZ<5(}3qLwc9nVk0Aiy$7;gRw}bUo4oph<k(1W#Pi!1G0Uku72Xi>5IW%R@0GB`z
z)$+Tyg=?Gh*({^;>*ZG@><<810+6|Pz*c}85q^<nvUmJe|9}AT?#rv>gHo#>Eahi)
zts3PbRf9DHC>u_c0qo0y`STStXl%RlVYMS@En0(|wP2fygv!`h8C(mwIE8Or4yL0H
zN~8IloYT7w-Il2}VEzn*GCsa)kj5y95RW@~LN$K#u1S+R?G_pRKg?%gl85BH#8>3l
zqE@&6>~Wv!I9F#9m+kw}W5*6S*se*FCb!fa<2Qoq<nz-q&P?!l;N|@M`9PLqPCL;O
zc+||v#X)ki;xL{OF%h65kVs@fDocR<2$t;FqsN4Zq}P~37m)oVVg+hBZii8<Lh0HL
zh*;(u<A0>V;s%k)wjrURBUtaq3y^WAFN%+K7)DUwn}PTXuRU=KZE1lp>c;h7<$X>a
zv6~pwY>veAVXG1kws@bxY{W~KnK`4_zR2!t&<6ux&^I?naeN5LDh6En3lb(p4hOV*
z0p|wN)&*|wpX{tYb56^xRiiDkt4jv)gxPMxrdM+bo++wNYymVSwg_+2>z<yTA`EgZ
zFDf@G4L++E4*h`6L(M+%#WKP6m{qf+gYT|~TYGQZRWWhey<;u70B3VXiqjo%u)acg
z+(?Eb36$&d{k5N36^V~&?oSxh4JYYPa)mBXKnekO#TpOX#^|)axU19Gv@@M{iPBg3
zVN5PS3BS&Z#L1W7y0Qof8a<=8wzaeOg|5i=P*Mnv#tRg%w))rC>#A;a!Og{36f*UA
z4VCix^$W;m5$K`aIHhYfCq^9vddh&FvB85++EreF;NjtNYKV%R2IS3^1N!O@TF^Li
zX5GlnZqu+~LyVoHfORQbqa5z7s($jZfK$PfTa&(523d=79p$$~6(A}R;C{#ui_?+L
zZL7`ZS@6YEo%WUI-B`JdbVA7l0N{w)t}%55Gjl~|HvHL#d6TR6-pNe(-eOm`pVM=T
zS4~^!eS2JkckP6glVmwWwAhUn1-Nyz?arK{*qlZwcI@2f3Hsm#TTOb{w|3U<+OlC2
zWo5y?@Zy~v{llck{tp)p^eie$TR!mcDz&__w_rs^b6#Mi^1%hZa^>2J_X{$A9PDq~
zd!4JELE;ayGr8wg)<6B;KBQ;m@vskP)6O}ruXDX$xM|XiZP_iGBE&$WA&fx8;2A6A
z65!g9u>wM=pD_p-RonXd2`oa*)2^hxwBd8*1XS$8i~tu@OEG=G<*reMuexWbcikLy
zs?WlmG3I&JZ9|O>)(<Id_H@$E;~7bxdcB>qDQaHS9o64N-w7q;j$&5f6e@4uh<|u&
zHOzU`ya`*bb(Vpeax-_9Q;?Mh2Iea_Ufv71%@xHcR5y%D`>}e`dOj{9wYrLUA`Pg|
zY|HzqSvokJy%cF7q<!4UeI~^<lB}6%?GYvj(1h1Ym_jqNRQmw_%wT0sd#rJm@6t}X
zyxu?aOZ%qCQ!X)DxkZ~c*+m7spHmsh9pTVFm>HjxkG5-7Fvd7cFbROokZ9>gI0%He
zAXdE&zxpcgme0<FtKM(rk87NG7TkT<iK=6xGaG2s0j=fix-Fpxd*VVYMy~_Ama7UK
zT!>$r8upfd`)#ua{6W82{dbR`s~AuF#SVp}0~z0{Wf7B*mE{-Oo^d}?>4-rZ351q7
z;%E@)rAOxl|Do&$%Owp6Q9NXa3eEi{WNdC;sSAa_*Sfv8mYZ0WUCSvaiOZ4US&|T$
z@M2l|kxS^!z*0rYF&`&}G$3+J;vaacWhL(h=Rxvvc(20-ZsvIlqW+mVGpp(T{F8(h
zS550t-d>6CkRg{V*a6NfpI09ri-4UcwyR(B?ElgNyae+E476A2{CiPsZe9gOEWdk*
z311HOSzM_!T()Z|+hCmLQy9WPdqOq@-1<G%w!u67B;O;7C530To`3Q?eQm?2X}~>K
z18Zv<A51$|(C+)o`Q8;-b8%<ux2|K=^heW;tlfQv!3qV3cI5LU&{73CbrxE^<y9Rg
z$U!LKsF?n=y=1Zm+`1*QH%?2LzpNOJL`SmFB4;(}?dOojaz8t@bMn#Z#<eZt^UkFN
zrj0Af!I!Zz>AgyOua@zf-|n7fe*BYIa*dhG1i@K=yko)D2Q``FF)=X_ZDdw)T3;}(
z8LdZmZ8<*RR^owCou+z^*W8<WoAB*-uLeJRJ=C>etwHJK9LLG{CU<VW_j(~@%AJ&y
zWXQy*n8~mlurtsR^(`$6^blt8uQfn#3I(oK)uVg&)YV&!%~54Z_KyEmyhwqwPOq(x
zOS79amCU$oJi+`{@s$M${nJ-B;M2UpeA`SRx)83h%|JgsdUVp_A^P$4Q!`@!*Z+F_
z#z~yuDgEiJ+o-DQea|>Tf~j-J{(P43+p~|my5Kac_>|{Mr6yBQ5wtjdow=yt{H4P$
zrt$$MBd5bEp-j#NCDm|D2vmK@S8gIzq!4XLC^dC-eOgMIU9`XFuO*7ve(;H}&(}5s
zuHs>;rc{+AHUo|p&p=LrWZy7Xv@wA(LcMWspXsKWk(nv(PqeV&00ClEeX0jtC{ri%
z&N2678_<GaAw<K;xV*)SPbO8^S)L&-1N&t@l?uNNyzV8RN;zinkeJ>D#cAC%PE^gh
zuw3J(=eL`VTl;i)UHERu;ilny<?G+ys7XLEB`uYtXldwD`?~*h5mKk$A)`EvJ@wkC
zH7IY-1cbMvSxMY?O=zkW5^yX@WE>gV>e@r~^-Cy+JNN2kl#(h>xqTL$A8S<Rek|4h
z${lOiWB)=l@%Js&*Ok;5W59#~Ryud>T358!E-sfdBe4pI%#!I8(wOm?q~0B??v_43
z_w>XTs2}8!b0>=kl%Hi)Z{cNyWPDnzb{yXv8iMm{%B<jV#4jv?JA`%W{;Hn;pQqrq
zX7-mh%~q@^(H_8ynw|Gz<*dNoH*zs{$&_0tm#vJc*xay{LNW8~>>J3LuoQY?T7Ww%
z8e%f7*`>0pCae}o`xU0LTetc`c&|sWBPx3*CnqT_U2CR=1%J}!CwDcCaLl`wU+VG=
zIj8QCAxO9|Y&XPUtQZa<C1i?JoFEJ0-$H~+>nwwn7z-V7H&Oy6C3=>;RmOtJXbg09
zqMs#7VxH|8Qu*XbFhxgLQs}`Kgj9_n@rWAb^okt<*HL8%zlg5tZ&U#P^z50)dy%jS
z%JaaaOy_kmZh9wBgmT8K&SM1_si91XoOR#kre9bDB#FI!L}o=ufKNoun|FsYl@an7
zt&M5;{)hANC-+-|Y#UW(<~TAX(=JgnGV5?vW_k9%i~5}!8ldaHZ^Fr($i>ICcG}y1
z8gRtbI8wv!9t}4CHM>m$@SzI3ZWdv^mqxZ98GZ0nZ%h4qkr{<2rq;gdb7l$Up6ui#
zUv*!Azr(}Uci`$8P7#ZvVI%SuX}2H^q>f{t(y33ohUgl&)SQxEwkN(wpjSt({6pq}
zH+VM%ZP>xsugb&*HzlP`qOI+7bk>zU7{Dp6t^KXCU?r`9_yow9^m<~&qh^1~+d6c6
zGDyE12!sN)q!ALEgIfDCa$N-n;P-OvC?xIu+pE2kYbKIk8Afj-D+faN;O71c&`o7*
z-15;SQf~SV>Oje@;@-RcCApxqB-6t21#g^Afiiv3(3sxI<NOTNE>Nh7a?ZHJMW8{^
zOF-B?{&9T<mnuOG$_i~gJrPa@*fmqE|Mk7w-a21FMWUmf_kL*HsL`%Hdul1pM@Nrl
zJ4Ee7ZJ_+768gqAqJ1EzM|}KsN59o8R(PB~y=3nPU|>nCQb=>ih``K)5>^U>ssMs&
zhCj&PZr-$Mo4Ak_l2LpyY1DMlgljGkz)fJ<C%^d3^jVDM95MtYPY2@|7KKf+-ar3r
z{T^s{U!+HW!u`f@4?02p)esH8^iv?&4wczkTGp1lja(8--^?y?dB5C1h<iG(DXV`)
zy)Kg8&3hqqz0LK!X6gpF5J{*$TgA)*wr0J}H&1=~Ck4F$D_(K_hgzEXQu9U6)OW%!
z$@D3H(M_T&)M>ct1N(?>2OknqfK4P!>db|&CtkPxyyJ_UY*M<Ns2w&eF+BL(kBb^o
z>;v5bjcNN<@w18zU?;Vl8)gE9OYYkE*<rxIq7L|m#6;ESG}0)p68^&3F`dTij+s8N
zUDF*tKH_La4G=PPHOc&T&tgpr+kAch<akU<4=~5!&-yMJ5_2|rt9=mpM-xtdzVfSR
z>EN7St8&`L``~jExs&t|oP~X7{t*AnI)<xA+4ZZcQk+IhuiA_>->KnuPOj>jZ_s`U
zJpEp0Cx%^>x#28QNIp5Rc-^9ZTLat#^R<RQ+tK$oXKI&ye#G(Xkf+R-2T9s+FXQkE
zQ&eIK;@snsR^(W@hl#mg)b1B9$Q9R@iNA#8`ahYQ)y8MX4iI%gO1UJ(Oh^o25|7=v
zL)@T(i|;P>1wl`JvN7y}QoE*OPp5nhi+S+mRCLnPN$$t4SiViWrd45)RrcsW@{HSe
z?leFSrW|AX&tFE?^_qIs?%er9tEWYNFST`a>`ENRyOrBI(+NgjxAmF&V4tZ^veneY
zH%A=j<P7+_ZkSI>VgKU8>8EoX4T9!$2pe>+EOU6n=dPj}p_4&`YGH5wovk>ZGEy#K
z&NmnGS0&JZu>76nI?F%XH5p`_H<#C6^8_DAa&`HtReuX6uy$={v{Q1%!VL4}m|J|m
z%Pf=f9-kO_!Ktj1Gn3OX>Ah)s*M(F3n!BocC2sWb@yQ(6xl<>JRDiw|aXiPaVgo|i
z7i`$U$+M<St3_^t$X7z*KCB~9N3y{qB1-~dB8`oVR(1$74)XbqM7b2S)mHx(;i{(=
z$=-v9@E>@rn#OO@vl)wlWd}!pN;}yYhT_;`z1XC5lr{DGbkZ5~&tFCQNyK^`oeo~L
z7LLC-%I9qHl*@5(iO-$|I<MV0{Aigw-MK|UB<UEzHyU?J@#aVsGr4h$6S2kSp2rR7
zF9PR!C|34f&;6IVv~B%v-A=0v>IR|-I_>ht{fO!5$)aLIrz9bUMiH&kBX+5h4S-~*
zuVXrc%o6B*he@OqsARH$LYP{Hj>%h7qD&#*;N3b=F5G{ax@F@=CAule9U+A7f`8qb
z9HLb!y=QNwP83oA`W@X6-jDlbVP^9<^+~t(epnT5vo%b!y>5NKd$$Rql!(Pw_yGBz
z;$Z}BAbQ0K@mAnMrf7E}%xD?o4&sRdyKj(m`?hr#muR1=XRGtxt=p0Pg_=Cj?6gry
zkG~o{jdgA;5E|S4U~<HnGm?hK#i8e7x-ap?{99XFh`WIs7miwP;nZHw^%9ORPX+S%
zhc}J*Jtf@>1k#-VE6vu;Yg)p!QX#SrWp3DggNR|?kBBU0d))Ak4hgsD*>aoeob(CN
z+Nq#+GIJri5Je2aR{tI6z}NYW=qoT>%cY@0h98CJeC`7oyaJUwD;f2ujG~o(3Mtxj
z+ittZt?Xc-x%b`59oZG3=3jPFe;%LYr(4MlC8VInNKfs%&zLA7A<FdYv#z>EXIjh^
zL4J~&<9XK=U9~)t!!v)E)LeI71S2Rp_n#@?4NYwU%@1ad+dLnMlRLF4J<K$`gjpG)
zGGrl)q#6>R67?sesHdl_m0+WlO+czP@7~=4EtWVP5#G=KpfVrsK@!LTD^wd!;k2dF
zUC6=iQ_<%?Sn{UO7U6q{@6)}8;!6UXiG8ktyKy=xoOK62NeT&zd-5nEYjPizWGlO-
zUdJEzU6_l3dZiUrT^v;oaL5hj1G8Ii!Av{7OY6$Nx>Rez9MKF7ckE*Cd9VB9-`|#Y
zo6Ha&o{*?3DKc;UyT4~Cr@tk72K1^6iV~Sr*nvx76y6^hSp|A31|5X%9-I*LGUZo?
zaEy$LGqJE}d}a3YS2Nva_ZHidc*3O$<Lxc|(zZi~rX0bDyR>zkUoRiTJlEZfQNEPt
z9(VU39gtkVGD*U0+;|#Xq@m_tlhNQ%3=>fpa`xAfV@9wwgra%_S_9W+_6@W8OMTGt
z&D%y}Yv+(L3avjUIrDX!ca;5`SI*!`48cwJlg2{=4v9Mejm|pb&bBLML^W=wu-leh
z{4XC~Uej9J_on&aGNVK3gttx2KKb?X$i??mXgOX)XU(<jP+00B--qQV$xClWN;7Eq
zi;byNyS9hsb(5Y?!)4#|D-;{RcWl431ppA_SJE3e$dI^7=mcj@N7P54CD#2;!5Q&C
z+T#3?q4eS)q(=i+Jj#`2?9j7$gPJ+_<Rgg$hz_9<VLL+WaqkLS-JM+Xtpb&UX$x^t
zO7$l89v&E^qmrW6-1s(B4^m(KX@(m9$}&1LC%UQa%m|MwSGHVBe0P5EoK0CgXSk{}
z=So2D#p<u>n*!g-pEx#nnEmE^``^_dGt(*Zxc9~aGVo0zqK9$_0o2O0w&-TTsCP&W
zK*{${pVQupWj#z3g{cCh*yl=InKjV2oG*`#k8LIbTq;$uFXE!OYa{AdAz+h|lE{SK
z0N#yK_fVRVvDxKi8fWKS--+Z@=G8+&CQ|8dgQi?roM~LxW*7l>0E!ZeCg&|MvPkp5
zkC_!BJDuu&uusIjZ?PP9*7K~cs&03-eSSMu_tVWyE#o6jJzm@m2CD&>CnI_eo36zh
zixR3JVPkFMhetX^cqd1s)$EZ|D6%0nNmfd9Rm-S<4Jj3Qulv_Lg;)!TtGi#ysP$zP
z6%G~8w5Z_~8(8EzIi86xB(Riy@IciA901fx{8pTElu7)TFLpNbBYom6FLn2j$SMfl
zt>nf=-tC*6v~>Rb4McM<08Lj-$(VO}HUST9=lK~l)QBiff6``lb*YW_S3HHf9qz5&
z=iO}8*h}ALwV!%=VoA}2;3*2{Ew8)Z$X#{h*8sjk>U5$IkEw?Z86`?*S3r_cM;$U9
z+)#ztxJk?U>WRD$g*GB?y|$D6$`?`->Q*^<RTb`CQhwrR!`}0}Z7-G{S@NZKY8|H9
z1I8*zELVB^NvAvQ@1E|i_j1~^afagx(fr9gAJW5E!AtGu&6>59v&SQ)&^F!gIx1C}
zUmlh5QbL0V?<E?9SEvlN^`~F?*58GR!36Xs{QSw~Pn#y-X{tTPlW6fFb|B2f*Q%mo
z((Q!S_*X>@=hk1_`s!`I!H>^$j+S^tWMZ>~8saIv2?P}3!(?*Mx~)_(!W!~4Z7#0k
z1z^Zl0(O71Y8wCcW}OtO>SbR)&Y>hMfXYf+JrHC?Q*^wjUPlOIM)mNixw(ng<dRwm
zNP`aM&X_tgs{m|$Npu8T6IpcH_SD^;TNf{n4jQUo|E>5(QvF1(4Ywk}#Q6Tt<Ia?r
zV5wM$B%-MWE<ad-@oW?zfVukms`A!<k`KIw${jjb<^KimjM@hKaufx}kER|v!HL~m
z=+%Tk6yX;{8Inb*l`R~vX1rq1{1_;PM6N6ytBJcaXLg=Cxx_#2k$Ows77T)fHA$z|
zbH<tBsNzHv8W0ekIf*&7*w|f)BQEI_y;p71MrtHvEsg~;xNz6bLK?T4rn5onqZh5>
zdtejcniw&Ac00+)wi>v*;-q=lk^E@qigXM7KgT9RePLC+2}mTg6&FcIiW~GSVCh13
z(yyq_8UBC?r||Ul9R~a^NeMkqwQd>T15D}LyVF5m+Sf}p?6Nzw%5Lds-CYee0@bKG
zi9XjK`fk&o8D=NGQ=SlHcl49?5i>DOaT)@qt_lY32|RpQl083ZZ&jHb*41WMV|%-D
zmBxSm5=YfoWccelcJ4efZ-KtxWNu!M99G6MGTcCa&aoSJ4qFjbu@z?pv>LBSltw_;
zpLYEzpHEE1g}@gIvFPTV_~Px56HP-$c5?xm-2UyGi)2v4>?k%MTsnU0)R@9+j1hU>
zsKuV7eKWGcs_YV4|6VPZkakF>92EMJ4qV)abJjj4lWijtd)lFX_add?zb4LZ)}L(f
z?UT~7HTr><EaF@mE}B;U+5efF+xb@?GHiUWUVT;lBGV^8vEi{(>lgX2z8IAn+`oVG
ztd>PksPE2@L`@e63W47gLez0}G*ZmMKa8?q<;TgPCNVBMPB;x(oKe*5(cc{m$90QF
ztHKbt>9#`%M8E;xn&Rvm7idlwJvLAKTnrkBZx;!&aB!g0LJxZ==X`>NsACrvXYTyE
zaR?|Lrt(d*de!s0HyYimV4H}!C{R8hs<ds2P6VHHWBLWHOaQyG<ttY?%wT%p1AQNa
zChY<PbvsOtytw+2%iNh6rlu)TZeueRw2B8<qyUj@W3aA6$y~DA(7+Vy>K&WoI(5sn
zk!+x=-jr-WLWJ2n;rZL4s@n*ZSdz!bSo|hG^GW}uhB?~Ih3$6akYVr!lOBs5BJ8)4
zwq(*{i;<>n)Ppi3&AlUf?C&>>aXtk&tH*%fGP2;y{VH8AoF$_^r}%f(O6~UUDUujY
z^{VP|cUGL?AuHv>_^JAG-y8LbZ2l~L+^b&ZEkNQ`NWNBlL%*5-@nchk@!3^F;0sah
z9*d7}$5ygep?YbaHdSj*>|abzv(DOfu25r4+Aq9a1$4kN^#J}>=zAap-Ut-#qvCD5
z3XQxbmFy|ZI-b_;Hy{*78I3yb2Di$cE~}%R!orsTwvBFB)AMzOLYmw8Qn$4=L}T~^
z%e32N%{yDqlPLm-;yU7XK^2?7{0i~fA|n3x-<$A+eHlO1piY!ZQ;ihwNvp@73eEK%
zSE+Ge<$a5IHiA~hrlTBfR#%NlO;RKzY&e_8?UEV<4>7!B^Wj}uzM#|_$)-T***02_
zidjMcfeWMw!sVcw?Z;`&>%ti*rUj<3HAHEK^*(`nT{ov1^S_KmK^ZSf2d);|Gp8@O
zlrAKFuo=E33Mr;rDB<|fUjJ(y@PprWv?d<p!&0P5)qsnmw)5%-sAlfGu70{^7WFz;
ztmMyOsK8x%8`FNk)gu!T3jjptGsi;6(1FX2-ndqjJZm1_0an@qlFxvFC3z4x$AiIE
zu^-iQf8WjH3&c-*y}GIyrC4p2I^R`-f<+hsG{=ABtuF(YGWXuPt8g5*bp;TOL_ibB
zb8V!CasHj|(j-2`MXRLX<t@mVOKB^@Uj>HZ89r$78(41-TlGoK1?e?7up+GNK&QpQ
zBNl$N>H{2}CUBn-rK28Fj+0d~Qw7l=UceKXv7Y1e9h70mE?=G%F<BublAtXTDUNsp
zd7+Gspfu0OVQ?W_Tfs@cxi{I$GBFuWR-)~|wq{J+mwC=n;A0fE8@{{KDdeYX)zjKN
zPNHbYhG`35(q^cuj+peo?U7czTh|1d;s1zo+m`$vQSSBwVv>#gsx1G2+15E^hRE<K
z8HDxZiJAREW|;}^K6UQgo%~}&bw#=}3>&Nsbl5%_A&q=t%uUCEqs0J4MwRGh-Mq#W
z?S0%yY4fg!4-yg-67awU#sEx`fxc?yuIDs^v{_=+Ne1-%kW5Ri+y%6t_l?M-tomCx
zMZ!IHV8TlP%aK?N9qX1guy)|!{cBR<m<c9kDbw&7B31_om&2CIayLqU+JpX<Cr_Nn
z4_y8H2IRZQR?u-EvX;mRNogi@X9MM0-w#+crh-IrX&wcM#T6jzWm5+9&2#@=7m1Q;
z<wew_JKFS>*vteDhzu38A%qM_x(;=@oOvNVtF(cBWsoPR7T%>ZzRy4}op9Aj|IF5_
z%VGsEr<3#Y_g9tBin|{`a~r}klPT7Baezc&i!)aYd0gYpEc^HB0g=u_m)KR>-GM&E
zWp<|2g{NKS#`fVVL0r^u#fmabU%+%L6^8-Z+8ZujT(S3qQpaJ#g2qE<<R;#^!;CA=
zB;@|YU8t#DjCBdxyEk!N9tBptdpoHB1){^xdgI2%w*2(gyW5ssnJpERB&?Csh#pZf
zoJ>JDfe0Tx{$WT&MBO5lI++ziq#1{jRBVtNq+PqHom4jWX76KSak}{oIcbr)U<les
zciunrSl>olHVv8-RyAc#wzUyb&5L=Q2s9%#pe~W^^~F^)<76iL`{4C(q>kwEjAPD)
zhHgZlQ?xUps4_|P3Ll3n1g0|eGwZQmGPC&&=3fCN`mQ!c7c`@C4%(K_uB**)B;@mq
zePYf<L@c)Yfoc=`&(ijIVe8Ogp&KTG<6Nd?B@MCFu<W18)>?rdTjdk6FTx1X#-jO>
z92Z4C83)Ovt~?^7vxnyMrvKQ6|0b-St^IV>i+Vy7anjLWb}G=xp?(%=2h@_1LQvUe
zyQTW~M<IPLes$;`V@Y>H_jB#%=ZTtkg@NYePQYJCzHTxgl|n%&263N=8t>qT5AWZL
zd53E})Q8l7Rh+7V(;7{KH<byYVgr+*jWu9o0GCJqyNf_Rzy@EUK9g_;HKP)}*DN_C
zl$Lvf!ra`cH)XCobk4hd`EI0oN)-VIrCAeCIN9KP9RwGydvl4jlQE!#9!{Bf@P9E4
zkNaB2t+z||Y+7gppRgR%<qx00x+c^3I7WywE#rCl!{>b7Z@lR%`@qQX0nEJZrkVRU
z*HNI3ozZZv=~~6IWy_Xo->{fF7gb~(!f9BIXf40y48Wmcv-deCG<zv{#)VOxc57P#
zil3iQV&M8^YuiK-Kz|y}PU!DR(;r^Pkve~`J{*0ff4&xpMbeJIt4y=a3uHaCD7(s;
z96tLHEFS+4pLUd{3*boqdbzhMu!xqNnEUcOcjB2(F6pLZjl!XY>fW$0H~-h9;ry}z
zC!$xz8odGgaH5pwBg&Lmo^^eNSm6=oZDL}DC|CYcQ7Iad{?JFt*aj3Yo2!2o4ue^{
z1->l<-MJ8IDLD0I#;&EMrRJ45=tQQZj!Ky->V);vjZf88*#ha)Q89kh2$QLy5fQh!
z6tu}!_G$JDCAiR5Xf2gcY|unY-jWRCN($^Ama?n|iyCn@EB3&c>-Fl@6DjVsA4NTV
zz8$%?tmy6A2I&u;r)3~50^F_blWZE))7h-?ssE-Hl8i=VR4W<3ST<1TA6hwGn?`v6
zmSSr8elYUcFg?qW$t#j#CLar|nZ%hsxVGt*OzlpuzNfykDhVU!YW|Ne?+!A!Zu|D_
zQhbOL3{qNLeWb8@$k+^~HQ<0dnbA4d{OavSG{F5WIA`a&yr?U}Z$N_i4%g%8VIL<w
zf8K#wb3u8quUy7GWjPEBx3a;=c;-e0S~iIaMlN&Xsq1@2<EHM!Ck+7d{R+G;crH5^
zF4$vC<$nl>kO#v#<EPu6c0D_&@kHd6BSK9&!3|3gFaffXPtI82qNjdED&?O4_BXSz
zxL9OK`in#Yq3_gWLMV?!e8TrDVmtJoQ!zO=`$1bl)7u)XbMe?V0aCNquWrp7P|zfV
ztvntUQM6W%b#`!?0;s|~7iGS-rGC|5d!wo+XMG+->5w!)${&v^|JAEo_jT3{K831)
z<2l^Ne*XM~9jTAr=X(N*^>woih(2R;T-C$Atud;_H=pVT^jpj&j;{3v)=z)aV>d({
z*9QH$k^S3>S;t0nY%a&1P1iQ6)7ESD$Q|-bxk(fHsE0+)4__}o)Hl1L&tHu+{`$4z
zA+cIIAK!+YbKfR{Pq1E__9OXqC~3-VCuC*W+rEuz`1}WkPjNZR%v&Ctm^}=BeX|sf
zOJvOm`~G%tw4Bfub<XDI`++y}xgAA_A;uci&c9VusMiya+LaqMVV#efY1i_I-E*(7
z@ao=)nX^6{?mrVA8oHq6IG?UtWLna?J}TZyDwm2_@HWO#Ix~sP1l2p#`27FydJMR%
zZ9aX`h3l;kJ1osZ^gHHdyaphlghD|Fv`|y?2ADYNFr%FMZKVPCXwd&4K&HB7uFs*e
zJyhr1wr%4*TDMX?Hqqj5=HvwCFu5ag!sm=+O-)WCu`t5Hv$5!*RsUclGz7;h)><uH
z8dByS7Z+EY{`~3F*@YyEy$znemN$exXKsY&>Lp)Dg&J@u|Mjwg%GV5{`s(PUXMg;l
z++x=Gb%PWMIUj@S3<^rgTGNCMX<zQIvH_1EovS}pR}Bj+up3Ws4DdphZIib<x~fK%
ze)$4)F%ZGsdBU!aeza+1ZMl3xNGroTk5TAJT`BpusmDf(EDzD&n7cBfKEhD9^BMQ{
zNvA`t0wWJC4w-N^EM|>=k0<jNExHwD-%BajJR!vWFwgA1)>~z_=o!mfJD+Uzhsqe&
zr6uZd$q&1URm1t?MBU?4Oyi;kVLUj{XWmXma3Z9BL74?#Bkn5>K<7UXId#X1q*Ie`
zPC~>b0hK%(pS(UCR%qI}otfZ*<8!2=)1IA)mmKFcQE$8TMd#XP4_;|L7zwpTM1hDt
zNcWJ*i(G`m=cM^X^|jXW9z(MhY5(LxU9kujSFrVsd-vm8ZhmVRy)}))K+S#yCI8Xh
z)+c;yJRehus;&D)vO5rv7e_zc42ZJmSgoU`_t-DvLCP-Llm>a)+&Xr4lHHgwo|)hL
z!zm^ZKeFZJCQo<R@9;42XqZI;C2d6MVZ-1@=AAG9f4dSGKeNE?{>mqum86OWK5Uz7
z&~OMw)HVZ`b>zi(QY(?d*tviIR&>ZBAfXUIHK4ulR)frwZK!~=-nEKH;e!CFr&y}M
zop}~72n?V-bRk0gpv8jmxzpJ1f)uC5vBx?w(bM@4q&n>J4t3dICpojV`}Mntw{f?h
z-|5nIBFrA9EAsI5oJQnWB;buqx5_`}Fyj+tr8bgWYChhN=vM`rx3RHF&v_oV+PC}H
z4Y5xGA%7BCW4bMO0>6Ny-NM<o^7;B0lk)b(S;I}OmpTow=d0AJRjY(Kvf8?OrYzPd
zhakpF!18o9Fqr$niLwMG=FzhcyX9J*y7w!xZ2%a~%7@t%<FkuV#g6gqAmoT>FF5Qa
z14PEF;d}<Xzj(H4VgqaS!S<U5aAeCihclMzn9>NDm?+{TkBn*n1xM=q1#8pl%~Hum
z9rd|cMy*a^%@zVX+_^||KpJ#6#e8JRVrfSUM(P12brTp<DHc%?Sw$~XF8Zh!irfh(
z@t~O>DzBOvP-ewgS61XY%HU{tB%@;&<KquoTg+o)vT6w+zxk!7-#rO1XUWGC#ig}u
z)GxCJA?|5rx&;*o<c+Q_JULB9DZD_4^+2^7igwJjqG&&4FuGF<8C>gZ@gIqNXxE#x
z)l%9K4MZM#q--x*;NZkTk55j%d4!4s%2pEnjYxZhM!U^rAPLel8ZRPpz;BY@=ppOA
zY2z3ZCO~FS(??4Rh>QuAm;yjd1QnUFwKS|3FBGeQM(iI;cuD&9OEnKnV02;T$i;xx
z!j{&C#Uln236rCic`3l`v!FkEan933xG<vaWY`p4UX2fERuMUOqLbv#M#il|+2lp`
zmJ4Ay_VyW8_7o~FeFi*Fpz>t~hZ1jMmW2g1LVKp6CedG;zj7HkWJqm=Xw)%!*LDt%
zO(A7MOeZn}K!z4b-b>@gmLG4KQA-dGbjS7Z|3AW~*2;muA9f8Fu<hfH9XlZ0m3bd;
zYSq(t_5JZX_q7Zo+JUr`hDCr!1kN(Ii=c*srWT;L&E4=VF)*!tSvep(Z2Y_rk*BXC
z>5)=SasZG-4R!qj@xJBSYIH>7>fd<1-0AbH@ltWq>n^e>m_!gn>ppXDA+tYS_WrK0
zqp0ykp@@Eb8)0t3M%!FmZ7R31G#r4RqC#)Mv~A!}JmEg0AB(_?Uw0Jc%w<Ly0jMrR
zxj=98`H-*2b{tH+qbXCSY=A-4?r;8oq<J*oaC!$OZ3kcnWtnehw*yj`dZ80Oj*PG}
zG<vrnEBi_u8sULZb1wAt6z&^qdh2W{mwZ2D$L=^cF^FTlt<%1+)m_>%QMLh0{_y3?
zFXWhJR`e#LkqS)SIgv|W+E4OY`5(Es)-2_!<EI{MSg*CY)9aaU3UkMA(^BoMabm?c
zSjR}K2dY>z7QDK%TkD6?IbGU&>BqKhdr%p$1bLl^IawlF9VFB&Z7v<8h-=uwByao9
zEm#_!bB;p|cbCnBn4{wFbKX+BRxKTg)Z1ORe*ON9+kXE3)2hnedTERCO}!o$t-jTB
zU!yxG*84dhA_+6b;aN^tLG%`A{hL^K01$3LIR|8Tjqo0u@i1_plVP4lbG^5jyyI@S
z3d>5ll7vP>%13JG7T$o&veLZNzh6H&?+}xm>Km|)l6TEq`=^&%us$F|H{=um1_Mek
zu5#=1g(KDaydQn(VJky(Z-nrX*Y-_V?6GZL+K<62H+`k>+v@8p%_=NxeZ}Q<-!CNf
zIAaH>z=)VGK4GD0tLwlO4EVp12?5UC>hNqn$5SCohD1q@?Kusmb%abDCDf*(gdwi)
z1{w0944!0X9}O$|o)Y&cr*rk&^9|<`O2SoAOF<zkBx)^%$OzaNft9vaT^py(PBV8K
zD~jsYWYo~FjL%yjES_uRvKNMS+~%ALqX?tL1mlEW6E#ylp@+e`CbR~JOa}E!b6S~Q
z=iKm)2}G1Or>3LpdsE6$Ap^`@eSDUlJ|Rhbq}Vrv6NRbacc(UnaY9cili?%-XI8Hy
zoDotuafZDeqR(INHT$}IdvHR<21c1M4W&q!1fb-pmpT<b789J`tgR0FUKUo*P?8Em
z?hceI2{f^4Kkdfy(W<<6?o6t#%BxmVP%tq%d;-tjPe0iPTmYGOl$uPj<#tCoX|}(K
zzy6x#0k=l~CNbXW+t}J7Rdmid<z7j*Y%=qb{k7y<w@w%*KbBmCit)!P4S~tiQdy{@
zMvW5BP2q>%|E|V~u3PYFahsmDC+F+XCl3z~vLehiE?|5=^n!ZR@nY$hZQ-93IXUtg
z?{*71@1Xwo-&;tai?K1{i;Np7L~xhY@H8bNJzz+PVgt9bXhabj8dXa6EQxj*79u&X
z`Kvi|PY2wfXB<K4HvjwQ*JpC5Y>>FC;?sgcUL2mx6WJUBJzPtH!&}5piUiaqGSNhS
zjC26S2FZL9gj1ebr2C)YXPbAuS-!&AYr{D=J%_mt9Uy-dijNK*n*0Bp?lH4V*v;dc
z#;WEH%pP*{>HYX8ktdg&VtTF4+RDP;{A>9p8f&}MI*xEf;OTYX@bL5KhTn-|D&DA0
z))#ytBb~yNQ*$oQz9uxqF1>EpFAp53$3Pnyj>T0rd`?wQCIGh<7m=`eyn4x<^*hK3
zt}~g?ZK7B_KxYJ}N1G%`)YLs(Q>mvSk#8NR=0L4dR{`w00QDhbe)5;oZ5_4AxfgkK
z#i$0>&Gfc7j>ziW+dTJb%YMprx3J<O)M64(hW9LcTJ2s{cX#gdjBKliXRfpv_U5Ej
zvYzF`mHOS+r@j9WH1<o@^@gDB$VF=_4h9UU{&$2MB13U0NlZK#yToPC2FiH7Mg0y<
zaw9MLCd=sVInc8izQ?26mgJl@fO?A(eUn0)vpG5$U7bP!6lmj{OPUBF*CQ4!I`Ocv
zx@?;uxxc}tDY_-)i?BM;Y7%vN6ZV~eE1SnIU0QC->CzIHd#FjyslHJ&KfIpxXb?W9
z&rwZ>dCWwAO_b676al~y8dL77J{uEiBHn0*l}H8~45az1f7kF}VHT7*IUsK50}TNE
zB?G>cLJo@H1yRQ0XY&xDD2~p%?oQ2O>tWxdiOyuFiM&l0Y$uyn8itK>u(l4(DG2+@
z(8XTge*6d<Isks>@X)m?5blI8*(NTPx0p=L(3e>YRJZHCq>iPi@T8mP5lK)8@iQX2
z!Y-4Mn^z7Tx)6X}r+-xY<15njps)Ec%q*yI>7Jjz>Hj|0pL{bIT<Sesbuyf#2I0==
zNP9y|gi&(?Fa%L4-`D%^dUzDlSV?$kB9|7qt1yq1WoBkC#y&)UQRCWiZEhab=?3qz
zFHFjtQg&(0CP5L#wKj~@ovo(p>(*UA*=g8sS$F8Af#v~^E?Zu=hFKm-=S^IU6A5<F
z$)>M2k-EB_ULkboOu)4C019T#sXY@;%A}#X3iKY5V1KZXhVi^jMvpHTh)GltY-T5U
zQGNrj#>XS>Kj|0QaLCG?=JSxeN_Y+0T5KK2CJ7USj$y#p4P5_u8ZrNDG2958C+Vj!
zSON%>cmzXFg+aJqpQwP((kZ|ej6{Vl&vEy-go09@V^gLi5<oc9_T2=-!JC?{-{y+}
z1nzYdvz!IlQm_&6r<v@Xr|$Npg{YbsmUEdLlJ;W*=$>X1k54ataDd~HoE4~_0PE|@
zwCW?yLB1WT=4>_h3-^p@G1j)#=g@k6C=gsfL}y8X!)JZd7vl|*um$L9?gzFD35b@E
z*xKcL5;*E|&H>9d1w)Z0D}7H~xlg-6Ed}c5qRDW$!B97`KOEH65oQIz(2$H3`}Kfa
z0I~prKWPkPZ^{U$^o->Uk=;OnF69;xf<7On&df9I*j(rGtOqQz{7;{n0si6jEnsQe
zq(EP(!m^ShuA2J`fd!o&N`_SHAcf>`iO}G}g-y&}L_PLwzV%>w2&#RtRzfkfA&}Rd
zAu-W<uu`7GJ3jiyQN&>km&4F){l#M5#KYc6`o)75u08C|Tr-()0^|<$A~8y8?M~}H
z+%1x&B@>y!J_f^|z+ePYP(p!XON);wONPLeJ&Es3{UJ4^gszjtEWkT#JcNt{e!<ns
z`s+6N%e*|3-hKtJ-8V&{DL+WHLr2S52E3O5w7gyN^!Mp?KGC`8ul_Dd3B+Mb^tQM6
zNLN?afIh?bieMc2#mVE+LlNacanX&-s4o1p)Rx1a^|EwJO3hhxLG`y;OLosX0{*0_
z$B}<1nX}lNEe7D!SP~PXqjg1PPg}-0QS{{+<;bXE*S@`MtVrNE_qP-G$V)l~at62&
zGQ>#TxpQq=Yfo&TXf5j_b9sp@N<u3xM%0kkRn_mGtzOq^?DJtIzFj89yq=rzEUB(t
zLQ?QZyT`R}MC~f<*telZaGfKKH+m=x+=I8oDe5<zoxgLlem%pG`oZ<)HQ2c8k<&k?
z+SpxLz07g>)Z9GvNWBAH&3X=K`_t|ET=$J9o&`QRG{<`Nsy9QVj-U-zqBg^ZOt%YO
zTnC||43$Hm3xjp^VvAAp)@9)-6Go7w1V1ZJ48@xV&NwM`6&oO3Em2TjX8OF?LL=Q_
z^kLGOM_`?95L(!BYLER58Z|Nm9FMWj^`vJYreS!OAD!cKvHP7TGZOA30ooIX!B_%2
zgmVM?(fr!1{OXy2?Y4QU&cTx>7H_p3<c85jym1N%%_P5s`|}o!=6Omjs}JFQop0&7
zS;jo^N4>BNQwd~2Vsi4>V%2c0*r2vBK49GIyp-q7-@(9n$zAyV{X2lgqXsL;B;CeK
zWKA*ga{=fwJ4N$eF~3}__Z$v%$~Nul6V8VojYD@cxDR*E19#t+pF6p}SX-hOWBpjO
zmw8mkv&SSaGC+#!#no@B*Qop-U;41u6gyNM9;tj^YFo>_z1K+;TelyRNx)_>I-+k@
zP=z{9gE%S3a>1Zf`plJ22%kumle6eM=;zI6U?iv@ly|CpDJe++9(!^{SE3QnYRzT}
z1-Ix;4tXq>+mXo5CqscQ(L9%qr=Bdc^o)!JAn3Fu&CumVIu(o|N(An?HW$B8s>BZ}
zzo}wJW-~zb)TF@e1PF+T5$h$v7YVe3z!?e!d_=7xwNSO(>+6RRB#M7|VY-2qd6Wp@
z$)>o$=|z0zRt^yjzB9DdB0pg=Ws8jVRuN;AbhG4;aOU9ZwqK@xWvNf>qthpyZ=F2M
zC<*}paRSHeK%s+4*=zmJf|_W}%xlXdI$Nd{#y-fbfKtO^nRxHs#KPJuiN_%9P{c@x
zAMFqLRver1kGwJ2kd>9Tx6i*Wbs@S>=EsqlmaDm-5ughKN0*BCZNYt`RllRXVq;S^
z6PxyILC%1|$2;eB80qfX;`3wg{C6)F7rl=^KiIh0TwP{fYpx%cyZuL}<cti{PyI0L
zDrESN&k^$<9A`k|l0E{Y%Slg>FcKF7xg5?0fi@4;VLmV~Jimj-RZ)$S|FT=99<Ziz
zy@6Oj78EoZu*g2RvU>ot&@6l67`QqzSCvE4%F^V=w{L9@pMmT=++Nk5qC_`%cW+@b
zf|BYZzRAgnz55!c%yxPr#Rjz9yb}M}j5mA%F!sKf*&O2w7r82*Jbn6N`O6nCl;j)u
z;c=0CkSXeKy6EUc-F-c^y5uN{pCP`r-{j{%di5Fk==)_uA<>{C66dQ@urKyH;xy<K
z`GZe1-WLVi8;6F6%OKZ{ditlheY3qykpvS1-PY)M*W`V?*tL_2(JY9qtHPQ4P}yLc
z<K4bY6+l-Ir0F;IxaIbVZr4MQA4WefK#+^+nLbrZU<R)H2F;o!b1H~dVec!_*Bh+n
zCSY4|jF0*HkL0+JRN}#QiTu>CPOYD>`>Y?qY1622<LEIj9xz)RcxSQS99A?AUb0Ic
zoccs8>H4bTd2;zSdGmstUa24{jGp8z;e$(MN{feL<`zCiU^@`NjiPrsIW4ktgNYL+
zU`e`J@#t0T<R*#vbl;*nPl|p+)P*p%`sCg?#+TMp$U(Sl)v6?R(34_SG(GEnygNPX
z{Cp)LVOS(4PX=Vx9AiZsW72(dXG!R5uJyQo{#r;cDUU%cV#JW#3~~Fv7<&(}ocq2H
z_)HX+k}V}kHVu?$Qg%a$(4awkXlkH{Ld&d3CDD+Ub{dijskBSGkj|E*XuqFt*L^?l
z^N!;^-urlt`?&7s;_U4I|BdhWv%caHzUdKGpLW<&N;x{8l;#r~7WYhH1{>M%DO&()
zlqjx4NHN@w^je>U8(7voGL%t<!hkjaa5XB0jHrKkS9%AMhzss|OUYt|$rM$d&Hv1v
zZEWMjpp`f_g=>ksr}{CJeJF^YOg0kojv5&)CW(aYIX49q*Y7`y>Hw7}2z?2sDU48h
zKP>rYXh@tpCLN?VAZG^hX1dpP@_^r5&TD%jSOibOXX5*SBSj_*GC>4mQw|Fpf@%mV
z+{&{dxFQf$QkH}N%FXNVMHz*BlNoM>TdO|<$tYcUz$W9d(0coET?Mpnbqm-)+05se
zZE@<viK1qVQCh7Mo>D#LHv0Jqvafo@q7+d*zvc&0WnQ=Qj_RMp4FSb!Cr|2%@^i{u
z8eME}pL=139ZJu1esOVe)OagZU2|-fcb&wPAP>D(h+Z+%e0k5Y`k#Tne*ZRql^rW&
zpUHMw{X=Of8^^oi<DtJZd3N!tM;|lK642q?OKyF8=6wyp!)D(*u$Z6>F|~P^nJI~5
zmFZpo+^W8C%goFp5gzB>482W{i15%k(4zX%1BL(Vthk-3nK<}|i(W?y=$LMM=&kHt
ze0!ybp0}^~aV@RlxXLP3QSs+TKM2RiSRV?YC^$ahv8Dub&I#^lacTebsss72zAe0$
ztXFR%6T#DWA>V87B_4mFS_3%FZcQvO^D_8n=v++U+B3b@N{8@sW?U)0aL_2fV?b2z
z?abJlu*F}qyU6VR{9Wl&GHCzc+Ko()89KQSk@mlCYznSD$Zz!F+S6=+N?R=OvW^ZP
z5hoC-n>78Fl$6wmgk~pVuF=7=qU<Gq5bOqUkLvPXrTmF1iO<WEjzIK}G&9QmhZFrp
z`Rs}WYvc`bP@a-2xF{3xr!?D6Ke0}ieU7lE=CaY`WZ(4ri}FQx+I*c&!7{PV;4$;D
z(MY6_dL=ET`+eZ=3`_M;NXLEVZOe@6SLQigtAV`qb@xoZ9l?J6<AMSd!FB-Sk<eV>
zc)MH`t!6{<!{VVV48tD%QeW3S;_;^iuz|U3H<M}S@X|QFB`TL)Rf3yY`ExR;IQwY6
z%U+#tnA(&R1YCTiBBcetL{v<SL2+(d_QHU|obQdUwk8+S?YoyY1Ww?xS!}Lz-efnz
z$Zvn(NGldwbMHf)SbC!%3^_ifN`(+cY|j+(Qj}8Hath;h6hWvYB%~eaz|Oi<Pe2V9
z6}?q9qT4J!yXI+D)aH!n$YkbvWv0jD2w<#aw1R?c4=$3&!NC}WaT^;HMJ^A`RW>%>
zKqINZY9ybM{o<a)s=O4_AJA=_UcoHfr{E))iK5C)j~i?oikj5|0xW2l@&T;>j;l-%
zIMvh>Hhb@XwE%k*YJY6B7d)a~B{PX0<OQOL^F2GDragOLS<bJIn;C8p7$tIM0l@S`
zha1FhFl<tAYstyi4nh#-h+4PrB~AjMbMpS?YTE}tn95zPt!}7Bcp8uV_tm}$dyM6&
zs-;D#UAo9V(Cn6T{K-H8X6}B@toBY~mu<bFnYu(Hk7?jpZZ^=O;KX#p&mI8GitKl*
zggb|kdcA6X@<;c(-{F#jC{VOcH%@f}T(QcJPRM`BpX+iB=+UtLQ$+tjS+$nYN)&>p
zB)7ZS5vfa6**_kQK0PK8Wphcs76P}}m6_;Kxg{)DkP{p3782qdK;YF_a*Ft%HCn4+
znLpJ2j1PuR&1AZ;dS1JB{>yq~#stfF@%7qs4MF}4%`}Vm4~u%_^1`IXI+V}N0ST;L
zW$UOrnA3NY@6UJ1)FYV3VVCeX0JQbeSMu*ES8>f%lY`%Djq;YINhasB56+oATUX?A
zI>$x;fx4Bgk{jjEpBHrzwmB8qwzXibhP|!t#xrNY1vxS-9F?Ni-`_38r>?#3u(6-P
z3$&;`TMLF<2KEWwIfxF}^862OrpwsBGkJ^8=3m8<IFE>qdEggN&T(s}f*e5MjLr`=
zJ`EQD#k6qIqAQ~<#aJ8q1MLNvmL8MW8||uPKlB;bGaK)!YM+YHx7N|TjUes_`308^
z+N1-zlRyNea(m>FQ<c1>+3Q3NL%}u8g6Ir)I2&dZ=V2ANjfNxw!UjrXnE&2-d0zEG
zq1d^XG&K*js;V$oo=3j{Sh@CfFiub4{;AG<%Y$SCESO0U20$95+&uv!b!uYEzr7q!
zZ3-wnvIxVfj^os&7r%As=_x3%XpF$yw@*J5en<IZ#t|v;4#nmVm**a}gWCHKW@0|8
zlSV)&dS^%{NIgLJlgnz%Xz9CvtK;v*oV<jK7z=FPz#7o}3in9QW$rKR_zy@U<aN+t
zpaNMe{&HPbW_o&pQ`^EBr4TEIn`itS9yW|G8fw~14T^caSWiad<N2KUo^1U>ydK7A
zrXTw%=m_wTjuEo(AKDNbgfyKIE{av%aP8hbYc3O#CiF=Jj4tvj=((`Cc0yQ01oJsy
ztp3{Q|2%HH@Sj49Z`Unam)zV-V^~N5g8?t@mt6n;Y}>FwRJibf8Pxsk%v|l?YDGRK
zK7VgQSUo+=3sEj81ya?L#0SawAKV{s|9-+$+N)RXD;iL}9q#;~8*g;!%5Gh|#x25U
z?gKen9bPH=Td+0|SeM!&l+V)PgJ&%)O1iw1N)QbAL(DtdZ3iA*W_f9yp7*Hl#!Ei}
z+v5aRs*60OkumV_9FUj4n&();kxyjk56t3<Y`cYnNlCG(vRZXZ9q4ldd^_@u1GhVS
zb`5lnG(JHeB#XIwArPva;+tN$<_2Z*8{Zn`QIKKZW58r%8S+V!R3hOTGSP3}xx)gG
zWWkF!ftT6`M^3_GB>T6Ol|TiYTJM0SoQGH(S1~1?nCHveIv<R&UhdyeG2Lp+ocGNR
z`};}G)?q8xv7o%(6bHWO+RjI`=qe!`2+q#nO`VJ8D|S$I)%+*myV0>IpmacUi+~mO
zE4$y++>s9hxwK7MTKY3Kf3}n8uyZpgxE~|yH9MS}!9~IqLsloSPGXXrILxt#lJd*3
z4Vu5o=I=H>sN<b+zz#$-lDyG-scC7&*iIn=l!T(Fr4#wq0SN3Ab%<)|Pfb4{CYEP-
zpt@=zCGpSST)%m(ms|ev<xAsPIC~~yoZQs9xr^%zUc^ebbg*+!hliRI?$5lOY48EW
z-u|hSW4J^ZV!Jv_Oied^Ud3jqh_B>Oqd6qHN&RcMxLxn?#mfUwwkx8Tq|5^2pG{4R
zp*NvmIb1>IH~jVk@-uKA>rG=}Ks#0b+8c^&WUG8oQ8|XMOuh%`lTmzE|A;MIyH1uz
zY)03uLs3UD6H%#VU(=kud<W7Jp-`5`5H7LvS;%cegaJy1H-P6j7(LmywvIs^?+Ld)
zvf)5{LeoRX_W>Y$MBYPh19ZaS*NLH-@to(3j3|v&-=K(T5Vb-xT==(%37_@G^*|F5
zYmL4+xPjU`UjAeu=KGo&f7?;*wECrhg-bk^qWqvBPu{zG_`2%>K0|(6X5p@507uky
zATrnySX)=XorrR<0)~W344R63U=zT#c9?k54<IdgOAG(M0#?I8)W$HZxS<WlcQzg>
zUH5h0Z+VHI4r_46?q0M^9&H1j6ALqcb~4-0KE?b<!-)3>6f+#19!kyy71LL>=+jVF
zrKfrwM^zcqYkz(;m1j!0<|Tnj$&UUI++k>L0CsC5Y{B*1&zS;QnUea@270S!qK81Y
zQilm+8a*HH-(T$Bjp=1@aL+(KQEbg)J}^b(ctt{6*n+!<CdTYA%m9)4v%#L4i-;Hu
zZ$$qOruYpMF5=CdgCpjVE3v!*qCy*{2aSzp*f3p=G@ngoS#|cM&6_rf`Kx;}K1Yf(
z<wN~z76jPP5`>_i>>#oGX}$rG$L?IRj4x5X0FgUUuq6pabyM~pq$(6c4ifzafB`b9
zqQ+D}9fOg0jfPIfRycCWZHUna2y7;s5ttnsZiMa+uZa#4^vue^)5$9!;K0g3R6uJ8
z%_z_bRjeO0JG!lJt~{1h1mhUa*6dIL)dJ~SVV#%kXlX`8h3*{okqwb`VjdJYhDry=
zGtZHQ4O0BEd-B5b@6WFoiLyHAv)toZ&Z+O)q_(4|s$A$*P>1|9<hG!@xe1g5-1ZEB
z&?sNXcZ06alXk7Oe6@@e?+)#49PZ%MlN9<gs7At~N&t`SAcbE9%H6i)Nl>w*ZpKmf
zN@hI@(Yl1Wz!a-aLhVHwM#x+)UV&D<4L0v{a@!3<DUAxwVm`ju1((|(6(qC}pFtyU
zupmK02;wo}VmTXthV?CglLgC{e;EF*UQ~0etbPUq+{m_*OJk3t#(w6T?q7*&9S4pV
zik^faZ20L>YH*-qPi=EPlM&%F+Hp;Bw9q_Bh~yHuXSl^p7eiUP72`7rAjF#E#)Qn%
zmq*U2Y_$qqPcN3D36PsjY!!GvdcgR89++#k1lxemP<|I~4ATG53nr-oMw-!_NW8fN
zpo#KCjjm;h1{y*HpvMq}Osb@5uO|{6KEW6-ipfmJ(_UYmaAsp9fob*m0}nzadT%zy
zAO-YXj5rz{2ykgDa4KCergeXSW<CVZ)1?&o>ebyjP;W3qk#f_bwU)n-Oi$bl4u&Dv
zq>?X%ig^M(XBta_N=oMf&mZeaD&uB&!>&GvKlVscYCnKWZ)BT5eH5;`VP|YdXJ-*Z
z_t{aCpiU@94|c1b3F;n2ORbMLEavZui<z6D1cUW<NS+Fj=7=Z^ecu(k-ZFFnm*zK1
z0OGjM@&@5MCeNjy=4WGcFqt2b{d<Gu4GOC5s7F5mTdn(shUp{YbEbqQmKLj9BF2m0
zk#XNSSoL)Xvh1<?3G~C5I(7+SZ4QP9f+T<^8v@<SXXrwUcJJ}yPnPP)Tn}|}c7F9|
zmwcFNN6w4P%twhSDd!lXsBn~^LS}Vy!ePFg@d;$^Z~uK})z@W(q6Oo={RS(yjlm7Y
zDc)t`xvj_`-4PGyO~Nop8%hpunRPf|9;I-~O9qh*9ATzLNU185zTxue{#T9EAN@Ty
zZF`1!Ci$8#B=2kBlQ9*ecV!4FEJApj%EE;U@rdxBq_py_D>ITQ{8|q|u#%qfDDN?s
zkgyn##h6lQvTdDzBHVxrAVEZHt_O-~I@uXnSva9^tFgo9Tl*-z2ST*EoJ&%~9ncU5
z!<Jm2+CV{QiWr8Dl68Z4s_T@Z0eUChp_c~*0;C&5@v(Qg)L$)Df4PFplt|Zs%?8Z4
zih>MOc|JJN!Rp)HST%*>CH)k}v*Fu|B1r`AN&&+oVTl7x{RR+m^16+qFe*|aG!Tq-
zaOFiPjwxb<1^@uoMfUT4g#F~%QIaUFMan>giU|`8tP$?8<8rLPnTrFnXdE3P<H*X3
z=UPM=b=58>EHLDbK@(A9Z`%6QN0L_@@CXSH2;`<-9gTR8)ssiC(BLykgT>el(u7c4
zT!xO~vssJ~6AM)hpyuU(uPNn`#BWn$xl`Luvi+au0`%LIstBl?5+dR2e2cs8JW5Q%
z{>`<?K8IQYEkF($y))opU)yesev)!&A|GZ;-30*O1|MGK!+x@PgO}Vr@yQLx9b^8o
zWu8b!NIg7?|3$Iy+%wiJ^&(jf?AqF9feKI$LIh0110VyyS;HC?I$(<x0f6Cl%~?q-
zzdV$5-uMAsKCEZWWxW%vT4WvSs&TyhKK)hvxz$1b;i;k{!Pi1_U6TeCW_M-DPL11=
z2M2utPMh?a68QrcU<}@2Nso=<&y2;J5%fgai-ZXx3`YF&W@?><F_+Czx_Ry*h2*tg
zZJaz;SRS#j8l@30>2Go7X>PqdN8k(X&5+GrlNaVZ4$kD(?o{g82s9c|@lM@3b#>!g
z=NLWxI%ghVQ+!IP*}X~j3HdPzsnDR*Nj|VySk-zBBa*MsM~9;jBv0|<UAEXvNg8Fx
z*$YKEfQWD+UJFyh)H{-B&LTPNplHLw1A%%S^%KRbkt7OtvgTSYC~Zh}m}=ERvRO*{
z&db}=&y@n#0uWI!tnW&jZ;sZR;fGNaK0v1(OYfioik!R9U{LSy)6u}YC_(DUeMbN&
zg+`Pg;BUi4a238>+_oSbDAU2`gGN=hF+C##yZa{}@$ETa-Z-mz0dj}FOn%#;QjGY}
zq(!F;JI^KF)7tT^>Q9k1bF4$C)WE1Bn#U1UEzot`NYT}|M<5$v3@PlbQgEOP#qbk1
zNY}9H(65g$yBHwDV}Ib7n70~RS_Ty#lsomObQ`wuN9Y^YM9)))lHezHGF>T-czaZK
z0r|!qVY~(L^3yI`TGI^z{8eQ6!nU1lA{G4N8{D5P3`k+cnngBkqRC&B6Nl+RK4mmA
z8;|n|x?(UGy^mT-4?nN)=2`uW?KCu163CL126jkBBjh1N3<4)R<?1870_eE=N{|2i
zm@icrBa{54iem)Am->}i;6JV-Gm-b3N|4ifo(0X(eQZKowz!SnEat%#$HJhQ1T(Je
z?%2LLre*!#qZ0G(rLI@sAbwgse3rnqaan!nFOCRmb4~vSP+0_kVB^hJ!54?yoWVd(
zu_N<J;Q7kv86#+zDG?f|18klQH#BrRr1Nn&rqq0R|DM5Ize=!IK_l|BqtFxlM*lt%
z`cymSG0rST0-|TICO}T`Kns!&(JPe&IA9QCeadWvEFcUt{6<$B9aJhd!Qmkm#`w>7
z!lM<)w9kyYgQvF_B5mUAP{J@6;8GGTf`)rsQ6EDXfB?snp25mP`;MJfjTp$Zt|N^;
zy%_x0TOTx0=H@RCRS07q!qgOIyM*CBkPJvlNhyG+wPCA~h`&nh8#4p~BeQ}s25C|j
zWkbqMbRCgY5XH_jpp=nC@pn>(w&Wh{5EmzR%zu7>m><ljEtA6%_nrm&`xj2SQa;+v
zy^;YW2^RmEx;xUWbru702naIRyKkR(2BcdPkh$hM6fYamnN%-XnY9=RtpD?Q^P42k
zv5Iun-MSTnSH%i(*nj@r!Qx$#QU;CpKa`xF`g_Iwc6GH$RxJQ}gLNvOGsbJB91xX0
zgCRQN({=v5-B~$k#CMLC<QR5dAHTxmc*s*-<G2}f%k*d6p7$NUC9*``9+8{4hT61S
zI|6GHM_Hhpt1JnS`<!OF5gQB9Ew)P8xdSAE4oA6znRiO}ie<RQc7d`ck2iqj{EEg0
z@PFbwCNdwlP5qyS7$u(C@<G*QNCC(%&11P2F%ezg9lFa|DdtLTdHt#W9UcA6*E1j|
z(j7Nb|6EXTeFED@jn!&%*L}_`{75RHNa#uQslCrUB^@OGd|7KXGb&0WkUAPXu4nFQ
zz(b3_>2mMy-^P)TAr}f^qD@23KyL$vJvAz{_TGn`;x}Zu#_BbkV8beaIe5W;E(#}S
zAMv>G&M;vl8`Uy>c$I>$tAtgG|HM(Z8nQtY+!3Y!l!tR{Qa8e@kz5mf>=tMRG>t@x
zfFY<lMd8TrL#M`#lpcxQbUx}JAMwIP^7}zsR4*K$_@m<DAZBobIuc7YWnCE1tOMxn
zQBu7TogQ-lqO6Ebp;zh^ksV*Z5(Nq@8_kKGkn-uB!EskEV9CLRLQ-R7UAUpUTMm#I
zpmRhiPxN|&Ey}^OD7+7<BTpKGwsq^BOZ}(^93dB}tHn`VAklJ>GO;l^ylU^RwbyPh
z^E&)|TW0wubd^vzTllY!!#Oi^#+B$KJyqL}>Eg9k$`LQdd~H76={$gVIb;zL5=`7{
zbu|rf#26GB)`_YPIl91!X{0c2ez@?$!0h1#z<>$$6<U%5SmBaA7qIc(VBps6M~^;!
z4<($N3-r9Pv9agKJm>%=tJOY<a*(}?L9G?ZDd~p5^MjmH^`7gqi4D<-!Sm%$B6Q*n
zma90=1$%&adBA28&@Va^Lkuf@gl5=h0X&Angz(IHD|S9v(x-<xsc=S$fZoNFxOwPb
zAdw>nYm}7|cApGzZur+cA!IjszKD;H6UMy<Og}XfG|_TXBV0809QPwC(>91vAOF%p
z6NTC^4-XFO0}9sJ5G6OBZ7#xyGIRlm(WfyoJzsZl;`XIHB2ZCR0s@xdTUMX_>lFcU
z&q_4WINvoxDJ&1C!{2__sh9_4-|t%k8AX?eNex`CuCC}IDeDA2%GwQsg)r~Vc6FUv
zSA0$J6h;D5e*O#w^)vV|AlJt3nVp&kGAm}eVahDkG!%%6DHGU-7}*t_A{$`qZZ^Ul
zl)ANxH%a+vD2^%8u0UCh^+rQ&01x8NyB~i37W5<q4@zUeH8K2nX;3>M*o+Ngen)5L
zw`dq)_)bSaB-nuRuNg+0AW@*iZoF(1pfg}JqCbXkAZRxL{1Ts_t^GJU>F~jrAID@I
z@iZ}mny5^m7lD#ed;NISBAxq4RU5p^mJNfJXaWHxHjzRB1_=cZj&jW!RO!Tkp&0gL
zmVkHQqhx;%;0UY?_%!ep$@M{B8qRqA0QC)v!N<HMDhJgrh11E&@p~;3nT}PQ9CxDA
z3T+9s@(4Ydg-~Y_MHuk<bcB_H(#6^DiHZhfItkLCTWv+50-m@w$O_^(WQ0F^{+u*R
z<Q||dfSx#!9jIZ&D}<o`!y2VIj(8Ac8%9!h9*mDaA)X&jgX~im4J!a>Pk@mqw<zHS
zK(ctu#q~;oZuoQU*xy&;<KqjBNNK1oT!d><>$x`!o@h29>Gb`#PoS|Q_>~ILj8lML
z>^OC2bZEmRv1;)@_WpBfU-e<-qm-kXm_Ve4h9Ex9FOTf4l=!+1OCX65JM0=J?LxB*
z1TO)pa|nQ&ftQ9=CEDI7`u*hBNB=IFie!lH1;oXdK>P$j4o~<)1itBFDjB{@YML3S
zTlhEdQ|rG0MTC$K=m_2khBDN&Ip2wdei=Uu@Ge9-S0C2%mz;r$6P*!c5r9I#{n9Y2
zLxYM3741L1&IP1`bN$(z<KKgs8LaR`J@4}w!(171?cy)N6j$?l9M1%EKH=VP_RUy>
zn?)H@1fb`I+i5S1Gf?{>I+Y^gAm^AucI>UfUs$z&p^c)^rohmt%EAt&iw~d(_M9Ks
zU^L2mAjSc+t{mZtysn;>Tz;>mLi&|u|05u2`kF~(*njeTfz-Ca+72^%0djB?ko<+}
z#^+mQWhKqJ04X8>1t>HI)V83KB6Gb%1|B-aOrieW3t$jfG)_etE+JtScNe@kaxW-h
zq;j+ezau>`l+R<Q6^sAE_Kc?m3Z*8h446$DBr=*#l;DSn1P~yp9y@jopJMg^?|l-I
z%K^lW%wRb5HQvOQ?!FZD<!+{x3E%&hbY`Zvw1oGrr5VpG%mH9V7o*+96xx?V0jOn!
z!;eO*s;lRlT-^^=9&jb}p-=E8T7N}_{?N(x%`Cd0PF%Bez@auZeuq61k?|5a5SZdY
zkszV3uT*x0aUdb4IDb<IY7pEbA9g$V-O&khnL=QZl_ee9ua*KzX2+32e+%2;*Lp|v
zhb~~B`0ekAI~skK{{|SZq!x)$QI34eGn*bjw8%3a6BoB1@0*Ce6DN$$J3Uk?6fP6z
zj6U`;PVn9b4&ivUkQE2a#9mqn3rbfNsdpH>h<<fk%yDOdsJGg=;&ZvY{uUr?BVQdC
zIdlAYLG*=8lnczAR-1!gtSy0yosbN#$9B@&SgspGbd4{E=>P_x4Jg<?(jGf)T7k?X
zQ32iXYUG!DtA#J#Vc9qvH6BSY{rz*xhOjLXk*>Ew+HpPwKmckBODu&J6yru1jPrJO
z?~s%<?RrO;8F~T+-B5I*(ELcuhT7ks>H)}v4P`M5_z!&4yI@2_v9_4(CQ!KH2D6VF
zL!LCGXYAqoee_#sZbk8#$*aJy;o(MR-C=`ZEV=|fO#q~1UPfm|;f-i9cV9X<3sfJ5
z#U*9`ftcqfMxoB_&l;T9`RX`kO$2T`_YX!=&$=$3w<ChK415OWTqQMII&fhoE|P{&
zUk^fa8odeY$Snj~nmk(qEi&4b;URYZi*67l8{a=<`|-<d|Ef!h%o}>#1ooq?tVv%<
z9}XnPkL2_)O^GraTQz01D?Ju;Mr>`UItRWNK(p(;a~eoMfFN}Bs#R}vkCf5`8|ZWy
z$cl)zJ$u}M#7Vq)(97d&04%W`^fLpg5!(UkazxNgHarFA{5BH*_~QuJgH&J;YO09+
zpb=b`*PP5ZD95BUY*K74^UfJK;6e94nSA{mR40kx5yFAL2Gjr{F=7=iklF?9r;wKR
z_KL5vi<d1E6cD&lqJT1oqwf!H%R06Ozz=V6r=kL&=}Eg+>G}z!U=AhzEGq*gu^M1w
zs9v%KTcJnMO2wC|OEC+`leV|Bx!%}{^3^ywcmz^1vHRUewY6XV39Q%$y}`BNLzwS}
z7=&Ftd&WTj6#|#zy!;2Nf()2Z{5N%9G`9x7U>ET#%11QZ4IE8f-Whu(**4PBosaFx
zVBjd(6|asN2~{mW7RtZ(I@~mNxaEujY8^B`<Z?=}`~}L5##93DSK90YU9DNx@XwzN
zCa*HjQHe}wHiMmnHg8ryMl%ijjy-e(xm3PZxf7`$`g`D$04#i{djib?%TV?_US27O
zmxX5b2GN;&FkWW{bZF$~ui0BUYU*D7Uo8N*0p~&e_cV+7!Qtn3&e(IE(rK_bv>XE@
zvwl}7@)SSBfSR}0=HTvssc|If#A&B}9Vx-Ve4n2veg5>y@~|Q^2Vy?Ru>m<j>1o{b
zrx~^_!n}NBCDQ%50+RK<W_+*@dM9dMc37eI7D{U*cU0y>>#-daPj-uY#2b4#wW&)_
zH-DNVP!^<efju3lFVPk_0E_Sld@lc^_=Tl%U`FyH`GYR%fBbPMbtX&aiz2y{pTWT3
zX_LyV1@bpG21`2~jj>i0TEE`2cgd~M1=KLn$asjIV^hY-LLM3#Uu|0qi+krV?;n-z
zVzw6{08r&mOW6wA7fFywABM2b&=)|O2$r0tdtn*2A@0Sw7A<$VV{EgCeYg>}{gd_P
z!C|MbE35r^lnsk3c@V0&y3-O8KJPNISNsT#;z^@?b#N+V!~@f1*4Qa7yeP$kGp}T0
zcSQ2DKCN(H<jPXw*4w;17CgCkcm+P>PwnP)vE){l^{B8^8P!gW_B*(y!gTNcOA|2)
zTR@b7$BSjNzf@JkUt2yCLMR=7MIPojcaVnw4UfeIwRy8U{zlL;552mTgB<w@#nDq1
zf&NPC=>IEuq1<<@$???CW#&JF(9qXqMdhmm1_gEVT`w0E!3Q9EZhPQ*`;%-{{H`86
zk9a*)Rpma5KGWMU4~3<B?g41?;2vX3YnWd<7}oa7$8U*21&U}WF3oL5Fw$}O%EEvN
z%fCPOOz0RM#L=$%?oVuDk&{-J9cLD_Iut2^j9b7l1CeqQa{$UKKwljs9d^bAq-=>%
zWDSc^E9!}1h<c{U1TtOiUA>iCZW1jch5Hz(tVjR!AUiv|G_1(@^LBORCdWnm{4HNu
zlW;xw@G@&M_o#CnPf1I&su@LSEXu|MxS$!7#6v%tl(xXqc9;b*|GH&b+BiwCgp1)(
ziVhqJBoGJJnTIrMaL&_7d!e#AMheP`A5Ze#joNC@x5Ibh{o2?X;cgtKW>w_~lqDe<
zPT5wJAvA=Z(Ta&4@6e%R@q2INS1|D^&=0Tde5HH|!{=cW2_L$Q1b5zTZ*b<J)v|iJ
zcM4@(nbWBh@MzpASzw2eb?T$(sThJX4nueS;7CEIOIVp?0Qfk4DGths#gf6uJk8p0
zuv>WP@U_I0-t0rxmCqMXWDRobr24o_gy*!Cp$&vo5ALQa$M+owhX_%?yi01)@Z5)X
zZ1CE<+DCf1vuLf_@e?PKxZ1DD=igiS3lf0NcqTrx+2HUpy!e@R{iQj5xN|YD;_NZN
zfoM&3dRCb!23u~qsH?jQNmjF~xCD108<f(J0fN(TA?K-sV|r2SfK3=f1AkOq)nU4e
zvHfupXF0Kq<hErP2d;*(=1Je1MEFQVIl}+=DOyE;SCw##a$#M*8LnoeMEr%72R#c4
z+5}ANioMNRsOmRgO=5Ry1VQ7lJi{@v4}}iAa0`|!sb+Tl;Qc3VUdEf}o*D+RIyFMk
zV>MBNpv~Y-C#x(z5eNP<P;3-12Qy>S58M2}Ut90*cw}v9Q@?JjcUdPe{*`FR6p*l8
zYF3QPgOVD7Mr1n=$s<It=9+I*|DHz>R-!BO;jZP|hlhXw3`i{tJY>xdBHsuuFDL@|
zwoP5*j;XXCK(5D50Lx0joE3{7wAC-2gR~NE`1q;mrlCgI|Bx&jc?!h>Da5-jz-d+Y
zO>X4ty?NqW;Q|bN6dNl9#Y$;iSd<?>(sXulW?0O#KT-*1RTshBJfTFA0|Z%*LnX|%
zJq%dJyMh8ec<nW<?X?zmAC>7U0j2#DNzf+5poz6t@R;FPKbfdMi)_xrA35gEH5(rO
zU0&M(u|^QItu^)c5;j_>(F63aQG+ND{SS#BAe8tITVDj=ajQemNknTS{@QkW$qSoo
zUi?pZsd8Myvjs|xGiS}Bxo8~ut3X&HhypjdYV<KWqhPFq*PA#+$Tr|f7m$(R1DpfM
zoW-w50hKxEZ&LoE-3xc&O(PK!B@Uw7hJqb`3i0Kf+#ba316v^NFAvY*wvsP(FV-q9
za2?yh&@D?0|Dnmx8<HHS@`saJk4KVo-!V>yA6fvwYI_hmwbRaBYawHr3yr|^SOA9`
zYnYx2y-lyv_TB4%QPb-Na*<#Spm~t>JMWmH)20Cl05^E^vr;mUQVxj{&J~nOo5kd3
z^?XU8e4BG9cr`FelOf5=uvg|NdVmN@m>ftFk9>TScOpuM4w(!)sI{Cw3k(~RZ`-gF
z3Xn_oQex6V5f~0V2{hY1E@b=!d^7ngAeTEkJ9Pt{I*Q#JK;nZrq*!e-B;=HJ4x*C0
zf?gCeK+pf@9mI6Jv{4T2-Z4i45x<BQhJNZg;)>vl!qY+-O1KN^SSZGCUOK3U+%IT9
zXy~?#9<X?N?5(Z3;hqC&k(Vb%Av@=RZw8%@p-WW*0^ESUK%H$d?B+KvAhMqQ!_8c|
zPipU%M3k1!jj2SrQ^vxi2ml`N{50_#Z5=U0#D1Xcrc_u+ENEVORXYBZGhySns7X7`
z+i)!Um_EoL^1st46+;uqyePL1uZW0`vCF7@NK^&SK=wV^7#C|CwE@8gJ@vQ|TEz=`
z&R8v{296yRmk^|RG^HNfQ~)ErX0i?r0KmTVj>rg)g{zgE9RU<L(MKG-K9)Qy<0A6+
z_FEO#z0xRRR*Cn#ZrfH8>yn({`@h)u)$RWmHvY{>qiDp%^axJDQe}hGL+J|;tkLy6
zU|JywD2&qdNX>#G#sC^x9_zS?i3t!RH^GxX(6PkQ3KZ3o_q1%>cWK2b`@m9|pI?pe
zL!t<yMznQ6mTJ_M0K1*uw?ckkPMI+ATR?Rwya4Ia{17GV2lqUriG0^}9K1U?C<_e0
z&0M@nB32=GLD^o2?gq3=;0|S1WEuacjO4-zJsX>e#%g03G2CeISHClw@oAwMU=s@L
zr5*%uDFpRCa8)yg!lhvc@F}gt$T=EGf`JSyINrFW9Yw*}0Lm@;9-GnZ*ue>Qas~rO
za_v)#t`00C97B|Q1L_pe0Y(pVQ*1dr!r^H~=BHvj!^@Dq4GN(8b<_768O@UifP9dY
z6S2Uo$g&~v0k|Iw^r8vL4*kuX+}2bYX;<!82jkBnZ$S+D27HLcN@oKekmPz^n$d@Z
zT9jT;Us>U#`VrjoN^l{Hlz{=g!G6pIFT4WHPI(tEg%?wyw{RgQ80s;5<}*}{*`3$_
z3u$MNUl(l_9s9^<xCRsvwk=Ow+G$Q-P|Cn`Ptyrm7*H=zR1WY3P<GW{Gm$hXPD>EC
z26=%g*9Z}X6t02QMySF2PoK;+a-b9Ve_@7wLR?D1Ldu(qiXuzeZ6x7zL9|AE_An2P
zQo%e&M1W0|^Zs(1Yw$4>*93B@SVQ0kLFu2At;uV4s44uqy%^yxn94vZ4r~ukz~!HS
zS)vT!@oCjA<23l_Y4)Y?8QcbR3$<?@z3jt5Vi#n@c!4ygD8V7yn!^9U*~afgf&8zb
zVCCT<fQ$n)vJ4<A49wzcDxq>jBmpm(%b@RqYz~vEQC7L;xdjg`db(Y>_{B1$0U-^-
z#z8?R1d1YIwc*?Toa_hp?>|Abu9e?hFbMDFKR7&VrC;qPL3wCy&@GZ<X?!T&)#Aap
zxERDEXnoLk##g5`t>DN{E<X6E=B41T*RU7vh&zwd6_cLKnOhcA7$RqtW;rt49{6yc
z=GpnAz9G~8+or`fCB9Cn!w$2%YL9gOiuC1E&?xNZa%&Enk~KHCs4qKqc&+95=vYQA
z;!V(bMAeTthsazW=f;gm0~)a9DqmXIWge<w=u%&dsSbJHz%SBxdeR`GHtoGDG2`dr
zlF%iAP-L$GuJ{)Atwh8;q7yK_xDIF-#F#XvDvDn}&E=0}*2wpubsvSWcRP=p$*w{q
z!8(TBqn7Yt)AP?g)34pGSsnZGT}r4(3UkNij~|CDlU_c%KQYzYEg`n<_?p<r%8A2Z
zlK%_W!gYHK%C+Iu&$g1Wc65|vkho`iLuyqNRXAb4pTU4yh;)+F0u9w;u2Nm04zOT!
zCZt!$ofr<9V{;fp+PEc(a#-Ycqr_7RUP9zf<qta<N0f_m85CseQa<pzOFg@o3+Oj+
zbl9zHnKJvnuqnTSVlbC$1TcjdMl^he+0Cbn2JmQS+RcG7mO|pOhsfD^<Hl?n<AtBF
z?Nkno)pjb#fz1O<Mso-Otv!^a(F8mR2?=iyKWlb3xf9MX7(H1TKsjDW$iBXQeV<fk
zPXUq55K|HRffxsy%THg<(*=&Sk29<Hd0b)DF7(~^fKUO$CCyRI&oas*j0z3hD|1K@
zU_V(J!DfQ3s{Ax<+kF6!-^b$O0qn8mz~`NVo+n}UD31<?ZtREZbX|~K@yCz1CR~9v
z<BBx6dok;W<Z)Ij<&u6*PgLZOYc~o_$Y|Vncq($!rhEE}?ITKOpmKYxH0+Z8sBqbC
zoMBMsf$SE7Qsc!yJ(_-N%1&Ii4Eg^lI)h}gwnhB0MSvaUcsLyL1h?iCV)sHr5ePYB
zH!QW0-<EsiL3PVuz^<tn^xbFtN>R66SlAqay<W~Yj&_L3qr<}5-+Y4RfnSWcEE#9H
zf(E-0MXK%$gu33>A1`D2s#X{O2^PK$6EO;a!J@oIxR2ku06B!RL7_@qJu|N%g!w`#
zm{orERvLK<>olgs_;Y4p2=oUJBfzS0HIk^O8H~h}jeX==q#u+V3gl}XlVZf}cX-oy
zD0L$i&|D`o?V_WT%9)ZnF%%(b-PgE_?jXQVh0yt=_)|V1P;$cK8E;yqCNC`dnmnlg
zb()c#wp-D_H@Z#9=nuOis{R6O8<?$Lri=rSo5_F)Z~GogO3cC6=!HhXbKMC`4|JvK
zMq$c!nTbCsYzkNJb%0nDQh|OKfb~qK=WFCQysfCXYtcA?*(0P?Thy6p8%7NJNE4FN
zNs<JzyA|@heGuB<Vx$o<vF9MNCx`%fmydKQ=a+u6<Q<h-_hk%>_4Gq^H3uQ;G*=m=
zGlKz3YR@(hcLgw>)fb+rxa@fNdUO(~7rv`lc{Iftdw>F?$^AX~J5sKTBky)_un+QR
zO_Z~+z33NnN)&pW%dNnkZO$|oGx=~vZtotwD}^pA`$Fo*)5;p8dq2K7yyfIezV^O-
z6&raSS#*)m2><F{1SQ~#0@h6vI^k-d3lnc;n#DAXe0Q+;UlvmoB2=(o^U+9swZRl|
zLe?~8CLk#%)(%1zvP_^LqR~`TF9mN}yw}c7jC{U$&;TbWe;8}?Km4=Sxu8NKMNpTF
zH7ZBd!j%TG%OV&BSbpE%V|KsS>!tN_haC~Jf2z_~6u(Q>9Mhh|#pS6j9NG3+()U7K
z?j5_3`4tX5=DnX&yu7e{o)s?I8-6+N<?#EojEsFjk$R@VlUmg!C7k41#gX`vrV5}T
z!V&hQP)GgfM`c>Q0u0cR49LhRpc*CtaKaP`PH9J2)8xmFDAj+hwqfz@;;Fxv4*$zg
z1AK2ZZ7`3wSH`eN6S2eHU;Jt|G%P#Sv8~5k=~7>`^)COLCMm-is{@LqhB9hZJ34~9
zt-gBMgn*oesw}+LfSV0?2}k~q#t6(AQ_F$esM8=hV{G^Au5?KYuG1CPTW@tv9SB8?
z`Gx;XHpk3*8c9vLz2Fdu*ucp~vJFt6G{FyG*H(LnhQ|M)yH})a+jQdB(<cTG?F_a3
zraDKY5jW<6QXlSDl==|W0P<|>?WLkC@L`vg=Ee;h-0-h~Lg1ps%Z9vV2D(`2*r;cP
zXbTMk{8A{EuAo1Qv^g4mjAnKsY8XS%z%@gvOPvECW<*b!_7D__=wPT@OZwa~hYf6?
z-EiERYtnL2S5A1M{G*BEB_+KX;{#58)rY3vMwcp`vY_J-9rmMVsJssNn%4agP;FD=
zWm60W_--O7hqpsRN=cDehWXIj&|~O9%I>}{7D&y+Mtt_EJHZ5jHq!NVXV<>khs7z`
zeopG41`#T`vG(5$GGeo57pb`PgH|=I9XlCUG&SfPLeXrfPZ34DA07}0JzZq{Api%;
zf+w6uSiOvg9)+5M%)DXoZV5nXk?-JJ(DdT}XryV;AsA6m4b!-t{I~K`3BLe5p}J>9
zfkg;aEu5PN6%`c>4c8&HbtaPX>1QIH0Zq*wq))<AQMWRm^Ut1IX)fIU3Xx{nSqPq{
zSTmH2v1NE`pl{p0ew~SWudc3+z~OZtj3NtNb3|HkMWl#wxKpwYx@j6JxD{?J2p>MQ
zo>2tLuoQR@4D%WLJ1iAQYdbYLLDr_&Hh;PCpRD*?^m;dc{yeXp<)~&hff9uxwqYD6
zeJ{Qk#6O-mULjwRH*<m<o@U`9_yBWXD1Qui`RPkT$BC;o8=*k$_^e<5;I601%{W?(
z#eI42(KK3QSU_b=HRu20$qyO?8UPT4a+)3VItk7}!|p_?Tx8+McQV7mogK7AF7BKS
zMq%L-z~oH_VIZRn<Gn4bcBr~mW&k8Yau9tb3@50_(9XA+_4#bQjnE0)8xRcRuOS1#
z>>CY%Ma=VCpwMT0pHH&KjcgkZ?9}ua<qlfu9d+jX&N3h6P{m!3+8fRK5;+(sW5_lS
zYa5NRN2<otSFes5Nop522VVO6NES-SdVU>AW`=%`HGIwkNT#A}c!=5aad4z~)JSxW
zcl>II?jMWQbH+pWgxUl#*AVH=tCIMOi7_i7n5LN?NDtL~zNIRCaz9h@$Y<XoU0ZUj
zU}aEkbojYEXpPS?Cq&4Dfs+}P6^u1AXyVM0UvQeY;?|(Ko3MR@T%r03hWBuYRdAkd
z*s_I&Ym!zSy$}jO0=l3Ae~awq-yRma+-zu9D7FQVIvoboFyR<Rc?)l}Jn+#5#zk~~
zB=O<Mr$I(NHuMdIgv0|^3OwuR273E{uOEp_MWyjz2`X;DZr7n+ESb5h_nm3<G~!1A
zSBV(A^(3dW2j%fr+^LM+DAdu}qGKx(p2CpZMM(Ftw9W+G&pzFID+^x$H~=-Qx}fDH
zTq^jp#2_IkhEP6QqBJMQ$9@UYbc&0o2frc+0qSA=enj$P;G&qNDEw(e2BY-GzN3^t
z3Ru8LK^^&QiB(@xnJgV9lv)i!Qvtj+s{4}}Hs_KCKZ{z@Fy?xXe$*lQ70(WtFGt&f
zu|$+3b_K-+RUfzlQLQcq@w+2am}3Mo-~r^dBxSqVeQOHb_OA|#Vm2@s>MF7G7kxI?
z!nD~6u^VMd;>4kxctFsm-*T;`_g^|^rj)^K@I;0K3Q=0P<sXZHjX*)U)$WIC_{WyO
zi_2ZItc4UVc^~51UnR0Lq_gl@S;~rru|~5GiNLPC(c?`Is<*el4w(;b;`s{~R`m35
zF&MsMHi}l+(ABOhs`FN2k<ZcScW4X<YQ@=2$2|!cFlM2?=aSok+~Coi7I4TQbBTe{
z&bhc8vJ|?4kQho(DWxi-=e!Q`2Y$=}i~}`MmaV;Od!yUxSsZs(uRX_(6Iifd)G@tn
zE)FlBkh*^<1%Q2YfgonAz6>bsz|S@Xv`qrY^uQ%&z6+oM2JXJaB@TH5<TVC=VC$QK
zUeoi~CUX6uc(VJx8DAGCETBULENn4nE+9j=yJ;s9Mhq$l8YzOYp!zVqb`Q-1hJ-S3
z$6>$?lDcZveLFbY#3gvsURF4c|Nr(Vw+$67ju@`qtzp2_1_bjV7M>>tnekGM`mlK@
zB9meah&HBV2+9tnFiAixb?tU#p>kY^I<m6rUl%Y4k%gxx``9SXzyBs~09>Ho6FenQ
zpc#~e0of-Z&pp4Wv4U3<cPo;BS>Uq4NlgPD(S<2s6REbubWt?HtFv=b&Fg0%p$CK}
z71PN5A&2X0W#nN&oFfPT9iMj{yA)?PIp#=@@dn`6g8B1x?N;)PLa2Z;*aJ1E0jL&2
zq{r$(<K4zYF02iGDA9Lca(@FM4~Z+__|h<&ZKMdgnr4<lh{ABoo#?ef@RQH}W*?zN
zjQ^&OP&N>RwNVyWZtReX)umet0+?n22!}m^oPuTW<ZIUIYJW`K_3euzKPn?^jGVDP
zsVu1Wd5Twp{X>a@KzGp><31(JYCB@ro~xT*``bJEj6tbwT$im-(Xi|v&7)k<hg^+`
z*-Q}j$`cUm6vA5V0cdQR($4JhsXrjx#d~i@?L>KGc*y3$@7*|;;6%gI#gKDBJQae|
zAinPaFoQ`{f#qYmJyz@OF6QbW917g<h7B9Y_lL4Z71J<rnFGYh!%BrNk><=pn%j!=
z6t)8sZ#R=z+o<Tl<iue~wj-gkKIm*rvqc_EmM2Cb=ogC=VuQ0CH6Op(K*mZ3H&&2g
zz=Ps%=10sSjKhXW-KwE;0yHgIC81XY<E@Ic0@Ujm<Bd`ZbtR_H&w`K#e9m52Q<2{G
z4+9lE0&eBu^8>AhJd66WDpV&lQ)z+5tMv51iN?IPsiX^?1IqVD_}|DKdX%jErNZm<
z=cSrYR5#ve{A|OodF7VeGM0_2uD-g}`t<0jT{CBC+wIxCKzqXs8%M@Jc1!%*g@1e$
zxVn3z%oFze;TI~OS(J`UBnHDhVYp7YVRvqStyitrx-#jq3n85k*9?BzUiSA50_C85
z+=qIHJXAx>bBR-U<qQVY4iH7yuvXYFS-P|xD-j3GEK~M7=rEYEwlp+Y!m$^F!7=Zi
zt<p@4C(aV?9}KTGNaAql)pN+@)#O^9vc0-K6~(-#r{@9pE}#A3r<z8`#)6bB6A%$6
zA70L2z@`b!OS7(Xu!<+l#0<KR<68uwGN2I461$hE=qz*b$<UbbDwT<H)HX-OYH2Cv
z1`cn;2U~vN0vwgy=DN(0+hc2toO`ny9zRm7PLIVN+mtbP9|HWcMlm#6{yEd{^?JC0
ze!{z0`<k;M62EQ8<x!LOXJ4@u;|A~iQucfMvy7A%EX@em`C&%_7OfSxA*5uOx~@SM
z)W1202-G9#mt^p~qtVIx+AZz1CwDro5kA%lAm~tJCCQi&rDs?yoFJ!P$M3@+s|)7I
zked;Pv)Jn87cQElXhv0=zI*qsZhSzqmgsy)DIhyJZ&vu>n+;9;hT@JQ+8`>{OI602
zzOap)8MRKEZvIybfC8Iz#<>5#sa$S>zyN2dK62|DKRv{?qL0&uw#cc=6ods0BV?`l
z6M_DRpVNs(a4Ai{1AKJVptT$NS5U8=D!ytin81JtP)G>@fXw4-ZA!J^l}Pvca;g?h
z{C2OV5XpxqzQ{0DAhN~}(8W_gK#<?JkE-IRqZW-gy10qt$kW|@AcH9%jAons7;^5L
z`4dGEil29MiEXhS=(6&CsUZ7pj>(y@qFtfC_^cbn3KEU1q);-WwKf0(Zn!=MfGsHp
zp<sW@xqnWyOU~H)w2L|C=GKRf?(9y~KcWpt(;eQ>6K9}?%!5aU@;;MTPUnp2mR&Gv
zDwC+eA)b-93riV=_Z$)x0V~b$dW!>*kYMtoB-u88pPanvHrY}q>xcrpukYC?Iw496
zT6pss?_(IvRAZ`<CSjO+(@v5PtyP92d{*hF8hTnEKoW>b0Vu<jFD$Ek_Pb-{0H?Ff
z?S$aziAIc?j)iRAq1AWD!~ilL8oF29t+7#2Zkn<Cp$5vRtqWv4!K(n*MtCbu8AzRU
z2VSc@KWgoM6szE~X-ZU7RDz})^h4x`;d!ah-iR_TL$<xE%f!}g+0gH3kE&kv9!CG~
zU1e=8kt$fF`ud;PlsMUB@KvEMbmaW$9}r+$<fXK(L;Q^vW-&j|EXei;euwh;%a&qP
z=y>31yWU{z#Vu&do($JnM6)z&r0ivN`L;n?y0K>na%RNBQ!NkjQ_RH*wxiv_F?!ET
zw!V;yTbS|D!WFE?`}`Z(qZZCjuYXy@QZ?zetvc00OM2pWCG|K3Jm=*TPP`lW^ilMA
zxiP8upa47h=_sa17XVQrO_!JWP~D<qOI>Dsyxh+ma;6B;GU-9@0&nj<?raYyN<=w`
zmczBH^$voft3bbLAj;CjhFg1X&`4xK!Tk0i)TgW|NxN5=fEY1!P1+9IPjBp^$_Ah7
zUr7TAQwsgu#|I0GiW;&*hpWt*v?5q&;I=4_kD*gyv2FdLP#Nobrj>7;f)=e!7hhZ1
zq5wf(-UIwDJ|^tYJmbuPKJBx9b~UwC@D;AqEgJZ6fq&$_$TRbtvC8>IG8)$k*FHEr
z@FU#sQT93%hZvKJ(&O0A3y-r~f@@@37_B$ghlq{^Wfw>E{W%BFTG!^X`4Js44ckA6
zSv2)JxaMlP@z;uN9IPICOuqTsdy{H@8MVa6n-AM-E$dkBv7pMqyz$y?_9HQm@`^(i
zx4gbk)-wdcl7#`ug!zD-`tYW6>T%r)Owq}$qG~!OBce#ZHQ}u5-%_a+uDi`GASlSp
z;q@w0enT4=y8tIV#b0)KP(1Ic#_{8GzQjdE-SEnLgchhB??!dg1H-h2>R_2SA+r<L
zFYCFfZD*IfFXU)+RnOB{9IFt$6QV!#VjNYBHOyRIkzRNwvjGTTD?>jD?KmH*K}wg0
zHhaTP5m{!?rMyB*tKh(SKj(gM5AIU?<Dg$PPou6l#5u*`&edR@0EplD<Pb)34TewL
zoL{#5?{^|lw5#P;+mnvM@}koQ4SJY=fF-tz;`D*rlk^odm{^0xkAaP6`7IuZB%90r
z_pMZ;6N_Or^^+al071M5Hfz$UxYXgi=Kj)A;kPvsX5JIZc`sZ_y_Wu+G%c0oUbf80
zJ`}U;(>-4xz61pM!xVlOvrEU1Kd86(hE7{s=XHAe?wD%i(qYb-mze?38QlEnsG$R~
zPP?0=cs?QYo@VUQZ4W;^_2FP+#Y!WVJSK@pMn<L$1L__eE*RQgIc(JNxqApl4E}p*
zS|O%O1nF?=8?Dsr?f@a|*miwcx@YyE&eJf2jRi@svb%{S){iA_mN!1H<*70i|0DHh
z&zYwb^aJ=O?cvgt3oNfLJ+pL;@k^1do?j_@E+%&B{Rw^CCJ4nzm*V={cuwTrhsoOb
z0@vhIP&9*8V>Ws8s)Ll<F}BFm@4)bkknn!H`A?5?{zNo~;D!yeQC>CnZa|cL<1Y>r
zQmJ+85bG9|y-o2EzmCkBYsnfwqfiCzbo%wIzPpF33FkyB{A6M-e%Z0U!4qa&!(Ro3
zgh~v*;Buy6e7Lt7xxS5FzWr*wE~@IR8g+xjq&wYX0;ZX4qOV(sl)>r5EKv_HuNx1u
zh{s=pa-khyG{X%iI8n4<p|3n)KRa-)duILu`_*f>vQKkWRi5s7banXrvFy>2kuVVU
z&_&7t-{lG~zX6#BHHukBwme!btgd|Ab7(<Dj=Zm}zzQ`#UrVpT)O@hdG*Au|VjGyR
zM%GxG{sO)Z(f@O0Wv};y$NswRUt;9S)BdvP*5`B>I%2=)&$~K1+r6;7C0IT{V_S7&
z%bAjT_D&S4z-CTnoj!T;J?Br9kzV3KFho#BJ7fo8DD_@D3g7zxH0IOu;*yf<fbYJf
zS(RmjjFdtTO|k+s^+3!zn~^If|F6V?+#}rF#x*(U`;pcZENL$f%&6p-cu6kyFEz5L
zr4U^ENngHqRXD%`054A}yHE`4FZBeUQ+vVI#^y6l<)oJ55pXU41bLFtEV)j|0OLrr
z*fAF=<B7G9*`<$5D`g7;ysGx8?vi(9M~y3hfyy@D_{Ub_rw?t<JB^=~;$q_?nDSyR
z52F==R9t6;fWAn(OA+i8;<eq#x89Sl`^?nnGTm(kx5PwTR8o<ws5j7PTh@k304k35
z)AtUpnT85RzAoxZJyaF)<MQN4r`v~B3E7A``X@!xBDK$7oFq1Y3kbI4oL7_Pz$?|^
z-tWz_G)Uz2C7lSf1xY6c0x*T4N#D5lj5=rE)8W)s(5O&la$w`(+5{XHTyXTCl7i_)
z;;@uMYHoqric|5n-#_I+?|K#WCAoU(OK&@O*sG|(^o8V0zuUtWFFG0*4E^|#YHTir
zdh!#d4?vHF(BJ~p6y=Eq2T1ddDeN<PcK$!thElGrQex4y>=tWZ*@&KUS{6<0y+iK>
z{b&{vsw-4NVnJ?aMY+#+*9f851f)spwb2C%2Mjrxjft-s8vgacYN+YlVfX~i1D*Z?
zZtnLfiFvxaPoG{1`v6`d!bJ4M<3MW=CJMJ9s*J`hryy!sj=w=uDIipMiuRS<d61WX
z8r~CMD~SG_&7B2Diw1wf{@G5V$(iLz5NMHdjrwgGLXO@KpQr~scd3cjty|P(B1)~2
zwMO&Cr<5+&PiV1|oN+LLi`o+RNdz&IsE8p@r6MH(gfI#8(yUqv1^uMoc8AU3e2A%3
zjPpsFFUz}!h>WKPh?Q>5H3co6clB@T#h)FGZjBkm0|2&wteP}OKRxcGOg*dqSE(MW
zOluc_LW+3u>pI~aV8IeDj@V#4c41GNf}+sS*Bg=x?y|cijspQ%+Qn>ce;*Lu{UCVc
zGO)67gn2RC@b%LiwO>X=<$KIFX^Gfb%BSY1T)q^yI_frn%6JI?B_YS?K*qyf*v`m+
zisTLyuY2r8&e8=RQ0x>?a>13^f&DLIUH#|XC`idSEiEnI-5r^xFZxx&G~34%4^CN`
z^P6*^=A^V`yiCx$l)s8r?+yl|;K^b+p)ogrFJ@^kKSbbwg#<AM0b30Rco9N)Xn3F{
z<3yM$xf3NU8(V+bL0AEt4LTijdy_d;^BlQ&#AXB6#CX*=@G?*&2W%3ngoHSur2J78
zy$G2D?gv&n%_JEVj#JQpHz6js`36#p^Jw(_26mo5tG?z{Gua$*Fb%*LZvYtvYWcS~
zUhjm2SlS1B<EYa6aCa>b0RyjhhZ)Pu?6qhMS3y{Y{sps}^LmCb-Afng*{G07o<?{X
zG~ob*$-?a;{09K62igw|=JbUCAjSj<R7B}nHc5lmBGnrjm8IJs!{LW4$Ez?ZY5WXc
z4rK{SKaR<fGFQ#U;b4539fJF}ETy;$1kl|mm=j%YycgM;KR<4cpS|hK4Se8}5G}`8
znfEVu?R=_o?JFdRWaI%XrPcTX`YDTA*`1lEaUA}Gs$?8M3XRX(YW#WCYvEgB$6X}$
z@$J}qaBN3`b>9tV^INILN>aeNo<i(GIfC`wu;2ib-Q8;xbUSm|E5C~$jr!Tz)OWN@
zMEfPJ+&uj5IAT>{cJ{gm)GWg*TwbjDLZqtDz{J{TQxd6MJ2WsrbiWVJb%@<pOG@(K
znxx$g@NDb(*4>yO`H{Je7e9hhr*8`=(1cIBp1K7%WcOi0yeN`8{u2_1H%H<{LX-hL
zbF7r+Mxm2D(zFI8CI$)Nete2#dgGQH89*SP5p71xj|MFqJGNML?8yiO%LYhl037=L
zV>hsR8dL}8nE2>CQTc<#ruM8^Q4?DUPt2N{91E1a<ec4oTY9+8fc)G8bq0RoF5X{V
z8;yGn<PUB(>SdXQ*F;b*J{Cz7c32K9bX4QMglww*$2PdI2v?8Q@WM2ZHaK)pNus`?
zdx{v>3(ZdKlsT`Y^fgrr0S<&^!n0<%4j@5W$S{=J`mfuaP~z)}a?^`GzkUFOiALz^
z9Q+&fXSxs`X>;HhBVZIil*UmvEK$(u`H<T6WV=+5Ir=NZBjr%IOHt-?kQ0=cBxVC$
zou;6Fb{DOaL>wU?K}qhSErTUZk)>@*teP?{%(FF%U9ql_xV9N@RVXCyO&Z?{zJTb$
znlM%QFSX@G8K1FK4B};nQx))&&}ZGgfB&q*0w(aex@40@z?9#XeC*KrBR_n{YUE<^
z(pPdd)UsM>1M%`W{CtLKzoFAW!7`RNeRFRp4>fJvtA4TcYLRneNi{oPUQrP{%9TR(
zBQn=1F&N@c@@r$s^JssEZ9@P_1K^c*ruEXFKnBAKbdH61#2K}Vkb`j<*7(sM;aU8u
zF!NI2J<`vi0%W0Czu@mkbmPV1g7XPW#-^)o8^23WqTwkvcyr%zOJUvlu)Kk7uRds&
z)4@x)kU1DgcY!t_uR-pMjK^5OZPCO-`l$Kl@n8Nc!0I+<-aJDTN{tn5(&xd9eYg}-
zJ5h^>*odsGM*Se5q1zeUaK&J~2HyE_RwI`SqlF4ijxB#$+r6BjboA)m>Wh$wKg90X
zhJM=#Y5mK=I(>(dSVB<>Rz)_3U_kLhoZj0{pFWLv7}RhMMF}vzvM~%qxw$>`0@3wt
zX9VI7{4U};GB=<P3MM0lDC#497y}-H_}Pw5`V(FZB2^I05eSAn#tE|ZZ4Aeu<6%QZ
zX)2`cWn<v}<@18tCr^}P5SsU!EN@WqZ^adelp2&{N;s<?;ySI-iXW<uz3psa%zu)o
zy4nEb%zSEIR54G!ec>0wi`@yb%GC$BLy9tP|GHjfpuWO|GcirX74dT%3>;(IDVGIU
zaUR9hY`pj2!MPJnbL$V}d>a(fdOvxR?v~r>D0GW3iuC~iSv~xz?KrFAa*%If>K~7q
zA~i><d+cINd=|yLFu1c%TMdswgZYl3du-DK9ecI3+`$RoWQQK{6a2{@K8qNO(8!!v
zc`f-FNJ3G9GAVFtMdK8ox~6K^{V5O5<#~K&kfUgSqJJ3@^4GOwp@Pb&McgL@h^E+8
zo&rgVXMh~>Bq=|QUG4$WJFiK9yMr3QfaYkUyQ5iIU<Jua2>OVIg5t(|QO$na67oYF
zW29lX1{hC-8McerZ;of^!e)c1+KW#mZH>~yZL=a(g0CLiloGfLIXT|=DyyM)g8pC+
zf~d%yhafQVO>F9fq~X9Iy&M{Uo9!V)H6gNO1u#PjC|w70C=h<Kn0xIK5sA;s3Toe>
zexT|mPBTCVq%49Ua_5D<6~Jw*F~juqkwd8(X}(b>)QJQ&wK~N)XHMcHk0Vk$C56>}
zq7__bo18ho-zm>b+~_A`)?-1uYb#JQ5>8P1DsDE?CPR9d?092A%k&*l+^Lm%dm^#$
zl;4st>wQ4sQPSgO?sT1$My2nE!gC>HD#@E2G}JF$_^0jtju-tqQz9}(SHYj%vsXzW
zG?kN*a}dN$VK=B2q-3}Kdh-RsT2yCqs+w)q{PsK)AGoq{iXZ-Oc%N+@y@5cs!6Ge1
z>du+$>;^wRhXKShn`4^712@{|E;|STEq-~%?UX{QE-(}rgr4uKEjHpV)N_A@;)5VZ
z$hMfb#~)uQH@<do&8ryS6{pf7k~M|2b#y2gCTplhYiYlfmXuh;JbpvvYB{DtZ|iuJ
z&*LMu{E_m678i5lRzLM}K=DQr<b0ba2mD+aDa#^)hfWs1e-ACdOjCZ@lYg#UQ_e2k
zS!^segF!?nj&BLYhZm_d@ss{Fn-nU`PCMqsNgK0?SeyD>S_omHGzVp+lhGT9E`T(7
zNV~|elOzU3953A{KtZz$+RZlstISsv$v!*a3?Ys8enB|rT2@;pt{5?pl#vMz(iu$v
z+~!cLrXC!XaO7oLs_*v434A>2o?J=a;})5+cOJ0#dNE37w#iD4{5Qs8lHYPc!D7tK
zYA`hD`Ijb2oLIpnRs3k}yPN6N8Lt#Daj^t86dDL#r?d<eH}E_Gn5e<BS|q)z(FSf!
z<5b!+Iz#etzcUz;!M?n|FSXxXuenpc2*7g@CeDH5yzW2mTUZ*-SvzLc91O`yaP<2!
zop8Z51AId=?$+n|j;+Z!pPwrjFBE&D`1sN5tjYCvVNc8NL%&M1#-FHeAfY48t45d*
z1rYAjl#`RAdV<KMa%1yUt>8!bB)-8{(#HY;O3V$75DmD65QfGn0F$%mn$JgU9rj+>
z>l2SYQ=Snx<i>25WdOZWs>Q12P3xQHEB!lz&Z0~7#=R#jr`xM$u+Gqu#E3$Ztq2ri
zXJ}_TKL|2$L}lq*$%W1k-3j1KH$T54Nd@&)*nU(H7kDk}z}*!U9i0cZ_b13o0bybM
zvbNk@Tz>xZ+QK@SOtdclWMTHG<7ii8Y($s?^n#EIaq;l9u2m_ie&BDs>RPf2j}IFG
z2Y42-&HF7m$KIWPaCk_S^XH7+D4Sq(CY4Y{*lqTch{aI!gqRB-awET-A!04eHI-S5
z)DuO>9;W?AS!o+LZ)P>M+WN1=_*(L5jL$*!BE+}<9tc9mDF4W$Wx*^9BlSNsKTSH=
z+nXhxgy3UsV%p6!nRYi%I768No`IEXz@7D9F_^$O4UB^!f<8j(^7vncK_%b9Cjyr3
z$4ocH;4^H)SrBZ~{S_ONZgdc2luK3e8|z4j=Ded+E9dxcP>jjG9~`%LK6vr#a8j;L
zleX5DrFiBhI8oj~a_9zcG_m98t-bd>LVruM$#J>OVi08~G=Jl?I!)l+sp}3u7!ntX
zl%H&VpL=~%%9kE-q^`e@5}vm?Tp-P6P*wf)GrkdiddB;7ju{T8X$xCT|C-N-KnoJ$
zPNouoZs=!{ty)$x5>T4aL5xhT?-lX*b#c{M<=vPgQ2G*ML-;TLWXT!%&f5snA=dYs
z?Pb#Lh<u{{X^RE;(;CEw#cDL@hL~sKOS3yqyhbZRBc&ldIj`v1eji=JbQme*%dmBR
zKW!B~d?D5Gkyg_r+-yPf^>dx9gyc7ru}cglGGhx4zW)8X-t6c@j?FdkwkMZ!o*Ebp
zdVf&S%>UKO#lO3bJie8g{I<tR6DcwkU$SAJBu^ORbh=_Z(6(|h$j%{=&O3j;8+t%w
zR=vC)zV_PA7_!%*Rih#DMBS;YyF=~Tk^QGx0YDnQuaW=h_*#X?%Icud&vvCMG1XsI
ze6lQ|IB0OSt2b{h?iunobv5#e%&6{-k3E+Q31pQY=0-<8%`%duVM?hdw_KawzJ=_*
z^vSs8D0=}HovtYNe25z%+QDH2#4*J#@Y?+K1)>})UB|7B#drt~Cr0^S=D`}suJqLS
zAjSP*^Q5y&{~AKmV=(@AK+G)Rflcc4qf<0%vXX!A-qt>HMor`B3<md3yV)?cX-Piw
zT*!$pb<`oPuEz=s;M#3D+$e$of<PLXb*W<S-NtI8WCdCWc$N9t2z*AmJ*Xrug8g^G
zVTQ7B%Z7d!K=6%g#&Ys>A!jo=HZCuA!O3^sn#)FYC$xH78XpUVe6P^_6Ey#qZe6m2
z(B;WLt_+6UqytM|Snu>V(P<Y{d4AX;=U{5gW}))weOtlZV<QX!Pul3@q9f8QjK9#R
zRO-q)XRkz&un)-=l)v`Gq4x-;R)v?(t`go{iJl21KKiYHaEcpOMsU*jX`Fe@TTdK6
z-WGhJ8AGwwKvcoT!R>|*xdQCLPQJL(Sd6%PV0~~c-GUZG^K1QlY;HsbJcS+$m@PM)
z6_-c5*2kFSB54yP_1mT<J~&Vz07^M1xaP78b65`fC=jG@LZ(Fu5HFFOgv?yd-dY-_
z09^uuMuwng#T#~$?9D}K3oQ#eP|dHdD2(vWK$=?O{wfXlcs{sy&mEKM-=aZ)jG4l9
z;k^?2_3fS|#O{XQ62*8pL2F_sSzX{gnV*)c*<d9wRw}VSr|9|ZSJ&^0tT}>n!ot`s
zVCREhM+YNCGcz)#7u>H^Kvnkb*>k1f5Fb{gJu(?$;$LTEv|&l3*s^zTD0{e_oyOC`
zAc18#^W>>9_4r!yDj$&^n28ptJm-=N0Dq*!B9O+f*64vo%sx;XBqcyO?~SoaC;{`K
zFS>(1$zS836W$fIyf9|M!3h|NMr6L!Dw?_?y4~>88<bZx8ITk`xl<D&&>3!FbGmo;
zE-A<{0%8?9pxp6abE&V!iI<=@FgLmZ|1clRlakbdX<r5RfX<n6v#$;V`Z!HYBI61E
zfg97*)3)3;89RcWRWi7g>2?8Gwsw-i_50Wa1|GKaaxtKJnTC&H*GF8(liKb?Xx6~k
z2FB_YOXw(=H8JfZy{?>b`LkfgSWHU#-@ZjH+}BqGpwBy+b6Xl_4e_VA4v$sw4D#<2
zPTtEZ@NCH#I+{crknks-%8Jy8!K~`JxSQ$Vgf437_VC#PGZ_-x{umGeROuv#e3-)=
zZhrny=Rs;hiK(PXW{{;@BxC_?7>9b6P6jMVvgpI>BnY>e+|+ndi6fl;H|U^33JHMr
z0_Z5zG&DTchhAJrXK<D2;2f5V#fmV=17oN7Q)t^(1J4K0^admXkvKutd=SeZ61foK
z6#&n$OI}B@PFexTq}wY<zewFa+DZDiKnDTBkh6&f6+m*-8z+)u0^9fi5~*lfsMqmW
z4FgzF{x-Bs(n#p$u_0zmd^?m@vlukJ2zIc9%$m(oF~cyZBsp54xVv%V2JY=At$&&H
zLOzAc(cm&#rH~onMEePwAnBseTr7O7a-*<EaK+~{BR%(SPrrw5^@AhI<@=E}AsQy`
zcvC%EZ?U0k@!3~Py)boE^9x*)WR5}pQ@MWYIh=|A2V?I6&voDT4gWNdw1*;<h87}H
zwkR5!N`+)pb_-D`Aq}BrWF{f2l074<va?4hAtWoK6wmvkj_Wwj=ee)@^<1y}y6)>Z
zucOlc|NDKvpU->NzlG<SU~K4TX2SIaRayMLY{!}U1a`q~J}2PHs01LjiBRhx3+Pwl
z*zYAI8-vSMZ1_=be_OF0@CeFj00fp1fpF9lr2k+b!Hxcxs^5m_ONTmt>5H>Je*8$=
z!;m+M;7A68I|g>bErG&gg>OPr|ASc{+=M3``T-H72b?wwpwS*$LcI?*FK-L~qXkH|
zsIN_+;{ct2<O~8`OD$vgfCzg7E8f&P_^Krr3L27D<C&uXcyQq_QqqT?9K+C$BZDsH
zZ#v?jama~AOgcJ~;2pa<bh(!z0s*JtxXm&UJMZuesuqZHbAcHa2hW=wkOmnaY;)1(
z?FVAIOURalCGYSv%}nU`@}>OtxqQgW{eC!oWW|c{W9a8+jmvs<Ul5<M4#q0o-ej%k
zZKKBlPhvvh6A-waq8KO(q6tVSn^CtNTx<}I>`$40$zov4WBF|i{y5#aRgQyM5wt{K
zk^@etIsxj)#VTa`eZNoJdy?e{2uxydcs#NlajM!Dg&H?}rVN9V-|B00o=NTlfBz!E
zAOKZU<m6tlmc;~cF;p`J`g@Vw<<v}5j2ek6WiYy%KHIH@OCCeiqYJ59ggiHSH5iC_
z3pzA+b$!sjRK6!$TfYK)T~<ZhN2&V*mt-K1+BcH)t)=B;3@3U*J|y9rX5r$>|Col?
z<=oieu!EWb_wrkyc{dQg;-quO(=+ckFE_W5^#@C*kMD3T?RMaT5JR6b!0c*&qa46^
z^A!gAoNG!!SCv$=(x4)Hd1)jCrz?a9`|8g8kXG^NQUzj;SNm$=_17QLZGGxy$OiE3
zzQ&3Nztp)vPIKk=6-NNtVOAjB7NQ)mMiqWd!&6I55{{r3P}@yH6^a!E)JYXAMgB5%
zy65I7_^LvGLJkKVpv}0`Nadc$;s%X6p=sdwVyi(Iv*Y1FecxQH>2Fg~;W+U6UFiSy
zLAmU<6h7e6v64*%LLxe*57{(Y*0}E^l{GHh=r>ZbvJIzHV4kIu3|!G+=Q;H#GsNed
zh4lmTJmzJhl5(a5LgkU`%Ii$1fP_Siz;wQWlj{Vtgo)t8m|-%@0lFa&nLVhA(8DE=
zI^_HuTz?Zhm7r|4L7u{p@n616+qYzx8=V1}xTV1`XTvj6$f`n#XqoMf247{x<6rgF
ztO^wAlM(t+8G5IWr%s7@3}i>~r~q#&U~L(GvP~vHxsr4qbw^U19}D^I$qP={n3%o6
zZ(E_hpMI2DzG;^J!>K`~Z-U!Kk|+zvR8R7e))L?}YD!>kkQ0*=A@E)pgNcjMM0T(3
zc_D1YvN>KRS)l=<oLTrqCHwsLfR>mT)azzd-Mvt`zpw98TI^hW+cr*O^JpN?SuNgu
zZv)TaWJ2Vzn@om$3I~((obY+;)R?%5;xJ<}c5wI5z=IqnyFH-}4?GfKazX|KQaO}i
zQG%KUipyRKvcV9y1(XN0yVoIWROjrQj<X!N0U(uWMSW5E$Z7g|=7RXDfH_<L7h>PQ
z^5R#4gBMyOs1U<PWZvAlOq_9gWeeU=7>gfWSUM$c#@NeKSUJgC3PDy4r!0Mk|7r+s
z<2F(^xq=unS~Qh{?&9d#2<-%ufgt*VX+4V}j6xnsgf(ApMZ)(9$Dt>D4Jja5psC&w
z-1-8E^G~+<^zY6)tTz-L{oB1*c8bgXH|O4l%3ky6{vy82@y_*WG}b=m@|p4FJu?{!
zeO-NuHug!=($u_-Lm0fh&Zq2~r@E&tkzGXbRd6#qf8gHAe(o}kMg+3UL(xj$BaVqh
z1i_Lq{j}XZ3#`W^DWyyb2!n67`CC7G+q@2OI7kMCE_^!kr9MfNK3;4*?3b2$ues2=
zAmxPpn!DRi^v8dM@XanSD(S1~qk#_|@1`pu;h`>dTb<U|sJwru@ZF)M23Fl%$ML;_
z7tVt^gNE%tlc+8vS@rTEcZP?HG$V`uo5m*HK0~<!1SeS<i#M9vY`n09r;hH*lJUE;
zL)GAii(wEXVr3orhc8}~!285m-tegbY6&e<)L@W+J>&BPXF+gT^|y1AxeOzF`}>>U
zHA&Uo+dad-t&})QP}mt}9xou+ILz8O^)jeH%`2^D#;FMjn+`kS-9$QWPxhR>Vu_G8
ze-=7>(aNev{c)x_e?*k+CT^Gd4RfCQRTvtN)!gKM8nsDwm)HwY`5ax5=duUjUn>+k
zw0f?Z{-G-XS%xurGt7BQi3tk!EdHV|wk^9`^OD!*RJHf%tk6azRSU^9%FaVzltU)-
zuye_e?&|2EhVFxSM{PIlS}|Me_nj`)Wq03nw4FVME3<y+X-W4~<?&(7e0INs`WrzV
z8X9!2sKCSeU$PoS=VSNp`Sb5)*Ad}#0A@kKp1AOD<&1y`fetQ*q_+q_hWaDsegL?q
z<{JF~V)n<47~W!GZ`Wt!<>f8Mpale!ytPSv1ZkzPS=HF;vWZ>O-FvX+e)i876DJeV
z(Wt+~`)!+Jsg{YYAdEF|jGuNe{PB(JIQ}#mkYNy|Rn^C5YMgrEw8v9io=1Zb#$a<D
zu-;^7Ug<TX<RPM|ai3xzd-}+HAk2E|nemrnp7dZl7Ll2AgYGon+iS}a6_gvh9|Jqo
zV**<GRRP<&F0N{<$$+P*r$8r=%N%Rvq$_*aU3LWSI})kifKy~yBB!kJTkm^sPs~5L
zxnsgX;}eqAUzS<SDsK>$NH7RD&DM=>+ECrBuxzwh`YEtIK}ny+4<FJ>xmNIAoxd5F
z!tpi^1HCi(crDr7k*0z{8zkWCLI#rA&mQ_-GS=}q;vqBk0P88Ao@k0G5tz~$VO}YR
zkZ<x8v4&+!-FJ@Xdb_(5jkufp3^Efi4P1v1iMYa34sTtOta6GDq*(Pn&aqYsDs8RJ
z$jlt`8OWAB@)m<#LJv42GV#?b8h2>S=s|la;OBYLU--yEj!zT%_{j7t*o8<SF!RN6
z1zKz&$s-W}lxQprxPh2;>A58aV$d_#<A{ao0(Mm`o8E1de~b7tw4`3efaCX!&thcn
zKc^B@o(&=#65|()<oT^UvjE|_Diz;QEv<Suuy<lAfZVVDo2+8GWuJb>xU3>ifPS4I
zh#hFjfzZeNOfNOa#=(F|=4iKRAl4*Du*10Ee7*~@42oC72<AMLKKj!*rQiMS2{f-}
zVH*`({Oh%BbQ<^}J}k2kd|l5;P6X(ZgGQAm%r5`(w5Mi-Zt03_z$qz>IR4!~c@F=}
z$!^nF=%R$XmJL3mTaSLvevFBQv4e$$SY=!{n^I(sZnts4<LZpen<LVXb_mR04!``<
zH0JdtUX@eQ>)dz7#C|=`;#!-RJ9t=jWCetad8<Dse%1(3=-9p;AJKN>Z9)3kVP6d7
zJ!dLxMhXPw5uJkzg~oGG5bRT@W`Q;ld=Fz4A9!#OXdwgnQ}|9b*bW^7LPcyS$(S;h
zX9+_pPD*)lY+5$jCNylBjlX83dK|~j;CmWtL)N_DdalB@2nJq|M@bs*y5jzWC~UyB
zc;!d_ag)>QFx+G4J#pfMHh!G?PUyxV@fE~SnpCHPJ<|mOGgzy9d=5P=HAuv2uwJYF
z>F*{zY@`HsVq{HP08tp_909;@#^k~kd;F^%sQ$RZmtf_|*GusYa-+k`f%+R-+6;W0
zv-e-NF@91G0Jdlh>Xg0!^2}1QyrE+TLyaV!-;fQUHnI<ooXBtYZrNVC2-iH3{oz3v
zz>WTz7q&b_C!RE`C`nE*u43VLRSq)D4lDPgG?1u=km+9@`)x_#pMT>#FepSlJA^=2
zL9+Q_q%*mD1h`%Vq$PC2K!t|OjSaH_oH{hS0n*PS|1RDsZp2m?Is8Z+{^SffKXq~p
z*{x7U@Te8YMfw71Q^KR744xdmS1O%lG63K^;whX`WQ~_-xGZ4Qk*YLMRf4}U#N=%B
zo+F~9w*E4Zb1VsLbn=yy?87s7z3wz90L^3!KeRPcdOwh6TDOs<LF_-VM`$mYoF`Nc
zK%U}Izh#O7u^}}oEM8PKfT?L`$>nQ~jtm11kz#gzT2}G%pU<BPZ2j^6-lJKo7DKd6
z;exP(+F^nw*a&OWGKl36e1q?k>%nve6D$3pRKZ4ho0uJlj>Uv4VY7s%?D&truz^Cy
z{WCul`c^n@Ub1)n`r+Z9uk)%rg>uFC#-Z0{g=6<-DyY0Yv4J=;LLRXTYIXlic9R)g
zKo2*O^o5I;^JV{#K?)HaPEg7q{uJE%RA4VuEr3$ySL81A$!EnEq=(!VC^k;T$nbWN
z$qG{x0R+$r5Y<e(9a0BK3E(Sf!GT=n3oU8*Qf&I(t~zUJnFs9+87UB?e$Kd5eCt|w
z@&?fE$&ma6%0HletjM=2eE+|Ye;zd{kyw!F>B2-oxGVl`5|6<$y0_=@<VlmLgpO*R
z5fEt*?F&->(P_NwaHCZ{9gh6SgD4WSv*3P+=RicvRybrqDqq9q`WQczeHrq8pf;RF
zd^b+{vycE0G%+@uC$}iw&EUk9mXy6VdUMZf_6cfiAOGC%RBq3%{(0e99^Q8ms1QJZ
zZ_Gbr(AQKA!p(ABCElgUzwhi;`8ogaSkmp0kx%aVyJs)wdtxy>(m3Y2sljR@W(k5?
zyXUMWUP2C|{oU<bjV(E{g>S1?9Nm>0RMzfodUQp{bJxSzT%*=40wo)ryPj`GGSZ-*
zgoI3BLee&~jb3G$o<EYoWqo$ts;aLW!&ICO-)k@jc>SZJ|5&1`^`>Gh+8;YR>jYZF
z-LX)9ysNx86LcAMuUETYcg_l0_GDK0Nihc;k)R#KkPrLZ*4NZ7dS$a=(Z=quSF;tk
zo1Kdz8x5PbOGeFKl9m$wHlMGOD3G<1H_d))Bhan&Jh&kG;_=$9t&KOG%Frb7oUOKe
zpj>tD#x0KH)C>NH?<~yHZLdoh6s`}H8pTnB4)X}&VQa}DOcqpgn)#z9>F4Y}bi-#p
z5#aa`mBpDa*j-du7!|B|mV;=lok_mQEAU%D9WZ_(4EN}OS8}(u#jb!o-lVkpFM~3}
zAKPkbu;S9f8%kw2R0F5bMY$H+;HNl@7jfGGoYUWWro!Syu?M8^H=1Ix<IjIE{+|v%
zUYMF4wqQ%A(!V+M<*NW8L2GuVU0A&wD0bO<zRlzNl@qJg6u#`axGGRffDl2*Q{Xsd
z#p^VQl~q9?BxUiRAIi`at1~&-6U-N+A|E0sB!rcLu9NfT)=zhA4DG5G`7NB?rYja}
zJ_)e?TuKf=Ax<VjpJ74x(vT*FD~qa;mf~o`hJnX72eYMy<iWo`cyu_Vo**#uaPH{{
zmw4?}4Wz$?u#(<5gC%$xNEvB$91Ov9;_1B(=z6?O8rBW8;l$)@|NC=Zlo|QH#jU|@
z<F2DA371dxs*g<nxIp%TNm`xF&r;=mTB;Km{ylyZj?2lV<d2;m^W3L)usZE`)@^_Y
zX80o=@2H&yQ;%hQI<piRK{#O*vD!e4gnkg<qx-QxQ7Z;%LPqgYkOPHoSv;?XHLRf7
zYzHb1djYv2_8S0TM1UynQCV483SlCwDf8sPg9n@N1!up|$-?<kXzYVGgHphu%miVp
z*8kKe_vmJ0O#mZG$RnZd6OZZF^!7Fi&%@D<^%uHiNiiW~(`wBw^#niPIJsF0-pZcJ
zk9)JTL%o9qj3=gE-Lhqx$Jjta+M7g==?wn%`1$MVj*Z@YaOFL7M&?Udz-BHdkTDAC
z&++h{K06eZ(C`{^yk5LsOQA&oP75d5HHAQV40%Y>y8%>iA!Z7sGZQj`iAntu7hZop
z>>l`mfYQP$@S*|8^7Kn-4bL|GO9MF%i39(Z^m~>m-4Swbei65)TEsQA2oR|S)n%Jh
z+Rc;sYgYZ&UxnRJ7O(Z@d9#QCG0efxz{$|Jft1=TV;`*kq<BTBEk#(2B6~<NaG%(@
zCwM_R+z)GILMJqt%hN;7W!#4j_K{Uzpl^rbjiA_rSAs5otN(t9YR2KaodnbY7aPtL
zD$#-9mGaARyJH%L!hYub`IK(n@B@?yaoj)xg%@z5P5|l43dp$My6o=JS{NJw@Svm_
z3DKlvqK&n$&&Xb8m7`dw$Jj}w8&h)S5BpEyi@L8AlnE^`3qu*ClTD?dbt5|>CLF~1
zt%Z)GWb1x>88Hz5G*hn*P+2U!)$czp1s9JE(`C<c$4u?B)n%`I4w&T`ig6pBjh^z;
zC&qJ^Nec9?MFDbd0If+wId0dQ3zQ5{3GT4#dqNdK_2MrD$uulCcBGVq8g2f>Sc{V<
zPEd$B31+bjU_Qg<L$Yw_i)e&|JQcVU9GHqo*MP$Y6yQYILhz{ob6pO7A2z?xX0JPJ
z3wYDAlS*b!!<(}fxwDKg()VD};%FEJ@;r%wp=PtRxB@#11(`1cu?94*%~om{0d;tz
z1pK`Vi*z**yH6JzlHFJs1ne7ljDjh`5QWQ|lq7iIoZ#kX;yPW?nzO(q_fo&`Zq+J8
zj=5p@!9~AZTzoztwE2bvQ&5Zu-Z?xY_SnbCiU5%X3%sAp;AoiFaIuY-h((01Q=Jot
zj|EO24;>$DT_v^$0$FNHK)e>dORy0Uj~49C=fQP{NG>pA6ae+;LWU)#3P2Oez}hTX
zVxEoRA*3q-lv=W7w=0!r3%EM{bSg;}U%*M@0)Ay_pdy<I!tsq>F~#XRE^(tM+DZS_
zw~K72_@GGxhY>4#om_g?!Q*l$WX}YF^f&X4a*43l&abO8V)M8oNqU>UYoFUW!3~^+
zsE16^NC6}nC7vgW9)mEIVhSi<5Gr{n4%mo$!8#2KFMXP@vD9mi(m!A#!EA}{X`+z;
z@n_W4@!*`z0zlORfejCSU-GZwpeCvq>I^_gK{$AE4iJz@g#i$cZq)cNQ(=)C9iY+O
z3CPxz?m2H}#=-z#i#HxYCmjnzk@(khx3Tj@iksGs`_2Ksz^9RE?Bi6)TG4Kfw~G>H
z#xrX#d$=r-^Y@86CB_|C*4_i(AQf*VB91Cy@Tp<LK+FE)KQOIh+sZ?;D}>=!k<T_6
zU-xUI92?BK0>h@i)u$;KHXh?=g%O`BSgW1~)Yy6dU%0yL7>}`~E9Cgj!IVY5<p{MX
zd$gZ|x4_glIjZ?tRVOh($kxH#1^k{9kRb&GlGF;MlpdnyiEKsedJm-kN2L?HvS!H9
z8Bd|DK<Jm^-9Xw-{rPPMkB(idjLX!{06p;n1+&27#_31=W`ywG^*Gb`^to2YU;=Pm
z0<=g|jr*quicT(o*5Ygzuf^VvJ0FAS%Oo>>%#t%e7(ugf1aA>};wrma%BIh(Uh>03
z@QDrIMahXUg^-5}N*P2xV3FZ%8h5iEKD|$!|NfR`cjaMG#s4DYNxV1NEubgDZBK@Z
zjaQKE4#5#Greh+iaQ5Ljr;`GcF?N%QamGiMfI-9WO#wDQ#v!sn7z-=x!C1p7UJjQv
z^_^iNXp9^Hz_^Hp{cELCjh?7A6uS-*hmnUM?rW{T?aN{C|BJY}fZlS0yOgx(a702X
z?6aF!6jP5MEqn~^pqtM&+2xqNrYQo4zzn>%*fnb0**anFrqz_vUtv*)q70OuE_4(W
zWt}9o_6USI!_ceXOudS&3qGSoxJ{KX$HS4KFZ2NiFvUCI(V@69TKLFFivXri=#vC(
zHviyyh5lw9eNY8h?;Vk{OVupkVbGb!N^vP*68l`&!y8K-v}$Ve@i-7e4=RFo8d6b`
z#06_*6Fxi~a8}JOn-Wd`B@h?ml>^FwbC2%4(ZPgQWV9WBVhX|RoJ9oz5^Bz#T*=S>
z31ahHaO8>R3L{Q}kUix9<q$#f&S|-BWEFz`f><vi)0C%f4~|@A$ToKF+6BeCD`MGf
z(7Q*~umDje#Z8<vs4RPOs$jqw|4L#kuo})EQ>iP{7M71v>ra-tx;ug#p2ti#Qe5j<
zu%?jaLZ3Asmkq863N=U8$IjqSr%Kcg9>gtVUsnmz0rEwvD93{odAYIyz!IiAW#k`#
zp4%tY?DLz=E~V!@sBgSpY>>$ak7$;64gg`o3ybHuUPNOFh=ndon9vF@O|T<(G|wQP
zFRMeQfLhrVobR}ki6zB<rM0V!;e<R_2dbh@1r*!cU<gda%c{!D&$c5Y>|)LnSfszN
zb$@s2C!h);TO2;mH>lLv74t3#87LaM`ZT1aca1}rs!hK30;gTYLkcd2i;UuGPswZy
z*?UM$t<pO5`Wmnu1N|O0O)H}^clqABQ3<U`?18Ut^B$_4sv3l>T{qfmXvS#;_GpxN
z<Wds}IR-F2;*8i5^FRvOR4{R{GTdqLU;lMNBZg=iv(z)6R#vz~bPByU&KG@f|Ni|_
z$l4#pPe;nX6HfMzco+d=O=NukrBoCA{)<wba)~`e=11>M>aF`y9laa?L*m2tNMRj@
zDTH+ETP%{0;`HtH6A(=NyS*KGr7mdWcE><7fx)P+i*)2+Bs3-sC_}D6qY)TxR_I3=
z2K7eNmkc343BFM`)KrpOgL@NideTeKboQ`xYlAeS?i+3WMuRsfdnGMIq(*{k=l_?l
z)%XxE(v>aDi3lB6Tr`k)T(_11xQAzp$jcryU~(9-rCSswm@ujNT=zmZlJXtTHcw;>
z$ah~U%jtAR85le*jKxws$%svljdOwa-4L*IfVG?x9zC+-X+jx^BLc>c2smMi#Rvq!
za7}trVj|g{!HyzJYKw}>?W3cZ{Givb@11@>C39bjPz)vpXh8QLIf8PXEAV$XOC#6i
zYHjmjIF;v2k3^CcwL>QkR~J8ULa>k7NlKo;z(7*~Wh?sB%^n@*41owAD$xlHm{pkJ
z-fKfA7daT_<TbbkJnSX8#v|%c9_x_w(&PO_SGcAM?_sS<;{vKB@V6a23j7#c98qZd
zq>koQ>TmyN8Jg&J5U736%K>2*Cu#N>qqu8_spJ%w>_^P-M)A%&D^CX}TvrO(g256Q
ztC`bQN8@{<L}b7aRk|Uwr4kEt+1Bqq6VQMF3;@3Z?@pY#*eW{p%>$VqHMdm54nBVL
zhyn2aVmSS#n;X~MC5&`$_MB5!5{#>+Gs4hk$5hp1;%%7fq(cZ6K$n`w)4=!B&4_^@
zIrIFrN*6@jqQK8+&>I0-0IjHv=JfoBGPXgoe>qU>%?7={rGY>p{x+f^^ImQ|Omfsj
z4ezW$G;2T277+36JuB4{WhQ$v=t9qpGfL4VD*tE!e%SPz^JBHZdGP6V8dSV+64z=8
z%m;I?9;qrc-#JrQB`t;}PF)b*i9(i57+oDQwfjFBVh#d=IHa;$!I7QA@cNA+zgeY8
zqAP^V)88K?8LklP<phrL{rZjv-|U-|?>*j4C~orN$hw#{cuOT#1mFmLG<fw`CG{;~
z>T-ZR9&CMLRj2H)d%OZ7yzJ-9%n<5=Sk-JYE&Yiu`Gl&53pqx6FUNm6q2Knf=qIe8
z8Wap0zUhK$$V6Wp{a68!2C;zz@X=j=ea1BIvIRR<?#2rkr6q3*am=}|)gmYe3dG9%
zBIrzcD84b_qsto~AnXh%Ugw9(SzWP+i6n_4Hct}EfkknmK+3wQrfhr`rz1?P-?iP>
zw?&yqg}B}VCm!MXH@BjgDq3Quk@tLjotB8!N74;U&v=6IF+kq|&HG1aHpdkLORM3&
z(!C%kW!3rZ+vzw?;GvLjL7ex^DrUZv)Ekp7`<Xm*R9OAzacoPwKTTz*t}<i8u4Q-6
z8rIm<Avq)NOxEZ5S1w%Xm_DkZmn?k)%7Q}ia-+EV3<;_HBtFh8uVV$D%~mPl&);MK
z53fLNf{;UI#=w-j+p`5zO7CE3fStQmC2xc2ne;h3hZoJTd3PjDC#`UUmFLOXm$q(o
zEKINVlFeBq=yT@7<%ZcL1mjYwE?`AiVKs#Y0k%yruJh3Bzd7@LGEL#6rvf}k1woms
zt_oh0*|8~YUhQ0|BM0HfG@e;}5V0Y;dV19U#K_~^BsnNj(&bAg7I?3f^QzC)uw>@P
ztqN9Pd*3B-_avJL%?A`C?6W7RwA}i<VC-q^0pz3I$HU2beQ4UC_{j?{Hhhap_n?m%
z<xcDAtcSY5Y(|m~@_At?5=!KCEKiDh3BPJn-;d8n1K?_)8)g1;E(p+VgP^}o^;p#8
zBaYA_e9d&(!kPdt<n;3C^R<%BPX>Bmk98YUY(hdpPR>GDl#n+~&n#9zwf>>OK?<B3
zO1W$M!rO2$rW=?)>X!x^$^e;Mir@6<F;OHXr!?qIj+@wG(|Mx;QMArg<oFSER`@zZ
zaas6rCK`1gc#%_Ar%}>&7gG8!s_{7GWO0a2_Wby%FE9Nv01^9m)3vzc<uQAzOL9Vs
zx_RN%gl9zS@B^13G1sv%$ciOw0lh-Mj}b1a$@H4g;ilDJ?E>2;x~N3QY>t{1Su-1|
zxc|55l^2Rb)5DRjMKdU4g}I-Z=aDgeE1~zCS>+?o10Sj?bcUp}O?demw{77^TZgob
zjs7Q9@WJEO#w@9bZx;IsWgL=z944DQpu&?99PUUa91{Qfh0UhVoS)V!!1mCLo&z;-
zdVbpW6`xr3Ay=U6!Ol)&@eB<B%ZT{Kq+g!GZwQGcMt1bN7JA&$X*ZW7Vh!xEB8<#r
zRs~`bk*9z~1&cNAhh<hTTxNb6cL8h=+~ns6>DGHl#C==Wit#!2P~tVzMwB^@fp=bP
zR2<9A){G(nNvv-;;&pmQu{%x*JE-3Me&WEIzot>B-JC7X{*Yamx~5+*dPZszVhQ}O
zj9!cxx~-%a6YD%q{6$v`Xzou5i9kvRUYUxat}q`^j~ceW!KGoty{^J<X4S|(eLAbr
zk_{VmUQdb{g2n9`8Csx8cqys(GP>&tVlEJ|f#p#i6$8@Ji?IJLBR2=MF-sA!az4q|
zQi?qcM|davrN8aY_}h3+GZ4aJwheA?NQ=)wQGj$fY7kWlnuU21;04ZgV1xv^aJ!I>
zUh6s1a##Eu?9%8y&%nKpZEgt^pcL}>CjkiqDQ3SQjPryxz7L*Cuj=Tpiwj(o$c2wL
z1)Lv#GqUCx#r?I7E+39@!>8`sIp@yIw1J3D+^jFmJT$D?`L^De#v@LZTUk&AMx_@Q
z>5AEUOEcdFJ6c<z(N9d=Orh;iVco5%bq0wI@-q`W?WFY*=+>K39y8Z2ip^=G@?`;x
zQS>h(2Ob*dO&WEgQc`u_Y_0+3BR@D~VH2X%j0b0~#fAngI~9UZuz%F)GkZT^d{~BA
z1!J06{wfF@DRH}Gk}Xj~Kny711!jx_j~Kud>tNKrj?qc|Lwt1C15l4x>LKo!1DVK*
zp=5N*p^_QORmw=Ar@Bp5JY~<S#=ak|?Isw;QNx^BSb1ozJNofJ3+LFeO#IU}15^kJ
zEQoJHGM62eIp8YJ*4uS$@n8$XhpCA)Du4xuvZO->Od0|#ZsF8hicl~V#(^?+z(Y0~
zkS68k9NsX-HmVDt!HM_5w)wLTmY0-l;L?2d=<>F*h{iLkwT4!v9aRsxx5@o6kB3J?
ztXP^*lB3ljf&BNUpBk=w<h1VH+8b}uyG64Ky4}3`eRSf@oOJd#6z4hQDV~yTi}CGe
zc{udLcPJ0ZdYCU%tP$*v%PbQOSj8sywzFT9+<%Mpeb{R5I{j6r#_{I&ml_%xNL5U0
zUsN~Sq)7~eEce3H3)pVR{A1MeU`x!;pO=D0UdOvaijI~`vIwg@E6dGQA`=vIyH;#Z
z(x*P$=><X6_(tg5fQC;hRrV0JOFXrtT<}UpO!CHOJ2cd>n{#t>lb5@dTWsS-GMrnX
zII02WnGlowPNT0$4`D4cGBUDjU6eqK#CWW?%0ktmX^Y*+|3xBKID2)q_5y*$T>%+e
zxMfU8rRJ(7OO_zy7M+AogKuCH%f@C&Q7rh+Dyy#MOX$~$Z!VZ9WN|?pqYpN#zsRT+
zIj5iq9sfs2lBr*eI4x}Wxo%6`uuqdfh}9uwC#6avz7T)lyK*^1dye~4Kxsbgf8E?m
zsi@?R0S86k0G<uf*Wp+1KkIo%pP6j6qwW1xo=M{+#nQWFe|}l+*S#A-$7pXLzb(#A
zXYva#T`Ip^KLq0+Lw;8fmvQ@lYirQbN?SFMq}L4%i}AywG=e<wGkJNVEEJb30@$Q>
z7bx@}A>NGSe5gDjL<q1TlkxIo{v98mI=+E0j_K2L*F8_Q$2SSIusPUM%RVk{BX*>w
z?(X6wSv{&BMtNEW_J1AZjXrv`bhD(SE-p8a?TYwosm8ZhY;3J3<P-ijWYN;A<?HK<
z`WR<?H{mUtHxASj-3svX2D&qNfuK{IiI-WVY1#TiQAvAHeMu@~@;B(|IUtgX+9Ih%
z0FWdgq~dnTvl4A4b*=AZJtd{qX=rHBItbVch>f0j2Nek5w#MyInsDYg@<z{IwwF)p
z!;+8=^AmZ?!bGpfp;xT7AIEk?M5k)Z4t~C^azA(Q`%Qgr)^PDp4tA|Ycf&E$oM33_
z*Fp9(1J}tmpu70Q9pLx;1b`_)i5Ig05gG`N@j7I$Ig%OgS}Z%l2^`P~8XqVRg-^Vh
ziORNyrxhtXabRu3$xPX~2%^@(4hdWkeaj0F>q@Qi=n^$QVjGUr7*5IK8R#YfaQ7l1
zArG_d!_?@mkAQhm!m$M=9;8-N-W-)|VWL)D7uuD4US<>6AYz2blK>i!=p=+)R(UVW
zd|;M^phEm^!yvj@7(y*UiQh$*G=39@yo6Rsa%@tBF`yZ0c!cCt0O<_?SB$op#{YdW
zJO`7?5v6|Y?!(t!ulyO35Mdh2#436_Ie)kRG0ZEk;aan6&GSwCCG!1seT6~tkFh1a
zyV8Sf6=DC<X@;^p$9tnOrL^M4Gq5=V70>&0SXq&j_uT{W%~3_6hPNb>Ig&J|i;BjC
zezLJ>>HH~nEHrc**tYwJt&nGkk^j81<;|IuB$J{zXr}37*R{o34PEa~e{x^P{aj+R
zMBrYjoXKa7y>fMGeISE!)ia*rb&5N8a{P61=s;_sRI$Y;@@`SE#Lyom^5WnEj&Bo$
zcQDTO+FxaMyXjkw|H-QqwvSx$?Pqs0TWoP;q$N05ed4PpwxwgUpQ!GC!!4Ekr$6}*
z#VEbeVDkEQ%@w)bzAAtCHPDWIk?;XuGJ0QZ3z=G??MU%=9~?{ym^LMocNMST=()FN
z+ZK4d{^u9?Aq7y!*to9rgLW`h75s}67;5xrXz04XJ~1MYIm7lAU=$4FXl$DTStV2~
zL7mWB(FqBKGTxYp_*T(>>(}Dr?F|pb_ew*+ujW_Ovm(Q$D0Tf*{+~ZfVLR7W39GuJ
z5YK~gWCLA6^Dg|Z^ui;<UhvK4)52&Lal(D6uzD^TOfWLwEprGCUP(noqBpGJ-$g=3
z+HoLT{M;g&Zt>Ti_$Nx8{+xJ&$!DbA6&L5370a+eCy(<;7wi<&T-msY$7QB~8G+a6
z9jXE?Pwxwz_|SKFq4bl?fIuI~olBn@f^M<0A!ED36*X^@2#UeGqiFoQwUt+F<CEC8
z+(h(_|Mj+O)TUUjhsd&_wF|6@@R$}Z6?C1{rS2A|4o6H7=H{DJoF=4p86Z{;AU&}g
zAkYd@Wug6`U^qhafw>Sx<6wc!bOuh=TsW#70YHSOB|yQS3W}Y;Owye|`W2COeBVBX
z!8?lR?N@Je^*L@idAoi?iNN0<>Wyd@Y!|?o<Z4rSD^_Jg1{JX>*uXkG>|*?<y$>D-
z`T?(O0)9ogZx|~p8k&2pNNvD0Eg2hTi};C<O(D(1N3-fDGsvz7NivHel+$G>K>Ff@
z@%(%v0n_kwy#tDW|3vyM`otd3OYK^*X3Z1>tz-GkhPf3Z@GPMD5Jt>-Xzq!sHJ`>{
zSZ84{opGaJrG<q-$D?`$>;L@V>=%<we>ACDSr3@pPv#Qvb3XoDi-oZnZTI~0vBBZu
z`rb<u;=Ag@&lMQ&YRvFjnzc8<EUqo%wB~S&{{5=Dh5PIh{8pQ}!tH-VpUm=~VC1pt
z`#c2<#)t4h;3^D5HN{rzF~c$1Cr?&{i~o5>FILh-di>R8*8HOy8bOhf1)sa{Yj`ja
zWP#35(MsS0YDXeg%nPwXJ1@=$0{IDPFAJnD2p~9JDp@Kr)_3`_-sD*dD(DTI?mB&s
z>4Ai#Ym@uWSW8*QU5L!7<yw3z=aYE58z$(tonHl3XVmQh49hSGg_X=l{S0)aW@Z}5
zJw)jkg9`A2&1t}oQR7W&5+oo~f$G<<UrTV;SXo(R6cpS(!~+Ej>J2Kaeglasc!kyp
z%C@Dm_M#ea<%;xf<E3&hRQUDmE08z<F2^%dwymu#=-IRB3^jBi-M2&VIC?HYI6smQ
zbP5d(%@oG~4-SxFpf|SLx4K*-9oX^06uhXwlfHX+=C@c4_we=G0{GY{{@yqYl8d+2
z&0oH-v$3(IwV;fL0^1*5%#uL{YFZ0R6F7JSlV4`{DRadkq9h_Bf{0!Q5>7mPqSZ>k
z<3V09l<zlx?bv<;^j^M;)J=$qafpeDY3Res0pZewTEjg@)zv%dzwA3?-1-2OUSRc}
zCD`OkK;o^}lf_acQBhKoAuXd_;H%qDtsv|olX19*9W;933)|oQ!IiUj_*9tQz-O^>
zP=Cj%Q~Z!Ixq{;-VHocRUN!ZiyU(mX;Awao#GH|OTUa=m^K;rE*1-opj}{VvfvAvd
zY!hN?OOUa~zq$+)mW!UPjZF{L+%#oeWv#2h`vY(9cMp6d8F+<p`<xQqT(-f_al?xs
z%0FCmdmXB=CLG5>$LeryU|>Ln2SsL)6H@p3OK(wBoXBK=u+bKN79M`f5GuADZ98P}
z2)D4j;CSHf;jy9rU46apA-lshPB;`J#L*Jxm4q@kec=oYCi$@+zkGSKHp2&!T>QPo
zLit4IHXSOG6JukaHrWd2^XB3(;A5u-T(KidhgMIr!kX+Az^2#wAdq<;FL82pM!v<K
z9V=!RnXQ$R3)o>(+-ga$hVD+Bi{P+ql`Vsgr47~^(`0Hm?~AI_GnBc{<mic?)53em
zcAcPm=2XS2yKV9C2m@6ILP1f}+}-_F()FmYg}F-}<HX0WUpc`+VbLb}6=<^sVg@s*
z=>?<L!Vfc45q$f#WN^HgV|&#}Q_~OO$`bh5Ib}~))X+LwRdu*+WN))qtq;t=L-G-U
zohZBDHT;!0rKj)v1m;FgCGQ76{rMk08Lz@vy>H5Qqae<&+#}ad;8q9F6S!|RIm!Lj
zO@ZU<21qR3>>IPF>Iw)A#pC-zcSNmHj5Z|g1wlQ@gjbwN?RCmMOyg!Qx)@>q7v8kV
zW__m8lTnA$tJ&rU6q6tj7zavGW~7~TN8ryxy=s#atLAb%>~m7vvNkC>c|+cdqvs6t
z^s*l4&nXud71etevpbBwuLt@uN&XlR-^Cm~I+X!;ukR`n`+9Dn+?;p#wUuiJzLYvh
zI2B%>+VTs&?~;^DT01QEL?1osJl|~R53^@(9ljf4)iv<2y5QgHzFt|hMtsRm#M9T>
z>;bASocH1fFWpk+Pmn4{{Zu%Fk^XZPB_cRT$S)1N((u6<6O(M<)d*!EuRor3<+jE~
zl6^uGjYIMp25+R%8ViQv<3r{?y)m>aW7sC+O56qt&X8X!kGe*zxK5A@P|uOPnb6^@
z@)^=1ITk8R4_fQAvDV<1S$#`tI}7C5Az<A>8JppmVEUZUoDK5w1=0sv*F*yh!%PuZ
zeqble2l#z6V64mt7h*`dNx^JT&e&|m{REfxW*6mwMJ^z4P(=ay0P<nBl0`i(w{F=Z
zgB@AI*HEBb*=FCh$nfHaYS)0S1@0Xk9n@@?SyjaaoQdE?+!C=KC+o!j44v)Lqv__8
zN8?VeqW@mRYruD0BXDm)*Vc>5aVNPa9KxH*;yv-CCf-L~tO_(;;KByKuYXVuKkh9^
zR7sgaq0g3<RYxxX$x=TqXZ~d-R@&lYQ~Z2=kF?E+pBKUQL~zp2C9hBe+vFQ@q|qvN
zmUE78-M-cXr}ClYYijNaM<GX>Z1j%bk9Mw}d;WA50v=T_;yrh0=T!eB`P|RX?<kv_
z&Szg;YqKrwM<1%<61ik{E1`U`*)5!X?c32O0`kpE%y2g1YpJ&bl<OzV_2bp*(7D%d
zg$SGC7fv^%0>5<xgrozqT*YYFA-d>Zo}7yEKVUn~@UAWcx<Q(3Ow{G?AZu9ypNY=t
z)yQe;rs^>K0rq%A<X7hPz$$nZALwfwRBzjAH+Zh|6k*AL^V1$GlL>`a5^(kDV6UMR
zCR#cj+Jv#V;f_^QRh@%(aQ}e=_P~HinoX%L09+~Azy1A6msf6-<O}A9bBV0)p<~$F
zFdvde6WbzU3E3fYjmw#1l8Z8NIG!8~%R;C?jq2UNnp5l-?#E`RfnQ^nEQJ>v`uOHE
zS4Ma+sEJr~9CYZv*|n(D!@;v~I7w3!(<2OKFyZP@L<KfELXFAdDj~5DR;7>ie;*dU
zW3wh()$PE!cD2>_UNR<#32T$AYOu#;QDh)`-r+sM=*azHFVF%EbDtn1B5=3O&cJwk
zM@L0^8~3l}<1^cn<@7G#<AIWE8<Lj5K9X~jWCRp20;BO{2J||EZ|f$}e%SsI=MCAv
z0>ULaIy)cQieX@lM)o>2!XQisU9V3gBC;_5kV%2C0NbD6F-f6lj#6r9e+PO>!v!8!
z%Cy1T71JK`9&djy#ztKzak(SZTo&Rs=BAQ5h?}WW?kcD|piz(s^}tZ92k?dUOyzSF
zGvfIpt(IBLPw=QYWTx{)UKAKBCZISEBP)rR<Mi}&b!oR>r2q68rYeeF&p>jc2xf6W
zce_mQZxm{Ni&Yh;H5A{ZngMnTAqF21OaP21O~Dquj$`W7k1Z|NK=Uu)?0M<8Jl8)T
zzu9?iJpgKG03eSx99b-dYov%@!K{z$lzSHy?(MGAEEcJgjFUib6(0BC-sBFakJF}2
zqj6>tFYh$jA5|D3Y@UHO(U04kk67Y|OH=XZp}6F`$Mf)T-9gt`H@8OWP6}UFhXB=S
zvMh%W9C#gR7oU)z9~bI$zU;(ex4cVxr?rfV;p=gTiyO`B2hjA;CGna9U&N|0T#a58
zotUbG7ktI@sthg=c6PMRPaE7J<SA<X?9s(pK7pYI$lUbe(btWxZW{2|zM?>N7Rs9|
zbp7DcYPC#2>t^P;s;VkcGp(gFQCi_~^u$vag+T4rrQQJ?;@_X>gPbu-gDoj@l_p}#
z!z1G4raj&TP`5wkgr1&W&y*@aWvBcT>a*HVnWiW1dH4`|#t{twBZ?mCY$979oEKT;
z)5vFf4I9e|oN<gDU~W8B$C_>NuYo*aA!Oriasz%KPQ(Br+=rAT$Of$MNJovs1XgiJ
z<Ee2u9dgx92gLrY2k+I$faP}Omtxxpf{MzM4SRtvc|BT(5_Eh?WL$z-enVyn4u4F-
z@YmZ<nM;Z>GCo6FpI1JB#28eSH9=l+>VP)vKXg|EQKUq1kdy2T#Rt*d+a%;&CT)wF
zha-sP#W+^(Z{jU?MHX^CW<m8T$1X3lf^wn=0ekI+OjOxUCD0f`oQ;(61WZLmp9i=h
z1wump{d9gksYc0b4JjWLfY3jZTN%U}zuh53I@3_Ka&=wu_m7eg;13FCZdX=ThR*mS
zFv!g~`1{N6gS{r~9rz&!3{BlmnI>)9jf+-#if93>1>~u6ma7mK1^S?I$Uw@<8GWRn
zy*|eDRa{j+y-lD9zx&5HIkDo@5;fdj{a{;v?gx`ipVU~ZaE)`a{sJKUP#k~xqm!-N
zTS|X@4#)D249+<(1_lSk7aWA$QH2LB(2?{@(oy9U!90IC4GkETJm=O;yo`rcFD_{W
z4u@WtKzJfkB2=gK)IxN|*@N*kTqf=`R+aMdY>C31c#b+c$21oi7#hA^n-Os-W@`nG
zweO30z@Fg9RAXttX!q!ceo)Ya)CUd;eiRdkyBRZ8aHVuAAO;2OFita;6s-+d>3u~+
z1`#26sq=upsm!iNGCBE<wRLn#zqu(_9ez5dEhQ0sCD#AXGo*STWJAe7@Ve)b8w|pq
zaDH1Or;Wg9J}^}f(o`LpJReBtuaxNMSqQp9lUTZv@EbIn5_k$<88!?`c|`O#6ibJe
zAK^il1!z6vN!FAH_fvBL_kAc5ASuQ_2A3X2l{G~<!-82-=k)O5Lh3H!;tIXnI0X1&
z80X-WarG;3dLZ}@hpwQ#P`>3?bkr^`y<Aby=W*TchN*z?0|zfbTT_y%yO%58Y%7;|
zbDLx?xPzyDW-D7hEp6f08Q<o=Ya}YoAMn(?5{`xox@cWOVhO=bK>Xm^IMHx_XE0@e
zFxqfu8|AH;%R!Bh0C>cwFOe;%=f#h9!afCJx`HJfPUxayhlbMFzY&BmX`2|t3DkT`
zd8O9M6=r*8<@+m#?~i|)cj8f)eG{Wgr_J%ITmBx-8y{RzT{rPHgcnRk<c&kY$r47m
zs7g|Qc#2c!B7*cDYA<D?!fG*HQdWjb6=oT8SC>w&2W~A5$kc&)s(iD!Ot0NPT7XeV
z%;?0{6sj6oNEa6VHjbF|LR22wBf@McitoEp3F2kw#73(*Q`kPyQGpM0fJ8;`RL|bk
z?K{_3o$uEX=b0SixBTY<Zn-LB{<$0&;wkX~?&<NOqdj|wNNJ0AF?_YLsvi-2&Xx0G
zuv-<p)1o56loTFKom;H|wkOvr#wT4#vFOZ=?9hube0fCRRIPT|s@Uo?PF+sr?*eqh
z+h>7^d?;RB?$+8U`tHzuCpE!!`K#-+RTmd-@_i7-@jdF+;g`|TnSbL_PM*wJyUn3v
zry5o;wKd$_JUmxhJM|3&5N~?)htcWNMUvYbRQM-)AFkr&wzjpExt~?{{{3tq?4<U@
z^=7d+=;y(WKEro9HrJS>7Pa+1EEh5X_#i5$X!0xpOF?ew%U?iG<GzLA5(ZL9(mgdF
zTNEpA==bQfQn`r(dNuMHaNAQOZx3O;qrZMorWbG`nx!BbR`K1uzsn05ai`cSK?2(E
z5R(Jy1Ozw(KrOE4lDcmg+9|Y=$)M?tD%r8kQb-8~Xspn#_+tN(N7Uj4p{}2})QseV
z^CHg3>$CoTCHALA(QsY(l4ZMh{J5%Fcl|h&bNoH(hYz!fu^GheXN{);B|~00c(ZD&
zCKZ@siV2I+)YA)Ukk6yV<I9&DAOk<bAy4AhlGIu&VCnz}3iyj)j6&zqO1x<ZVSxZV
zWWJrYs^^;ApcAW`vL@F_Ja!B0HA_2%$qAaN*O2To3`v&A4U6_7lvKE=)-Ahx!)zof
z$sfTc<!@>@8Hm$V?5fbvS+gl`CbSF5k+Z9I-|y<`k{FXV>zY&Lk|-!BD3Opk5T*?{
zqsVkY%q&Y3%-B0PNbn8`p!Ba|VbexY(dJaSxm7OZV=DR?rv5Q$Z-bHFW5MdPYPa`^
zd#gSso&MIKwtAbQ)NJ-i<K^7&K%r!vhyFHHG6m3uC|jWFWy4u8hF(othXGH%)uP3Z
z3)(Koh0~6Qzin0q4`}kT2T8WNn7T-q*n=L<Of+i)a?b;{a<uWTXz;uF5>|rOLdG1t
z|GeF<Pc#KXyhHc#oLm}<in7iTU~TgK*AKU}O+L9VF>TG^ZKiPc;o5KW3_N5ay?6$M
zJ(46uge4J|7@Wdv_~1da=)EJr9&W;Iz$Y+Vm6Vj&_4J$#>M+XhUEqwFIj>|m$yEfI
zRDdq%!G<Oqpw<htUY_Ec<r%M&EpfVV0KT4Qc6eG<VA+H-12uAB#J0jO3uX<xP@u)z
zlzpqW;=Dn23mKqV2XG-D>b~L_j{m>&_`_zBaD#9l-0F|8ajT9FAO0vJzVVRXZ?J8B
z{489n%UXQb&p*%OMeo&ze;^~mI5qU=>K7Y=)fE)%D=*Q$dgRCvnwx5Fe8I@W-2%*w
zdbvcC4%FGBan(xNzj9HOK~b5vyL)hgD~2WL$rMpKRZzf*tJ(^(2?C0nnzVgotUZp_
zNM7ta!O*pJw)fLEhNfAU@QA3~GH#Q$8$u?z>u29Q;uZ}>5J{tN-_8bOhZXJg!(Coo
zNp0s{HLPWj*0b&ruy&D6o0Ne#0u?Y2euVBTXQ;;>#@galIs!lvln&W-BI;L~M!kHg
z*)<ZMynN3%bh%v_UE;KcfwzAB+2!E7H*d5@O>vDNv5kV{jlyvMgb!h^_mVK<$5sgd
z7VI%4bh#8ZQW*>Ofas!TLr&}{h<hc1y^MVrCSd}d5vzQyfl3TZ5ey7b`_Z6p_<dQ{
z8qUNq%{P~W%2_Fl0HvDy^5+x;Ka1zyYy84*dGd>easkEqGD<B%L$!;mPAAnYsxvEY
zS(Gq%X|}@l!}2Uz?+wd0g@s-#@l?L~_eLtOuD%fx!k>0L7f(Ml%9Ei=&z^JwBQ#Zr
zg}FPsx+)zz#_NH&^~im|jkBTDX6$Mvx#l(N;LRcr4309dW>r;HT7=M<6aOQ3Jt%%<
zJiBPkD*7H57B&U{oQKh#zVl8Q)Mv9{KEcLo6vmIef^2}elDu~;&ui|5htU-;!%_qY
z)x4;!PA>7yuYS;MD2SV!C%#!<N~z_RZUqnfzbvzh5znF`5l_h}e13be&?497#D<5t
zySM8~&TU%hrxc_h^Jtf)&L4D3xdtuRsIMNl9H)6SUk%5L*2R{2kS%6Gq`N-t`)q8c
z$f}wIy<>>x|AlaM!9y&1lI*@1s6SzTyAC%fWs8M8gKqKtB>k<gD0w;4elk@jVqB#7
zEKI1#ahZ$=cA!C})~NtlupeV-nuZ!lZ3wdomLtLCp+nd3+C{hr;L(9-nvvHFG7_>_
zV<{T|&IAqZrZ0PA=i?BbR@WQD8w80hxP)`m3=c0hCR&mjVDJYYJC&hQUc74(e%A3_
zW_~=5cZyovB}7X;)UkX><L%kJJ|QTjzu+;mzFq8u;Q5)Vw;&1wzStLe%f$sGr40+n
z(G@FKW`b}=zTc_7_O-T%w57awJc!h-OVxSMg%orgTZy3_|LR95{U<R*sFJ%E%N8_q
zbthGT{K;b=BxnPB4J<LA;Kd-1VMH?lIJhIUQSY!FV+Dks^99u%BO?XC`i8mlA~A`9
z7sF@Ddv4XkcvVROgp2iSb+n?X#ZlwDGIhN3;wE{;7jFOmU|`Q0`Rp3e>{#Kr;Ch$r
zUd>m`wVgaAKm0H0*S?f1<`=zoH>eKbGI-eS@tcKxkl_D}pOdT&CBKdw&6kt4xT(3w
zS#(Ef!I0dj`Qm2+qgMF1j>-Q1u1fhz*p{JXQAx2~O87z|JW|4<qCu;wH-mqA5gVJ0
zovOD;s@v<yo@)xFhR<0FXO*p6yH>Z4!wuHs<$3G*ijy%MHY5D(j8DRS$kU<FX5=x;
zPMWaAC*f@_d{c<pZF2p_%8;IjqSRoIB1AE$LwM80=j<Z)9G{QoaYR9oO|SPc7B}S?
zYUkZSCo?Ltx~g9aGF83lNBbT~Upcp6`Ix~+pR(^O#{X?cgIlhyP5}W+uxQql9z*we
z=m9JbwJnwgAjV_ttH3xTMWaciYs>1n7tFa-ZY?u-Qg)}OVq&U?hX-#Z*I#O`4SCe|
zNDTPg*47N=4PbsK4UM9-($r1O-o|f$c8Y&qHlBPmU4Cday|C@OFI&6V6Tt&wUS%AA
znbI2^m?8-~i>uPgf>vBInaG9hq4KlD-W!UkKV`F$GMg^&_s<)V!sVrGNx7-*=BqPy
zNlsg6y68{R=Qx&rz{0}9z`)?ko%rAbYx=v9DvI|lHhGV-a-LdMZ{+w#(IJRmIDWE^
z|B{`AN5elN`K0Unck|Xp#8h3AUutm37OPpJdHcO7mj+D&Z*aGZ6|~&?PX>^kZ|^+y
z%w3W$>boxg`HAp&;K%&@u;pXdWF8HvJ$VNYsV(a+UG7r#K}K4;t?ALw5$jc}%B!j-
zGLUO%pVrBY&bzX6_l~QnVQ-)6rA#0OS@Ct~?uzjr?oCV}LN)Q2S#?Gw5Ef)<Zh2!m
zAjd>V+Q@irZ9U@pZTdP4V@ha-rYb~+YCWN#bj2pyV{m+q!VQ)fPl4S2_33Zgy|SS(
zuet8CQESY2SYXm*H&LB#^O2Frnx~OZ4LsxhmA=;nm~*?UnZFtjrFE3%!A1u@|HOu)
z%9hu@PyqoZ*NB!IU%z53!2jz{Q{h*w%f5l{s!EY=o1WE^VS*aJYR>kdXY3=$yVy%`
zDklr8Z`dI@P2wa{T*n{P8{if!_7%;wcu`SaK1nuUu&^Xl=K1w|P=i9WT^F|-AEV#Y
z51;-ZcaC>1P~R|MzQ-u<!>+^vMgSW>d{iuAY;4FVZ2J23HZF5Me*WCg7O6l@XR=Q+
zFpA#4TTgE}aBO@c&*Gxc%j4zalLdwWx_U5u?v)~=NX(=;9C8|e>u(Z^VCdnPJfb{+
z3d0)!3bHaYS@8Ljz6JX~3j-<;zcIk9z}$2HJ93^_*{^FL9g-N5EG90l3jiFFtgE;(
z@S+6`XhLu(zjP-0MPp<=HFKXry>}gQ?-}Xo=>+%#Er$LiBu=0%J}I_1cn_X(qkkXP
zE_wd@f$$$=P0fOkgdjQ~L=<a5Gdwm}OuZ!I0YdPr>&z8B@%I}J@&4?KLugJM6RN_m
z2ylTc5W~Y@H?8CY_hAj02v_@4CHY>|s%8S+?}c`Z?neTLaRDJ{vbf_H=CdH=eb5n(
zD+@uEFKkCHOqw(aO}OOK!mvs2Np3VH+2V;I^Z{%FJ11v0;upNGAjdIZSaf8}ZQjd;
zZfl-}i32$&UI4DO8(5;D*CN3-%*SQw@huxqie46~QdxGFgt(N20MKO?N{w;GP>V6<
z2&K5fU|S&z8gjPcvAekXvjE4q%xY%ecZ_Bk0K)I!IK_uRE>*;k0NNxwCtzyoO8xwH
z4#*Fj;32y^w5fuU^uxdyD7pm41@TdER1%2qbsnDrxvG#}KtZrL9m*>!bs;ezZwY2p
zoflD1Q;>oaqKxqZ7H8-2*g<gLq&UE*3xNZLSiAhiWoLlFRkL8pfK%iuxTuK?y05N5
zjU>`OFY$J$;Ill87-wO`qR2#y;Z%(c9u=vz(3&=On7y^Q4MvYJX{37rji$tP>NlYD
zC&(c&(@Mk@#@WV?e>mGDpg}9$D#ba6&dxjfjT98v*2czw0^ACz5-`0RFvU@+3`1$*
zc;K5Oo~J_!c@6*dwo|(kKR;1-j2*^5JaeR~z#l4thD0U<*BjqVb_!`ge>4jN)z1s#
zDj~|KC6NOqw)>gF{q+k~Fz6RqKIh3~J5UBxG)*8jHP8@)fUZEa7dj;U2w0*IaSZ|b
z`1;O=?^^P#S!7@UG2r&l_G7^DI+T-kY2Y=`eODCOw0B9>^_neg2ERytJd9N$1HC_N
z2)5a5JtzQiG3M{_ISAA6GmV7lcrv4MJHsO?z>VKUtQLs#)=TXNFnmG2IFSJ+g0hS!
z3DOK@7X#X07)*uLH8t&m$RWguXTsP4SLyE$-3#4He>Y$PJVR+MOK?sAon-7F;KS@Q
z;PL$V^E}WiS6hw8U@2elRTWC5`$}2h!q7W?Dc=(u^Hp1n+b~*YXk+cKw^JF^7vRIA
zzU=OH{TdwGslB;81|}wx82kaJzO|#!tNg(G1<4l-_UJ#>JhXl;2U^3!5>i33pT5^)
zC5_u{fk{wuviHSX3aB1<HN)a<LBWa9WBd0TzPOB5hJr0(p*|u|%{V4CprzbM?s+XV
z!=IrGDQw0+i&wLH7k2|NfMJ$%^cBDRPHfF=V#ONx865^-0^bt2F9T^WcoEYZBqlR(
z{E^sWTuOkGF)~-jhz4&bg@)|MbN`(EdlMjn46|(nSIcBT%w4WEgPoo6mYbCDYo_rn
z$!c$59sAm$cgztirzMNnk$0Mdmr=}qCE!zne)srp4&0PseCiY{=22>pK~xCTk0%>a
z5ZK(`O}G}wE4-pAJ<3V<Hj9ZpP(7BsKy<OHmK!E0{LMZhtXXx}!azwS+v?i^oeM0#
zRtJ`HiLosUoc&A3`dd<LV{Io|(d=N7OgjmRv)umtPZAtxXN$iw#~x#*k}q9c3-mc8
zHL&U4y_eE;m%~`o6J6`}3GWY!j=o!#=w8KAn0Z;8Fqcst4h7VIVGWK6A3Dmfe@?LE
zm-Wt@g$K_V8R@#}n5Cg0Th%|G)gS6^1PMgu^F)M%IQ19b>pP{W{0M`Y^gqXYGE>30
zuMOPybTBP_Mrz1YhbWJTw+{?CA1tv!Lmi<4$ehu`7RH?iI*bSWCq~YK=vuhVsFw--
zyp(t(tY5SU0@tk4(z|O!A%<n-aZ=1_a&dNc*tgmdLdMW!^N$-Auf(T?$~C4W+35c6
zu}s}xn$I8N{|<l_WYl75D?3cYeBk16+W6Lqhy1S8M6Pf<2Mt^^IFO$}2a1q<-Oe|z
zPqPj}ufs6l9I`>1S$G(mn3Ruf{{G*fQ^Bbdh?RYfO$wfxrq7=pG2N1P9RgNI6jB3T
zMAcQfeV$3jh8(L0Di>uQtjkbmz3$$%89zj177!4vnD_4wpswR1=tag3zg!RPEpVB;
z=jNU<DUzug)G(82gvFh_Da`{o$bWT@jjk{8o!*4DkWIsh=|)GN)_)bwZfp!VTKcAS
z);VB`r&LT1zg#<#J%}Ls8Pq&bB2uM3o8y`ak4ak!W@FKS9Np;5KD;ORPYzF!8>%-e
zDCAeH+Pix0CyIOyor{)i-yKqRtv(V?xzCfi>*1Q!PUftbJI$nw9UV5pX18|Va>9W2
z30bOaOm#K)9u^&hzI((6rM|#$jmqtUX<I5Jtj4|_JGe3Qn$>6ym)8Td@y=&ebud^C
z7nXkU;-WR&*kX?sDxSQ>ioHusjt6g^iX?~lE5%w4=!`>klq8U4LX%bdOq0HK9glz6
zrJ?TK1}a6Rs;uT-N<q^jXzTlzSrnN-pcn=Q_?A!E&jH&+xfb+%0qzP43}iu|sZvFS
zN3)2>hD6o<`*qBDopP^0PGFa{Kv7YV@GHs{r6v|LSD>JwvzktteM|ChuGB2X_ZxI&
zY?S-?y8hzF5<PmWk3SVT>24m?Q<($EO1q)bBXJM{nCLxYcC8u9tHi56-#fH&COapq
zqQ*;5e8)WUngg+D`*vJWds?(5e$%vrZEQYX>CZY727a&pb)lV$$;gpAo$;Y;qCxs%
z&DKY=-SQMM`fgs5=YbPAv=6c$Uhe8cx|S(7?_7Z5^6xmxtMG7d)dJwcAasFYRQ$XQ
zYzt7^_sUWLBDmJUdV4*5m<d9L6zCAKA>idUWD^_Kv8+N?B12aeia^*I3ZUF?5m#ZQ
zu9!5-Jzg{Ld{TW=oUZ6)V4Rn3V+ugQmI8#q;KClwUv!|b^liN9LzUe2|3qajzFG0l
zz@l(FSo2tL@;Ygz^{%BJj&mwmYP22lQ_9$j2KsAUS{dMhzgN<S4wYjQ2zPD^?2jD#
zE(+8Gpi~H)iOD7l7$p@%#kt#Sb{IwU#ySCQph^c6#k;uN!e57Taq+2N;5Y{sxVwd-
z($XN-0_OZ8Q=f+q&q?O}B}ewzd75*_+W$!wx)V<ki%fNz^``z>Oe%yA#f3LuW2Hv~
zW*L1oRq_9p^QYyF3=i9Z?gq8^*7HWII~=Ruvf+Y!IPD!kF{nGE@iar`LY#I~_*4ex
z=Wa>=%Kw0(K68P5J!#?hM@h2o7ILoNeBXw|3V_ur;cU-L9kL`^53GXR<`7*wZ2#=O
zPj>$wrS%Npm#A<{NN#DXCk_*g9L|tTuvQd^iXjNP{qKv{&t!jLJ%sR<a}&kI#c2q~
zG%yKbI%s!v@JAc9peTc~gHNgCRt6Wup;|^0uB6%%&4y8cDO3C2twZUS6COc|;Q*Mp
z90E;v$}YK`l)4<RH`h6bEEgo=z?Y~EOE{?}u&F^F&cZ-!Pe~)8X2tQg&&(XKEiJ<y
zSmZDv_0K95i=vAEMYL`i(J2eX0(MtYEUs9wBBQKq4oSBm5IMs83ek1sRl${^2;V-y
zr5Er<LUWN>T)a2xE3RgS0p6;<DO*S5zrhxVqk%@ZdSlf<)={+<20T9eg>Vd#M-wt4
z&c*o`ewpM!9HNCpk;iZ!;hE(P+2co;I*3^k697U|+OubR=>P07w+iNlk2#zp?<~#O
zKzGWLm#=J3hJ(rxnW+qe#5M|HSIbKq{+V~@ci+Z-+djWK(@JIRm*eO>OyVS_MX`Nt
z*TjW5@cLr0dPG_t>{h`#Dj+BreX;*kYJZUwf0%0EKJ!}5XSj4B%JhrBvKLF@GXCd1
zcq_!t|C-$=P=oj+d~OV*s`W!|_-v}kz1}94%hJDHayy%k`%f5a^DndhS~){DBTeH<
zuBGgUw5fYOB%d?VFl^M9+IRdg8|EX;Wb?OAB_$)UnNV&9nyRy~@VbO&)qhS~mc9KJ
zzu`Zqz#X9$L(Xp#<k6PRXhX5g9}y$Uw%0w$9>{cDy4<<<CoFysrF8c;cBfZq@DJ&T
z_jf<P&~-|DWJSofv0ZnH6b}?Xp803GdJ+MQ_CIl(nF^AN@EZ7Cf1vi~UY@C7J{9$I
z7(5a9gW8vyzzATMCQf*|v2o{g!x0LXMBEj_;G{=gz0o1t*a=6Z#=!>&>Hn-!3g3ZY
zZN<K*W!<$A7MY-+6cpj?y@q{B55Y{l?+2kcIfvv-CIh%Pp*wuPtt$892w8k!=bs7(
z1@_V|{NH##OvbpdXpyYV0w_TkyTvk>%*`$RkJ1Z#{YhYu^}pzif=YHpMO;uy$x1ql
z7H@b*VTFI*f|>9A@g0Mpl1%2|r^61K6$|rCwY+oz1{ys5_C;tMco_d5=4;@?3b5(h
zp=_D80*_V72|LfKQr@MGTh{%(D93--kB(}1(G>aHzbyWbUwc&O!?9W7dLBDEA8hB2
zhVt2^>i5X^tl@KWG>%<)9T$;!VL9xB(uFh&O}iEw9h5?Ufjgn{SkypWhYrZ#kRf#f
zC1(T$1(li-w)L-$dHjvf!W6@I*K07u{67F2E7KvnW7e@{T6%`N{sYvJ&-aQgoMzhM
zzp?n|uzFC7@5Rb5-u+nU?x*?5ggA{>Jh7;(tT39^b7`<=!wwB3lOL1K*fl14zbMH6
z{P;7GEeySC*un6Dh$l}-d`~+Ss3@1y%SA=!5OfREK)*Nku!BZF9~>|}bVm8l4>9}Z
zR>2#M=T&N3o#M|Id+XcUM3tT|k~TIkyhs2I+vGo=<{MQ*9AYpU8XZUa5+#lhs*TMU
zxS^DLm6r81#1%ZL@LUuvbEf#`58KK-W%m86Jo9_MlrAR}`F-%NVZiElQ4pD&ob39I
zTOafx;Ne^{f&=O%bt|q`-VcHbDz^@Z-P!Z@_UKuHF>kQo?dE=7Q(aBbIype$5d!H<
zJ8ix1=;z4{92JnXUch6r8)U(-&E<$r9qR^_i<fs-SgLLq_|<+}@9S+_-CaVS32TQN
zYMZ63tsQ+3wSbd!DMSzWGieh8*aR^&l-U#lfav>_pcKP9$e_SoNT3+T=y6037!sfg
zvMB+ml8uuaa(w!@@#zy)@Y$Vz9?s8eazJ^^o;@K+^I5b)D3~^C7YE>lWtBSn79rpZ
z`PXY=a7>eVgAqoa9@apZD8efG@d?o$dEvr^BdJy*cz<IDH>jjn{i6kFL9*Ibuo$GH
zK#L3@p2}uz>qztiz?<>0y?qK77Z<ey;lvyc82hn=68P3|&6>q>Rh~)GEJ_XeKPpdM
z#H)QbJ*$o{9MnWb4kr|W(86znQkQrYtpIR$lwV4@65qdP17C=BRUSxf_u;86I-@c0
z8;1o#e#N0@Ae#`>FIZG9+KV_V!i;T9+R(jzx68&fLL$gHPz2xj7HE&~E4cHtlFg?;
zeu&T*2c^cm+Y2Yk2F$=u4;Eme!SqK;+qPxa)GP!VprE31Tiq7sI8@Ro1Kq(V3Am*c
zVb(+#baBW0uzh3UiWpAuHa9ONXlE&ni57eazj5x_18sz;7?oO7O$w$0={Fb54-X){
ze>^G&-YG2pSg?t!1xw`!0u}H0=y3B4D5>H*#x^KOPNh@xZEbGIKc|4#2Km5bh5s9T
zrIG~n0C}#CKZYRLjf_{P_ACm$!8n9@1pv@=ZHOYduRgzVgB=XpGKvTR^ns6$>ZReI
z$<`jLSsbx%wJ?ZL+Wzp_kv$MAh8r}5u$_D<Tsf7&ke-Ldnqq4bVf{l~P#$i-@(EEU
z#XN^iYFp0>fjxtelG0Eyl!<1tn^r0F-`4>*Mb<KkJ(CS^!z-u<WRCmyL76i@aSZg=
z&9{YuSs6bwg)t;1CnG^g2X`L|4}>LW2>3|Gxwv)yGIkR=|M0o#M*Z6|dDMpqK_l!p
zVy<K01gS$14|a$gH8Nx2M=Td7#Cn8dRfHcRJOBab5Y<e<ISLg;6GWswwftxQxD*ud
zh+}dGjzxwFf<CcV>b_g})Ar3t*MyK)rXt-pFF>k;1PVv_rFK~2K*%kf$qsy@>DxC~
zJfffQj1lAqVq7vgMKU-vl&p;a+sODG6to<31X$NnSU2E$3_`XF0>cow+V$m|Y}u&U
z+WBKYf2o4)Fs@&^EQDb1H5Qg;-T%DcE77E}WC_|C2<U4afCn1$7N4x_VkBUZ@B)TG
z5qxVVACk$f!`M->Xo}@nQdvKERZ2CBlRt+Wfi`d~9*=Nb19K+Y^~|f^xes>0AOQ!g
zF#JwfE=ik%9WEEBEeTK&Y1;l(&QtHouU=(B-H0{V2Xq<h4Vp(NFd1-DiM4+^N{kjs
zoS%f%dn=TTkjW4vialGW^zk;|VjeD-F)uB4-*UaY{2)4tfIUutHLx^M7PMXlHhdbg
zfLbp{JQK>dsi^Pr<@BP<Jl;Ivwc*{g&_mFPJ@_n!1LS2k1LE6)L2yriY~c^}*pNNX
z`k%|>zOhJ;AMpLSAq0=ReQ{eeM6}#_dD+?M3On9qsjj5NiqFFdKRO<OM_A;bB*(Bq
zm5vx{=+woh0~lQ!BcYMSC{)rfUc8V8poG88ehm$0JZt0wr*TzVyU5J1Ija4PaqCw6
z7^I8owZJ<OqJk>{&%lw%&h94xupAj11vaSXijY#){9VBZLjE7V-UF`3zWp0Ntq76K
zXjsW8N;H(DtW;K{DT?;c79xs}3Z<b?Qlur-S9>TWZKGWx6{V%6BLDXh*L~f;=l48+
zulv4V_jO%deZS}V8OQNn$5}u@k>Wnh_-7j|Bu(B{RlV{#?3oG%0w5UToiD(Ypedq@
z5zY|Ap<~Au+<xp>NX7e>C)mkS9$Jpef&uj?K03O~+&sF%%Fw`IDTKh7acZ(^0k$4=
z@_0Ry0YlIt$F*Y8y0<?vC1oMN*CM|d>nFNUiYX5P;CE?82N3Wbr&d(WO4wUGBch@<
zsAr_51+CIOaNt2t1}nJ^pYCj&MzfH2f&cd>mC^&BMb8dbtI0Ti-?l8PcDW^>C@3t}
zH`^8W0%J7LhCA$Zeso79oJGm=%+C2P=en9#dSGX;>+y%Gw{}<kIyr_JKYoG~>n)=_
zGm$x>ht*-i-v-MX1Mu|vA_-4N54^blA*nQb>VbTN5SVj|2G^)Xkr5nmj#o&|z)+4c
z?lJ<1A)ICy2;zZzXBQN((>xlKLx~lo%~b?VpammwyjTUKCnT?c|0}A&3P%|mFw`p4
z?KV*uLLmFITY+c;JfxEtY*Fnf7H+Oz(uDOy1D1mwCk_5py(YHLo;!CEyy7s5i;N?h
z=p#6ed|;lE9^3s`UsB~mvXCR;N_OT+!}vp&iRT8eL1K<a=UCx;^jmbbyvqHDYAS3l
zclZP87BR2ca290_a?F%RyFpN6gB?T4#X`HT>iEf+MawY%Q=tmhWE{<=q1%u~(E@LP
zMn+_8#iJKxj#^lFLf}F?IekF{dqg{E=M`D$51U$8G(3@_07mTDZC&c%p~iwEhK8~+
zvg2qqVwM@_+qdTeZVM;)I{l0vdoahs4OgUEQ&w8~FwU@PU|RHa7X7J~4-k<!L&(+e
z9Y#&$piN*-ojDT<#APK0Ss*M5XXS?Pb(o?68SI4Otrn2R;MKR8H{t!j+0QT(kcsaB
zB>VR5k2rrA28T3x0WBRj#7e5k>U4i{U-zP>W+7-;sN~#WgAtRs4PEF-<i}w4B!7|V
zlgO6^_d&DSq?ZS|)tB94!{;gLx5s@9Gg`Jiwj@&9Ib`pb{<J6)ginfnl*aGuD17??
zM<I@z_3D00u!(3;cy^{0-<EUN!Spak`=uX7mrccStWjpLl6nSFxq#URhfY{n%z(m-
z1|$Oi0Z9_6IRtsAQ871Oz?8=i?L>~lG-L9iw)$7HgY$P63owSrcc(Zxw7n6x?QCsx
zagxlMJv-EH<8(iL2)?P0|2>lV`|Xdijc7#D_3>onAby}R5FW)R0AwP%_E`8H)w>cg
z=8Mko2?7fmNMCRC5D5u~4pI6XL=+pxFF@30P~HP2t^RrlpMYG!IQ#b1tTXhWdhh@K
zTz6Q%eS`gIV*dgU+@zZgldc<Mz7{ZF6a9TL%fU?}tmR!s7n0F~n-{K};QA!85KFTg
zQD%|1<Doymk=ldAzjmEC!*4tie{aV10m>aM7uQ0NqoSg+&{%8Jw1|VyJ?;sNy0!L_
zf%I1&&4v?Wq^mLX@}?6_p=73fQVc>-*T3ZR|3Zjs-m^3(7DwtZ!AVIAj2dq23WOz`
zq#{Vx(7TwkTif%HYzqeYocFNCK?y^Dg%n%!lW5dI9)o|yt}SQA9@mz0KNpC6l3}IE
zzRoYWUHJ$XAmcwb@X7S(C;Ej|Qb!+mF;8K)2I}wv+^lN2a(0lE4Zym(CmD-WAFVIT
zv-|tNIz$btJQb5o#BujRI>Fp|yJyNJdEDrx@6ftOCZ_w|f27%<JkA+|T=tk=y9seD
zjS_;P<888aag~&X*?BE7K}ev;!|73pS?9BJTYFC7nkfsg-hLZ`<@l+Zb{i99!CEM5
zh?}n*36=5Xc%PAwHY<b!vMkU2_bfy9O}QW!Y}dO9+bA3s$qRCeMSdiwhvo@`#lcn4
zA9tS5YckGuWG>(m!~u<|j}g%`4u^sCvF2zWUL<#jRILxPplTg|LC$gr+ARcfi2l;=
zJ*bR&wpRjLlQ+X+F}rcLf0ac|zR{ux-2!3cCb8U4;ss!r;K^w@t#IN<oVijWhfZ%@
zPL9h^Iy^l7b2G6DAn`c52z3~mH$W>*KjU-l>&!2VPea?nr4_dl!KKg@nxVVX<kzSO
zgB$_XcF|=2f8qDJ?H;$XKDoqySGW^1)BUycHYca_9nP@UJmG(pxIgCl+Z{tw!qU^z
z8{)EDTRlu`qourtS8lO2m9|9dh9X=~$8W9qdj|$p=L9%JXx0JV=w-)*Q@}}R_EY4{
z*zm`yjxO2$LzVM>?O)%2_R`gs>9wLbz(T-DQ@MyR-@g;gAXJRLFhM+fvJOzHgp8Q!
zKflyYFZkze5|%}is~5)3pF3A{)daZnXU&<D`(zz@s;I)%af%y*Ul$jvwDy-@GQE#x
z0vR2V19xzM!hL+x&+pOA5{2Zi3-7#@hDqLK)%>c5zDCuul~!q(ExIrk*4YOjO^C6D
zw4lag5^ewrK7?FhD+&|zYt+rT?9GsA#}W7LKsx~ExZMH4Phk7fv<&~EZ>nDy!3!S3
z<pe=49)e;uw~$`d5ZKmniF#f|vt0ggI@i8xCb3NgBmEbMN0uI?@)LW$E4fVbb>prR
zC&DTQ3kwU!@n$bwo4>knwXUh+hsy%5-*0=p*h4m4*3zujqS*0WK2GU<>>Hk6Vzo)%
zIXpRibhw>jC2^NRs4xNdB_>@WBD(1D=t5OPQ(S?4J0pb8m8gi)Ndyj@(pka$(BMFn
z6bQ&&RL@$QB0k|jjLCb6Ew$F3rf=u<-ZOdz7zxe7N5luY&Ez+!Ul9HDDJK|K3Teie
zV7ESR4p6O>ihwt&`8yLqgFDd6K|+}r<0<&rc&x-`SC?%g&WZEI4wx>;kOpZzl$4L7
zcdzgF%m+R_nL*QT<n$XEBq!{bSQf{jVg#jXuFop~OAScg_HUuqC;VCKiNQn&D1r{F
zR<wGU|5=a{PWvrxJ>1=Yv!lW&GDAXKd?{K2G#pr?IIZ%)iW9O4Uabj$R+Rk{iS2iz
z1J0Xo%wr8pS#|_{+!PKpYyv}hBG8tMXLfk<7~B4~OLpuHuc(AV3i_0&!MN|NJ54nS
zFVh55HiJp}Qu}-z78YS>GC1{6$58qR<kkChbjHAbgth41eehrzx;N5`f2ocaMCU^z
zO^^bI%6b`Qv%<G6AA2NS1T)#a!&;X-Tt*1o?mQVyn08$`AGoi-AI!>E+#4i$><tTz
z4N><1;?(u$HvPunmYfGq27X`2Y@M33*kRVL<#L?sb~UJMRov(GrTp7Sajl~8%Wr<X
zLT_I+UHZ-w7o{H&R;_qe=8m7fae%MpW8+n~!u1|cxVCuw#kww=V=aUE4hH;}xeW&R
z4E*<RiEZt1JYHI3>QvESS~gTvV+x^9Be(!W*{76a5>`+lc&SWWqm=s;(BEte^!=Fi
zWi#mCpdf*==h##BHwY<5jxZaV%$m<BaptMZD$D0!fH5qma1<3b&g#S7(RlZWWj6N=
zmJA36(nF*-;ar1Y-+7W)t#EYwvZm5`Jz1B#dObb4W1OyCxk42YG&Yx^y3~g^IjZ8<
z=O;k09XjjErPZMFfcy&qHb+DpW1{ntL|F?4A36byz`u}q0m^Ac#ood<8Ofn|DU@dA
z{n-KL=lx&+C<Yh{H{{@I{D;&zxK)lK=Y&!L(sP#Qu3nj~#B(d&`|I{sV|-mwS#g%>
z-GlFpHc;#Va+u8b9r5!q3QW>lgU}-n4-Yuvj~n<xC`=(m4238~2903i=7X@XNk}5o
z<Jya}Te*4w5*liSp{oI{am=&u_<G9KXJ=;z+7lGJXc4t^!bwQ$q9PBX(&D~IusE<?
zvqk>M)-$VGIsh62)j4SFc7c76<581`5h>R+0+Z43Kn4RDXu|t8v%)t$hLNYtM32~Q
zT5pgvaJ95mAC4`mS<!hxSWy*Hr4oR;c{VIH6d=Pe_?U0pC-2Bhi*C4YZg*Unvk*Ur
z@bdGcz4t|7iXcJ=c+IX|yS{XXd)^g>&>bJ#U*y<$Xx!YvWROvk{_=F+mt$SsQ{oox
z?&fks-Fr9j4lQ!|V5gv;*efrfU!jOO2n2!=>OlA*^pXq%xhYB6gYTZ0%7D_#1Gpgh
zd_IJwNG70wGMY9INF6N&tqL@~C<F3Ww<4Q}ABu7^JR=vMa=%b2E5Hzl5nrF9Scppx
z2Ge<}Be4?Z*ekej#o9T~K`%Q`M}k$AP+#6|v>U88CC>p?S_**&6i+Vbj3GfPj(PdY
z9uM6fYZ>0jPryGZx(xumCz<nxhC=YnLG~NIhSX9IeiJ<3Ks@}&v9(-flJ>wwKkt^T
z3T&u=ei&!ae>R?e9xw<4X~r&?fH^O&7|VuXpw5RP)xMgae%Ijt{QXFLwC8p9H=@AC
zaf{QiT$HMi&-~g~;)@ha0620uO3Gl}MVa*0Z~W(*3hYNd-Ox@=s*y(I4>$xpSq!O#
zmL&Melc~h`Ahr8CwjfRBqh<-^sSBDRxzvo~D2Hhh1}rDsLIPQNFlAVMDO-XLZ1J*X
z6A()Cf@~GVCQa6zfGw}1^?87lBZ$bj-p&V;YUX<UItFb4Y=bX$8bkAHV0VY^43FBV
z+)m%!yp}iJGn}lul#C5x08bV8zy(ai@~=<6S6KCQ`!|ZU!l_I%mY}NV#*?C$EU;kL
zJGYgr;Lb@bsSx5l&$^aBujx?!8&Q45`*4$uGxC0#GBGmI@$~7Xz{X)#6aw#i1>Lzm
zct98+J&s7J>Nlu$m9jRiUomCLKyy>ag!5a@d$*tV_G~g18v3x>ne5e~qO;)EP#ZDu
zlvzlwN64#ZmC3Sl2KWcZ-TW1h`6z8<#yzJ_sX&bHaFn5XKNqXuV^^0KNy%|v4$Qd9
z`JMBP(T17z*99=LA!Oebj1?gj*1(MXj}h-pKo*w_R+s0~WeJ4W(0*QQ;^rAmrEf%Y
zeC0!&y}f5sI35hBwQc<~F&t;=r59iFtimOS{aYVIMP1j3m4aCEsE4jRpCE2ylu?tI
zq}@8@nYuoQ+A(J>`;z<R!1eb1J=@id)ZN<GtrU{m-?LdR&~2fFg|=J!H)TXI?Su}6
z;;oKiiOze!B;oWyS9SU5ijn=>K71oWC3mZ&^?H5CcTiVs+|^i*p^u2K#y*JGGMhPD
zN#dgQ=eThl28JmC!K`jcuPvr#d<MIB<oy<q-}LNWoG?DBpsTI6lfvc_ogay+HK+F-
zxjeb~#N9U)p`N}*i)Vyv4|z9It@!QubM>)vfN!2oifuL&VlfiE$`J?kKEAlbIHO3Y
z65XV?7x*5*=J(=1s#<LrHKSlbK)Sh5JsNjm*tx6BM9RRO@;zlo`h9`5zR{Ib`f`4k
z)Btfj3SW^vDil1VOOV^2!z^&<06*c+IzG1FBv>IZXDS(1U}3L{;#hdTVm{1v@IZv3
zc>!kj)IrkqM!u`Nd*QNP>3sN#8<tc}5;%2L0TcAfMl@DF3_%bVQb0VBaY1pMIOHh{
z3`RM+Tqq|mH#fIRT1IbYObqKof`8uqeyV*T1BF)MiHf1Fpn$;EdV>HN`B#|y(~2nj
zs)l1XW&{f@w3?!_(eCsXcjK#nl>TZoP6DbYAcfV^bG|Ql&||2Bld{9de)PLvrtNN$
zU2mF<m~S5W9GBJU_wU|a^<M;>{1qvmMu8b2Ay3r03<U*o2#-}{WB;Q!14B!A<jsvQ
zHI%|tZ4vh$KJ9b)K4eZoFVF<RAi503nKU9mh~TwJP0v~O=T4N@QK{ulE3v2H&13I!
z$`UU&WJqGv=@(uS>BCH~$bRJet6+hJR}>ycT2_n%l(tq2&->?h<CF8v`}6Vk#|*|_
z!TO+Dd5P<72u#nys!W$R&Ixik!Dxxnsjwl$4&twcgyWDB##2C0P(f&(LbO`#Mpi}C
zELDgDXez;bcujKE=rz#f5{!Y|4*gwvj^pwD7yTq%hYfBwl#aiU_0Ko{{`3drMr5m}
zI1rp5c#SJDU<j%z0I4M2M^^9|H)9(bIZLzB0|N_rkUfojPJ53?Vbk1)`&RnfZHqnx
zjYHLIsO~&zVlP>x?*s;F1Z@@CiX)lxv&k<3ZdZ9tzCT_xPF2cxLGs@Q2?-hsorm+Z
z48j(Opy-H1Bp>CM;}K%1cYK1k?IqIlvs$;*@SoxLJBQb3%zvW%io%T?#V$jPEB(8j
zrdZj+qet-;Oe1AHwFDUYSK<*;s1MjBI?J(%Xo>(LRC$+)@4=H366Ho<9Y~1C_mywn
ze7doBp=RoRdN;CtbNQ{N?7lGSz|q%%@dxy7Ixr4q3QktH=MATo)+N{8p(~4b(?o_+
z5F1R9h)U6tP!MAbq$noGKt;Tc&F<-{LGq~!%?k|++uoI=y_RC<k>r)j17bW6{zh_T
z;uE0CkioDtd~@&ejmIf731$pLKFvW{4_k6RTG)QyyL2UtC7zrb{`x3DT{&811fG_~
znDT;K682yWsg_t7k6;C=SU_K6_J$i!5GXxbO7DBs)M&cvHOsH{qj_szDxw!504EN<
zyimKu%2IN(qlX1CMzT9{nWUs_a_K#f;7wXd_!yLYh41gd$_YFFHvy4-$!9ACjSZhT
z=Zu#)<MxLcf2E+02Apw5$>A?Vg$piv9*oMh9y%`}eh(sTWZ_;wmdpfXOb|j2S;O>O
z&&^&3XiiXzIM^m|?)Gj08omJA8`O~@m}&~NFB^n3K$_|4HV`O6Hr|7WN||sF{E?NC
zKp&J?;dKZ_CEBy4i^{d^rk)y)HBjWCK%=}v@PMQz!)l-{cDfnCspxG*VEQ1%IJ#j>
zHNf1;9Kf3WTu;pOPXJd3x%81V)xgN1AMSp=CiW<9gdK&f4nwG5P?0=^ssuIA*02K~
z5v2|>)x(D2becLvVTz!6Ni2&4kf2mJ&MeB)8{hmB7UH#xO2gn5sXxdC7TNyD_dj6l
zL?VV4(BuH&Es#et&hQK;#GiM7^cpoF$XgoO3Ur-XJ;1{Ly;Z(VZFR{uu*VX#hmVCm
z_vz8>@$}T81UmMbSLPHa27_RA<U@i?a33j!Kso$qbeRb$9mD)+*ZqAB#ohRt+(GoD
z^_{aXP$nY;0u+WqKNI{3KvKdlD6Z4nJM{1<KxgWzK@Pwuhrtsb2seQa^(XXuLHd}6
zEzChv*KvvyuLJsvxJ=WEsI7{MSF3xy5$cI^2lO@Rn4p}>!oXju#|+dQKzx8Vx(tr9
zWwfnxj42#zowkUZ+nC=szne$FBiT@wag0Xkr24LUTUST9ksz7;(^)_Y0a_}+tA$l$
z%93G_(7V1BFjIiAE>f|4GBHRgQrxx~X;;AqB8q&$<r%&{tfqEHI2MQ5cL-IZ^j3qm
zD2a$g6fleq_g3j1Jed2$1K*&9C#K+>xLB5!re<G^a3J>(z+h#^82u+tpWZn1O*{V}
z7%J6_goJlyjvhNgE4x^GjTZDaW={z8qNpwR{DW|`Mv5nSM&N6QpG9*MDQ<RwD=l})
z%1O8l{#*ebIOQL|HzCUP&=78)2@G<*7zV<}OVB}JtV34Rs&*_82+x;dA;8BQ^jbp?
z*=tZ~jZau4WC<<4UYxq+<u;U)0E}F*Z$$E8D>-uhmz`Xoa-#{fl`wDo!1+Y!TMR|;
z;+=Ca>d!6iW=o-a0zo}FL8;)YZ)(!Xj~9#ebdaP0uk+2IprYlG3K(%iW7eg$G<SxU
zHv(-&+vf_yh|O6@T*u<ObW-#B4hhZsE6<;la3+Mzg#h9yo1=9!&OA_Q*A%;T|7ZbV
zdfU(fpDcOSr;j4(;Of{L#aakevP=L+xgYXp{8a^eDAWrrJmLjtu{VXw)W(IbLF{4H
zYVO0BCE{b~?Q>xDZDmK6;wxxx@!NE*;b>XkT)lkBI3gd5q_k(JndzAr8<(fD%w$e=
zJFhBTaW!hudF?!UFE)E@Y~riGL!(l)oJBI|QT&-GJ(JDhF!C4sS{fNO%7kN~2iO$-
z%9!Nb)@HJ2uvU#)<cr$5@eWh|ygivYwo1Ziz2aAQb#~4$GMuD#_^`qX?~j-gEnX#p
z?`HgCF!MJ@$i7!8K>~oQ@11U2m9CH25`RgnrDWJGx`E*R*<$TU8|&R{^-bIFEcGj-
zX)s8ylwW%DOL`0XbaW#$=7UYF51q=R!FbTdOz?vT7o`k})d3eHP0tm-{bK&x-$6b%
zZX6vr3)~hFMgVITE*Dyd2I2hACVm`t<dC3b@{_`F6i^BS$o!L?{DG_xm6~U?(Ws=1
z|E7CfK1zAQKNZxA%j!aG3GtV!$^>2l`{=Fe#i2X>HChloAaRT{L<nllLglf#TSApq
zQe;0Ae)6Cx@OvRR2|*6Ix3>#Ow52#vpt>V2P1xkHglPv3KJnpU8bg6dF;uMU6|ERF
zn(jxWbiPH;&aSwANJ30g!W0iC%Fu=)Lebihg3x<~3d{>>&S0-wej*DRWd&hP7!0*u
zatr<spekJOjlDxb!-g^DN;QV6hG)Q;^mjhm15OoYs~JX_M^WENgZ3(}c(i6(hu>T?
z$g4d9be;rCSzJ4~`b14xv4!}~!+bT>?^|F+6470plpOG3_Ri&8qUKNQ=cJ<>EZS&?
zMWl6P=}i&yr^$QDWDf<8T)fcntL&JK`Ow*nhQt#&(ChBm^Lx<enQ*?3g?&Vfuqu8c
z;w5rj#r3ZhZyxOHJN~-bTKVfa!RNoMUoC)!q(p!9;>FI~s%9ClY#kh0&zhT$offaQ
zZwLtbQ+kR+5|Y(Q9CN{ww6z+G5P7ov(q<VM<$F1-AOSdGJ@_h})GIen)!x7V0<P3L
z1l%)7O}I?K(G*52u}Ey^prmDc{!gv90@Dz!cYVb~l{kZ_lzvqYfn^ZFMh4UpOO8$g
zNQ?(+5sL9+oN4G4bQDF)>gkY&da5RR=UP9&yC>W3C$n_4wI!dIhL;SW!g14ZOG!j*
z(~9ghH+SJj8+^JpI^XZvCUo^dug{!1b@E)7l}7=zLO5%q3f*7om6)GYxo-dU&CAAV
zb8P?}OWn4ty(c9fa>FLGc7(T*j~<LURZBsyZFp5>_)1&!Ul$ZywOItcPta?JJby^w
z4SsF&5Z<rmAvsatK21=hGW`V`lBck~us?E5gkO#snkcxaXmsOq`z5v3;R(-$ww`G}
z-sUS)CLj(RloIpd49Dm~8b`B8+$;<Ditx3aesUI;=C5GiNFPXkibutLl(#y{4fHZB
zp$f36gY>7mijuq0bCPI8ObJt8Z)3}KjF#(&t^FldDXv+3;CQWUAnJ9`Rj?Kyc_Rw|
zyn6BH6}}-8(c2aOGc}d70;c1M44Jr8d@(zGDcUYk48w&n#=xK&Jn;0_ps7BA+NJ=7
zc6^N^p{&FK12co;$4LtWXc;a{C``6NWI1+VRIRV{@UgWOD}>hidd4VpJ#tXFpH4{H
zXqw;F)HH#C65(04sYzl=efyJTIDZf!8FqjjaR8|0C{~;VDG)cFgv=fx*AP)J(m6c8
zCeaRes^p-TN-ajDklm)%P)l%0uIMkA`*LuY#-UWQ?WlR*@KNU3%E%o&7vKEk+WP0%
z<cwaW8V)rDKa4Uc3WLhmLvL~l_|!NCt?A?h(`hgT0(Q$eDuyS^LO3J`Lb*@q%!oD^
zVbwEAgA1IXBy5K8-LoMM7bs01T9?*6KDfSzx5jQotx5XO{P{R#5s;fr7G{Icl!^|1
z_$MJJG)$lO&@)kguKP>+rx0>Xw~xuRYxk}0#V>;B0MmToZS5VxKOUE0nt=S186<k&
zdP^-@jS{ko{HACR=ZBl)lm!o*GUjJ{)pULAUJp68pW^qh`o?#p+#tsxa})`~bk*%c
zh6{$D;VL723kk%Ka(2txJAfk<NqdN>n}n7UvqVJtF>3i=fHbOboE#+k%+dCWXys%o
zqydcXNbSK8VsDL6)S<^<?U#!TC`i$Wik@ABzJ#zj-ykCuKOb`wll6}Ea1WA^0M(j7
z#Uo7q!Lnl*^t=*|8KT<yCzip!P51+Bq^gDr3N8hvPiA9)EE2L%Vy6p;LyGwqZ=DLH
z8GmjkQd=pjphYtwA>k-811W~D;cTx;$uIp{@r24dDr2_x8~Jm`aD+Jvb?>|Xc_w;w
z=$E=d`zkEwqOd_qQTJW&1NhQ)G*cu?$CrF7Z3zYbmk*g+h-rdaQ4h}?L+%*nEEY&m
zA*K-Wk?l23ZxRlRa`3j`TUn{%CeU6%Fh)Qd^0Hpw^d(8qlS7eZkle|orv9Rt-e?GE
z8iq|v(N>TX6hID1$T)`5>M=J|PFjqcp`<P}GYMykq0L1CgDVb>9Ke?7w@BX&$d^*2
zAeBFvF_Ru2>|Ui?e#_t~(?A9+RkAEXdsMX9$P3LMa4rfkq#@%tOm_l<|M~OhK2)Y=
zLgL^%C`*%2TmbWk6W~B~O0OP3ngX0re`=@y@if_@1GfFaAP!TuKZc~Bi*rF@6go=E
z!3VSl!Ng1Kdk#L``B%5$9LUAwxN!^sKmSd3sNMq_W;B}>!sh}N1z9s`HUWxD%FL&%
zl^yy6eSM@Pq={GBX{To4>*u2`jl<}YbZd%ZLuTDRC}_JfMrJ@fmq%IJP@EJ6NzF$0
zfm425u#T<-nr|8#48h8D-`hs!Q2A398my8xfJJR}Vf*-_r)gesL*jBZ7L5(2N#^Kj
z@z}b%yR(qy=ih%*h0v`%_hhaTU61M?NW}{@KsWsS@|Ty7BU<syz_+ach3Q`jww_cJ
z=;{)a{<~exed|B$SL*sN>N(G;P?1lEydM~LHN}Ykf?h9wSV&jrrSFlyON5fn3Kxcp
za>?_}9Ge^lj{+Vlp=)gX^XRebIpJWkTozLSMGM;{`D3;dL`}<$Hh9(Di2`^i!`F4Q
z-ET9FU(xod(2FgI>QUn3MN7LeC9MEFaDd%w_P9p`z5!ywi(e-KA%}Y#r{?C>@<Jiu
z0t^4VS;sjEesvu}UV2<U?Rl22H-4iG&4JT~lqZ+u$3gc9AYKOJum@&%a7K>Z9t$_p
zTwZ*@v9qTgH&PDp8*1jxxXE8BE~|9TUDAE<KTXHjZ(DN*#jJ~_ao3c;e^A^Pc_SNz
zONpIRAKScT8zid!pzDw`d;Dr%cRp^c@_n_;4ac46<2eJK&VPO+{ap2ObO$J2XE2Z+
zO_@6N%^4lcEqnItmYoyQAwl-2dG#LK35aNcrV+=#LG(9vUOf;2pPl7kKZXcV1!3nz
zv*kc$M$f4abrIKC=6T_tOMUG7&CVF<l!o-DMtZuNf*8QcvXwQz#ohk>Br-LqlV)+v
zdGS;G-~rb+&qhw40JIL&$;j;s&!XZ+0ne?`2@Vs-9@pntgvmfjN>1!QxCf2@1ynHo
zHS}bhKn&Mf-MJMMgjBxiiJJ8h9aRr>wE6`wYZS|$WpbDGrd+-D-Up4Q9IlyfotC&t
z`flCJkvDSww_vZJcnkL#x4yDq-0;)e=AGI3=N<8l!_bKxAQO>9yO~ss(CA@*(vgP4
zfr5n6zvE8b9oD2G{v|4@hNbUrZ~lHh-u<OdAJyBUL*rPRKkUyJ=6g1R=ixOKwR(QA
zb<yuXMdA&&d{T^oaB@M7f-Fx;90i!T4>PhcKnKS9u`8*&@j|-7#`b>!x1apgd|?@q
zQ&Yl`%~J8^%|#Ho$B!Rx#6qG>2iya^-xbFzMI2lW5$|I3qBIN-<<EwN9_cxHmoR#9
z#@0|X*5B6X^OEp`t9!Qm?%l#T7ZoWZDDsM9F>ZXK^=KRi06`%nOJO5^d-nqqo*y08
zHZod?6WttyzGx-9$n&O5ytv2X>#@O=9y8^yG!-Hz5G^iv7NVf(A77HEXwdcRxsbJu
zr|F!Yli-gLdq6~J(lzi!pa2Ls&B1vL)f>6BVSrB0*>E+ZbD8JX^5cIO{wsxV=tiL&
zC#a+ER%)~dGU<o~BJ@kPVHCaGQXq2}x`<RkS-99eDte*OBU-R%I0J7|-2dZ$fp9c*
z80|Nq*&vw4)wp)iAd03Y7bHAjN$>3XIZ&bg9^Dv#W8_MXN7fhat51Hq=R~i-_c*I7
zt#<|<!(SaTVzT|-DnCTlDY62Z0Ku@u-bJn|iad;eO|L#(GZ5CFh>3u}{WdrjLkL8{
zt^kuL>=~*Hu^KOj4^9+y478N371j1x45~YwUwuUBX<Dncp83q$UJnl%Q~kfuz|fAL
zP+AlFxNYfHTiXrbP&r<YDJ0!rBE%#MqTkCX^O}&z$x4bfiwTJkPww;L)Z2`NuN(k!
z{6C@gB5xwrUU0b)c{s7ZgCwoS>pJz%#eV$8`{d=h-@hEFdc&nfGgnXvVS*IRP{y4}
z37?Rg4$OGFeC(*YRbjMrY|HxAQbPVwXg*gIIzcW0*(sY>{#$8GzqzjuVuiHQVCVpA
zQ{I<$<~be9mJZ5u0PH;$UV0UxJj!c;8hJA9!SJBnVe$|&QB#caj~Lp&C?rk7L4@Q{
z#IzwQX)EX~Vk5HNzrS4_^dx<w8Vru^I;r-g`KODLhTzPe?LuV+P};JX3kx9y0aitZ
z2~-*{kx?RQS|LP~AFcsII~62MUV%l6vT~98%vQNS;@~IhjJ3St?9M^lOK@e-<}>r~
zFW;%6;!$oIb0C7GI@%Z{2ZZHIP2w7=#Q@m_nm^hYcEI??j}tb+***8-JrdK~j~nYE
z$+A4v!4jn!LRH73X#Vjf)axX|zo6^=V!qrTB@nD4ppJv_3Mjx0y>d_XWd5TN(I`tL
zYI6LYG>oMQWo%JRiuTCsfrYHf6@H_$<fBLehx?6mIe2dXjrtobp$q_}_7e2K3?%7d
zi()=ZPK>C{ZM0WZzv7jWS_J4oKqb{e8X8xzE>OP!N1@dFm>!lJj{AH$u|Ec$U6X0e
zAkLkVb2U1z3f{{i*nLi0shuKdbcntYnR=a~Mj*n4$RKlcc#Y}`ba3QdgsBZ!E~RHc
z1_bx@SZD|KN7&(36fQKydI?T4im(Mv2+V?_A3&po;Cs>WhLbZ_FYbtSu|HVrqgdE*
z(CXE!Tf6o$=m2m=`99J})xuh)Ob>ARjcAr3D|~^AkRD=iNXUiv8#_?T;zS^6txF|9
z6Bk$~kURAZIdRx{jcxBEabHp(9uPvV*|Rxf9I<jq$B8iU5cqI_QBvF|6?y=~+-um}
zObeRtIc%u#47%^*le>H=9d{ImNn&=?U!W;+eNs<;9FKeyFOVq@GBWsZy+A#v0M8^r
ztyumpDTYANfn1M+4H5Nb6VIB;L611`>H%zF9f<L>&}uRansheZaCCjtS~8@Q1>~EW
zTRjWELk_})d+@QPe~$57YYYs)tk2<LDBX2aO#Ye8qP`0fbf`IyjKU^nQ$38LhdBHe
z!+PaLJTH-a)W?&dH29N^JBNW2;wiG*@B?X}G+Vz#EsFzG=)?qmE<5&TppIUYmd+d-
zT|wi%AeZVWk(=ZUMLp32XxVX$^$!l_qvWM`jzz%`&HfKA+B$XO_%uk#n5@&@QqXv#
z3vER8y)WWmQ4cy0jQyAnV~XLSy=PzuNBr9i1~?LRMrAE`VzIigMWNg<Fz5$Z-3W$N
z)G@l~*csvEx?aybldnT)zObL=^z<GadExY5=qQgCZgx1ZXtw-{n;6&mLdb=LqAtU9
zd{9<(jaBYOHKhiL=-SV{Ls27!MPxGJ;W6trMDwE@04fKwo$ei4dgSl`gsarASAS<t
zVoY;O5Rl3E=B^K!r9Os%B5!YR!4?*KTjPV3CDIJ4-jsC~Xq@UOg+0$z#Xhz1?A&}}
zH|}lKQEbPf@AJsK5d;H)5$h~osOr7qTjNpWcGW!7um<bswhBTLr+6|N_<slXBdW#6
zU~Rq#Ldu>#>ukH|%N2?d<wlh6)Zzrr1Y>}%`+_={G?u)EaGhx&VPc;LWt+yPU#2pj
zG!e*`@^4BJJ<++@tO%DbS;6C*;XVmC2l>*Xv?CBY;rhk}-!iC*BbW9Us{S-Tx>SL~
zgbv2{`vbIZshe*aLY!-3;^Xd<-+(ex1<d2IHso_hhq)CmHuC$CYUBl#t>6j%^)wYJ
z(~v6VG%CdGTbYYEij$b{jL|mxyAnaffBE$@clKU}fzBetK?YSt*@{TSXm34><{10)
zN9~7EgQrWXF#-f}G%t#FHxPHX%>fupu+#K=(esa-8izfI>Bb5GfyOxp91Fa0BP)eZ
z-<Dm5SLcP6L-Ajz)}g^MKX?nTabQQ{XX{ewg{C`8Al-#jmh$;P#Er+H*^-C?c>Q^d
zS|Jh#N~P(;P5%PHAg$iIAH#%-dHE(^N5bs_%|<HkVX4FS6DSm_7Fk|b7d5#3a`Lm3
zswzt1kTPpHkLzFAQM^EtQFb(nO_jn#W+^w7%#91dUv$jo?7&2<A}MpfU83Q3nKYvy
z5t7w_^AoDHwQlcznULM$?pN{4Akr=~*>e>u(8`)uHF~7jUip1R1)VJmJ@P(p#&^K|
zfFDQ43}A#KnRZGu6T{k{Yp*q&yVn$F*vd6)#I$1i{nv|2KTl5PM);^@+N#^TTkf<q
zD2>11tl^LpUC|{V-c<JoWe*RiPr+DaD)<w^Mj8tY2Gq`^*nZXPOX|$g!Cei?`6kqz
zkPvbEVD61PT49W;ZJ69wQSi8;=3JRspph9IBvTlXqgjkgB?53LSu{Nb-x)#d6k|j)
z0MJ+Bxd#`BU*|@ICbM|+=DUjBFgS6-%`pZ#Y}l}2fzbQzyhWCc04^jn9BNHtQY}(U
zV*+~~1vb}ewrd>Kj!UN%AJyf!Q@`WOkY#%RJoLzoU%y@j>Pl)oIc@ErhUAoh+K%Wz
zBs3dQU?u>sqWL@`-(Rx(sS7)ONH9mz$&@9e)(AKsSmb0JtQHCI9LJ%$U<B`leJf$~
zlnC@k-KhMVhk7%8Bdq+f^?v09M&3cgY1Y?M$fA&I0PbiI(>)Fp73aP!E=jg-zE-f)
z2=i(il(#GST{lu1RF?=WHpOqnz&;OP^WzyXn<8-92^~pWfJ3zLT@0=_Vmm399RiMM
zwI-yQ1KGUO;P!(h_OYr+Q*7s9w>XW%fMY*#*P6u&_?>AeSC>9LiZ++Dh42Fsa4dQD
zhe+^U_EaQt;Yb(JbxwG0@ocR|&AWFG{Te%VMfqk}3VppFVS|vn6E00KAQTnJBiUd&
zk7SwA{hBwoC%IMo!fb3{FtLqqf<zbr@A_S4hJ|f<7wy@n+knv~E1Yw)nFbDybnz-m
ztmQHc?~L^XL-qD@^15`};lZxb<^0tNkS~>9w-XQSWxph!elC9KE+71gfDMB04k~Iv
zoAXg7e7Nsxj-&F|jDz}OvVg)#@&<5`5+srFi5;|~%N>lqX7pvC<HY1pQ}f>c)~@9X
zBu_4m{83v2nK*<LMZN$=*Gq+?2PJ1Ih{VSSSK~0V?P|E|sN)cg*0K?&cAtY4T0oni
z;$N(})WI$09Jkuui~Ua*`^MK|;FzPMRQyMGQB;cgpl0_UIZ|RQXjbo*2KfC`*ZJS6
zhxA(H{jMApQ<8!U6FW@u+)v4SjxF3P25|CO?dcw$g`=Yz*%%?%Z-_Kd0CI-{ry#A9
zuAs<ZeUF*RDHG=hZd6h+9{I7p_MfuiUTJG@4<ho&b4Q~(XeK<xz<~fzKmaT8R?9PM
zv|{6-v^#3mm6U4|J;)I_nkXnEQ}@NJslv-)+9i2!#g2+3pZCUH>=y<wIZmx0zalAb
z`ORJG0^96R4Yi8f>R%FuoG#g8;NC4GFX&z$eb|c>5F2IfQS+^W9;*Ze1uF`Vd;tNo
z7^g9=`d0-7Tk%Mseeb_jht~o6+Zi)*A%(%y$xm6Af>AR;awC%_1=<&RtO9@u*(h37
zkwwvoPP!{G`hD))kDxL!OY_QBUhRCJXc?K_@Pg22JL|(mqx{@s(1Luh&YF1)AgM|V
zZlJZO5tm9z`uBd@i&MFK=%>Ae%xYob6N$&X!a%l+>#+D>jy#-yz<9uq<o-iF4+q*F
z%kX)%cJpiRfazlzQIJSt?8F!vT-1)hWtfCcpHEsLdA(eq&BOe7m|7&qt=ApKTAu%2
z3poCZ*qrctQPk1YC7N-J$f=LuG*Gd0_xDd?(4Q}eZbk2SVb}8Yfg(*@ksBK%%tm%j
ziOYbMjPlH=Lm)FBI8~5kNRlFP7&==7Y3?|BbTLgfK;}V}^zjl(L`9DZwJt#tP;l)I
zL)y@IaP^QRxKJ2pf7{ESBOk922|8nM&bhaf|2#6LF}d}g3$RYqp9N8IH5K@>2<kGl
z72aoqh7MSy8^FSn#rvPAMItoxC<Z>h`t+lieaaMPIB%hAon9ed2aztwZU&nl9+Gi2
zyqUF|PBuDkH{ZE!8>V7$qyNeQ3Q5@tCGq(}@T9}^<buwNJId=6o=a$RNd1Dx8|#~X
zZW=p`N;Xz_4kRXG??L<D1W@U8{?J)j%G5CDYJjZ|ifLkI&7Vf>2Z`^=xA_`cvp}b9
zK}t~-r#?LG;9rtUGw?{&@o390cA(1O&<TMAlDHp8K1eAdJJA~E;+W%qJ@^3MA1{XR
zXd)9eUVp_RGBPod2XiY5jS6gNRSh9`wgLV+`D5wGUUR{*yN#^l^UXeF@Zw41pv%KF
z1%SiTu=G&sQ~bbPhlZ_~2h%6Myh#s`IbIu$Onp(bzzi}-++J2)4Vx)hPObC(d4$aU
zKZVvs-(FV~D0%E~!57fSSupSUX5_#lADsgZVo>@C2V#qq{!4=W=#K>Z^44mA%S;x^
z^Z+vNx^2~9RdVrkN%^Tu*Hl)Z<3r5{96jp!H=?63TjM$mm7=D5!^6V~)`GKG`8>kg
z!8>R7Ro2AW_c;wY+=IYjF{8L`>p^IXTshlqtgR_B@CYse8BBzLjEQ(VSe3o-QuM)Z
zEI{;bd;bI8o5<Bf-QxmH4`hc&A&{pbRNi%bk|%?6Ecsfi%5Uc1$_f}6Uh}7L7)!8*
znt`T1)!CG0qZC9R2_+IK697h^M0&OR7FSPX79ksdDMfZVI5<!UGltR8v5RUC!wD?=
zWfNJ)3%JN%NCTg7?a(@g$c$kqOHeMgy5#@LE))sr(%c(JQmjqR;C3l0`T{avE8Vi=
zs=Ip*KqN0Hx)D`Lk;^o77O?wUhr9SKImN{s@TkfG62txFg;#)a6ve?u#Q*ErqoH6~
z)L(;ABn1CmAbVC4cMpkd&@Es^-Cecl|6fM#_EP8LJky;YB(yQ8GmhpN3*I|5{`~OU
z``yrr5F3M+2iu<l+TIS{dxxOuB{l{hmHs9=NxA|JRAqM(q<eDPKj88_%|w&A2n-js
zY>>ey7f_8fW0RtqD4U>UQFQ`lXI(~5M}PzAv~UWM2pOwiGkRtRChYug&VoB$Um8Oo
zhg!)b0GNpIU*&r;+{grefmbAgl?-1N8`}7w8a|3?Uc7Q8y=Jw}0ApNQIT-vj6^aPp
zkho16aT>6hF4_J6VnEpJo~8RXo+_=*t0~W5B|=#hB;&|MLofgmlh^_gk2G}zU*X*v
zgl6v`K`i=Pe3vYYeWD3z74Z(jR}(|K7OnlY4D#Apd*9lZFY%@5Td;5_JL5-vy?!1B
z|A#EdkI0*YBCusva7j#QjTMsxkgo@A^7xva&Q(8m9+NwMzNJe?fvqw3xwJIT<Rz~E
zhaTJEwVNf*;m`Xa&ZO)*fCe8Svt()Uj4={k?>Bi7XW^Y(k|&_174#v1FI?RY4tW2>
zCz!@Gz2b*}tF0&2x;_N*jI-=4CsE@U$6FHbpO?BcEtA_~x+9t-P}>eBK`dLe|8(l<
zp6!_gg`wqs-kDwFxo1WIzM;24gMFmYUB@;dT&D47l7CpghV%?mYwtJH|DydoiV#pc
z;V^k`vQE1~F2T@xK97Fl%b~kpk6pXX`6e0KviPAB(6Zy9ytAlomx#CeWneG3WAIDx
zL`%WdumK}a4UrfksN~6X0)Gi<+9-ZF6h3_PUkQ#*6`naYXM<kan+E<VrSw+*FFEzL
z(c#L>PZ0M{WY9D%Te<SFLjSMK--Ew5Y3GMt;Yzz}ghK~<U^dklB&B+yJ8YzoaD<KH
zLNM37pV;(N{P29k)P%nYK;{2iR1I0FD8Lh%7og$CzyA1?%^QzS6gVZVhzbVR%~aUR
z$eMRL7Kb+FsZxl>htUBEfT1M@^xU+*w2T;i%lgyPzRD@VrG>9c|Gpj^W!2I=c<EjX
z;Djp*lqhXDCeybvyH*iSgK^|3*)Xi4aO#vKydft?e>$Rv@vr|@^kz7U*G)Ld-s)?g
z_@6|+C71D0Ou9I!kI-W)Htt*w;rwI<<@eM2^(g|hJk8bU+fI7Ew$tADPx#u{Q@!EF
zQQX+E98yPy*|H_mz7R_U(zR*R5H?`oJsicIkE0ttBMwhP!`}`V@)W*;>>oIQp+`P`
z-*)-hJxH&C?)+;&8rCU#Y=!}$j_{6nJE+j1!C+vuiYTaGG)@jA63r_`PDW#=h3pVi
zG#tPs8L+V~`;3OoD<B6%)QjB6l%Qmdx4>YijVOh#7<qatN{x+-oVC|ZW+0!yf4rLy
zt)3fSD1g)){3`g){JCr=F*toMshgdSm%+djAcb<NIh3JX#6nikBp4i+sJw%KQP6Xa
zW6&9jbE~4^D;jZb_eY6@4<c53In-+vu5wA~z$4PBq59VfyZ)>`yrry{?(->IKQ&=9
z2s!eLg>~f5qP0NIAC&_Mk05tsoB=@9M#-Ay^#C>xO-rjg9)J>nCg(!^AQq_IG<oM%
zr`FXA@3bL@_PBUKG$L^T0LcYj*N<PN4gDkfIm|wRt^ia>!FTkRNa7uzpje2&0DmH2
z7bzdGf+@_v`vszd2v7`8AHYmFQH>+fYGS|GoMBvxXU_zmv3^i&`SRRvp58!mm7!tU
zgX`C|;CtCQ$6zJx9WXb;WORH8I7|!wLJ-!N)DU0}36{H}EhZAxnLS7Wcnb6DBnH>T
z1E9kz3bi(cqc$GL;FJV5FQL(JOVkc$$%=qsNsm>KJ)klrkNlrYn^XSP`MWsH!S2Q>
z)k5;?)7Y^sVVi<rd2zSc#tAACP5b>`#cij3u4(_siTz&h@8ZU9CVKXw|L~u59A+A^
zS>jK~h_N1C@LU^W(mAWk^POR_BpgWPiKd5Lny{IQhSO0F07f_9AMBJ~&Q*UmZJ&5|
zX?I_b$jGTJ31UKazeEpb?rG2OD=yyWDn2V_jb-O7Ti&bNUYH#<G+UO+vaOsV)aBHE
zF24Em=iDG9Vix&Y)?11Xj;}fQJRw1()yCLGB{PIKGle_NtkbdGv@rZbbgXdT8Mhgp
zt8Bz>*deyb-}H4$4FDCR{?u_ovC-;6Ew&Jl$+r4K%>HOZWvs)8StY5*TQGFzuBR_s
z{d|2IyGMU7;Nf4mnQOlbcHKEi+bM%nJMCBQ(c|9TZGQ{5iCn6$!5Le{{m7UKe2;HU
zJV<ChT!eqUKJ&I9(aF32%h~vL{;1bqk*^XLqlNY@NYxQ0U&yV%E(g6H85s$5iptyw
znTu37S_4;C*QaaO-QF$roPwnGG$X9bq;J}I)hZ#wGEL|yXaZ_QvoN<RR^g|E4YTrn
z<Q6ybsG^AJ9dYGAX<uUA*IwzRO8H!_>IMLKDUNF_lX}UL?8bhe7NicQ$)gCFFz=?R
z&HG`{z*37h{avetP7i;JL}|6sbKIYLY=PLD->GjFxexYgKg~-Xc<jgdV$Zu9TTfkU
zGHk1`wyYDm)5yUwGhgid!kJ9c$pImOQys47BJH_!xE@1KSie5HvM2#r`kqKOq(}Yz
z<f2WsD}8ng0JOC1)`VaXrq<wttE~`z<5CH^9rm6p2A#R>?UYY#y=d{a*=<(sk3*|(
z7d#R-w~N`_Myg)+hCQLRa;uju%lXod;|Mnh8$<dv@kre7e@_KI3+x5v=!OexFlo6%
z^*B_M>ref{)z3HsudbEZ^P+)#y$B)E*r3W<)>;#KM80VM>f6Zrc?pt<lC4R24Pf$a
zN>^DE>UeVs-d0na_efBMU7{~A(Ho~YClovPodyERErp?zxQ4<|=2zcaQdt(a#oP1z
z*>5-80^t}yR}Gs`cGUUUP%U3!6n`2T8f@%0G5mtbLZ;y+Q9Y3}TV$flc+cZ?;uaJ(
z`B0b4x^|UQi^v|7tP~_pOch0bf_tXFB>oV*MuNi?uQc@H(CN@5#(usnuF_biTxjs|
z4uWAPgbeI-EU-W%SB+<&JB%K@vm=yD06j%wrC)x1ic64HD*YR`FOxzc@!)@bZr}dO
zy9F){6!P6~l-7-7fa~=gsT6RYJbgj1!%q-?a-c1-k4VUK@18|j3r5>ai&HndVdhiV
zOqys=VgCBKE_0p7O(1>}eFbnDvNXdq21a<}W3B&XpMXUs)G3ov!i(=r!hJ;o5eSw5
zd+J-gcg96T8IMyqI4+^LMQcbZI2VjQ$rMHI6Mm-h3k-ZSFKxEIpD{XY3Sb16b0ffc
z1z`1K<M*Nux56bxxk;daNmmOSXswa1wDtX0O<$p&B?}J$pZH1W92P@{uveyMA|gvD
zlL*A)gU63wHH2`zIY!4OekM@NEa|Q^7yMCu>^K^eh{zS<S82i>^p2I%t>~=Dgo6|$
z4?y$@IsJB@Wybb{{HxGa5I8Xem$vu=Y(W+u%_1;}{{$!U*Smv|7ho#mViA!WQof)P
zq2UK4u^97WTTAcnL{CRSIO#Sj(FIQquI}f$22}&jUGmF<O)gJA<AA3L!xzn2J*}vK
zt{3rF`pBwz92@I~m{@eW7a>_#kl+n62k{9sG#XSs&<}cbu-O~iM{B`BkNzlaq%*a<
zr)Mi{tueZpo*MB6NQkZuj%8iR+-x2zTscsCuW&#X)<gh7KWT6+`X%rR{_RFl_b_0X
z6Z|H5F*u>&`b!Q&>Q=d?^5)1kzP+eW5E8fzmn<qO1lUbPF3HVZ_}^;)5M_FDXVY~`
zRJJDn9hpAYff<cfOEF>zLJrEM!|hJ_Vhm;MkP?u_MWC5Vw*e*%AOrL4%HUF%QCWGj
z_wRA<6^^$xfq}X6mCO!gyiWlM3%#?XwDdASVPO6@L$Nj`L9_CEfZ@Gk6^GLQCHu%4
zfXJa8s8&{hTL$BO3?eV(Xa133grFK94~7(I1|-sy`bwb_bdHeLk->}j2>2Gh0AA8t
z6A;wgd@Rd(9Kje^o&lTwx;OyCBHw2=(~UmkQ?=p!mq0BL`_*UPB=sLQv_xosIW#21
zu%R1}+r9f*L!J%m%7=NCr3=ZwL(C(J5>p4MD|p`Ym|)xaQF?xK0-i0tF8rh)0l4qf
z(_0GBF1u~FY(4xWNQPp|r}?OvxL@hyKrv>(zvOVlK~$3y6HHZ85jLOB2QO5S9SyeF
zXS8W<D4s7;VHwa9_*<g2JK6)Pphq&O+yoh+!L23~IEeaKuaN=UI%R;O4$U`F%*Xx$
zY|A(!83ovy#t~GZm>xw{sJeNb`r*U*ML&FZhjC7s=xlF%8MPb4HaYqEZmVJsMELuI
zNydN3WK2=zFg*sJgF{3@wGNIuGi7*IlwVY-=|6$nWJvFFvHy7HlKfRniw6lFkK#e}
zfxg9A=b;y^nY#0kOn^Za30nAcm%d7P<%?BY;)bT>Gssmb`f?y0?r9S*kSxCI^lUQC
zFQIu+UVU2++&ArLR9%RRL%?FSh5UQKXaMu%(10LUc$2p7&i@>#r){r#KpLwL&Yy9H
zYLx?1`+7l;Jw@Uzb|`+D0edp`cKAJb9!ZD-4Y4Z_+h7%05`>U73Dn$E{rnm3P0C72
z`dqn4Il$oPJhX}zh~{N~4FwA(J}#g+m91dLIM04B5Tsl|RX?2cS7!^-w7wO_kK0?R
zg+>~b_thN8tCp@@xy$m`F<`l{tnfmX@e;0v>0AdZu76O&LTdSkJ~h=V!UtG9&%FnF
zCQ(I@kgkx@HsWK_q$31W<<0o;h?FoWS=Mt~0>q|~&dA$qE8Zay4h{>}1hxeBa2!8@
zLNiG$scgMr_3EiWX;&0(fp<h>#!C>)xFHQ{%T(&dJI`9S!(ZhUOv8Q}cK|lLGRg=^
z#0v~3tA|M%IC6y0O<(}|xRA3;2+^aLz|hF0ivPgaiei=(D)$uun=toy1x={93RQ4A
zG8E7TDZlstu_?M2j21^_eW`pR7Sy{`7S1Pw-W6ob#1mjBJDNPI1>ez!Bw0644gTdM
zg_17*7j1}Jp?f#?`)u9ffpBI82m(Nx(LihO6X3q!F7m$*g08qi7eBY9WkpLXZtK9!
zZqYzT@S^e2VXR^ppuy8nJ;sabWS-+KYv^JXvUA%Y?zBb88N2<?uLmKN&ZcH|&aG;y
zaQRU|gngskg(+$MNMSZs`L}giYx^~t{&S6x<QGQ?_0Tro(0Nk?Kw5hGxQ-nrNEKC_
z`c-y=7mpNIS&80mP@eF&aTu7x#C8n%S5@(bGan@$uYDuHVF$qo<E&T(5F7X06x=}@
zf1G{0X%<Jvb<w|;mnelZICtTA`7Od-2>t0|?}F)`kdpyV$u;NB6evox6o&d<WGDp-
z4;c(1@EN+qEuAx!ME2yNd-}p@sh1!G@V9rLupbDMTP&xfG|jtg+2>>1|0+^&N*{gm
zQ1Qpehu@wcatKob^qq|j<6mi~lm{}2o%A}$2@Jj)Fu5a2fguB@N5z(ZxjzKVJYSi|
z^19T{r%);2L4i0~U)FQx8K`9RS7-s|qM1BY=n$`r9>~B@U4A4X4Dz-=F@aaHnn*<i
z$m3hWkL2~08Hryq<{0h$pTua|JroM)B4IECc`>725cNuE{M`2@8*wzvG>Ucc_V%tm
zzBqy1u>PAm%S+FF+tL;u8zn)xK-+h>4SQE^u91;1HTdqkZ{_ilW4))&qti*=^ECKF
z$_O*#*E|$6_0c2IXAWlcub+@2Zu_gbcL<;fikyzM0%a*5%n~zLW-*8Ghrkz0$R4*b
zw~Ir0Rnc61-M!)mOfs7h^YFZ2^rXgvt&hYSl=R~KlP$GBD4Bg&#FAtGk3f`W8wS^x
zJt1r8nl*20&sIa*jSV7c^=TprwkW8tveHknqD}Vk^TEeY*K$|Bdw1OGVkJ+#kWBq|
z&gh;~GA_AI|2JatLJEjU@Ai#q2M@wkd#BoeEvY(TFhc1mWqzgN<_~_gl!vw8+x&OF
z^|)Es44R%dWiLjf<dVYL6FGCDN^D8V$;T+GMGx4qp^lncvHr(suyLaCk=36Ts)B8*
z7QCcoV`cTRjwWGA-X84fDc4%(=Yf5+aC0YS-k^ylvn2M9^7^vELiZt^hoFAF3%Vgl
zNJu^Fx@-BTx-N)@_B2cDBNQw%lRq+S9(H^VD{2qG+5QH1h^;cJ!xks$#cL~?`d(ip
zu5>)10^Jyt%Pw!FSaM2#=WxqUIVLNqi*<uPHsSj6z4ZK`|CmPw4h*0OZS$|R6>GF@
zU&acnefL`Z`?HPuN{gO`1&#~X8tnZY@6c+s|IqK>S-M5cvsW$hpDy3g^3_`K$>M6^
zPSv-sz0~fE-04?T5?&L3Kfw39s9#;#dX?h(SM|@{9qiUmI-A8Iv2Ehq>RAg6O1iva
zZMHNPWeGWmpW4hnlW(T~nNq1PF^86urDr;WwC16xgHU2J+=)q;7fYf>aC}t1!g07S
z;skD0BCbE-FeUvuUJ%BJUI49CD|GJ6v(J*ou@6T+d+%NBD)J-LNkzjt=&|1S!NHxF
zvWNqOirKrBk(}d!Z$Ca5!PN$ZYs+xfHcwCXBj_dYOCFdsG%6>NAxp7<0C?Ueua3Y1
zBsn7_>FkM`9uTT?HILh+Bq8$^T6X-FHN9*5e4{2urXNr`Gdt?jhJ`9uQ!q7%KpGSh
zpug$<Ba1#XX(VI;e!M{2uNPMi8Bri!L;LK-rTa@59h2hE0Ru%LEC}qU)0HB5O?KB~
zbu(wEpTZZR{PfTr9*r<MpkE=52<1D?D@3^sK%wlR;_L+TTHnFtOCRM=T!%Ijzp^nU
zhxCakE8{Q-2@I8rp@%}p+;oC+nP3Q2->`OLz6<u!4}4)bF!w#s1{*<YHS$&;xLcFR
zeLv00cc6W+1!vdJM|<v4abWDUI~u3}*ZB`Hnn5s@hky_z8OUqOB2)!~efsAkHF2C=
z*P-(@fi4=-=d|MTtPZT419E3FxC&dYPiT+LYZ9`t?wDy_vk#de?uMM|G{16#Mcsni
zs}Eh=9d;4HH22O88w2TQqsf@KHad7sl3Cch<nzbtGUQn@B$BO!zrwGv5O;Eb!o4gU
z_wypn)<k<WHYEPBZsxhG9k@2Oqm~E8AgptUhV6`1+p@!7mKp1I+~P#x@5eU+g54dl
z#2+5WlG2QuoBPMN_|k`gZp|$%ZSS-)&pDxyP`)p#qq${)S1g|n=9_JD=rrrTELF3{
zJQFFl5gbb<l4KA4OVx6G^9rkB^ep5tL$j<DI~Plc@<fs1L%xioAFY;J<GG@bCv2tf
zTlH+ZLQu@pU)n-z{H{Mxkq_3q+`qzp;|RW6{Ik>keCLUbQDw~CzSM!XSD%1h;KCm-
zlW}LTAT<i}g=T?$T}$LDFlYRmW?@;8_?U_(RN{&wc*(Qz_b4Q1>@gdakB6x8`6-q6
zs9ioH*_*E8{qY7WSXT3})pjT=D_;%dPez*zMv{VqsVjoX5Yoa2J3JNwh<E|b^K=%1
z_dQT3k;KlZE=d%giOX2{dZ>zse`9FQCK@N!`%lm3n({sZ_(9?Q6uCta72t&*U`hT!
zcS8KB@z6{q&)GM9Q8qv@>4vK@_}s5ybbN6RyS6>kyXc2L7rz(=tFU6^j1|<jlNf|0
z?d%$$0}jp4HmDp4$H9693_3v|rr~8%;4=tLL$D~QfrHxL8BMevGGXQoj`hmu+tH(e
zX`X;b#{>Tun0SGKMMyrOW{>cDa>zrKD1rnC=!|*svV{^oph-H0R}bwL`F(J3xuJH3
zSY(RA<>}ejAw=Xy=2=dwklBM)ip)CY*t8^fwrhJw4*I8<KE^eIW9!P5)#K-f{3>y~
zN=pfU6n3MS+(%Eh!6=mjbm2n2Zu#l&+dQwlQbR|0Wv1*81|N|Co|5Ai@Hlwfc89ID
zieI0MDjp}0WXGmxmgjjKc=%_~PE=^PaI4AeVIkaWvoS~Wp1YgA!m@IZ4~yz3nlr|k
z&0|qcWohZR6l)vnll`Bz-_19zi1EnE=a&`G&TtKi`mGiE6Y7JlNIJ%cn@LRAskCtW
zlP8~6i~1by>FFU;-z;dLK?(;Pa*Lcbf^(XFY1~e{b$R%fPf2b?kDncfy5p`+u0z`@
zW3}UUrxz$H?%K2a+3xz|a6s*a0ATFbGbtz=VUxNmE4pXf!JFMv|DM$krgoylH>!Rn
z0H{tV<HEvOQkNm9uj%;3gFGfb`!%aqoVrR&1J6<t{5Qxd+=>&xt|I$p+j4rdfN*N1
zr#~C_?{7@Ph?QUhQjT%Xy2<U|C=j^`Koj(3={bv+3Q8at>|mr?G5yn7v-V5*AVwI$
zURBc5A6S3I>weRPOzKsYhv$G}ScZ_owdP{kO3$FgCKUxf<w>MX;ZeP6STgIT_Zcuc
z0gVE5?Ns9xWkKJHx|b}MG|Tmj+qQ$9!rDn?4NKH-KYH}2o+~X)ctt^{heU%M>dUdC
zaO&6-opsX_8)F#~tT3kiTDroLrOy_`;)81-ucX{m(gSIaC`uS-O!{7kHKOKOB1&K@
z;6ZMNekeHPM2B7C@sNB$6FKW2*KoD;@<2R3Iq<#AQEX2f&w=R*&jR3s`-E4{ID-uh
zfqQYs?MMtQyuA$bb(TK3_Jw2%D0ccR+&w(DBQu=A<yRHqXGJ0y920B|4um*WT~?3a
z<bSjPacAK#jXILP5c8@XZHVt~SQ5|y-i>agam<n>OMX;lj-KecQN+Te9^%ZHvJ9s)
z7)q3l-)IhnNEK%LF~QhUC>eeBH&eXXnw<s<W~UwBh^9&W^oLKDKXGV;K$fR=?R6X|
z;~l%ZPosK|e%|m5-`xh=U{&@a9v<zG;S8vd@&6Yt)C?Jxg*0weHhh)38%!&T@Yd+9
zPpU|^F?6X#J@Ig&B`&I<;k0_>bMEeDV`JN7r+aYqZJ$PTTh?tEK*V;jH*B(NleE)?
z;WKQaCGhc}lyv|+4qSW=$Kamw`D@GN{;u<F+de<Uf}`~?_RIQ!ayQqXH;n(4TJ@hr
ziUl1s$MV^Z8RYe7XlP&<$p5=en&rXog8pv~6AZx)^0>lL$vn&$85#m-7>Fv!Z@Ba8
zSNe^sMMU<s*=#=@I~)Jvo4pYSBW~F?97!}ebjD487C?rGb%&lDjB}8j*VEVMGt9-w
zd9>@K(1Ei@zIAn3;{Sc>=+Lof`GE_;clc9#yJ^?S>!0AyXqlr8h!J<I2hN*@Nq2Dw
zy}?vylEV`#PbpaF%D(ab{Z2QRBLkO`%n<WJ0g&-+ja(f6c@*3BT>GacqlY2*@}Gi+
zKC$<IPJKLzSBj$l6ae_spC4*>@G`e&qP~DAfYLk<L?k?p!7Z>DoB`g}3uLu$ASyi4
zH-2Z7CN>7A{DQ<-#O!J7{O%HtREtCer+V8c+Kd%1^`=VQGnE={yq=$ygj^3)TuZRa
zWMyR^t3H~66N#f!m(Memr@y)>iJs*Cp^qzjxuw*b>eQ!Ppg=2N+~`$t!#+M7%Qt6g
z`;lcvu|tU)UpZ>Ka+2+U#G(iwg>HgoA61-ETIQfjBR}ulnF@OMU9?u^ll-mxJ74K)
zToO@5;taDGX6&Nqi?6RDvy}3T#>rvFmVY}9UG~yD;^wOvpC3cmDgs!8-<pM%91h%Q
z@~$i*r>qpW@<VHB`)@04egoKd>v$gKbh1T%tt%($Y(2n)@<`-)Ct>wz8w$YS=h}{v
zoxbl3_atKg(iDcE7!omkR;SbS@z0Y!L)Sz5w?{sGy6-HGFc@yN9yMU^6*Np1K2sXY
z8@EmqPK?MY)2F|Sh=nKtlOnQLJC)r^U5DcR{a1J7W~-;zM!+Pm_BF@*+uh9d%h#^m
z!*CMDlApWgRKZf%NwY~-gOBRomJA&tA<xj#=m9NWAahCtWH)g~<O8DP66&2LT2g0n
zpX>!~gn!HaHd)M!qv<i=Hp24pM8`4E^X&k<-0$l>qAHSn-_&NL@mu~+ls7Q<TmZZc
z8b-wS%RZ3V%OI(wkUoG$ZU|C<CaC@F9d*2CvGmG+9{^93l|>M&A9x8!cn!w2y#NDy
zHCbj?QM%BGLhK-`9~u7AnZE9I`6n<zVu6GO52S-}2-~7rR4||5T**hgUis5w3dHuJ
zgs}aQL4!*$D&~@#+hlSLkdO)1lNb1;J(Lu+Vg(y20Pl4Em(|kePuR5>Ty~sf`7*rR
zC>O?oiOhbKr<9-ME;|l1yjR2CHMqJ^b-OW(A>Lf7C8bpLOa;M9crcd&0|ODD<IQ(}
zqp2^m7?ATyCF#5ysUKPfva7uuw{E>~3MqQqZEX`uSg=UrSx*}$N09NR+QJ;G=ly5y
z9v(ZMZk`Mty3X^t(_J&#w`B=1zl7@dJ}<VmR9fGRUx#?Z7dRUkPQ017)uY&Z?93+|
zy0?T!KYjjIh{Jgz1J$qqcEWBm1dhR@t<7+vicXDkDzo<adC2&4(nUY^BkK#18nleL
z0>$E$v@{vxlcKF0*03k68VA|C9Se=E2E&Q>>m3#(U{>FPgw}~n4<w&Hr_MqyGTs?o
zl5g1EXX>FJTKYnk^7LXkOl2s+b5m{5ucX?h^pZd4i`470J7!dTDMV!!7VB4gE3_!K
zd9i~oqJ1Ya<>+kEYvZ3i!(+x->8^kNR-KbZ=DBnRoygKYi%f^Ed(2{TPrK_^=Fj?m
z_{7a+*{1b#pGDMRxafG)DsHGv2|vflWKcO5#U+D95!A4~7P!K7llM<PBG3#xUt$uz
zM0Gglcr;~`7~(esSoO&^lNp!^_4hBzd$uF9_^X2BmKJ6oYICM+<o!~^EuAHOM+EUX
z9(>Q)-y&xHbcZekzwC@ZUT#sjZ^?eVj~l7$0BbVNkg2BEA{1dFgLBOD(#>{k&GVx}
ze{Rsz&!NuZ<=tcp?K%&i>{%nLpL?xcf8G{Gn`O`gyceZ)*!+YzhbQj@aey<F?7paO
zr!*%*TD7!GxL9UTxBbo~j}@fzbQ9sjV^Y%7avhChVdnzhx_v)aG;T3A8H1Ta58{if
zz5b7#^|`YUA-JTQAdfs`1Ad}!o=Ls_sX4=XPyi`g!kP(LP?=+f82jB0xOOHpG+KM!
zw&?qHPKruO6PSZB+SQG+07@BWsII)5gx!~Yjy1c$e&ksy+9-xGdH49($@?hRCNLHs
zGR&=TbeMR8ulS?syK4?~0Ks<zuF?=V$R<h9PW{Sum)lD!zC4$5TG^G)!;lt@kpn$q
z$C4VcatI&Xggao?ZGO-7_iwt?IY}Y&j9Cn8F+;vya-b2u^c(izneYlZ?+CrQYWNn<
ziF3k}P5YVwa?_6AWc$VW{O^$uh+LL~V+Z3Joq_&b2Rw=+;daF2-?$*4yu6@yIK4eI
zy{w>oU6s}6<On;9ue*+q?{IG@EiG-7r4pb!w{muou%VZ_iXCJ?h|GQf#jFva%d6(N
z8G&*$&gcNY`b6ty%F?W7B#-#lk9vRZe4gyIvJ_vA`hT2$Bz!ED=~cdJ-uy91`t2E9
z>Q}D3p1Cui^wrkR4DD2Vj#EDm{U(Vxd@B@Ii>`?{O5E$yQMXB+?t7eI=B&+g?40m?
z)4nkD85HA2=s591WKo6^2DNM)<{UyHus8gR5*8!YOU{Gs?!*kF7jD2nJ9dau*cgDB
z)0RG$-FYwY`_v=Pq;Kz@#Na)#4O%%kOm^ln4fEG3;fuS|hIqE|475>u4FlyvvcQ=?
z#Tc5O=w@l|3dN>U#4;MeB&#MiI<T?J0{-i}|AgX|Wc@g+d+-6<%+}4!uMcQwr_$u$
zIY~MotOuKcIAt|SeL*pNa_IXBO!QuiR7Ei76pl%Ovxtd!RWVt3r@wCDOHC&v1-!Y%
zwfgq%;gfKwPCF22?VY*~Fwiy}S(GhL3P;RuUjjQc-EG)G%a$+SwmYrx*OH5MJdPHx
zaVvHvST??46YN+NdLV4~JU>(~;dJ>8bk@70aeRPpaeC0W{K2J?N7+DiwgW(jvu@sf
z9ibUWd&)(t3ip!U|JU4i$8-7j|6XYi(k>&TXpzWBMvF?K60#{HBV>;<DwWD83K_{P
zE7_Y$_KNJ4D9RQg!g;>c_dI_0Ip=X6zxzDS^3VOfOME`p=el0+*Xy}nicAH{^EEJg
z>nKj+5FF3+N6Cf(h|24-f|FQmF8}EY|MoJ)ZN4bC9cJ%(QJOePA?wsasu&V#4h;j!
z??9QXX{h?zv2)Z&SSgGLUCD8x+=x>V_>GN9M1f95Yk(;%SbMkJaq{z09H(J8A^CoG
z&K0+9g0}uQxIUOS^5KvsHXVZiJO(5o7^fl0eo$0qQI_=(75cg%TgL_zH9v)@-uw>n
zW-BF#ak=-BTbvi^c^xC6-@5~sk!XALTktNt8Z(HExH6;0*n7-Qv>R;>85D7G5w_tD
zJi>p&i~UUh-J9JsP<IEf^j^EmW86BpK`06$4nYwUhro-D##|ZGUKsfB(T8E~Nw*$g
zAbT~jO+5KZ(^{NaBE(^dfj`;wlg79%ZLFDE^o4j2uwk7*X$t{EkAOD41d|x~sDT-A
z#VuBEZcg=h^Qb5C^srZ(0Z#$X086m7I_rQ6r_W=2!iDhOA43QX8u@)!I}RsZlHa)=
z3WLR4Kdn)qOLcrj`!gh)c@&pD6m&01V}anII@8K1smt<C*Wl?PE7CjfKE$LezHMp7
zbj|Z43p>ncDj(v0q4CVK+qvVipC+U9U~a5h##(CrY5Oz`%Pl*3K7oaG;bo_?nyc3z
z+zXL9Dmu|7kVM8%91#eRlz@&E!upQBzP{lL!v}w~J|-m=7lFq4?azBI^yD_7D1a(S
z>49RG1nTifQxt+kGRCB$h0&GW;M+JWcw6k5k%Pm|GexKz%%JW_coC5!2pTslF~x-a
z_c!FG0h!LqXJSTIgVAC?ciaA2I8fKocvZ&qkKnoAOXo+(ThnCsDdmhy@4q35E!8oP
zKseMsOc^m6y4*|eJ_}f}U-J0z6%_m4y_K!IUot^8;7+j$nzZVhH^Djp{o-vz1^{8M
zZQ9DnxQB8@kt~s`Tw8x-zmU)c0DvT$N5_v@+ulBtV#QVm7nL}{tY@8nZ#dy@XzN1(
zqM@(84{1i}T5`8t%Wl}jrNm0PB3+5d$iq6f&mY(Y<iN(^jJ5zeNp*k{eQ{@&v1wc%
zkz8;Xe(!?L$p;p9ex)SfWW;*ZMHE?ahWCh5kH(oSoRLg8RN^zV34NVEzX3Z%5`P}_
z?Y|?276fwhb(92VhZ0_c5me1BY>SBRot&)XL>se>IQ3*7jOjYMx)-r0fTI+MwHsm(
zsA16iPa5@1gr)sHMus>ZGC-+<b^r>Fqfx;>I5{~PejA<!B}KIR)Ej`<&(9W*ZWNlx
z5UjO9bIn!66fq`-ni!{&@(zF|HX!bCpi%PqakVw*dyXtAreE)dm0xB`;6c8^B*NoK
zm<ta5BVpB#ccwu*L&B2W83g#I9obTQrOgT}<G-@oXp(I<)Egk7Kifa_q5P2aWL*9a
zsm7{z2OpsE7A6=ocDK0^%PCN`M=2~m53x&S-F&Oev*B|24pwl{G|m(&B4af!siM)<
zZ%&S)Siz0I2mMokN+g+LSht8G=p(Fnmc9IGfgGoelad{c;Q1k=*@DB!S4*lLhpuiH
z<%+}86>jVMcn0qwK*$xq{z@=H3l0;<F;V5bn%NsS<TLhmi-(<o19maJ7&CB$wo`fr
zs)^@ZRZqjl+FGZ$b`|ZJ$E!{r_;9_jdI}RYy5|~&Oh9bw5LdqzlTE18+$@bVxYpek
z&w6({tl_MTXHMZUOi{u9M^%a*)qDdQ@bxI5^f)m$f=Ay7w<SdnDZIsGF4TmOm=ypl
zDgfF^(t!$aJGo#EpjZ)Vzi0Z%U_`=&*=T-qwYO!9cuL;+2^E8x2zRnZ(yw>rIOH0<
zNG13#6Fzf-5Al<7pZdFfl3aEf`_!4Txe$5MVhYHRCtr)KjLZi3=GOz&=V)$trbPlF
z5_cR(IqiUw)m9^ZV2YO~ZU{~0hwn)t!8n^veJ#y56G{`8F8w>qtDuWA3(wVuZU;>j
zX%VDro1lq8F4Grz2VS&ILc-IbiZSHfNTOe`(zpz93%j^@yTMdw#TK!qyLj^|DXXU%
z_dDsjGA-BG#OAWEr}Pul2wFe^VPx>$_?2_FG<0(?eX_Pioe#!@FM55F9tsh%G|>K$
zu|=8kyGB9x?`csk^k~Zob3n180fl-bu}+iW{IkBADH8)0reA)yirU!P`*7*sFJ-yz
zjK`a#J_QIG^F`CnSKz!q6}YPUHa*M&&d8hq5<#qh5ye(b*Ise#z{vuB07N`W3&dSe
zKCA39S2rJ%Ty_IPFs!4?@RX2N*x}dO3}p-1>2NSWqSO6Ju5NDcu|KgqI^TLoo}IY%
zaguXeu|RVP5w5>lTE<tdi&p{)$6DkeLVBr|!MFRuD=wp)!hK&e$W1x^_K~NsZEIQG
zzqWQ3mXI{#g%qz3GUyj;@bYeT7#SXR2CR9y`ss0x+&_3_6=xq?S%JTWiqX}6OJUOe
z{`uPS{y&Y@{O19$e;5x^I{r;?$Qr5HO1<OgWMY*PXn68|^C6$})&&i_#Q=l%Wo8E^
z`~jyQ{X&C!8@Sv>==$u_T8v~1GgzN2eKk(!?7BM=#pxpEcoL?GsU{)`3w0T*#3q=K
z+z?TOo^K)8+<)3FP#R=DNj7SzjQ;z>oASv_gxrbmqJs#9!PD0T4cCplygc<w`|Chn
z5LyieJOamXq&6AJDn9GJ30OrEnM9=Offh?IpFzzxC1OkRixC!6);fpg4e6~k1oKSB
z3|T(&;QoDL??v}c;&YFH%mY;U0U!flhY-+^xIqLmo#YP!Pb}UGHjbizpnjb8w1mM5
zelQ6rWysobd5nSZUP5w5Icqx2Y@mqr^ufAp-GWf#kI;F9T9aA8)%7%4ZVQzEN#+?S
zK*?^eFvNTbK9DxittctKST0)0FwmS2>25?r3Chn1q&bh3_3*7k1OyuTlKf2@)F%?L
zg9J8py_+_ms7Wq5tVGgnLrCaBA8GW|_<R$&SF*@b4|M2vF>{<58b`%jK;_PitCCks
zVQU1p@uw^s9w0&_3r<DM2j@01NkQ@!Qegb&?tB`JKP1@>)M410!2lBX?ELykGzmCy
z-a)nRo@xGZ%@ctg`|t9O9xzHaT8)R9hC-_cWzJ{R_m^@J_ba;SkCIwKVxy_&=paY{
zB>5N+6U~Nu10)FIr&$P>Z29fO+iqT;3jlN<siQsIilMS0iyIgat}vxFyodvHBI=i-
zCkKv9lBNa^f{ww1`3qbukD^tZ@otP{sAGDlPS8G~h6U;vV~X%n2R4RroC@%Mn%`&_
z*gDXBi=GA{cMGWUAdyWZYZ9;nFL1qpq92%mq)Xa}E)ci}1aZ3(s*<Y29}s;NI*;Ez
zv2M2B0_;mCu>Kf1*(7}W=(4af1AYb_J-t)+<jzBaN8nFDrkuVTj<O<g74r^IJiU}S
zjv{AMDN*9WNp~OxkDc8JcUb{iL5jjVIcjd6c<w8TD6%PKWj)^7MRHiY106$JC;=9(
z2fU*qM&(_}7X67n%sXX9u&d>hFDJ4xckSAR9W|TCif~ccu~LP+1ayrA-lE<BHzRK=
z=!=ONO6l$taSIE!4I5A+=w8APgb(_qauO%n=Ccu}HUJNpM}DEJj!r6`#T~VViOG%p
zel$JgT7aKH7CH)}8%j4;*H*qfG6JvIw~sGYqql%El!APQH2HWmb|?LYY-$8J7|8~N
zCKXwRx8$NEF!zwSzZ~A7wyX>GSZYK|QfQQv{Kk}U)RSKbAvDyEr|q+Gh%cj~E5z)`
z<#9rj3mRx*a|H2(@K1Et;Ri6{kKi<2NRd@D_-EG<brc~kM3l9lv>EvI-()4vqv(&W
zFbEiKpV9cuS*OzcOBRV4F7!mji3c9p;;YfA1o3rOF27B$l;<dBSRZ>2L&&y2Cm7aO
zav>HJVFd=SV{?U@yxpFD+9Kk)KcWpJ`Na><kNi!D2U2)0$KHyVt`O3vJH87BIFfTN
zXsmLtagRc@3!Z7Rw?K=z@I$2}D#?Qv&+J!V_9u!7BBno=6FG&ewSYo~3cAOe%2s~<
zD?{ck-t;Ygi)Is~tQKfVLLDPd%7aMtuow>~%~J9ZV|0DIlO$_HAl?ZS1`x(_#aN{q
zkOTvg-i(cs5SXqYetnRU8MSB5Hh?X~=%58y24L|*POO3<TsE>(Ht*A-B4vXNU-!cs
z{{cf1vA{z9f|yZAPPr)xW7HjB)9ro_G%sF9P7*?#quw8y1v(#sV}U^agz_mmFDFN~
zha#u?L09#rGAI~{&;y4p4TTXV;sD;vEoknjH`ozHz7c3xc5RO}#%5AnZEe#lEsl^g
zOhV^X=LM<E%-&b;qk@Kr1Q*oVTf9fIK@Afc*_eal#(P}AM)qa3J_)l}g@q~b*GK{Z
z=7pk+H1R~$jT?fDIt=}{a)p)AC=a;>6>VuT0r`g4PMqw7PA4mYNFXL2?Rkwm9uLT+
zH4K9Zq0MjHy#CYmQCjl-a19X|Hw4?Lo07n+hEWi1Rmm7YWIv3Fj)o$#T#6LCvs@!a
zD~~RlA0NpOv2%Uo3&x6)+h=*-okq7x<BT?If7%N6nl1?2H&PNP^Vmgukzi2Av0v!{
zvZ>a3P7JhMpy(+H_!LkaWbX|SvoX{!G|o!7*E8C!&jgEeQ_H~Ez!t1|sReHb3c@ss
zmFP%H(Yu{BbixCMHv42V)Lh{<fgcL9LE!}%L%AX607lim3*z1l)ccTQcey-XBrE0q
z>V~)M^iw`bQ#d76<l|HnD0(dJt<5@vSpbL<hAKqUwmTFb-1r~IwQF~e&3|lO=vOg0
z^7Gerh~ud8?mh%@?T4mUK||QaaG0V8##sB*XvNAN7knriDHlWwB0`iSRrm7c2gAcx
zcNOBUW^g<ADX={7m!9(I0qy1)RQ5#_<lyh|S1v#XQ!!`U2dA){de_i!0ri(5meCn$
z?VIqPI8t|TauOHeg2P<OiQ%;eX#Zf)05v$pe7wE4rJ$iVdfmSx`1++<i*NyKWnOR?
znP;k(TL8C*lf2{d-Uq@N4}W0z{Y$$vyv69j<Th@!9-CM{C_B|dH3Znt?QPaY!$Wn5
zS>8Khi#q-qR_H&&mxe*KdQ1=8weR*<iJrftiVi#hE77Po*!Hb?DS(~^(hI@i$vs<4
zbANF@J{6Hn0wHlka*8hWNw5+HoPe~oMbvKyDeJSUFf@cH#*nNgoK|J^g7fFJ!WG<t
zxcz7-oSqAytPn^aLR*z!(IbmR6^cVSvgJYR4EY)GK`yq<-hLFxboU0pNfyF>U(p+!
z(;ZOIqMmzQ82h=-b)GT+)C5d$HD2g3^!$*(7`@&Ltr<Q`pULC9DNK}0Nje>MvTtWV
zFA_p23eHw#`*KJQaoLto#EcPZFJskH$^3l+C9X8~AXmw4dS883^8?LJvgrN!?YOrN
zWFDT%hK`eRMzaZG=y4-aSRlJ$0W~|~>GX~J6D(-dVT_vhB)_g(x6TYJpQ(M<f|!My
zwmn1RxxjO7SD)%tCnR=L&QJ-xtd;EkqhT1s(0T-|AmIm*O^eH?1!P74!}A4L3AWnC
zW`fV#g&^Z_PK4S7;Cw~vDCM+z3!j9kGUUAI5{ru4XFYLQ@Poi8(NNhVg^czSkeEJN
z8w=S2Z@M^^{>_}rFxSm#X-+d;O-(`xWn}I~a=(GM6V8K}^f^)u*5#tb;LXsmSnnhT
z4XmHSU5*TT8p}<W+1RrfP-by{k+YC^$tJP#o9wzmf0k382+0?3dIciDp+g|~BE-2;
z=+2~m&Ho=Q0HbtE^2(Z}ix+QE+k~U$%oMRkVLsnQ5j2o;#(B5-_X}O=an~1UcFD3r
z)B)12?s`f`p&vomBMtTay!@Xo<UFbj6mtyrX(?#Bpc_1nVOlyaY*O^Alu!Bcc=fC$
zqPb4@Jv}|&3T9dl+y^B4n4!zNU*J<Vf-i?B?_7*XiE*GKF;qT2ikHN+<Deej%oQ(l
z$KPVxAY9=KDdKYp5-^2)YbB0b{64}T0=(qUa6G?0uj;hHMC2}aTpm!?#h(%Wh7olu
zuv#1kdx(3Tr00_O2NXyTD1P5+H6<>|#Vh^{l(A;Va}}hmQDlAcCN%tNPkA@a^@PQh
z`I)C`#oLnoDlpDnM?eUNY+&+^eLP4qnB$#)cpfRE_aS*}9cFxS04w$DW4Azu;DVpL
zMd1V<0$P(K9a&@mu39f`%^(+}$Q{<nRjDzW1x|9IXU~U-k)J=GI66%ZhqNz>AHQQd
z<u^@Rg(HvHH7GKu0G)Z?k&Ep8cjzCD6-5en=UN^G7^YmtujI9o4YLK4m#^<8tU@81
z@3a6Ok#?Ve)p)m|XfdGe20gik)`3H!>_ZS2l*z-B!M?l;ogg8SZR+e}ZSv0PT)uMU
zF+efIvqZRtdHk0st`|@c;I&1L%#Us7ei2{v(;SaDpO9ayT!HHh;~OolxWx+OfA2n{
z-DKqWB1qMDJn-|qMR6f}E|l48l$+snCl(J9&y<{fM;yinZbG)^^fkY8*<+G2{ayML
zZb7L$%^dr>ek<FGire(0wT#0pHwkMS5sZ_yKbys`A0P=LsOr#gow_`?Bfi|B2CG+`
zcP!Cdueu0Dk3kLV;Rbd*ZjK*o49Mw3+!lEvMu5SQ&D@jtAz=H-k|cP~y09)V(Li**
zIglt!)RBG3eGNyw+(cP<Ilu7UYZXJu7erve_07~ANx3YY^){$A*{BbU#%8SS`IzO2
z=<(*1W9WzA#*Ix%O|&jqJ)n}WTdnam&NPZ1(<#4U?J_{O1*ni@i14144~)v7@06)`
z4ZJBAMZ9p~PJ79EA>@oT4Ux2jK1A<4Di!@zw`G7?jt9;OI5D2CS4CNNE}AQc$o2GY
zBij1=FJfcD&Rvm-0s3_~too$UP%f-+_Cs=igqdHKaXnopsQ=;c?wFit`CxqbPnmop
zMn$x{&8GB+l)X{TfVtk9HZav1gq0#`k*S>AS^&$$?~l>}PQGs@3JQvOLX%uFt6qb;
z=nWJR^XY07#2|zYQ3Vt-GhVhk5OlM{tmYTyx{VX-IKLcqg^uI}xx)cxP-nQV#-~{I
zDdu=Cd;|0l``_E!4;t~*#cOJ2M%$n`a2C{JQbu;=g2PX3j8hh8bXD)Q5BB$0>COeC
ze*^`Rv}fpyKEqPDkitn+-+5quOmmGb@Djd=D-+JK9=u_=4FQyX0|VzQ(jEL=`PeQ{
zoWPMLr^i9>Ft(b)Xh=Dexj7^7mfHM6K|%)Co`F<w3k??a#OGP&j!a|Y#Tzt&VC<lr
zw21l%y$4AQhVkZjl3vBw;C|%dfR({je~v~;ukAqP!$fm7;4hEzg|Cfsp^Ch)X#qHi
zm`XZek~uaf6{6(Hrdca-fgIF0SjCo6lf5zVmi0X5SP%oQW;Nv}YR89!k?0BUN?`0k
zF)l!sXJ>|-7E+|shFv_!9J7F)19K6Rc~O{pxf3y=mXz$Wn8UURiXu!mS^I@ge5&^8
zpBZ+IGT#nMD=Px7E+9K6KmIDsQ;D13HTLMyMRT?8%AwOiMPemrJP=8-lp>rbWY^pY
z>dS86P``xVfS$0%Yk7U7VoZd9*A5wf`R-YC1(N9Wu{@wYZchS030kyVODGbHL&hM)
z$yzpJSL?$0U0Q(3_$LSC@f;8p14V4_WTwOYZXZvFl!dq0z$PuKC$-lleF-if%Bm)5
zYN7lhzBzKzVj+VW5(}BcuWz>feJ?s!vilIKKT?fI<cKZ&Q0L+`#E=;xH?><5uIN>>
z>qy=ysVt<-BU$(;;DupNr{4>4FTDjSPE@HxJqk1C32Yj@mq@}!z>wSn92APd1gd4S
zKbGvp!38BlKd9rwm_lQJ1d2&j+9Iiq?2y6|;M$lAtC30ccw@;N)XJ9Qt7&N^wX~jF
z@{x!P)sVY?ipmcZ6@p1zLMILZt7OnYt6{`_g+R(jj$?f0D`3UMxu%V=824@E@5)--
zx9Hi4XAslVeLyc^pclzlj#Ke9wqIZ}I~sD9<ev(vKKl%Yh(M(%;&1mD1Y@q?C)Q1{
z^dwQ|l$ybBApM=FEK14xYsk%nX`VP)kwZ_`!;uJ;T}MQ~17f;MuT2F>=7tgQIffpx
zg%fWVHGyFL5Mk2d79@Ezh)jjLNwKs0s?V3B-{!=v_DTN-sg&#3fotQj-)f`0spq56
z0W*?I9>fPB1||Z?=L>+uNTP?iXAW5xS!@%EUFzlI2uP6tgKE{RxO06dRKkS8t5=l`
zt9HekzlT%w`(xv_%oNB<>KR7?b6ED478gG(#nPVd#ytG|@)SJ;MTrNV)m&P0)D>qc
z(OW-xsSO|0lHhz)$&=H%2=Hq4j5S5x2#qr`)7PTQOoc9$d~r5qSiArs0khVtu-IRf
z-sp5;qa{4&=fOdZP&IIf1vp0~6cwS|L8!LAxm#qV(CtSUCY}Yr&qw0Cf~?_9<N-KI
z>$5*{5JSl_3R10>4)Fr?jnWi7o)ketCpH&kTdl==$I#>WJH6cjb(bv0{@Kk2BiJ6m
zn%b7vjxHM+{(Ow=)O3aIHOC4!!387+rb}$GFlYNJiBl93m~GoCUgb;rWps|=<JgDs
z6kYF@nWzP8ww#HOV*^?QsH^7Y0B~&kiGhZA7Sgw&5@(M$8@=hWGJw=%8447^@c>kF
zro=18kc2Blw}qI0Ed|K3-(H5zxCPN<c-C}Zt<*6#FYt!a0GK)IA9So$9TrmnAGMJ=
zM0Q93<I}x3rgIOcaVPPiVo3yX!{OtDuywvWe4#ZMcov(ja2EjNe5sPM8}qxads*$T
zpoG!%MC8K|aFC=kp#=%wbF<7F1s!vG2sVX79!Ma-0;82ZU%n9PAF2wF9RR%~L8z@7
zG(EL9Je$N>5rhpUWh^bmz)<=r=F!W*4|F<_*XzP2djz1yskt=Pjghm4eL6BR**K$Q
zgJ*m^4=@+$wiC<P;tmO*(|FMG=k+Eowdird<U}lf3YTtfnLj`a;YK-(W1ss~(W8va
z%u>gG0%HN@Va&-B{&Tjw>Og9A3tr?9(lLJsxs{X8_NP4#HJ>-WJOvJ?1;mGP1~_Oo
zV)N%O*%J>EZK}=<!o&lCg*ELh{L;=istMb4YV03vs`5Gl&2J<p*6T|JD>_D0&f`JL
zhhnWlBR%-f+2~rl9@yuTg{YScVMGL^rZseScOagG4$N8*!yLgiPUl)U-@L9_)3hEx
z^6OzBd;QEuZ-0eG5;PB#+J9eJ^mBq#{@1U6Lmc{lg>Lv$YyUktFS&&OeEpm3#y|h;
z-%$_$r+=Nf3|39XXw6e!c!s;{fnDjy;Bcd%;4FY9^T*tVKNANtsH@4Tspe$p3vYue
z3`jNb0C&}4_NuN`IA$fqW})`aE0vJ&f{>43bhz~uP5Xez+J1d<A^eMhT@g}!|M@D8
zhIPjkL{~s^3hL(y@7Bft;dSe*1QPu<n)$?Mq<|O=ZT{z9xxQqwhw1W>m)E7VaR^b3
zPW$thm<+VsjPdpL&Cw=Wz4@<D^OGU3hiGU}kaiCJI`hA(!~PGi=-;KK{<$3gM?Sd!
znMC@3%Tw^xaESV^*W9%jNixFovKyg)xgtdF0_WbtF(Du&|6F=@b2-7Rn?m47uowO<
z3|nLq9E#9$6eu$5AGGdw#h;Q!<EqtH$$~;GoF-mig8Uwu`@Ie)2?2u1>LC(8Haj~j
z%W@3hcj}d<n?$Y$2>^JBZ;c7tNXjMHDCGaOg-;>co9tbIjEtb!fHZ`JIfZ*!yRNpr
z7QdGL%$V1@MA-Uj>s4_kkuNsDbbn(~BYr6a<0HEJCNbpBO%@&62Z^pZul0d<!bJd`
zi4@x>;}hDSYBcEHogs;4{SSSyW*%_kYuFW-M66c<B?LwA9#GjWH00U{(jz7a1?{Uo
z7IYScjcD|U@{_Yj1P44I0)I7+v99!sj7OIdIuTKK#0N)$d}302R`oEMKMMdA^BE6K
zBd9L29Uh1K;Rv}9JR^ifBioZ6Oi2r}U(v$2GB3Z~`SS(PMEf!$eg^cCtXSN^{myCy
zp5nFm-2#(pnHd>G1}AK_3U>I7xSmV;fQ_U=;^bfb6Me1byRtIr-A-_Uduw}z(eZgi
zblvG_JjeMjS536%;RU2U;Nn9b*aMdJQr&^wrNVc4<JGB4;QRC@o{V)hvazJ7a3i0d
z2b6x<b@3M<08@I(w^RyqZPA8m+qBkp<ja}k$L~lwM9zIqzbJzD#x>HO!g=CQ&%^1~
ze&AT_`}Td%62lZJkQHB07pr_edOs)Uai4=_KQP9eGTw`jQBLFL8gyl+w(=MW*`i~Q
zN<iO2R#?N)wGA5j_aHPm(6Nx=VKJ3}Z9h3yVD^?%>>L~g&7Tjt!z?0IWrPjF!0gup
z-la!uXegRS8fR22_FcP5c4-qyG3RkKt{cytJBNQKiMZf1Ou(c04QeDPWXNx27Z%Q_
z&BVpVI^SoNhjmYRNqQVCwl#NSys3Tn$oO?J)ocA;M*SV>p95_x+2UdBMt8SYMvvHk
z%%;W+fDG)0ZA4OD{xPPH8)JdNqINZg3Gyx$bQs764q-OtwHZ{qbqF7tU`DXc<{~;f
zxGbdX#Z^-bIlzK!U%Phgd-RnOhKAA537x$6r343g+}gWg!4^sf3t<J%t2GOqJcH>e
zLJ1LW6^^psoer2ZR>B3inBdV@ygIDDSkPruRgmrC2A_a2-$#tQ6p&34-toxScL$W)
zrJ@tO8`ow3NV1zZwa_?YY{|tu$eXTdV)6<a^N(2|A2hn&T!4Sj&5{D*{~{>|fIYaM
zWmG5HO3TvC$-yTDt~8z<>SI52NE5s{kZH2^LFo*B#F%FtMC9OWD!y!gz-0B%;LB#A
zMMZQu{lDx}3mQ3!8sD1b+l>Ssa|G#d38=u5_(`C?k6tRopgjiG8oTzGq;L-?Ca-ZK
z(k_`{9mPZX34Q8&u*9oOF2vM@LSPGJ&B(|aPy+SiL0<#&H)_X1(%$>9@04g~`o%p^
zZ(^K5XHOJtb?dS#uU4`(q?<wg85K_^F9#;awd5ig+?)VGEO!i=ZvGe#U}oRF8R)Lh
zotXA~4xVYrgs{U@@>nZuvmc`_!6@ikI)e#q$pn(JQ<H;p_b7vulQ?_!A;k)rG8;I5
zHi#Bw{I8UpTMFTY*|DqrAIjR}FB^HPo4YYH2?cJd67Ju@$0uh$2%DL1d|qgv1(ev*
z)4t=MIZ7UA-<=^!%o+=xi8Q*P&w=?$J{}3AVwJ*D7O-LAllNx%*5bdutXpt+6^pnU
zidA{<a*fLediEVx?z(e4Yz^dFYpM4ny23DD{wwd>YY^Qhpl#>HWJfe?L?bl<<K_d5
z;`g8d_Ib+hgat#5iMkI6QdWRakM~&G@g8ghJ5KYRSWs1VWv*mVyuyVk%-oM0SJTl6
zX8cCg9)aKt;xCE<TzMM;3#oZkJe2&skN3do4Lbvoy<w;zvjluC|3swX%y`l?4kDZ|
z0|<vbj4h7X+Cx+x#5O{`A^#EpBAQl4!XhZ0fg*x%i%=?c;*=poF~NS32m!LC;Y*JM
zw*7&}1pa<#wz0zzROU;?I0i6H&L{vRmVzTS_>j|9!f@atx`9LLY4XW*HQvo@H*MN6
z^J-|)u<n&QBQVx{e9Im1RhPQUpK_$ahs6e#A32Lf_*G;t5U`c7OAnLi_}nA9mXzbr
z*(}uYMtiA_y;y*VwbHd#=hNCti)ecaEa>{&7z@d!^YBlwRl_f#a7m;}*v<RVBeofh
zC^-)AQm<eAyw;>JXoHA6mA7f*M(CD@dYB4#8@X|pA)kVx@SajiWYYK0JhN!jYbEwo
z*x*R&5fX~9B6^*!Pt-E}gPlgLWLq~BmMbY@(<%9R!4hJ|&0_Y+WA?n^U<ff#bK#;Z
zbdI3CBPSFQIw@hNH;#82vczcpirZ#4O8m0x&;5?kBg{fnKY;#I8iuhwmvQ+Hegy*;
zz_;lj(AAW;*2qwU+QaW{BA&nvc*rb4pq8K6dY@;}>=#dPXK*d;`!0+zIy`)Plt-e;
zIEjK@a2U#T&C3CTj#MXELnOWG4o>AG_8z5}O4uk@kOqBi8cIq#0Re^QVM&JR4+Tr}
z+9QQYI<5429?U)-jPI5g(s4^`R&{$Y*KZYaaht!^?{Tf&8Fib|OoM9`t*uilt)E$J
zp2$6f(U>BNaCeVjuFHK1VeCN`k9C2%%BkWIt!dI&Qz`6ru70zsKX{pwx~8q^7Zs<&
zAM`NQvk0Nn-A*25$`w@t0U`wWKD_0^u;OD2kHp6UCh$AL9|MF*ug=uthxy_mY~^v!
zytk8riJDt}x5B+2aEi0Dvc7rw!!}#K$lFR^%x-b~BLl5b%)QUyXr!F6Hm?)!__!~K
z&iP0toC@C)0GoBfL|Ib>q%tOrbma{i*JuGAk;rA748u_3oJ=5!X)rw1ErVo5<2=8M
z!T$kJmkvyDpi|y4nr7d?YYKyG0VYoQs1X=*iG!PRMxEPs@E`-)n|#z!LN}ndBI$Eb
zg78lXFK|Z4@%IUgZoZlE<LA=>VFekZybcYex*egzBfK-k5wslUh6l|L(N#R-n12ko
zn1wnPY2WpUh`(%6M0bKg=xiY=@WLa3`SxAMm1lpCo`G^lCFLxbs-OYUo_&=#;;85E
z71VSbQn!lit`R<AfHD&iQv+lKr0MD)hu!bchLkdBOYa98N*zW(YSG3m_<2)Zg%Ei^
z^ztHcRnK9Opjg#M)co(8SDy1pEOf_Fx11s|`pWAsOM3hI-i7P{3N40|IXqDvy;FI&
z*I}3?cobN5#<p(21-#b%N?q1I9Jkf)Lx8m_pdTkXLM#kwRo1G^0YZ)VQ5Y*T3Pv4y
z@liIa8XC{YnL`n8zf~&EcR52NfKc{*wm;ShIRuM>^^j?v&0l@!+m!}={!W;G=_#z}
z^=&^BOWTxlElu8LBt~Ma31ZZ@4g|Bp6fFT#h8+5CCYqI-R|%VSy2uxPNT_8v<d`k>
zv&?J*bNu@xap@x3i=B-ZOVZf_b^7<cuWd-m(7v03KP*VN)pcjIL-?+7p?fFZT(~3n
z1*IWzT;@TC_#F0LROt=4ht|R(M~^Xo9gEZN1Sp{52xTs%&Az3|NBe}Nc+ZFWYlztX
zzC?j&Qo+mFPsF+QFrDGka}x-k<V~%sGvBbhN;p=?L_c2Kxbbz!PA;y;TRrdJU%Rtw
zMf-jG%5!SUUB_|s%TM6ucCr1<9g2SXJ~T3}U;tf_-*g|f{E7X-Bj=yo-D8^~G__~;
zBjoJ_-5bHxC0#C{uL1xTto}l#uSP5GVbRd37uQj%B!E}JU_%dSKqtf*k6&r<An0Q5
z_lX@_g9L1=<}8?dWd|3{dC<C5vIdX^a2PrE&;F`%0yIbub9_E}b;1Cl^tJa32+~3r
z`C)qb@SN$;;H|Le!AVc2ldOmLv7KRk7%Dj+afE(_{NvC`?-xEdH{Lj?=+u1QV};w{
z>Vxdu2Rrw%D(zafxZioyd$;RzzT3lE=XP2-^U)k#;<AzEiOb!GOB1hnwTa~&@^2l~
z6Dk?hvq=)N%hR*w?0<G`sHv>&jBM`2V1`#MywEe)O;!tnGzCm(UIPKnrF!A|!-lXY
z&V9BWH<t3lx$b}sa8~1=k$#^eFD6umXMJ$O9(PMJ`rkP9EJeDz#{g9}AoTI3y%S>Z
z9&hK^zMbFpiR%QDsO*dLkn6sl6@v}INIm4fJWFp4p7*{UphMp-<V`8OvBo&Al5k0K
znHc05&$hfQy>BGbMk{@{c9o3XblJ<;7TFB80G$)zE&<Hdm9&qT7dRyKJ)TX~b2@L6
zU(~j4@a$Z0`$lkqO}j+=rA>%n0co{}tFAeJMr^}tT5XRGY(mnAeik+aL-DFcPBAfc
zM9GB!MBZAL`}@<}#Okk_W)fD6A<ewM=3ZJIjd>X>;?q=JQ^<rJBN3PeIMZ6<U}SWF
zXEi!DRvu&}qO9Y=jnrox+HA4NH^Isdf5jn%K^7$T;YNpL_3e2>b>-hbhUEsGbqn5i
zWWR>rmM2#GyIo~X)^FO>7jk#2*LJ|b3ZS)@8`u>x(e@m#5W2H{7dN*IcC-#pJMrw?
zc@{S@gyyeMP`{y)&}JQ~xZN^P4-L}wXVRxn*Iv+c_-avL1Nqoba7EJ&$3?e#Ry~=H
z&gFB$pNR(#TpP?dfd}qKOAj_E-a(qjSyPsEqO!hjEiONo+!7!nI1w9JFbKZUJC+7x
z(#S=fF~W?A;DVF#X+v>(J2tFeKPjqt?pz47i1o=j-*XK@_skr;^|RFWl8@bsKj)d|
zRXAiaKi^^8E`{H<B6AKCrWgFn;Y(WVHCGIpXWVQ*etv>YKtKVRlP>Zzb#DIgn`-wJ
z>H{RtoEdL%5MOrWjHvPXCmR~%6+XL$ov?gRVO>nyv{S@ig)8wv|1h=$;V<sT+%0Mw
zZs=pkr)#QKGbe49{~@Q`?ccFdD{6FgRiS>4rI{Z^ZCpTh1W0LnKwQ=bzM>vEG08B_
z8e?Y@Sp0h6n?H-Ud7i8j`bi>++wFRSTuna(B$b94DGa9DOQ>e+oUmZDAU=PV>%%uN
zZpNb@tjT=Zj8Ym44X%2eie#?-C$&N*_mW1mtty&Dny<Bv$d6kXZ}n8gf%YKg1k$AC
zKUrjdx#M5rm%lj-;VosKuj)gYk%{)@m`ZMff4<)~zgFxEOmKAUytiWvDUycR{=FNe
ztn8~-{^t+qF0DSw2t{RQtD#p|nWf(zh?i7oYIp~Co#26>6Yjw3V`5{=dosZPi+9YO
zgE~?`K*}ARXZl<1IIDni9ggjiPS4(8HiT7A15r&Mt_|h90AKSRPkS?H0(f=8z`A{d
zAKL934~u6}7*?=j5p7gx=uv!#fKrs+qzYV+Ia#yXX)@ow;aVAYTJ0|tD4hg^+{(Vz
zAb}LhAPC%cVeXoqSn6b@Ak(;Lw5;@+hyTf0wzUt$xioXtPIxd?ADa7hhL-uDJm^OT
zJ{?!Jy}oG9BWr5pWC9PIgB&*=V!PO}u`$0@YW{Mk>@O=Q91O^0^!f65?4{|pc8g<z
zr5=$RWBS=euBRBZs5I!hSf>r>OB-%#ezN(q>qNgkU4|LrG5U7bU`+ZB9i&IGy^#^_
zf!NsCgF}&`h+^PYPLv=5C8(tM?SANhOgkmiwm!(TL)Y@8SJ-azO-makW0S%-+=NjT
zUgcpQIbY#RTzq)A>T!Wz0p2!fO4L0iB~^`=4u10a>w1=x7Mm^j1+{T$cPeaZas1&O
z?2>#X-OH;v8W00-lj3VRx$~j9w1rHw?<(AJ4XVghU$S~@ELsM9=k`X2@-<gXY&VgJ
z#g}ODCe_T^QIXrM8q{IuBwnjXd?GoZIi?Uzw|Q*(iU)$EzE8A5A9zzk{KI8EENih!
zWf3+HClaAaQ5!fmtpH36y7Swwy}vJ`%>*H+M!6F?L&~F)ME4M{e!$`*95;nMgZFN1
z6>7bDtIE9H;u}^;X0{vKeiBF;Z)F3)7h{ukV=G1$TKhb^s)WzhS$)Q2X+ccfM5}=N
zh8S^qd3isGH`+jH{OXSfkCFtxWtVkU1vpyw?fY$qJD%5<sQHg?ykg?FuRDL`4k0)9
z#IU^y2cKy_Rew}TXP!U1*+N242t43vIXSO$Vxum{Pf4>_)TUJe7ZYahlV}Tl#LR71
z{&@gu-EP=rlj@z|F>;M-snn)hcf*Yvd%+NeaokHe)6$^0EX-`Pg^MpEj*;Vaqgym#
z?9mF)>h3UJCE_W3w8S0S9Ho})b2Afo8~WAk+e*SCBkzoDU6?S<lx<WWJ0eoJnnj#x
z><4(QaaAJnwMunD=IZZeCjeOD#=pBeXfs=xF5(w5pk&}9h(h~ngc@t#7ng=WAD-I|
z-Qb4LfToo*Fo@c20J4sJr%CQFy@(etVsItxC)#WlR_^I?@F|=&pFZ3h;2F_vbi#`t
zWS2UA3+;GT_nuX6hjXW+K7N-sH*dS3=YW49_cNEyYS7>I+c8`vV001C!9@Z6z)DWs
ziRVxQ<C2sBNTqE54u}7$(Y80To!mk?Vg34s3FUhe7p4@bS~m&{kyi{@Dk?Ei?tJ9w
zk$?wWHJdFa$pMKQDH#cz$fq*I<&~-T^<k}Pdh6Ha1mqyaZd5NKJiNGTL>~_RdW9G6
zx;?xnw)0t|Q2nqKJ$3TrvIA#!hd$sfnp+gWXp*Lisn};bhSti<jx9=weSX_&TTe>;
z`t)P~$^fSvW20C5uz*QHO6nde%jVi^S3C?OhllI&I6e1WZK-SPP$c`jfguxXuUBLX
zG^ZNJK)bG_$=j0Wl=t&9-E9;_s7$_GJx~iTk1-sXTHl}~+O9Qh$MPkmr|xCNz^iLp
z_YM6d8oH^ss~zdzW$A;-Z;K6%SIdwy{YGRqUY0G!k8+;MoF9W6{TLLH;#GE{Ta7>E
zsz>KuXXZYEihboVvrJqR&K>78+Y-90?n2aC%OYl*;_sHEfvtnhD@?p#xi*~>!fUGm
zq57@nxx@a5*-iL2HXj<*zC1V8r#Tmm%WypQjZp)+V5y0THJKsxpssJ_8ZfbY@}3xm
z@Xoa2!fzZzwRWuy>z^W?))J4sauo(!0rr{KU4GjFTM@}5{b6Qw)@fqzk2HxJKavrS
z^6l&X{rhL0#Rwa$bDgLdHpHD=pM}(vBsaK8uqE|k90{Q^KQMbO?Gv2Tv6cHzS<Ni9
zO*@3?C8y=bv&Wf&LuWzP+TNO>s{=Y~S}_8#)dkZQXxgLf2V=VHmr6Bf*~AA{&Q(gO
zUZXMOcN%%fn0U^bKgG5Ltrxe3sLowxi>^;|vuT6rU39`sbTUB)6`*QyFZo?J;>A4j
zm{CCJLrG<&NsVJNzk3B;A=8bz^#%;PMQ^ifm>*+98^<>epLOE+SImzL)fN&o=+8E5
z#|AUg-Nib`h2q)q8e@*)a-AF^p+Ssb>g%Uh@yExefkEx-@o{|j66!?9Dgnzt+m)7b
zUw}SXr5?g^+p`YC8#gW17}~!#WntMg_48SAW~#qXu%DmOkaJr1EhSaJ$Z<*QV9}Mz
znGr)?M^L=3o}P+`1LDYjzS_Zr53Ih*aeRN2{n{UyAg6i(Mtg<Bll8jk=iANNZ8@U}
zMO#OOlU#g4a$C<nY^rW-eTbSCRdCG2wV%@VcuhJ65bKI|b{V8;!{I2RC$z)uZolh`
zR%E2s^M0(our}?_6`6J81)Azx`^H`G=!Vr9A@qsc<hJyU*z!ZxCspIJeBI5d6=Ha@
zPF%%69Roey60$?cUtWy+=EUfuwX2)fJZJ^HTzhP_*mfzEL>;Ld<B9&JZ3<`7&VAZY
z!F}X;06mdW9pPd5R+x^eM%P$BVBLoXX3TEyo;@+2dG*8ewwO2<buEu>!;F4WX2tc-
z?-mC)7;?7w=s;<w+H(-mVEQR!7h+>K*7*HwPw#-ljL+-z(W5ta?cN<dD)tk=TYE_e
z0A!P#T#OB@<A?RLcNA127OD|=(=%SL)943JD=0kq)Q5gZ568;IGbU|~;{1BkNJ}PJ
zO(*$~m(Y#7{0L83g;%HMcJus2PnjPT(`twHFWX+*mBAa_7Fqe~YSOLN*GmhTs!QB~
zk@z8lVp|CbH#stN+0;##VYYQ?mL_y~#zsM40aj=7yi|=%4C8PYX2=vi=MIcb*?|b>
zW;9HfYkQp3x#Esny-y4w7X1A`pI>wnzBG|`KTXRdvA$N@qi~GPC&O%V_+p;}P`iCv
z0a%e~Bm&9may!P9rWsBnymDx&r`jg3z~{A~#R+mKJ#1|iuY&BR()1S|xpLDcH0E{j
zZNNZgn(J~)RODA?<DL~j%PK%dUZgt&R<$iw-|JgaRV4&*W?&sImtCPL^IqSw%$1<o
zkGX&RzA8NqE*;fA%&;C+Z{7`92jC9%|MEr08;Y{a!;mL8K|aGE6^TplXk&qNsAxM3
zX%5y}&lh6HUMbQ5Kdz`o^s*n`La-p(x^Dd7Zfus^hxoc9gVl4W;q`6%uZ?~Cy#Gyn
zY%D+Y){%92n2i^!-JNj!lx%SRWQ(T+%grXof#>+%Zaivwh%Bv71U$Ve>eG6||Ir85
zv?iwizNPg;#_>0P+kRx2jrS|}i4g@c!G~cG4}E)Lpo&%7<IbT8yo^YwbMydrcEg(_
zK7-8F$k(qwOog5CvI7l!taoKEnm&E>afpPjCYD7#P=bY@uTJJ@U~%w99*rwMd3Nk5
zN>|&PXI+6_D#|Uc$)sYg2j|4m<_pW1wLKbV;yStgXiU$03g7&GEc;6#-h}&h$Od(=
zSZVNMx8~fG@AroI;72-OOJsnSot);S&)Q_`&MTgmmCa()X637#ZJVH+VIE-z5ldWE
z33pm+XHxdWJ?5bz1dm%J+5a?50y~7BM=P%Z(YP|o%1<?&E@^bcwIvWE`sD@LEP1m(
z;}<@g*;nX?NC}kzM8cF4<0Vit^1rcyxT&!0yVlSb#51IYmmRTBONC(Afsml<=iG|K
zqc%MjgHKuyp^oID9k07AhrMa79XIU2tXZH9nTp*bBGNL}TUQ1^X<F0?W$odFv6)%I
z_vVzC6pyhnOT0Mk8Lx*rh(VO^X$tgTDC_lWz|@KH$2eol-8-p6)>$&>GPa1?zTP5Y
z9h-sdm^jT-0!t?$**P}2|F-|+rl{6^M^L}b^BZ@H7_R(jYjVX}LL>{F+Lt?G`?_7P
zn;@!$wSzS|<ivnOncuc_8q<N9MnwF5gn0zl{LAQXZBMa{uQ%Tre!p8prN$4EwJqXh
zk>TNf5YhxJSh&OwxOQ*@YMv2dsn%W<`}3g+6#AiqQjSN^Ob^@89+x}3C;Dd<M>d|m
zkrq|NMEjHNnR)(q%PA&Rh*|a@nyU6Zjau*E(TBSEdR*JMBIJ##f`TqQC^y@OQdVY1
z)+_vaD^BB3R`5Rajq2I6S`eo(4D;zJ1pnNEPWl)2X!WXl7IpDLdNv_E%<&XLv&m$f
z(0qBnDon47+j^abLaQ8Q%e`eFweoSDkp4ltAK_I^-u^`>l@(|hP6M_&yHpRgkK8#K
zP|GjOodb05kLS@v>?9AmR{&t6Ej*aN1<`><O~}ol@oh51j)=<Fsh|AB`I8;eT9~f}
zTMf205ij_~;3J`D*Uhm%VdO0Cca=BU%&2=e?W||dNu3k?NqZQYj+itq)b=RuIUYXi
zs#U(Gbv-SC*3m+0zR5+M@W*@b%fobk@#kWz2fEhDk9murS**aJbsFxB-tv3o696Mw
zpVi(i-tDRr63Hn7YJ*{H{g!1c|MDC$^d1xadR)e7cI?HWU(98sf>)nl5rD&NwZ-Zw
zBf7;J=N-JlpcOtcFmV|F$EY-JYAe1=4|A;5c!c$hv?n9Ex);|A2gyEp{J8k@)#A-z
zDCxmj6N~c)tYe-(KaANElDM-fDk?bxfwC`PTR7J4VH0dLDBl_(y$LT;)=OSRB^tfg
zDKF%GJjY8HN`HkiM;Oe#zS0fqHbEflYtdOgoL=u*_Eiol#9Pw^0YzQa$burj?I(b~
zZii)h(*UP%g8887#=Q@jyUq;Db6r;nlb!Qdx%ha~8jq8DCz8<*;*9SOU^HByO$RUt
z`PkS<SK2kQ6YMQMY~HK{b?Lvw3bj5eIawZ8Kmms(2D#@DS*E~B_EvX`#T&ZK7W=xF
zrK}$DXM{%|3#k=xO4fo%dF{kg#~~u3`pID;XuAOl9m9dES`rN`#YhTHfk8&9hGlkm
zsTgb&Yn0<LeWNvbj$vL|sCCrG&+jEzGC9b4YiNa9>8~=w2U<PkI3bb))313}n|M6A
zYS6$Z1J>IBEpb$(V=O=c%$&)y%t<t+7VwbAqZ2siMK}~^oBV~&&Bb(72)hvFa;u-B
z-)+9Ix;Ah{b~_eES)x?{-YtXkNKr+_A6NS3-iOyjx8h^_4s-=`!C8n4@u}q>O!zb$
zb$+r0xm1pWKs3C^7dtyVWTuWvno!G@HZ5S4|FVQW6Caac@2K4Q+KSRrLrYEeYZWOK
zm|PVU73*8N(QYH1&VRcB+_HvvX58$TvT8i~*B|kLZ+j(<N!jx@>>2>q_Mo*#sZ4pc
zzYwcSv7SI0J9cj!_h8y}H=j)kgN5o@i`SHX`+xKPMo1<cb{}|e7lZkvw6fA;O0XeL
z^#$BGQZ*>kWafWqnzV~|X}m9bQ8bT&jjupFk(quF`&<0RRvx~-O@wl}-%8QbB`!St
zvY7V$t?q+MO+_!#Y4Bj!IOsH+jje@IFk2jjWzHQ$*`604qd<kVSyYK-;-eA(FAuv(
zZdBGvR@wjNLVe)7x|oNW|8joWv7-H37pKvm|4V6|UvInM&wsy(+&N71_q$}ybu9k-
z9RZ<lR{s4@LCG7N{{H8c`!s)lBGtHlvBuv&eZ6Z%_TT^fzw?7t=3q8A=oSC_OZD%x
zn%H=o<fW%8RpiVzo$@m1mHp?h+TlNV0SxE3g9u>D28aqmbAxQ(uEz$B`LEEu5Kbr?
z@6b5=&s(P#K<$|-;BgFTG0EcQYyl?PgLuR=zbkvTCi4}#2uS7=k_%+mcXa%{Fz9}{
zwuCyo4f&~O(QV2D(6a&-(97fWOZvRc^7-FV2LCNo(>!33PMs0|rK8l>vEu*T)f2K_
z3=5*$=`$4N`Wm8vz8+L3z_qbb(+kPVbxu23tH`L{;Q%b3p$+dK)ENHa=0sh*pFiK=
zYN+eR`<crYQmO%{&VN9m$NCOA2Vv3C(UF5o=HK&z8R(Z3;j&aJyQBWk#ijVUO%nh7
hJw^S0a%0W0T(A&t(@PP;@*Ikil9ZKrbyVx-e*yCCee3`L

literal 0
HcmV?d00001

diff --git a/articles/preface.html b/articles/preface.html
index 86ac21a..fed504a 100644
--- a/articles/preface.html
+++ b/articles/preface.html
@@ -26,7 +26,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -1200,7 +1200,7 @@ <h2 class="unnumbered" id="references">References<a class="anchor" aria-label="a
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/articles/training.html b/articles/training.html
index 06c0d3b..af8141f 100644
--- a/articles/training.html
+++ b/articles/training.html
@@ -26,7 +26,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -112,28 +112,28 @@ <h3 id="red-hat-enterprise-linux-rocky-linux-alma-linux-and-centos-stream">Red H
 <code>epel-release</code> RPM. Then R can be installed as described in
 the <strong>Fedora</strong> section.</p>
 <ol style="list-style-type: decimal">
-<li>CentOS Stream 9</li>
-</ol>
+<li>
+<p>CentOS Stream 9</p>
 <div class="sourceCode" id="cb1"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb1-1"><a href="#cb1-1" tabindex="-1"></a><span class="fu">sudo</span> dnf config-manager <span class="at">--set-enabled</span> crb</span>
 <span id="cb1-2"><a href="#cb1-2" tabindex="-1"></a><span class="fu">sudo</span> dnf install epel-release epel-next-release <span class="at">-y</span></span></code></pre></div>
-<ol start="2" style="list-style-type: decimal">
-<li>Rocky Linux 9, AlmaLinux 9, and RHEL 9</li>
-</ol>
+</li>
+<li>
+<p>Rocky Linux 9, AlmaLinux 9, and RHEL 9</p>
 <div class="sourceCode" id="cb2"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb2-1"><a href="#cb2-1" tabindex="-1"></a><span class="fu">sudo</span> dnf config-manager <span class="at">--set-enabled</span> crb</span>
 <span id="cb2-2"><a href="#cb2-2" tabindex="-1"></a><span class="fu">sudo</span> dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm <span class="at">-y</span></span></code></pre></div>
-<ol start="3" style="list-style-type: decimal">
-<li>CentOS Stream 8</li>
-</ol>
+</li>
+<li>
+<p>CentOS Stream 8</p>
 <div class="sourceCode" id="cb3"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb3-1"><a href="#cb3-1" tabindex="-1"></a><span class="fu">sudo</span> dnf config-manager <span class="at">--set-enabled</span> powertools</span>
 <span id="cb3-2"><a href="#cb3-2" tabindex="-1"></a><span class="fu">sudo</span> dnf install epel-release epel-next-release <span class="at">-y</span></span></code></pre></div>
-<ol start="4" style="list-style-type: decimal">
-<li>Rocky Linux 8, AlmaLinux 8, and RHEL 8</li>
-</ol>
+</li>
+<li>
+<p>Rocky Linux 8, AlmaLinux 8, and RHEL 8</p>
 <div class="sourceCode" id="cb4"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb4-1"><a href="#cb4-1" tabindex="-1"></a><span class="fu">sudo</span> dnf config-manager <span class="at">--set-enabled</span> powertools</span>
 <span id="cb4-2"><a href="#cb4-2" tabindex="-1"></a><span class="fu">sudo</span> dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm <span class="at">-y</span></span></code></pre></div>
-<ol start="5" style="list-style-type: decimal">
-<li>CentOS 7 and RHEL 7</li>
-</ol>
+</li>
+<li>
+<p>CentOS 7 and RHEL 7</p>
 <div class="sourceCode" id="cb5"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb5-1"><a href="#cb5-1" tabindex="-1"></a><span class="fu">sudo</span> yum-config-manager <span class="at">--enable</span> extras</span>
 <span id="cb5-2"><a href="#cb5-2" tabindex="-1"></a><span class="fu">sudo</span> yum install dnf <span class="at">-y</span></span>
 <span id="cb5-3"><a href="#cb5-3" tabindex="-1"></a><span class="fu">sudo</span> dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm <span class="at">-y</span></span>
@@ -146,7 +146,7 @@ <h3 id="red-hat-enterprise-linux-rocky-linux-alma-linux-and-centos-stream">Red H
 <span id="cb5-10"><a href="#cb5-10" tabindex="-1"></a><span class="st">[puias_computational]</span></span>
 <span id="cb5-11"><a href="#cb5-11" tabindex="-1"></a><span class="st">name=PUIAS computational Base </span><span class="va">$releasever</span><span class="st"> - </span><span class="va">$basearch</span></span>
 <span id="cb5-12"><a href="#cb5-12" tabindex="-1"></a><span class="st">mirrorlist=http://puias.math.ias.edu/data/puias/computational/</span><span class="va">$releasever</span><span class="st">/</span><span class="va">$basearch</span><span class="st">/mirrorlist</span></span>
-<span id="cb5-13"><a href="#cb5-13" tabindex="-1"></a><span class="st">#baseurl=http://puias.math.ias.edu/data/puias/computational/</span><span class="va">$releasever</span><span class="st">/</span><span class="va">$basearch</span></span>
+<span id="cb5-13"><a href="#cb5-13" tabindex="-1"></a><span class="st"># baseurl=http://puias.math.ias.edu/data/puias/computational/</span><span class="va">$releasever</span><span class="st">/</span><span class="va">$basearch</span></span>
 <span id="cb5-14"><a href="#cb5-14" tabindex="-1"></a><span class="st">enabled=1</span></span>
 <span id="cb5-15"><a href="#cb5-15" tabindex="-1"></a><span class="st">gpgcheck=1</span></span>
 <span id="cb5-16"><a href="#cb5-16" tabindex="-1"></a><span class="st">gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-puias</span></span>
@@ -154,6 +154,8 @@ <h3 id="red-hat-enterprise-linux-rocky-linux-alma-linux-and-centos-stream">Red H
 <span id="cb5-18"><a href="#cb5-18" tabindex="-1"></a><span class="fu">sudo</span> mv /tmp/puias_computational.repo /etc/yum.repos.d</span>
 <span id="cb5-19"><a href="#cb5-19" tabindex="-1"></a></span>
 <span id="cb5-20"><a href="#cb5-20" tabindex="-1"></a><span class="fu">sudo</span> dnf update <span class="at">-y</span></span></code></pre></div>
+</li>
+</ol>
 </div>
 <div class="section level3">
 <h3 id="fedora">Fedora<a class="anchor" aria-label="anchor" href="#fedora"></a>
@@ -171,24 +173,26 @@ <h3 id="python-3">Python 3<a class="anchor" aria-label="anchor" href="#python-3"
 <p>This R package relies on several Python packages. Users can install
 these packages using one of the following methods:</p>
 <ol style="list-style-type: decimal">
-<li>System package manager on Linux</li>
-</ol>
+<li>
+<p>System package manager on Linux</p>
 <div class="sourceCode" id="cb8"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb8-1"><a href="#cb8-1" tabindex="-1"></a><span class="co"># On Fedora, you can run:</span></span>
 <span id="cb8-2"><a href="#cb8-2" tabindex="-1"></a><span class="fu">sudo</span> dnf install python3-numpy python3-sklearn python3-scipy python3-seaborn python3-tqdm python3-beautifulsoup4 python3-selenium python3-lxml </span>
 <span id="cb8-3"><a href="#cb8-3" tabindex="-1"></a></span>
 <span id="cb8-4"><a href="#cb8-4" tabindex="-1"></a><span class="co"># On Debian, you can run:</span></span>
 <span id="cb8-5"><a href="#cb8-5" tabindex="-1"></a><span class="fu">sudo</span> apt install python3-numpy python3-sklearn-pandas python3-scipy python3-seaborn python3-tqdm python3-bs4 python3-selenium python3-lxml</span></code></pre></div>
-<ol start="2" style="list-style-type: decimal">
-<li>Python pip</li>
-</ol>
+</li>
+<li>
+<p>Python pip</p>
 <div class="sourceCode" id="cb9"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb9-1"><a href="#cb9-1" tabindex="-1"></a><span class="ex">pip3</span> install numpy scikit-learn scipy seaborn tqdm beautifulsoup4 selenium lxml</span></code></pre></div>
-<ol start="3" style="list-style-type: decimal">
-<li>Anaconda / Miniconda</li>
-</ol>
+</li>
+<li>
+<p>Anaconda / Miniconda</p>
 <div class="sourceCode" id="cb10"><pre class="sourceCode bash"><code class="sourceCode bash"><span id="cb10-1"><a href="#cb10-1" tabindex="-1"></a><span class="ex">conda</span> create <span class="at">-n</span> labyrinth numpy scikit-learn scipy seaborn tqdm beautifulsoup4 selenium lxml r-tidyverse r-devtools r-rcppeigen r-rcppprogress r-fastmatch r-rpca r-future.apply r-pbapply r-dbi r-rsqlite r-bursts r-patchwork r-furrr r-datapreparation r-tokenizers r-reticulate r-knitr r-progressr r-future.callr r-hrbrthemes r-proc r-ggthemes r-meta r-ggally r-matrixtests r-corrplot r-statix bioconductor-tcgabiolinks bioconductor-clusterprofiler bioconductor-fgsea bioconductor-deseq2 bioconductor-m3c <span class="at">-c</span> bioconda <span class="at">-c</span> conda-forge</span>
 <span id="cb10-2"><a href="#cb10-2" tabindex="-1"></a><span class="ex">conda</span> activate labyrinth</span>
 <span id="cb10-3"><a href="#cb10-3" tabindex="-1"></a><span class="ex">Rscript</span> <span class="at">-e</span> <span class="st">"devtools::install_version('dbparser', version = '1.2.0')"</span></span>
 <span id="cb10-4"><a href="#cb10-4" tabindex="-1"></a><span class="ex">Rscript</span> <span class="at">-e</span> <span class="st">"remotes::install_github('randef1ned/word2vec')"</span></span></code></pre></div>
+</li>
+</ol>
 </div>
 <div class="section level3">
 <h3 id="r">R<a class="anchor" aria-label="anchor" href="#r"></a>
@@ -198,6 +202,11 @@ <h3 id="r">R<a class="anchor" aria-label="anchor" href="#r"></a>
 <span><span class="fu">devtools</span><span class="fu">::</span><span class="fu"><a href="https://remotes.r-lib.org/reference/install_version.html" class="external-link">install_version</a></span><span class="op">(</span><span class="st">'dbparser'</span>, version <span class="op">=</span> <span class="st">'1.2.0'</span><span class="op">)</span></span>
 <span><span class="fu">remotes</span><span class="fu">::</span><span class="fu"><a href="https://remotes.r-lib.org/reference/install_github.html" class="external-link">install_github</a></span><span class="op">(</span><span class="st">'randef1ned/word2vec'</span><span class="op">)</span></span>
 <span><span class="fu">remotes</span><span class="fu">::</span><span class="fu"><a href="https://remotes.r-lib.org/reference/install_github.html" class="external-link">install_github</a></span><span class="op">(</span><span class="st">'randef1ned/diffusr'</span><span class="op">)</span></span></code></pre></div>
+<p>This project uses a <a href="https://github.com/randef1ned/diffusr" class="external-link">custom fork</a> of the
+diffusr package, which is maintained by <span class="citation">@randef1ned</span>. This fork optimizes the
+computational execution of the package, providing improved performance
+compared to the original version. <a href="https://diffusr.yinchun.su/" class="external-link">View the documentation
+online</a>.</p>
 </div>
 </div>
 <div class="section level2">
@@ -210,16 +219,16 @@ <h2 id="folder-structure">Folder structure<a class="anchor" aria-label="anchor"
 <h3 id="preprocessing-steps">Preprocessing steps<a class="anchor" aria-label="anchor" href="#preprocessing-steps"></a>
 </h3>
 <ol style="list-style-type: decimal">
-<li>Parse the drug names:</li>
-</ol>
+<li>
+<p>Parse the drug names:</p>
 <ul>
 <li>
 <code>extract_names.Rmd</code>: This R Markdown file is used for
 parsing and extracting names from the data.</li>
 </ul>
-<ol start="2" style="list-style-type: decimal">
-<li>Download and read the data:</li>
-</ol>
+</li>
+<li>
+<p>Download and read the data:</p>
 <ul>
 <li>
 <code>parse_wos.py</code>: This Python script reads and parses the
@@ -237,9 +246,9 @@ <h3 id="preprocessing-steps">Preprocessing steps<a class="anchor" aria-label="an
 <code>clean_wos.py</code>: This Python script cleans and
 preprocesses the WoS data.</li>
 </ul>
-<ol start="3" style="list-style-type: decimal">
-<li>Identify stopwords and run word2vec:</li>
-</ol>
+</li>
+<li>
+<p>Identify stopwords and run word2vec:</p>
 <ul>
 <li>
 <code>identify_stopwords.Rmd</code>: This R Markdown file is used to
@@ -251,63 +260,107 @@ <h3 id="preprocessing-steps">Preprocessing steps<a class="anchor" aria-label="an
 <code>compute_distance.R</code>: This R script computes the
 distances between the word2vec embeddings for each drug.</li>
 </ul>
-<ol start="4" style="list-style-type: decimal">
-<li>Construct large knowledge network:</li>
-</ol>
+</li>
+<li>
+<p>Construct large knowledge network:</p>
 <ul>
 <li>
 <code>build_network.R</code>: This R script builds the large network
 by linking the clinical trials IDs and paper IDs.</li>
 </ul>
-<ol start="5" style="list-style-type: decimal">
-<li>Extract the stages of the drugs:</li>
-</ol>
+</li>
+<li>
+<p>Extract the stages of the drugs:</p>
 <ul>
 <li>
 <code>drug_status.R</code>: This R script extracts the stages of the
 drugs from the data.</li>
 </ul>
+</li>
+</ol>
 </div>
 <div class="section level3">
 <h3 id="construct-the-model-of-labyrinth-and-validate">Construct the model of <code>labyrinth</code> and validate<a class="anchor" aria-label="anchor" href="#construct-the-model-of-labyrinth-and-validate"></a>
 </h3>
 <ol style="list-style-type: decimal">
-<li>Construct the main knowledge network:</li>
-</ol>
+<li>
+<p>Construct the main knowledge network:</p>
 <ul>
 <li>
 <code>main_network.R</code>: This R script constructs the main
 network for the project.</li>
 </ul>
-<ol start="2" style="list-style-type: decimal">
-<li>Compute the details in the citation network:</li>
-</ol>
+</li>
+<li>
+<p>Compute the details in the citation network:</p>
 <ul>
 <li>
 <code>citation_metric.R</code>: This R script computes various
 metrics and details in the citation network.</li>
 </ul>
-<ol start="3" style="list-style-type: decimal">
-<li>Download MeSH structures for validation:</li>
-</ol>
+</li>
+<li>
+<p>Download MeSH structures for validation:</p>
 <ul>
 <li>
 <code>download_mesh_structure.py</code>: This Python script
 downloads the MeSH structure data for validation purposes.</li>
 </ul>
-<ol start="4" style="list-style-type: decimal">
-<li>Download TCGA and process the data for validation:</li>
-</ol>
+</li>
+<li>
+<p>Download TCGA and process the data for validation:</p>
 <ul>
 <li>
 <code>download_tcga.R</code>: This R script downloads and processes
 the TCGA (The Cancer Genome Atlas) data for validation.</li>
 </ul>
+</li>
+</ol>
 </div>
 </div>
 <div class="section level2">
 <h2 id="data-preparation">Data preparation<a class="anchor" aria-label="anchor" href="#data-preparation"></a>
 </h2>
+<table class="table">
+<colgroup>
+<col width="50%">
+<col width="50%">
+</colgroup>
+<thead><tr class="header">
+<th>Data</th>
+<th>Link</th>
+</tr></thead>
+<tbody>
+<tr class="odd">
+<td>ChEMBL</td>
+<td><a href="http://doi.org/10.6019/CHEMBL.database.32" class="external-link uri">http://doi.org/10.6019/CHEMBL.database.32</a></td>
+</tr>
+<tr class="even">
+<td>UniChEM</td>
+<td><a href="https://ftp.ebi.ac.uk/pub/databases/chembl/UniChem/data/table_dumps/" class="external-link uri">https://ftp.ebi.ac.uk/pub/databases/chembl/UniChem/data/table_dumps/</a></td>
+</tr>
+<tr class="odd">
+<td>DrugBank</td>
+<td><a href="https://go.drugbank.com/releases/5-1-10" class="external-link uri">https://go.drugbank.com/releases/5-1-10</a></td>
+</tr>
+<tr class="even">
+<td>Comparative Toxicogenomic Database</td>
+<td><a href="https://ctdbase.org/downloads/" class="external-link uri">https://ctdbase.org/downloads/</a></td>
+</tr>
+<tr class="odd">
+<td>Cochrane Library</td>
+<td><a href="https://www.cochranelibrary.com/central" class="external-link uri">https://www.cochranelibrary.com/central</a></td>
+</tr>
+<tr class="even">
+<td>GSE65185</td>
+<td><a href="https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE65185" class="external-link uri">https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE65185</a></td>
+</tr>
+<tr class="odd">
+<td>DisGENET</td>
+<td><a href="https://www.disgenet.org/download/sqlite/current/2020" class="external-link uri">https://www.disgenet.org/download/sqlite/current/2020</a></td>
+</tr>
+</tbody>
+</table>
 </div>
   </main><aside class="col-md-3"><nav id="toc"><h2>On this page</h2>
     </nav></aside>
@@ -316,7 +369,7 @@ <h2 id="data-preparation">Data preparation<a class="anchor" aria-label="anchor"
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/authors.html b/authors.html
index 46995a9..ad00db0 100644
--- a/authors.html
+++ b/authors.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -57,11 +57,11 @@
       <h2>Authors</h2>
       
       <ul class="list-unstyled"><li>
-          <p><strong>Yinchun Su</strong>. Author, maintainer. 
+          <p><strong>Junwei Han</strong>. Author, maintainer, contributor. 
           </p>
         </li>
         <li>
-          <p><strong>Junwei Han</strong>. Contributor. 
+          <p><strong>Yinchun Su</strong>. Author. 
           </p>
         </li>
       </ul></div>
@@ -70,15 +70,15 @@ <h2>Authors</h2>
       <h2 id="citation">Citation</h2>
       <p><small class="dont-index">Source: <a href="https://github.com/randef1ned/labyrinth/blob/HEAD/DESCRIPTION" class="external-link"><code>DESCRIPTION</code></a></small></p>
 
-      <p>Su Y (2024).
+      <p>Han J, Su Y (2024).
 <em>labyrinth: Labyrinth: A Simulation of Knowledge Networks</em>.
-R package version 0.2.3, <a href="https://github.com/randef1ned/labyrinth" class="external-link">https://github.com/randef1ned/labyrinth</a>. 
+R package version 0.3.0, <a href="https://github.com/randef1ned/labyrinth" class="external-link">https://github.com/randef1ned/labyrinth</a>. 
 </p>
       <pre>@Manual{,
   title = {labyrinth: Labyrinth: A Simulation of Knowledge Networks},
-  author = {Yinchun Su},
+  author = {Junwei Han and Yinchun Su},
   year = {2024},
-  note = {R package version 0.2.3},
+  note = {R package version 0.3.0},
   url = {https://github.com/randef1ned/labyrinth},
 }</pre>
     </div>
@@ -87,7 +87,7 @@ <h2 id="citation">Citation</h2>
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/index.html b/index.html
index 4dc3316..51efd9b 100644
--- a/index.html
+++ b/index.html
@@ -26,7 +26,7 @@
     
     <a class="navbar-brand me-2" href="index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -73,7 +73,7 @@
   <main id="main" class="col-md-9"><div class="section level1">
 <div class="page-header"><h1 id="labyrinth">Labyrinth<a class="anchor" aria-label="anchor" href="#labyrinth"></a>
 </h1></div>
-<p><a href="https://github.com/randef1ned/labyrinth" class="external-link"><img src="https://img.shields.io/badge/devel%20version-0.2.3-blue.svg" alt="Current devel version: 0.2.3"></a> <a href="https://www.repostatus.org/#wip" class="external-link"><img src="https://www.repostatus.org/badges/latest/wip.svg" alt="Project Status: WIP – Initial development is in progress, but there has not yet been a stable, usable release suitable for the public."></a> <a href="https://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html" class="external-link"><img src="https://img.shields.io/badge/License-GPL_v2-blue.svg" alt="License: GPL v2"></a> <a href="https://app.codacy.com/gh/randef1ned/labyrinth/dashboard?utm_source=gh&amp;utm_medium=referral&amp;utm_content=&amp;utm_campaign=Badge_grade" class="external-link"><img src="https://app.codacy.com/project/badge/Grade/09b138b2fa9242229f081cd180f6fc91" alt="Codacy Badge"></a> <a href="https://makeapullrequest.com" class="external-link"><img src="https://img.shields.io/badge/PRs-welcome-brightgreen.svg" alt="PRs Welcome"></a> <a href="https://github.com/randef1ned/labyrinth/commits/master" class="external-link"><img src="https://img.shields.io/github/last-commit/randef1ned/labyrinth.svg" alt="Last commit"></a> <a href="https://github.com/randef1ned/labyrinth" class="external-link"><img src="https://img.shields.io/github/languages/code-size/randef1ned/labyrinth.svg" alt="Code size"></a></p>
+<p><a href="https://github.com/randef1ned/labyrinth" class="external-link"><img src="https://img.shields.io/badge/devel%20version-0.3.0-blue.svg" alt="Current devel version: 0.3.0"></a> <a href="https://www.repostatus.org/#wip" class="external-link"><img src="https://www.repostatus.org/badges/latest/wip.svg" alt="Project Status: WIP – Initial development is in progress, but there has not yet been a stable, usable release suitable for the public."></a> <a href="https://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html" class="external-link"><img src="https://img.shields.io/badge/License-GPL_v2-blue.svg" alt="License: GPL v2"></a> <a href="https://app.codacy.com/gh/randef1ned/labyrinth/dashboard?utm_source=gh&amp;utm_medium=referral&amp;utm_content=&amp;utm_campaign=Badge_grade" class="external-link"><img src="https://app.codacy.com/project/badge/Grade/09b138b2fa9242229f081cd180f6fc91" alt="Codacy Badge"></a> <a href="https://makeapullrequest.com" class="external-link"><img src="https://img.shields.io/badge/PRs-welcome-brightgreen.svg" alt="PRs Welcome"></a> <a href="https://github.com/randef1ned/labyrinth/commits/master" class="external-link"><img src="https://img.shields.io/github/last-commit/randef1ned/labyrinth.svg" alt="Last commit"></a> <a href="https://github.com/randef1ned/labyrinth" class="external-link"><img src="https://img.shields.io/github/languages/code-size/randef1ned/labyrinth.svg" alt="Code size"></a></p>
 <p>Labyrinth is a human-inspired computational framework for personalized drug prioritization.</p>
 <div class="section level2">
 <h2 id="introduction">Introduction<a class="anchor" aria-label="anchor" href="#introduction"></a>
@@ -98,7 +98,7 @@ <h2 id="a-notice-on-operating-system-compatibility">A notice on operating system
 <p>We recommended these dependencies to be installed:</p>
 <ul>
 <li>
-<strong>R (≥ 4.3.0)</strong>: We developed this R package using R version 4.3.3.</li>
+<strong>R (≥ 4.3.0)</strong>: We developed this R package using R version 4.3.x.</li>
 <li>
 <strong>Python</strong>: Python is required for drawing plots in demos. It is recommended to have Python and <code>seaborn</code> installed, as the <code>reticulate</code> package will use the system’s Python installation.</li>
 <li>
@@ -106,7 +106,7 @@ <h2 id="a-notice-on-operating-system-compatibility">A notice on operating system
 <li>
 <strong>Intel oneAPI Math Kernel Library (oneMKL)</strong>: This library can further enhance mathematical performance, especially on Intel processors. oneMKL is not required but highly recommended.</li>
 </ul>
-<p>If you encounter any issues while running this package on other operating system, please open an <a href="https://github.com/randef1ned/labyrinth/issues" class="external-link">issue</a>.</p>
+<p>It would takes less than ten minutes to install this package. If you encounter any issues while running this package on other operating system, please open an issue.</p>
 </div>
 <div class="section level2">
 <h2 id="before-installation">Before installation<a class="anchor" aria-label="anchor" href="#before-installation"></a>
@@ -152,7 +152,7 @@ <h2 id="installation">Installation<a class="anchor" aria-label="anchor" href="#i
 <div class="section level2">
 <h2 id="usage">Usage<a class="anchor" aria-label="anchor" href="#usage"></a>
 </h2>
-<p>Load the package using <code><a href="https://github.com/randef1ned/labyrinth" class="external-link">library(labyrinth)</a></code>. We provide a vignette for the package that can be called using: <code><a href="articles/labyrinth.html">vignette("labyrinth")</a></code>. Alternatively, you can view the online version on <a href="https://labyrinth.yinchun.su/articles/labyrinth">website</a> or <a href="doc/labyrinth_knit.html">GitHub</a>. Basically that is all you have to know.</p>
+<p>Load the package using <code><a href="https://github.com/randef1ned/labyrinth" class="external-link">library(labyrinth)</a></code>. We provide a vignette for the package that can be called using: <code><a href="articles/labyrinth.html">vignette("labyrinth")</a></code>. Alternatively, you can view the online version on <a href="doc/labyrinth_knit.html">GitHub</a>, or <a href="https://labyrinth.yinchun.su/articles/labyrinth"><code>pkgdown</code> documentation</a>. The examples I provided would take several minutes to run on a normal desktop computer. Basically that is all you have to know.</p>
 <p><a href="doc/training_knit.html">This documentation</a> contains information about the contents and the necessary information for training the model used in this project. The <code>tools/</code> folder contains all the code and scripts required for constructing your own model, so that you can understand the technical details. Besides, you can refer to <a href="doc/preface_knit.html">this documentation</a> for the background and inspirations behind the overall workflow of `labyrinth.</p>
 </div>
 <div class="section level2">
@@ -199,8 +199,8 @@ <h2 data-toc-skip>Citation</h2>
 <div class="developers">
 <h2 data-toc-skip>Developers</h2>
 <ul class="list-unstyled">
-<li>Yinchun Su <br><small class="roles"> Author, maintainer </small>  </li>
-<li><a href="authors.html">More about authors...</a></li>
+<li>Junwei Han <br><small class="roles"> Author, maintainer, contributor </small>  </li>
+<li>Yinchun Su <br><small class="roles"> Author </small>  </li>
 </ul>
 </div>
 
@@ -211,7 +211,7 @@ <h2 data-toc-skip>Developers</h2>
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/news/index.html b/news/index.html
index 0b50cbb..f96e327 100644
--- a/news/index.html
+++ b/news/index.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -54,6 +54,10 @@
       <small>Source: <a href="https://github.com/randef1ned/labyrinth/blob/HEAD/NEWS.md" class="external-link"><code>NEWS.md</code></a></small>
     </div>
 
+    <div class="section level2">
+<h2 class="pkg-version" data-toc-text="0.3.0" id="labyrinth-v030">labyrinth v0.3.0<a class="anchor" aria-label="anchor" href="#labyrinth-v030"></a></h2>
+<ul><li>Updated docs</li>
+</ul></div>
     <div class="section level2">
 <h2 class="pkg-version" data-toc-text="0.2.3" id="labyrinth-v023">labyrinth v0.2.3<a class="anchor" aria-label="anchor" href="#labyrinth-v023"></a></h2>
 <ul><li>Added <code><a href="https://rdrr.io/r/stats/predict.html" class="external-link">predict()</a></code> function</li>
@@ -113,7 +117,7 @@ <h2 class="pkg-version" data-toc-text="0.0.1" id="labyrinth-v001">labyrinth v0.0
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/pkgdown.yml b/pkgdown.yml
index f139f08..874f380 100644
--- a/pkgdown.yml
+++ b/pkgdown.yml
@@ -5,7 +5,7 @@ articles:
   labyrinth: labyrinth.html
   preface: preface.html
   training: training.html
-last_built: 2024-04-26T12:44Z
+last_built: 2024-05-04T19:47Z
 urls:
   reference: https://labyrinth.yinchun.su/reference
   article: https://labyrinth.yinchun.su/articles
diff --git a/reference/activation_rate.html b/reference/activation_rate.html
index c59a924..43d6810 100644
--- a/reference/activation_rate.html
+++ b/reference/activation_rate.html
@@ -24,7 +24,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -181,7 +181,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/alias2SymbolUsingNCBI.html b/reference/alias2SymbolUsingNCBI.html
index 2ec3dc9..e99e91d 100644
--- a/reference/alias2SymbolUsingNCBI.html
+++ b/reference/alias2SymbolUsingNCBI.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -107,7 +107,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/assert_dgCMatrix.html b/reference/assert_dgCMatrix.html
index 877dac0..7808125 100644
--- a/reference/assert_dgCMatrix.html
+++ b/reference/assert_dgCMatrix.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -76,7 +76,7 @@ <h2 id="arguments">Arguments<a class="anchor" aria-label="anchor" href="#argumen
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/disease_ids.html b/reference/disease_ids.html
index e456085..d3bdbff 100644
--- a/reference/disease_ids.html
+++ b/reference/disease_ids.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -106,7 +106,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/disease_impact_score.html b/reference/disease_impact_score.html
index bea2fdf..21abd50 100644
--- a/reference/disease_impact_score.html
+++ b/reference/disease_impact_score.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -99,7 +99,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/drug_annot.html b/reference/drug_annot.html
index e605b80..f2a8d76 100644
--- a/reference/drug_annot.html
+++ b/reference/drug_annot.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -144,7 +144,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/gene_disease.html b/reference/gene_disease.html
index 3f0ef06..421610e 100644
--- a/reference/gene_disease.html
+++ b/reference/gene_disease.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -112,7 +112,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/get_neighbors.html b/reference/get_neighbors.html
index c9a5ada..9fac150 100644
--- a/reference/get_neighbors.html
+++ b/reference/get_neighbors.html
@@ -14,7 +14,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -149,7 +149,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/gradient.html b/reference/gradient.html
index 3401de3..c4bdc97 100644
--- a/reference/gradient.html
+++ b/reference/gradient.html
@@ -20,7 +20,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -133,7 +133,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/graph.html b/reference/graph.html
index 64fe57a..91d5bb4 100644
--- a/reference/graph.html
+++ b/reference/graph.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -115,7 +115,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/index.html b/reference/index.html
index e6abf18..85584da 100644
--- a/reference/index.html
+++ b/reference/index.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -146,11 +146,6 @@ <h2 id="all-functions">All functions<a class="anchor" aria-label="anchor" href="
         <dd>MeSH synonyms</dd>
       </dl><dl><dt>
           
-          <code><a href="model.html">model</a></code> 
-        </dt>
-        <dd>Pre-trained model of labyrinth</dd>
-      </dl><dl><dt>
-          
           <code><a href="ncbi_info.html">ncbi_info</a></code> 
         </dt>
         <dd>NCBI Gene Information</dd>
@@ -204,7 +199,7 @@ <h2 id="all-functions">All functions<a class="anchor" aria-label="anchor" href="
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/is.dgCMatrix.html b/reference/is.dgCMatrix.html
index 409c721..3803613 100644
--- a/reference/is.dgCMatrix.html
+++ b/reference/is.dgCMatrix.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -82,7 +82,7 @@ <h2 id="value">Value<a class="anchor" aria-label="anchor" href="#value"></a></h2
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/labyrinth-package.html b/reference/labyrinth-package.html
index d93459f..d32781c 100644
--- a/reference/labyrinth-package.html
+++ b/reference/labyrinth-package.html
@@ -14,7 +14,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -85,7 +85,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/load_data.html b/reference/load_data.html
index dbf4e7f..eef862a 100644
--- a/reference/load_data.html
+++ b/reference/load_data.html
@@ -18,7 +18,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -115,7 +115,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/mesh_annot.html b/reference/mesh_annot.html
index 6ad3200..04eb9b4 100644
--- a/reference/mesh_annot.html
+++ b/reference/mesh_annot.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -116,7 +116,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/mesh_hierarchy.html b/reference/mesh_hierarchy.html
index 8b5a41d..6db0c32 100644
--- a/reference/mesh_hierarchy.html
+++ b/reference/mesh_hierarchy.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -123,7 +123,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/mesh_synonyms.html b/reference/mesh_synonyms.html
index 1aa146c..6d227dc 100644
--- a/reference/mesh_synonyms.html
+++ b/reference/mesh_synonyms.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -118,7 +118,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/ncbi_info.html b/reference/ncbi_info.html
index 360dcb2..265d541 100644
--- a/reference/ncbi_info.html
+++ b/reference/ncbi_info.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -109,7 +109,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/ppi.html b/reference/ppi.html
index 228dc92..f24e204 100644
--- a/reference/ppi.html
+++ b/reference/ppi.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -120,7 +120,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/predict_drug.html b/reference/predict_drug.html
index 4466271..9803eab 100644
--- a/reference/predict_drug.html
+++ b/reference/predict_drug.html
@@ -18,7 +18,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -162,10 +162,6 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 <span class="r-in"><span><span class="co"># \donttest{</span></span></span>
 <span class="r-in"><span><span class="co"># Load models to the environment</span></span></span>
 <span class="r-in"><span><span class="va">model</span> <span class="op">&lt;-</span> <span class="fu"><a href="load_data.html">load_data</a></span><span class="op">(</span><span class="st">"model"</span><span class="op">)</span></span></span>
-<span class="r-wrn co"><span class="r-pr">#&gt;</span> <span class="warning">Warning: </span>No cached model file. Downloading...</span>
-<span class="r-wrn co"><span class="r-pr">#&gt;</span> <span class="warning">Warning: </span>downloaded length 0 != reported length 14</span>
-<span class="r-wrn co"><span class="r-pr">#&gt;</span> <span class="warning">Warning: </span>cannot open URL 'https://raw.githubusercontent.com/randef1ned/labyrinth/master/extdata/model.rda': HTTP status was '404 Not Found'</span>
-<span class="r-err co"><span class="r-pr">#&gt;</span> <span class="error">Error in download.file(paste0("https://raw.githubusercontent.com/", "randef1ned/labyrinth/master/extdata/",     required_file), model_path):</span> cannot open URL 'https://raw.githubusercontent.com/randef1ned/labyrinth/master/extdata/model.rda'</span>
 <span class="r-in"><span></span></span>
 <span class="r-in"><span><span class="co"># Construct disease weights</span></span></span>
 <span class="r-in"><span><span class="va">disease_weights</span> <span class="op">&lt;-</span> <span class="fu"><a href="https://rdrr.io/r/base/sample.html" class="external-link">sample</a></span><span class="op">(</span><span class="fu"><a href="https://rdrr.io/r/base/c.html" class="external-link">c</a></span><span class="op">(</span><span class="fu"><a href="https://rdrr.io/r/base/rep.html" class="external-link">rep</a></span><span class="op">(</span><span class="fl">0</span>, <span class="fl">50</span><span class="op">)</span>, <span class="fu"><a href="https://rdrr.io/r/base/rep.html" class="external-link">rep</a></span><span class="op">(</span><span class="fl">1</span>, <span class="fl">2</span><span class="op">)</span><span class="op">)</span>, <span class="fl">1098</span>, replace <span class="op">=</span> <span class="cn">TRUE</span><span class="op">)</span></span></span>
@@ -188,7 +184,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/random_walk.html b/reference/random_walk.html
index d6bc696..00059f7 100644
--- a/reference/random_walk.html
+++ b/reference/random_walk.html
@@ -18,7 +18,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -183,7 +183,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/sigmoid.html b/reference/sigmoid.html
index 0845a08..6e25687 100644
--- a/reference/sigmoid.html
+++ b/reference/sigmoid.html
@@ -18,7 +18,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -111,7 +111,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/speech.html b/reference/speech.html
index 827aede..897056a 100644
--- a/reference/speech.html
+++ b/reference/speech.html
@@ -14,7 +14,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -111,7 +111,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/spread_gram.html b/reference/spread_gram.html
index 9c7d517..425719c 100644
--- a/reference/spread_gram.html
+++ b/reference/spread_gram.html
@@ -18,7 +18,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -148,7 +148,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/spread_gram_1.html b/reference/spread_gram_1.html
index 2dca5cb..b9b4485 100644
--- a/reference/spread_gram_1.html
+++ b/reference/spread_gram_1.html
@@ -18,7 +18,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -123,7 +123,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/transfer_activation.html b/reference/transfer_activation.html
index c83e665..4fd1c32 100644
--- a/reference/transfer_activation.html
+++ b/reference/transfer_activation.html
@@ -16,7 +16,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -126,7 +126,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/reference/update_gene_symbol.html b/reference/update_gene_symbol.html
index 532ff14..82a03c5 100644
--- a/reference/update_gene_symbol.html
+++ b/reference/update_gene_symbol.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">labyrinth</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.2.3</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.3.0</small>
 
     
     <button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbar" aria-controls="navbar" aria-expanded="false" aria-label="Toggle navigation">
@@ -113,7 +113,7 @@ <h2 id="ref-examples">Examples<a class="anchor" aria-label="anchor" href="#ref-e
 
 
     <footer><div class="pkgdown-footer-left">
-  <p>Developed by Yinchun Su.</p>
+  <p>Developed by Junwei Han, Yinchun Su.</p>
 </div>
 
 <div class="pkgdown-footer-right">
diff --git a/search.json b/search.json
index d7d019f..00f3697 100644
--- a/search.json
+++ b/search.json
@@ -1 +1 @@
-[{"path":[]},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"our-pledge","dir":"","previous_headings":"","what":"Our Pledge","title":"Contributor Covenant Code of Conduct","text":"members, contributors, leaders pledge make participation community harassment-free experience everyone, regardless age, body size, visible invisible disability, ethnicity, sex characteristics, gender identity expression, level experience, education, socio-economic status, nationality, personal appearance, race, caste, color, religion, sexual identity orientation. pledge act interact ways contribute open, welcoming, diverse, inclusive, healthy community.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"our-standards","dir":"","previous_headings":"","what":"Our Standards","title":"Contributor Covenant Code of Conduct","text":"Examples behavior contributes positive environment community include: Demonstrating empathy kindness toward people respectful differing opinions, viewpoints, experiences Giving gracefully accepting constructive feedback Accepting responsibility apologizing affected mistakes, learning experience Focusing best just us individuals, overall community Examples unacceptable behavior include: use sexualized language imagery, sexual attention advances kind Trolling, insulting derogatory comments, personal political attacks Public private harassment Publishing others’ private information, physical email address, without explicit permission conduct reasonably considered inappropriate professional setting","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"enforcement-responsibilities","dir":"","previous_headings":"","what":"Enforcement Responsibilities","title":"Contributor Covenant Code of Conduct","text":"Community leaders responsible clarifying enforcing standards acceptable behavior take appropriate fair corrective action response behavior deem inappropriate, threatening, offensive, harmful. Community leaders right responsibility remove, edit, reject comments, commits, code, wiki edits, issues, contributions aligned Code Conduct, communicate reasons moderation decisions appropriate.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"scope","dir":"","previous_headings":"","what":"Scope","title":"Contributor Covenant Code of Conduct","text":"Code Conduct applies within community spaces, also applies individual officially representing community public spaces. Examples representing community include using official email address, posting via official social media account, acting appointed representative online offline event.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"enforcement","dir":"","previous_headings":"","what":"Enforcement","title":"Contributor Covenant Code of Conduct","text":"Instances abusive, harassing, otherwise unacceptable behavior may reported community leaders responsible enforcement [INSERT CONTACT METHOD]. complaints reviewed investigated promptly fairly. community leaders obligated respect privacy security reporter incident.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"enforcement-guidelines","dir":"","previous_headings":"","what":"Enforcement Guidelines","title":"Contributor Covenant Code of Conduct","text":"Community leaders follow Community Impact Guidelines determining consequences action deem violation Code Conduct:","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_1-correction","dir":"","previous_headings":"Enforcement Guidelines","what":"1. Correction","title":"Contributor Covenant Code of Conduct","text":"Community Impact: Use inappropriate language behavior deemed unprofessional unwelcome community. Consequence: private, written warning community leaders, providing clarity around nature violation explanation behavior inappropriate. public apology may requested.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_2-warning","dir":"","previous_headings":"Enforcement Guidelines","what":"2. Warning","title":"Contributor Covenant Code of Conduct","text":"Community Impact: violation single incident series actions. Consequence: warning consequences continued behavior. interaction people involved, including unsolicited interaction enforcing Code Conduct, specified period time. includes avoiding interactions community spaces well external channels like social media. Violating terms may lead temporary permanent ban.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_3-temporary-ban","dir":"","previous_headings":"Enforcement Guidelines","what":"3. Temporary Ban","title":"Contributor Covenant Code of Conduct","text":"Community Impact: serious violation community standards, including sustained inappropriate behavior. Consequence: temporary ban sort interaction public communication community specified period time. public private interaction people involved, including unsolicited interaction enforcing Code Conduct, allowed period. Violating terms may lead permanent ban.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_4-permanent-ban","dir":"","previous_headings":"Enforcement Guidelines","what":"4. Permanent Ban","title":"Contributor Covenant Code of Conduct","text":"Community Impact: Demonstrating pattern violation community standards, including sustained inappropriate behavior, harassment individual, aggression toward disparagement classes individuals. Consequence: permanent ban sort public interaction within community.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"attribution","dir":"","previous_headings":"","what":"Attribution","title":"Contributor Covenant Code of Conduct","text":"Code Conduct adapted Contributor Covenant, version 2.1, available https://www.contributor-covenant.org/version/2/1/code_of_conduct.html. Community Impact Guidelines inspired Mozilla’s code conduct enforcement ladder. answers common questions code conduct, see FAQ https://www.contributor-covenant.org/faq. Translations available https://www.contributor-covenant.org/translations.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":null,"dir":"","previous_headings":"","what":"GNU General Public License","title":"GNU General Public License","text":"Version 2, June 1991Copyright © 1989, 1991 Free Software Foundation, Inc.,51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA Everyone permitted copy distribute verbatim copies license document, changing allowed.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"preamble","dir":"","previous_headings":"","what":"Preamble","title":"GNU General Public License","text":"licenses software designed take away freedom share change . contrast, GNU General Public License intended guarantee freedom share change free software–make sure software free users. General Public License applies Free Software Foundation’s software program whose authors commit using . (Free Software Foundation software covered GNU Lesser General Public License instead.) can apply programs, . speak free software, referring freedom, price. General Public Licenses designed make sure freedom distribute copies free software (charge service wish), receive source code can get want , can change software use pieces new free programs; know can things. protect rights, need make restrictions forbid anyone deny rights ask surrender rights. restrictions translate certain responsibilities distribute copies software, modify . example, distribute copies program, whether gratis fee, must give recipients rights . must make sure , , receive can get source code. must show terms know rights. protect rights two steps: (1) copyright software, (2) offer license gives legal permission copy, distribute /modify software. Also, author’s protection , want make certain everyone understands warranty free software. software modified someone else passed , want recipients know original, problems introduced others reflect original authors’ reputations. Finally, free program threatened constantly software patents. wish avoid danger redistributors free program individually obtain patent licenses, effect making program proprietary. prevent , made clear patent must licensed everyone’s free use licensed . precise terms conditions copying, distribution modification follow.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"terms-and-conditions-for-copying-distribution-and-modification","dir":"","previous_headings":"","what":"TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION","title":"GNU General Public License","text":"0. License applies program work contains notice placed copyright holder saying may distributed terms General Public License. “Program”, , refers program work, “work based Program” means either Program derivative work copyright law: say, work containing Program portion , either verbatim modifications /translated another language. (Hereinafter, translation included without limitation term “modification”.) licensee addressed “”. Activities copying, distribution modification covered License; outside scope. act running Program restricted, output Program covered contents constitute work based Program (independent made running Program). Whether true depends Program . 1. may copy distribute verbatim copies Program’s source code receive , medium, provided conspicuously appropriately publish copy appropriate copyright notice disclaimer warranty; keep intact notices refer License absence warranty; give recipients Program copy License along Program. may charge fee physical act transferring copy, may option offer warranty protection exchange fee. 2. may modify copy copies Program portion , thus forming work based Program, copy distribute modifications work terms Section 1 , provided also meet conditions: ) must cause modified files carry prominent notices stating changed files date change. b) must cause work distribute publish, whole part contains derived Program part thereof, licensed whole charge third parties terms License. c) modified program normally reads commands interactively run, must cause , started running interactive use ordinary way, print display announcement including appropriate copyright notice notice warranty (else, saying provide warranty) users may redistribute program conditions, telling user view copy License. (Exception: Program interactive normally print announcement, work based Program required print announcement.) requirements apply modified work whole. identifiable sections work derived Program, can reasonably considered independent separate works , License, terms, apply sections distribute separate works. distribute sections part whole work based Program, distribution whole must terms License, whose permissions licensees extend entire whole, thus every part regardless wrote . Thus, intent section claim rights contest rights work written entirely ; rather, intent exercise right control distribution derivative collective works based Program. addition, mere aggregation another work based Program Program (work based Program) volume storage distribution medium bring work scope License. 3. may copy distribute Program (work based , Section 2) object code executable form terms Sections 1 2 provided also one following: ) Accompany complete corresponding machine-readable source code, must distributed terms Sections 1 2 medium customarily used software interchange; , b) Accompany written offer, valid least three years, give third party, charge cost physically performing source distribution, complete machine-readable copy corresponding source code, distributed terms Sections 1 2 medium customarily used software interchange; , c) Accompany information received offer distribute corresponding source code. (alternative allowed noncommercial distribution received program object code executable form offer, accord Subsection b .) source code work means preferred form work making modifications . executable work, complete source code means source code modules contains, plus associated interface definition files, plus scripts used control compilation installation executable. However, special exception, source code distributed need include anything normally distributed (either source binary form) major components (compiler, kernel, ) operating system executable runs, unless component accompanies executable. distribution executable object code made offering access copy designated place, offering equivalent access copy source code place counts distribution source code, even though third parties compelled copy source along object code. 4. may copy, modify, sublicense, distribute Program except expressly provided License. attempt otherwise copy, modify, sublicense distribute Program void, automatically terminate rights License. However, parties received copies, rights, License licenses terminated long parties remain full compliance. 5. required accept License, since signed . However, nothing else grants permission modify distribute Program derivative works. actions prohibited law accept License. Therefore, modifying distributing Program (work based Program), indicate acceptance License , terms conditions copying, distributing modifying Program works based . 6. time redistribute Program (work based Program), recipient automatically receives license original licensor copy, distribute modify Program subject terms conditions. may impose restrictions recipients’ exercise rights granted herein. responsible enforcing compliance third parties License. 7. , consequence court judgment allegation patent infringement reason (limited patent issues), conditions imposed (whether court order, agreement otherwise) contradict conditions License, excuse conditions License. distribute satisfy simultaneously obligations License pertinent obligations, consequence may distribute Program . example, patent license permit royalty-free redistribution Program receive copies directly indirectly , way satisfy License refrain entirely distribution Program. portion section held invalid unenforceable particular circumstance, balance section intended apply section whole intended apply circumstances. purpose section induce infringe patents property right claims contest validity claims; section sole purpose protecting integrity free software distribution system, implemented public license practices. Many people made generous contributions wide range software distributed system reliance consistent application system; author/donor decide willing distribute software system licensee impose choice. section intended make thoroughly clear believed consequence rest License. 8. distribution /use Program restricted certain countries either patents copyrighted interfaces, original copyright holder places Program License may add explicit geographical distribution limitation excluding countries, distribution permitted among countries thus excluded. case, License incorporates limitation written body License. 9. Free Software Foundation may publish revised /new versions General Public License time time. new versions similar spirit present version, may differ detail address new problems concerns. version given distinguishing version number. Program specifies version number License applies “later version”, option following terms conditions either version later version published Free Software Foundation. Program specify version number License, may choose version ever published Free Software Foundation. 10. wish incorporate parts Program free programs whose distribution conditions different, write author ask permission. software copyrighted Free Software Foundation, write Free Software Foundation; sometimes make exceptions . decision guided two goals preserving free status derivatives free software promoting sharing reuse software generally.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"no-warranty","dir":"","previous_headings":"","what":"NO WARRANTY","title":"GNU General Public License","text":"11. PROGRAM LICENSED FREE CHARGE, WARRANTY PROGRAM, EXTENT PERMITTED APPLICABLE LAW. EXCEPT OTHERWISE STATED WRITING COPYRIGHT HOLDERS /PARTIES PROVIDE PROGRAM “” WITHOUT WARRANTY KIND, EITHER EXPRESSED IMPLIED, INCLUDING, LIMITED , IMPLIED WARRANTIES MERCHANTABILITY FITNESS PARTICULAR PURPOSE. ENTIRE RISK QUALITY PERFORMANCE PROGRAM . PROGRAM PROVE DEFECTIVE, ASSUME COST NECESSARY SERVICING, REPAIR CORRECTION. 12. EVENT UNLESS REQUIRED APPLICABLE LAW AGREED WRITING COPYRIGHT HOLDER, PARTY MAY MODIFY /REDISTRIBUTE PROGRAM PERMITTED , LIABLE DAMAGES, INCLUDING GENERAL, SPECIAL, INCIDENTAL CONSEQUENTIAL DAMAGES ARISING USE INABILITY USE PROGRAM (INCLUDING LIMITED LOSS DATA DATA RENDERED INACCURATE LOSSES SUSTAINED THIRD PARTIES FAILURE PROGRAM OPERATE PROGRAMS), EVEN HOLDER PARTY ADVISED POSSIBILITY DAMAGES. END TERMS CONDITIONS","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"how-to-apply-these-terms-to-your-new-programs","dir":"","previous_headings":"","what":"How to Apply These Terms to Your New Programs","title":"GNU General Public License","text":"develop new program, want greatest possible use public, best way achieve make free software everyone can redistribute change terms. , attach following notices program. safest attach start source file effectively convey exclusion warranty; file least “copyright” line pointer full notice found. Also add information contact electronic paper mail. program interactive, make output short notice like starts interactive mode: hypothetical commands show w show c show appropriate parts General Public License. course, commands use may called something show w show c; even mouse-clicks menu items–whatever suits program. also get employer (work programmer) school, , sign “copyright disclaimer” program, necessary. sample; alter names: General Public License permit incorporating program proprietary programs. program subroutine library, may consider useful permit linking proprietary applications library. want , use GNU Lesser General Public License instead License.","code":"<one line to give the program's name and a brief idea of what it does.> Copyright (C) <year>  <name of author>  This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.  This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more details.  You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. Gnomovision version 69, Copyright (C) year name of author Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. This is free software, and you are welcome to redistribute it under certain conditions; type `show c' for details. Yoyodyne, Inc., hereby disclaims all copyright interest in the program `Gnomovision' (which makes passes at compilers) written by James Hacker.  <signature of Ty Coon>, 1 April 1989 Ty Coon, President of Vice"},{"path":"https://labyrinth.yinchun.su/articles/labyrinth.html","id":"introduction","dir":"Articles","previous_headings":"","what":"Introduction","title":"The introduction to labyrinth","text":"ab","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"from-human-memory-to-activation-diffusion-network","dir":"Articles","previous_headings":"","what":"1. From human memory to activation diffusion network","title":"Introduction and inspirations of labyrinth","text":"Memory fundamental cognitive process allows brain store, acquire, recall information. serves temporary storage system sensory cues disappear (Benjamin 2007). Memory plays crucial role encoding, storing, retaining, recalling everything simple sensory data complex knowledge experiences. Additionally, memory basis learning, planning, decision-making (Benjamin 2007; Nussenbaum, Prentis, Hartley 2020). Specifically, enables us learn past experiences simulate potential future outcomes, thereby influencing current behavior future actions (Schacter et al. 2012). formation memories, recall, reasoning based involve combination systems physiological processes allow humans adapt well environment (Schacter et al. 2012; Camina Güell 2017; Nairne Pandeirada 2016).Memory formation comprises three stages: information perception, encoding, storage (Atkinson Shiffrin 1968). stages correspond three types memory: (1) sensory memory (Atkinson Shiffrin 1968) briefly stores raw physical stimuli primary receptors vision hearing; (2) short-term memory (STM) (Baddeley 2000) involves transient storage manipulation information, allowing individuals temporarily memorize small amounts data perform current future tasks; (3) long-term memory (LTM) (Camina Güell 2017) long-term storage information, divided episodic implicit memory. Episodic memory consists knowledge processed recalled conscious level, personal experiences specialized knowledge, implicit memory encompasses skills habits expressed without conscious awareness, fear, riding bicycle, heart rate regulation, conditioned reflexes (Smith Grossman 2008). contrast limited-capacity sensory memory STM, LTM complex cognitive system unlimited capacity long-term storage retrieval wide range information, including factual knowledge personal experiences. Memory can also categorized contextual memory, referring individual’s personal experiences, semantic memory, referring textual knowledge concepts (Renoult et al. 2019). Storing contextual semantic knowledge allows individuals construct new knowledge based past experiences, facilitating survival (Kazanas Altarriba 2015). addition storing information memories, LTM plays important role learning reasoning. can automatically relate relationships attributes objects (Nairne Pandeirada 2016), giving individuals ability use stored skills concepts make rational decisions evaluating different choices various environments predicting possible outcomes (Camina Güell 2017). formation consolidation LTM involve several brain regions, including prefrontal lobe, associated working memory situational memory LTM (Blumenfeld Ranganath 2019), temporal lobe, associated semantic memory LTM (Simmons Martin 2009). hippocampus acts relay station information (Squire Zola-Morgan 1991) can integrate situational memory semantic memory knowledge network stored LTM (Renoult et al. 2019). Consequently, even concept knowledge, knowledge network formed different individuals can vary. individual unique experiences backgrounds, leading different understandings reactions interpreting information. LTM stored vast complex semantic network, includes various types interconnected nodes, concepts, memories, experiences (Collins Loftus 1975). kinds memories experiences also integrated network; example, individual’s representation abstract concepts (e.g., time) may based physical sensations (Casasanto Boroditsky 2008). cases, individuals associate time experiences, forming knowledge network. form organization named semantic networks knowledge networks, emphasizing information interconnected organized according meaning (Lehmann 1992). article, use term semantic network represent form memory storage organization LTM, knowledge network refers artificially built knowledge network. semantic network, concepts represented nodes, concept--concept relationships represented edges nodes, edge weights indicating strength association. higher edge weight implies closer relationship two nodes, typically resulting higher recall rate receiving stimulus (Anderson 1983). Learning memorizing new knowledge experiences involve building new edges reinforcing existing ones. organization facilitates information retrieval enabling individuals jump one concept another network simultaneously activate neighboring nodes form knowledge connections, even direct correlation (Lehmann 1992). interesting example semantic network police officers, black people produce strong association weapons, association even stronger police officer sleep-deprived (James 2018). Another example individual’s preference Coca-Cola McDonald’s determined attitude reflected semantic network (Lee Kim 2013; Karpinski, Steinman, Hilton 2005). Homogeneous networks consist nodes similar properties characteristics (Mhatre Rosenberg 2004). Nodes represented kind elements edges connected nodes high correlations, jointly creating homogeneous network. two examples mentioned illustrate different individuals form different LTMs, memory contents stored LTMs satisfy homogeneity. Moving one concept unrelated concept impossible homogeneous network. process individuals store memories semantic network retrieve information LTM often described spreading activation, network also called spreading activation network (Sharifian Samani 1997). network model, initial node activated, activation state spreads nodes along connected edges. diffusion process can quickly span multiple network layers, extensively activating concepts memories associated initial node. node receives activation certain threshold, fully activated like neurons. Otherwise, activated. may lead recall specific memories, formation decisions, generation problem-solving strategies. mentioned earlier, unrelated concepts semantic network may relatively strong associations. implicit association test (IAT) paradigm proposed Greenwald can effectively test edge connections nodes individual semantic network (Greenwald, McGhee, Schwartz 1998; Greenwald et al. 2009). paradigm tests strength association human brain two nodes, .e., edge weights. mechanisms association activation activation diffusion networks depend strength association nodes. strength high, probability activation relatively high; strength low, higher probability non-activation. theory partly explains forgetting phenomenon occurs human memory. Additionally, activation diffusion networks enable individuals retrieve necessary information, reorganize memories, apply knowledge different situations. summary, activation diffusion networks effectively account dynamic nature memory retrieval use.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"unique-advantages-of-human-cognitive-abilities","dir":"Articles","previous_headings":"","what":"2. Unique advantages of human cognitive abilities","title":"Introduction and inspirations of labyrinth","text":"Compared computer programs, humans possess ability think problems different perspectives exhibit greater flexibility knowledge association (Lehmann 1992). Therefore, humans advantage applying knowledge one domain another seemingly unrelated domain. example, concepts biology can transferred economics (Lawlor et al. 2008), economic models field electronic information (Han et al. 2019), linguistic concepts neuroscience (Mayberry et al. 2018) computer science (H. Zhang et al. 2023). characteristic led humans create many cross-disciplinary fields, artificial intelligence, computational biology, neuropolitics, bioinformatics. Humans can use intuition creative thinking solve problems, ability think across domains allows make new connections different areas, thereby building new disciplinary knowledge. human brain contains approximately 100 billion neurons roughly number glial cells (Bartheld, Bahney, Herculano-Houzel 2016), connected thousands others via synapses (Herculano-Houzel 2009). Neurons glial cells form extremely complex networks. Neurons communicate via --none principle (Pareti 2007), glial cells play crucial roles central nervous system formation, neuronal differentiation, synapse formation (Allen Lyons 2018), regulation neuroinflammatory immunity (Yang Zhou 2019), neurological diseases like dementia (Kim, Choi, Yoon 2020), addition traditional supportive functions (Wolosker et al. 2008). complexity lays foundation individual’s ability process information, experience emotions, maintain awareness, exhibit creativity. Drawing fundamentals human cognition, artificial neural networks simulated using computers mimic brain’s information processing. emulate human cognitive abilities extent, excelling tasks like learning, decision-making, pattern recognition humans naturally proficient (Agatonovic-Kustrin Beresford 2000; Parisi 1997). simulation human cognitive abilities shown great potential (Parisi 1997; Zahedi 1991). However, neurons used deep learning artificial neural networks highly abstract, architecture unable account neurons (Cichy Kaiser 2019). Therefore, field focused attention fitting data rather interpreting (Pichler Hartig 2023). Currently, field deep learning concerned fitting data, effect fitting used guiding criterion field, rather integrating cognitive mechanisms discovered neuroscience (Chavlis Poirazi 2021). Much progress deep learning recent decades can attributed application backpropagation, often used optimization methods update weights minimize loss function. However, neural networks deep learning biologically inspired approaches, biological rationality backpropagation remains questionable, activated neurons acquire features backpropagation (Whittington Bogacz 2019; Lillicrap et al. 2020; Aru, Suzuki, Larkum 2020). Currently, two mainstream learning mechanisms identified human brain using electrophysiological methods: Hebbian learning (Munakata Pfaffly 2004) reinforcement learning (Botvinick et al. 2019). Additionally, synaptic pruning may related learning (Halassa Haydon 2010), epigenetic mechanisms also play important role (Changeux, Courrège, Danchin 1973). Although Hebbian learning, reinforcement learning, attempts migrate human cognitive mechanisms applied deep learning years, still perfectly reproduce human learning features (Volzhenin, Changeux, Dumas 2022). Comparatively, energy consumption using neural networks reasoning huge (Desislavov, Martínez-Plumed, Hernández-Orallo 2023), contrast human brain’s lower energy usage training reasoning (Attwell Laughlin 2001). Another example attention mechanism neural networks, inspired human attention (Vaswani et al. 2023). Attention cognitive ability selectively receives information limited resources (Nelson Cowan et al. 2005). ’s complex biological process involving multiple brain regions, encompassing selective attention also coordinated consciousness, memory, cognition. Selective attention mechanisms associated short-term memory, 3-5 chunks original stimuli can enter single session (N. Cowan 2001), attention lasting just seconds minutes (Polti, Martin, Van Wassenhove 2018). selective mechanism allows humans focus targets limited resources, reducing cognitive resource consumption (Buschman Kastner 2015), refining elements deposited memory (Chun Turk-Browne 2007), delimiting problem space, narrowing memory retrieval problem-solving situations (Wiley Jarosz 2012). Human intuition numbers may also relate attention (Kutter et al. 2023). Thus, selective attention crucial cognitive activities like perception, memory, decision-making. Attention mechanisms deep learning, inspired human selective attention, successfully integrated various frameworks (Niu, Zhong, Yu 2021), greatly improving performance tasks like natural language processing (NLP), computer vision, speech recognition (B. Zhang, Xiong, Su 2020; Guo et al. 2022; Ding et al. 2021). recent years, Transformer model, relying self-attention mechanisms process data, demonstrated superior performance across various tasks (Vaswani et al. 2023; Khan et al. 2022). multi-head attention mechanism performs multiple parallel self-attention computations different parameters, allowing model capture information different subspaces improving fitting efficiency accuracy (Liu, Liu, Han 2021). Practically, Transformer, neural networks made significant progress areas like NLP vision tasks. Attention mechanisms deep learning implemented mathematical functions assign weights different elements input data (Niu, Zhong, Yu 2021; De Santana Correia Colombini 2022). However, subset studies found attention mechanism deep learning fully simulate human attention lacks cognitive emotional context human attention encompasses (Lai et al. 2021). Despite differences, artificial neural networks successfully applied several fields, including image speech recognition, natural language processing, robot control, gaming, decision support systems. applications demonstrate power artificial neural networks dealing complex problems simulating certain human cognitive processes highlighting unique advantages models simulate human cognitive abilities.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"simulating-human-cognitive-abilities-the-way-forward","dir":"Articles","previous_headings":"","what":"3. Simulating Human Cognitive Abilities: The Way Forward","title":"Introduction and inspirations of labyrinth","text":"recent years, Transformer model excelled various tasks rely self-attentive mechanisms data processing (Vaswani et al. 2023; Khan et al. 2022). departs traditional recurrent neural networks (RNNs) convolutional neural networks (CNNs), favoring comprehensive utilization attentional mechanisms process sequential data. Transformer’s attention model primarily applied self-attention multi-head attention mechanisms. self-attention mechanism considers elements sequence processing input element, enabling model capture long-range dependencies within sequence (Vig Belinkov 2019). element transformed query (q), key (k), value (v) vectors, representing current lexical element, lexical elements, information contained lexical element, respectively. attention score computed calculating similarity scores q k, weighted summing v. Recently, LLMs employed Transformer’s framework, demonstrating improved simulation human cognition. LLMs large-scale simulations human cognitive functions (Binz Schulz 2023), emergence mark significant advancement computers’ ability simulate human cognition. LLMs possess enhanced reasoning capabilities, Claude 3, released month Anthropic, exhibits self-awareness contextual understanding needle---haystack task (Anthropic 2024; Kuratov et al. 2024). zero-shot problem scenarios, LLMs’ reasoning abilities without prior knowledge surpass humans, rely analogies reasoning (Webb, Holyoak, Lu 2023). Furthermore, LLMs can comprehend others’ beliefs, goals, mental states accuracy 80%. Notably, GPT-4, considered advanced LLM, can achieve 100% accuracy theory mind (ToM) tasks suitable prompting, indicating human-like level ToM (Thaler 1988). LLMs can also simulate human behavior observed experiments, ultimatum game (Thaler 1988), garden-path sentences (Ferreira, Christianson, Hollingworth 2001), loss aversion (Kimball 1993), reactions Milgram electric shock experiment (Blass 1999; Aher, Arriaga, Kalai 2022). Additionally, LLMs exhibit cognitive biases errors humans typically demonstrate, additive bias (Winter et al. 2023), individuals default adding modifying existing content rather deleting pushing back problem-solving (Adams et al. 2021). LLMs produce various human cognitive effects, including priming effects biases (Koo et al. 2023; Shaki, Kraus, Wooldridge 2023), suggesting LLMs mimicking human cognitive processes may possess cognitive abilities approaching human level. specific domains, LLMs closely mimic human-specific abilities. instance, ChatGPT’s accuracy medical diagnosis providing feasible medical advice complex situations comparable human physicians (Hopkins et al. 2023). performance metrics show diagnoses 93.3% common clinical cases correctly (Hirosawa et al. 2023). Furthermore, standardized clinical decision-making tasks, ChatGPT achieves accuracy rate close 70% (Rao et al. 2023), similar expected level third-year medical students United States (Gilson et al. 2023). Due GPT-4’s superior ToM, correctly answered 90% soft skill questions (Brin et al. 2023), demonstrating excellent clinical skills. However, ChatGPT’s ability handle complex questions remains unsatisfactory compared widely used technologies like Google search (Hopkins et al. 2023). fully replicate professional clinicians’ decision-making abilities faced complex problems, primarily due text-based training data, resulting less satisfactory performance non-text-based tasks (Y. Zhang et al. 2024). Furthermore, patient care medically related domains, sometimes generates false misleading information, potentially causing doctors, nurses, caregivers make erroneous decisions, endangering patients’ lives (Z. Ji et al. 2023). LLMs often contain billions hundreds billions parameters, making difficult implement debugging understand decision-making processes (Z. Ji et al. 2023; Khullar, Wang, Wang 2024). Therefore, developing relatively interpretable models viable alternative moment. models trained specific areas expertise, possessing prior knowledge learning exclusively samples. Recently, life2vec model successfully predicted relationship early mortality aspects individual’s personality traits, demonstrating relatively good predictive efficacy (Savcisens et al. 2023). model provides clinicians family physicians insights assistance can help patients better manage lifespan, showcasing potential specialized models.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"simulating-human-knowledge-representation-using-large-scale-medical-knowledge-networks","dir":"Articles","previous_headings":"","what":"4. Simulating human knowledge representation using large-scale medical knowledge networks","title":"Introduction and inspirations of labyrinth","text":"summary, found computer models simulating human cognitive abilities tend achieve good model fitting results, Transformer-based neural network models like LLMs. LLMs perform satisfactorily wide range contexts, multiple aspects addressed adequately addition aforementioned medical issues. LLMs require enormous number parameters vast amount training data, consuming substantial computational resources training process (S. Zhang et al. 2022). Even training, reasoning LLMs consumes significant computational resources (Samsi et al. 2023). Furthermore, LLMs produce large carbon footprint (S. Zhang et al. 2022; Faiz et al. 2023) require considerable water consumption cooling (George, .S.Hovan George, .S.Gabrio Martin 2023), exacerbating environmental concerns extreme climate events. computational power concentrated labs, LLM also exacerbates inequality issues prevents labs gaining LLMs (S. Zhang et al. 2022). LLMs often considered black boxes, making difficult understand explain operating mechanisms. Recently, OpenAI demonstrated early forms artificial intelligence LLMs increasing parameters training sample size (OpenAI et al. 2023; Bubeck et al. 2023; Schaeffer, Miranda, Koyejo 2023; Wei et al. 2022), challenging many scholars’ perceptions. argued resembles Chinese room problem, LLMs emerge intelligence rather acquire deeper features language, consciousness may special form language (Hamid 2023). Others contend emergent intelligence LLMs merely wishful thinking researchers (Schaeffer, Miranda, Koyejo 2023). Alternatively, proposed LLMs resemble human societies, large number individuals collectively exhibit abilities individuals possess, emergent capabilities resulting complex relationships numerous data points, akin ant colony algorithms. suspect possible reasons emergence regarding capabilities demonstrated LLMs. noted earlier, developing relatively interpretable specialized models maximize usability transparency, making safer clinical applications. past decades, humans accumulated substantial historical experience fighting diseases large number low-practice-value papers monographs (Hanson et al. 2023). Translating experience clinical resources bulk become important issue modern medical research. observed clinicians tend treat patients automatically based diseases considering comorbidities patients may , e.g., heart disease, high cholesterol, bacterial infections. aimed develop model simulate ability maintaining model interpretability. Therefore, adapted original spreading activation model replacing LTM knowledge network substituting memory search inference random walk approach simulate human abilities. LLMs often trained using knowledge publications, promising life2vec model uses medical information Danish citizens. , use medical texts build knowledge networks train models. knowledge network large-scale, graph-structured database abstracts core concepts relationships reality, allowing AI systems understand complex relationships reason . can integrate various data sources types represent relationships elements properties. Knowledge networks abstract real world AI systems (Martin Baggio 2020), enabling solve complex tasks reason world (S. Ji et al. 2022). Biomedical knowledge characterized using formalism, abstraction process human brain model systems formally mathematically (Phillips 2020). Although biomedical knowledge use formulas describe biological processes like mathematics, physics, chemistry, knowledge networks can establish mechanisms involved biological processes (Martin Baggio 2020). example, biologists usually use nodes represent genes edges represent regulatory relationships genes. knowledge network constructed, can simulate humans utilize LTM choosing random walk approach. Numerous studies shown random walk can effectively simulate human semantic cognition (S. Ji et al. 2022; Kumar, Steyvers, Balota 2021) consistent human memory retrieval process. Compared outputs spreading activation, computer-simulated random walks showed higher correlation spreading activation model’s results (Abbott, Austerweil, Griffiths 2015; Zemla Austerweil 2018; Siew 2019). Furthermore, brain scientists used random walk algorithms explore theoretical concepts (Abbott, Austerweil, Griffiths 2015; Siew 2019) simulate specific human cognitive behaviors reduce experimental errors introduced external environments (Abbott, Austerweil, Griffiths 2015; Zemla Austerweil 2018). Similar repeated, random selection various possible solutions human brain, random walk simulates random events exists individual problem-solving decision-making processes. diffusion model, applicable wide range situations, even computer-simulated human societies (Park et al. 2023), demonstrating broad applicability computer models many different biological scenarios.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"conclusion","dir":"Articles","previous_headings":"","what":"5. Conclusion","title":"Introduction and inspirations of labyrinth","text":"Humans excel employing existing problem-solving strategies (Kaula 1995). rapid advancement computer technology, surge research articles drug repositioning aided computational biology bioinformatics. Figure \\(\\ref{fig:banana}\\) demonstrates relative number articles drug repositioning included PubMed, shows increasing trend years significant rise recent years. calculation method also exhibits increasing trend. banana metric proven effective quantifying analyzing research interest trends across various fields, defined number articles retrieved using banana keyword per year (Dalmaijer et al. 2021). Bibliometric analysis drug repurposing. Drug repurposing gains significant attention since 2010. adopted banana scale depict trend. observed clinicians tend treat patients symptomatically based diseases considering comorbidities patients may , example, heart disease, hyperlipidemia, bacterial infections. aimed develop model simulate ability maintaining interpretability. Therefore, adapted original spreading activation theory replacing LTM knowledge network substituting memory search inference random walk approach simulate human abilities.","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"setting-up-environment","dir":"Articles","previous_headings":"","what":"Setting up environment","title":"Training the model","text":"developed tested labyrinth Fedora Linux version 38 39. package contain operating system-specific code, tested operating systems. theory, labyrinth work Unix-like operating systems long required dependencies installed. need rebuild model, strongly recommended use Fedora Linux downstream operating system (.e., Red Hat Enterprise Linux, CentOS Stream, Rocky Linux, Alma Linux) follow instructions . Using different environment may lead unexpected issues differences results.","code":""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"red-hat-enterprise-linux-rocky-linux-alma-linux-and-centos-stream","dir":"Articles","previous_headings":"Setting up environment","what":"Red Hat Enterprise Linux, Rocky Linux, Alma Linux, and CentOS Stream","title":"Training the model","text":"Fedora RPMs R ported CentOS/RHEL project Extra Packages Enterprise Linux (EPEL). RPMs also compatible distributions derived CentOS/RHEL. use EPEL repository, sufficient download install appropriate epel-release RPM. R can installed described Fedora section. CentOS Stream 9 Rocky Linux 9, AlmaLinux 9, RHEL 9 CentOS Stream 8 Rocky Linux 8, AlmaLinux 8, RHEL 8 CentOS 7 RHEL 7","code":"sudo dnf config-manager --set-enabled crb sudo dnf install epel-release epel-next-release -y sudo dnf config-manager --set-enabled crb sudo dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm -y sudo dnf config-manager --set-enabled powertools sudo dnf install epel-release epel-next-release -y sudo dnf config-manager --set-enabled powertools sudo dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm -y sudo yum-config-manager --enable extras sudo yum install dnf -y sudo dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm -y  # Use puias computational repo curl -o /tmp/RPM-GPG-KEY-springdale http://springdale.math.ias.edu/data/puias/7/x86_64/os/RPM-GPG-KEY-springdale sudo mv /tmp/RPM-GPG-KEY-springdale /etc/pki/rpm-gpg/RPM-GPG-KEY-springdale  tee > /tmp/puias_computational.repo << EOF [puias_computational] name=PUIAS computational Base $releasever - $basearch mirrorlist=http://puias.math.ias.edu/data/puias/computational/$releasever/$basearch/mirrorlist #baseurl=http://puias.math.ias.edu/data/puias/computational/$releasever/$basearch enabled=1 gpgcheck=1 gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-puias EOF sudo mv /tmp/puias_computational.repo /etc/yum.repos.d  sudo dnf update -y"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"fedora","dir":"Articles","previous_headings":"Setting up environment","what":"Fedora","title":"Training the model","text":"","code":"sudo dnf install R @development-tools automake libxml2-devel libcurl-devel fontconfig-devel openssl-devel harfbuzz-devel fribidi-devel libgit2-devel freetype-devel libpng-devel libtiff-devel libjpeg-devel python3-pip ghostscript pandoc R-flexiblas flexiblas-*"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"debian-based-system-and-ubuntu","dir":"Articles","previous_headings":"Setting up environment","what":"Debian-based system and Ubuntu","title":"Training the model","text":"","code":"sudo apt install r-base-dev build-essential make automake curl openssl libxml2-dev libcurl4-openssl-dev libfontconfig1-dev libssl-dev libharfbuzz-dev libfribidi-dev libgit2-dev libfreetype6-dev libpng-dev libtiff5-dev libjpeg-dev python3-pip ghostscript pandoc libopenblas-dev libopenblas-openmp-dev"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"python-3","dir":"Articles","previous_headings":"Setting up environment","what":"Python 3","title":"Training the model","text":"R package relies several Python packages. Users can install packages using one following methods: System package manager Linux Python pip Anaconda / Miniconda","code":"# On Fedora, you can run: sudo dnf install python3-numpy python3-sklearn python3-scipy python3-seaborn python3-tqdm python3-beautifulsoup4 python3-selenium python3-lxml   # On Debian, you can run: sudo apt install python3-numpy python3-sklearn-pandas python3-scipy python3-seaborn python3-tqdm python3-bs4 python3-selenium python3-lxml pip3 install numpy scikit-learn scipy seaborn tqdm beautifulsoup4 selenium lxml conda create -n labyrinth numpy scikit-learn scipy seaborn tqdm beautifulsoup4 selenium lxml r-tidyverse r-devtools r-rcppeigen r-rcppprogress r-fastmatch r-rpca r-future.apply r-pbapply r-dbi r-rsqlite r-bursts r-patchwork r-furrr r-datapreparation r-tokenizers r-reticulate r-knitr r-progressr r-future.callr r-hrbrthemes r-proc r-ggthemes r-meta r-ggally r-matrixtests r-corrplot r-statix bioconductor-tcgabiolinks bioconductor-clusterprofiler bioconductor-fgsea bioconductor-deseq2 bioconductor-m3c -c bioconda -c conda-forge conda activate labyrinth Rscript -e \"devtools::install_version('dbparser', version = '1.2.0')\" Rscript -e \"remotes::install_github('randef1ned/word2vec')\""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"r","dir":"Articles","previous_headings":"Setting up environment","what":"R","title":"Training the model","text":"","code":"install.packages(c('tidyverse', 'devtools', 'igraph', 'future.apply', 'pbapply', 'DBI', 'RSQLite', 'bursts', 'patchwork', 'furrr', 'M3C', 'dataPreparation', 'tokenizers', 'reticulate', 'magrittr', 'knitr', 'progressr', 'future.callr', 'hrbrthemes', 'pROC', 'ggthemes', 'meta', 'GGally', 'matrixTests', 'rstatix', 'corrplot', 'BiocManager')) devtools::install_version('dbparser', version = '1.2.0') remotes::install_github('randef1ned/word2vec') remotes::install_github('randef1ned/diffusr')"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"folder-structure","dir":"Articles","previous_headings":"","what":"Folder structure","title":"Training the model","text":"tools/ folder GitHub repository contains code scripts required pre-processing processing steps labyrinth.","code":""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"preprocessing-steps","dir":"Articles","previous_headings":"Folder structure","what":"Preprocessing steps","title":"Training the model","text":"Parse drug names: extract_names.Rmd: R Markdown file used parsing extracting names data. Download read data: parse_wos.py: Python script reads parses Web Science (WoS) data. export_edge_tsv.py: Python script exports edge data TSV (tab-separated values) file. download_mesh_synonyms.py: Python script downloads processes MeSH (Medical Subject Headings) synonyms. clinical_trials_parser.py: Python script parses clinical trials data. clean_wos.py: Python script cleans preprocesses WoS data. Identify stopwords run word2vec: identify_stopwords.Rmd: R Markdown file used identify handle stopwords data. run_word2vec.R: R script runs word2vec algorithm tokenized data. compute_distance.R: R script computes distances word2vec embeddings drug. Construct large knowledge network: build_network.R: R script builds large network linking clinical trials IDs paper IDs. Extract stages drugs: drug_status.R: R script extracts stages drugs data.","code":""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"construct-the-model-of-labyrinth-and-validate","dir":"Articles","previous_headings":"Folder structure","what":"Construct the model of labyrinth and validate","title":"Training the model","text":"Construct main knowledge network: main_network.R: R script constructs main network project. Compute details citation network: citation_metric.R: R script computes various metrics details citation network. Download MeSH structures validation: download_mesh_structure.py: Python script downloads MeSH structure data validation purposes. Download TCGA process data validation: download_tcga.R: R script downloads processes TCGA (Cancer Genome Atlas) data validation.","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/authors.html","id":null,"dir":"","previous_headings":"","what":"Authors","title":"Authors and Citation","text":"Yinchun Su. Author, maintainer. Junwei Han. Contributor.","code":""},{"path":"https://labyrinth.yinchun.su/authors.html","id":"citation","dir":"","previous_headings":"","what":"Citation","title":"Authors and Citation","text":"Su Y (2024). labyrinth: Labyrinth: Simulation Knowledge Networks. R package version 0.2.3, https://github.com/randef1ned/labyrinth.","code":"@Manual{,   title = {labyrinth: Labyrinth: A Simulation of Knowledge Networks},   author = {Yinchun Su},   year = {2024},   note = {R package version 0.2.3},   url = {https://github.com/randef1ned/labyrinth}, }"},{"path":"https://labyrinth.yinchun.su/index.html","id":"labyrinth","dir":"","previous_headings":"","what":"Labyrinth: A Simulation of Knowledge Networks","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Labyrinth human-inspired computational framework personalized drug prioritization.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"introduction","dir":"","previous_headings":"","what":"Introduction","title":"Labyrinth: A Simulation of Knowledge Networks","text":"labyrinth computational framework designed drug prioritization population individual levels. integrates multiple sources prior medical knowledge, including: Clinical trial information. Co-occurrence networks literature citations. Drug-target interactions. Drug-indication importance. Disease-drug distance. labyrinth emphasizes importance aligning computational models intuitive human reasoning. Thus, employs human-like knowledge retrieval methodology identify potential drug candidates various diseases. labyrinth aims strike balance predictive accuracy model interpretability, demonstrated robust performance across diverse diseases, evaluated using ROC-AUC metrics.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"a-notice-on-operating-system-compatibility","dir":"","previous_headings":"","what":"A notice on operating system compatibility","title":"Labyrinth: A Simulation of Knowledge Networks","text":"developed tested labyrinth Fedora Linux version 38 39. package contain operating system-specific code, tested operating systems. theory, labyrinth work Unix-like operating systems long required dependencies installed. recommended dependencies installed: R (≥ 4.3.0): developed R package using R version 4.3.3. Python: Python required drawing plots demos. recommended Python seaborn installed, reticulate package use system’s Python installation. OpenMP: package uses OpenMP parallelization multithreading OpenMP exists. OpenMP installed can significantly improve performance. Intel oneAPI Math Kernel Library (oneMKL): library can enhance mathematical performance, especially Intel processors. oneMKL required highly recommended. encounter issues running package operating system, please open issue.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"before-installation","dir":"","previous_headings":"","what":"Before installation","title":"Labyrinth: A Simulation of Knowledge Networks","text":"installation, recommended install Intel oneAPI Math Kernel Library (oneMKL) optimize computational performance linear algebra. Windows users can download oneMKL Intel’s website install default directory. default directory : C:\\Program Files (x86)\\Intel\\oneAPI. Debian Ubuntu users can download oneMKL using apt non-free repo: using Intel repo: Fedora users can download oneMKL using dnf:","code":"# Install oneMKL version 2020.4.304-4 sudo apt install intel-mkl-full # Set up the repository and signed the entry wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \\ | gpg --dearmor | sudo tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null echo \"deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main\" | sudo tee /etc/apt/sources.list.d/oneAPI.list # Update the package list sudo apt update # Install the latest oneMKL (version 2024.1) sudo apt install intel-oneapi-mkl # Create dnf repository file tee > /tmp/oneAPI.repo << EOF [oneAPI] name=Intel® oneAPI repository baseurl=https://yum.repos.intel.com/oneapi enabled=1 gpgcheck=1 repo_gpgcheck=1 gpgkey=https://yum.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB EOF sudo mv /tmp/oneAPI.repo /etc/yum.repos.d # Install the latest oneMKL (version 2024.0) sudo dnf install intel-oneapi-mkl"},{"path":"https://labyrinth.yinchun.su/index.html","id":"installation","dir":"","previous_headings":"","what":"Installation","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Install labyrinth using: can download pre-built binary packages Releases.","code":"install.packages(c('devtools', 'BiocManager')) remotes::install_github(\"randef1ned/labyrinth\")"},{"path":"https://labyrinth.yinchun.su/index.html","id":"usage","dir":"","previous_headings":"","what":"Usage","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Load package using library(labyrinth). provide vignette package can called using: vignette(\"labyrinth\"). Alternatively, can view online version website GitHub. Basically know. documentation contains information contents necessary information training model used project. tools/ folder contains code scripts required constructing model, can understand technical details. Besides, can refer documentation background inspirations behind overall workflow `labyrinth.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"changelog","dir":"","previous_headings":"","what":"Changelog","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Changelog: see ","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the next-time ACT activation rate — activation_rate","title":"Calculate the next-time ACT activation rate — activation_rate","text":"function calculates activation rate node graph based connectivity nodes, *relative strength* connections, global loose factor. ACT spreading activation formula represented Equation 1:   $$(y) = \\sum_x {f(x,y) \\cdot (x)} + c(y)$$,  c(y) represents baseline activation y.   \\(\\alpha\\) represents proximity node network.   \\(t\\) represents iteration number","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the next-time ACT activation rate — activation_rate","text":"","code":"activation_rate(   graph,   strength,   stm,   loose = 1,   threads = 0,   remove_first = FALSE,   display_progress = TRUE )"},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the next-time ACT activation rate — activation_rate","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. strength vector containing *relative strength* connections node graph, last time activation rates nodes. sequence matrix. stm binary vector indicating whether node activated, short-term memory loose scalar numeric 0 1 determines loose (weight) calculation process. threads scalar numeric indicating parallel threads. Default 0 (auto-detected). remove_first logical value indicating whether exclude first node calculation. display_progress logical value indicating whether show progress.","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the next-time ACT activation rate — activation_rate","text":"vector containing activation rate node graph","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the next-time ACT activation rate — activation_rate","text":"","code":"library(magrittr) #>  #> Attaching package: 'magrittr' #> The following objects are masked from 'package:testthat': #>  #>     equals, is_less_than, not  graph <- matrix(nrow = 7, ncol=7, data=c(   0, 0, 0, 0, 0, 0, 0,   1, 0, 0, 0, 0, 0, 0,   1, 0, 0, 0, 0, 0, 0,   0, 1, 0, 0, 0, 0, 0,   0, 1, 1, 0, 0, 0, 0,   0, 0, 1, 0, 0, 0, 0,   0, 0, 0, 1, 1, 1, 0)) diag(graph) <- 0 colnames(graph) <- rownames(graph) <- seq_len(nrow(graph)) %>%   subtract(1) %>% as.character()  initial_info <- data.frame(node = colnames(graph),                            strength = c(2, 4, 3, 2, 2, 1, 5),                            in_stm = c(rep(1, 3), rep(0, 4)))  activation_rate(graph, initial_info$strength, initial_info$in_stm,   loose = 0.8, remove_first = TRUE) #> Number of threads: 4, max threads: 4.  #> Solving activation patterns... #> [1] 7.485100 5.012514 4.741765 6.746771 2.210533 5.461641"},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":null,"dir":"Reference","previous_headings":"","what":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"Maps gene alias names official gene symbols.","code":""},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"","code":"alias2SymbolUsingNCBI(   alias,   gene.info.file,   required.columns = c(\"GeneID\", \"Symbol\", \"description\") )"},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"alias character vector gene aliases gene.info.file name gene information file downloaded NCBI. required.columns character vector columns gene information file required output.","code":""},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"data frame latest gene symbols","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"","code":"library(labyrinth) update_gene_symbol('IL-6') #> [1] \"IL6\""},{"path":"https://labyrinth.yinchun.su/reference/assert_dgCMatrix.html","id":null,"dir":"Reference","previous_headings":"","what":"Check if the sparse matrix valid — assert_dgCMatrix","title":"Check if the sparse matrix valid — assert_dgCMatrix","text":"Check sparse matrix valid","code":""},{"path":"https://labyrinth.yinchun.su/reference/assert_dgCMatrix.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Check if the sparse matrix valid — assert_dgCMatrix","text":"","code":"assert_dgCMatrix(adj_matrix)"},{"path":"https://labyrinth.yinchun.su/reference/assert_dgCMatrix.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Check if the sparse matrix valid — assert_dgCMatrix","text":"adj_matrix sparse matrix check","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":null,"dir":"Reference","previous_headings":"","what":"Disease MeSH identifiers — disease_ids","title":"Disease MeSH identifiers — disease_ids","text":"character vector containing Medical Subject Headings (MeSH) identifiers   human diseases.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Disease MeSH identifiers — disease_ids","text":"","code":"disease_ids"},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Disease MeSH identifiers — disease_ids","text":"character vector.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Disease MeSH identifiers — disease_ids","text":"disease MeSH identifiers obtained National Library   Medicine's Medical Subject Headings (MeSH) database   (https://www.nlm.nih.gov/mesh/).","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Disease MeSH identifiers — disease_ids","text":"`disease_ids` object character vector contains MeSH identifiers   (MeSH IDs) various diseases. element vector represents   unique MeSH identifier, can used identify reference   specific diseases MeSH taxonomy. MeSH identifiers hierarchically organized provide standardized way   representing categorizing diseases medical concepts.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Disease MeSH identifiers — disease_ids","text":"","code":"library(labyrinth)  data(disease_ids, package = \"labyrinth\")  length(disease_ids) #> [1] 1098  # Check if a specific MeSH identifier is present \"D000544\" %in% disease_ids  # Alzheimer's disease #> [1] TRUE"},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":null,"dir":"Reference","previous_headings":"","what":"Get disease impact score — disease_impact_score","title":"Get disease impact score — disease_impact_score","text":"Get disease impact score","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Get disease impact score — disease_impact_score","text":"","code":"disease_impact_score(expr, control, max.iter = 10000)"},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Get disease impact score — disease_impact_score","text":"expr expression matrix control vector scalar index normal samples. max.iter Maximal number iterations augumented Lagrange multiplier algorithm, used robust PCA algorithm.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Get disease impact score — disease_impact_score","text":"vector disease impact score","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Get disease impact score — disease_impact_score","text":"","code":"# no examples # TODO: fill the examples"},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":null,"dir":"Reference","previous_headings":"","what":"Drug annotation dataset — drug_annot","title":"Drug annotation dataset — drug_annot","text":"data.table containing drug identifiers corresponding names.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Drug annotation dataset — drug_annot","text":"","code":"drug_annot"},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Drug annotation dataset — drug_annot","text":"data.table following columns: drug_id character vector representing drug identifier. drug_name character vector representing name drug.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Drug annotation dataset — drug_annot","text":"drug annotations obtained ChEMBL, DrugBank, CTD.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Drug annotation dataset — drug_annot","text":"`drug_annot` data set provides mapping drug identifiers   corresponding names. row represents unique drug identifier   (`drug_id`) associated name (`drug_name`). drug identifiers names data set sourced ChEMBL,   DrugBank, CTD extract synonyms drug.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Drug annotation dataset — drug_annot","text":"","code":"library(labyrinth)  data(drug_annot, package = \"labyrinth\")  # Access drug names for specific identifiers drug_annot[drug_annot$drug_id == \"2838\", ]$drug_name  # Imatinib #>  [1] \"ALPHA-(4-METHYL-1-PIPERAZINYL)-3'-((4-(3-PYRIDYL)-2-PYRIMIDINYL)AMINO)-P-TOLU-P-TOLUIDIDE\" #>  [2] \"CELONIB\"                                                                                   #>  [3] \"CGP 57148\"                                                                                 #>  [4] \"CGP-57148\"                                                                                 #>  [5] \"CGP57148\"                                                                                  #>  [6] \"CGP57148B\"                                                                                 #>  [7] \"D01441\"                                                                                    #>  [8] \"ENLIVEN\"                                                                                   #>  [9] \"GLAMOX\"                                                                                    #> [10] \"GLEEVEC\"                                                                                   #> [11] \"GLIVEC\"                                                                                    #> [12] \"IMATIB\"                                                                                    #> [13] \"IMATINIB\"                                                                                  #> [14] \"IMATINIB MESILATE\"                                                                         #> [15] \"IMATINIB MESYLATE\"                                                                         #> [16] \"IMATINIB METHANESULFONATE\"                                                                 #> [17] \"MESYLATE, IMATINIB\"                                                                        #> [18] \"MESYLONIB\"                                                                                 #> [19] \"METHANESULFONATE, IMATINIB\"                                                                #> [20] \"MITINAB\"                                                                                   #> [21] \"NSC-716051\"                                                                                #> [22] \"PLIVATINIB\"                                                                                #> [23] \"QTI-571\"                                                                                   #> [24] \"QTI571\"                                                                                    #> [25] \"SHANTINIB\"                                                                                 #> [26] \"ST 1571\"                                                                                   #> [27] \"ST1571\"                                                                                    #> [28] \"STI 571\"                                                                                   #> [29] \"STI-571\"                                                                                   #> [30] \"STI571\"                                                                                    #> [31] \"CHEMBL1642\"                                                                                #> [32] \"CHEMBL941\"                                                                                 #> [33] \"DB00619\"                                                                                    # Count the number of unique drugs length(unique(drug_annot$drug_id)) #> [1] 16716  # Visualize the data drug_id <- drug_annot$drug_id hist(table(drug_id))"},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":null,"dir":"Reference","previous_headings":"","what":"Gene-disease associations — gene_disease","title":"Gene-disease associations — gene_disease","text":"list containing gene-disease associations obtained DisGeNET   confidence level 60","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gene-disease associations — gene_disease","text":"","code":"gene_disease"},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Gene-disease associations — gene_disease","text":"named list, element   character vector representing genes associated specific MeSH   disease identifier (MeSH ID). names list elements correspond   MeSH IDs.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Gene-disease associations — gene_disease","text":"gene-disease associations obtained DisGeNET database   (https://www.disgenet.org/), integrates information various   sources provide gene-disease associations.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Gene-disease associations — gene_disease","text":"`gene_disease` data set contains gene-disease associations extracted   DisGeNET database confidence level 60   list represents MeSH disease identifier, corresponding   character vector contains gene identifiers associated   disease.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Gene-disease associations — gene_disease","text":"","code":"library(labyrinth)  data(gene_disease, package = \"labyrinth\")  # Preview the data gene_disease[[\"D000168\"]] #> [1] \"TWIST1\" \"FGFR2\"  \"FGFR1\"   # Visualize the pattern # \\donttest{ disease_gene_count <- purrr::map_int(gene_disease, ~length(.x)) hist(disease_gene_count)  # }"},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":null,"dir":"Reference","previous_headings":"","what":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"returns neighboring nodes specified node graph. neighboring nodes can returned various formats, depending `return_type` parameter.","code":""},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"","code":"get_neighbors(   adj_matrix,   node_id,   neighbor_type = c(\"both\", \"forward\", \"backward\"),   return_type = c(\"binary\", \"name\", \"id\") )"},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"adj_matrix square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. node_id ID node whose neighbors retrieved. Node name also acceptable. neighbor_type `neighbor_type` parameter specifies type neighbors include output: - ****: return neighbors specific node, regarding graph   undirected. - **forward**: return neighbors reachable   specified node, specifically means downstream neighbors directed   networks. - **backward**: return upstream neighbors specific nodes.   Specifically, reverse direction edges graph   return downstream neighbors. return_type `return_type` paramater specifies type return: - **binary**: return binary vector length equal total number   nodes graph. elements vector set 1 nodes   neighbors specific nodes, 0 nodes. - **name**: return character vector containing names neighbor   nodes. - **id**: return integer vector containing IDs neighbor nodes.","code":""},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"vector based `return_type`.","code":""},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"https://www.ibm.com/docs/en/cognos-analytics/11.1.0? topic=terms-modified-z-score","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"","code":"# make an adjacency matrix and randomly fill some cells with 1s mat <- matrix(sample(c(0,1), 100, replace = TRUE), 10, 10) diag(mat) <- 0 # remove self-loops  library(labyrinth) # Get neighbors get_neighbors(mat, 2) #>  [1] 1 0 0 0 0 1 0 1 1 1"},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":null,"dir":"Reference","previous_headings":"","what":"Compute gradient of Spreadgram — gradient","title":"Compute gradient of Spreadgram — gradient","text":"function calculates gradient computed activation rates   given background graph. gradient computed based Sigmoid   function log likelihood, iteratively computes error   permutation. function requires background graph computed activation rates   input.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Compute gradient of Spreadgram — gradient","text":"","code":"gradient(graph, activation, threads = 0, verbose = TRUE)"},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Compute gradient of Spreadgram — gradient","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. activation numeric vector representing computed activation rates node graph. length vector equal number nodes graph. vector contain activation rate node. threads scalar numeric indicating parallel threads. Default 0 (auto-detected). verbose Show verbose message","code":""},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Compute gradient of Spreadgram — gradient","text":"scalar representing gradient computed activation rates.   gradient represents rate change activation rate.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Compute gradient of Spreadgram — gradient","text":"","code":"# The graph G data(\"graph\", package = \"labyrinth\")  # The initial info last_activation = c(2, 4, 3, 2, 2, 1, 5)  gradient(graph, last_activation) #> [1] 0.03732267"},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":null,"dir":"Reference","previous_headings":"","what":"A small example implemented from Anderson (1983) — graph","title":"A small example implemented from Anderson (1983) — graph","text":"small example implemented Anderson (1983)","code":""},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"A small example implemented from Anderson (1983) — graph","text":"","code":"graph"},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"A small example implemented from Anderson (1983) — graph","text":"Adjacency dgCMatrix   representing unweighted directed graph. Inside adjacency matrix,   row column matrix represents node graph.   7 nodes 9 edges, self-loops inside graph.","code":""},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"A small example implemented from Anderson (1983) — graph","text":"Anderson, J. R. (1983). spreading activation theory memory. *Journal verbal learning verbal behavior*, *22*(3), 261-295. https://doi.org/10.1016/S0022-5371(83)90201-3.","code":""},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"A small example implemented from Anderson (1983) — graph","text":"","code":"library(labyrinth)  data(\"graph\", package = \"labyrinth\")  dim(graph) #> [1] 7 7  # Draw the network plots using igraph # \\donttest{ library(igraph) #>  #> Attaching package: 'igraph' #> The following object is masked from 'package:labyrinth': #>  #>     random_walk #> The following object is masked from 'package:testthat': #>  #>     compare #> The following objects are masked from 'package:stats': #>  #>     decompose, spectrum #> The following object is masked from 'package:base': #>  #>     union graph <- graph_from_adjacency_matrix(as.matrix(graph)) plot(graph)  # }"},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":null,"dir":"Reference","previous_headings":"","what":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"Check matrix `dgCMatrix` class","code":""},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"","code":"is.dgCMatrix(mat)"},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"mat matrix check","code":""},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"boolean variable","code":""},{"path":"https://labyrinth.yinchun.su/reference/labyrinth-package.html","id":null,"dir":"Reference","previous_headings":"","what":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","title":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","text":"One paragraph description package   one full sentences. can use Config/testthat/parallel   true boost test efficiency.","code":""},{"path":"https://labyrinth.yinchun.su/reference/labyrinth-package.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","text":"labyrinth","code":""},{"path":"https://labyrinth.yinchun.su/reference/labyrinth-package.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","text":"","code":"# if (!\"devtools\" %in% as.data.frame(installed.packages())$Package) #   install.packages(\"devtools\") # devtools::install_github(\"randef1ned/labyrinth\")"},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":null,"dir":"Reference","previous_headings":"","what":"Load External Data — load_data","title":"Load External Data — load_data","text":"function loads external data R environment. required   file found locally, attempts download file author's   GitHub. package, used download external data     folder, pre-trained model file, drug annotation file, etc.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Load External Data — load_data","text":"","code":"load_data(required_file)"},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Load External Data — load_data","text":"required_file Required file name without extensions GitHub repo.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Load External Data — load_data","text":"content required data.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Load External Data — load_data","text":"function first checks required file exists labyrinth   directory (defined tools::R_user_dir('labyrinth')).   required file found, attempts download file GitHub   repo five times. download fails five attempts, error   thrown. file successfully downloaded already exists   locally, function return file content.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Load External Data — load_data","text":"","code":"if (FALSE) { # Load the trained model model <- load_data(\"model\")  # Load the drug annotation data drug_annot <- load_data(\"drug_annot\") }"},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":null,"dir":"Reference","previous_headings":"","what":"MeSH annotation data — mesh_annot","title":"MeSH annotation data — mesh_annot","text":"data frame containing annotated Medical Subject Headings (MeSH) terms   associated group information.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"MeSH annotation data — mesh_annot","text":"","code":"mesh_annot"},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"MeSH annotation data — mesh_annot","text":"data.table following   columns: mesh_id character vector representing MeSH identifier. mesh_term character vector representing MeSH term. group_id character vector representing group identifier   obtained NCBI. group_name character vector representing group name obtained   NCBI.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"MeSH annotation data — mesh_annot","text":"MeSH terms obtained National Library Medicine's Medical   Subject Headings (MeSH) database (https://www.nlm.nih.gov/mesh/),   group information obtained National Center Biotechnology   Information (NCBI).","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"MeSH annotation data — mesh_annot","text":"","code":"library(labyrinth)  data(mesh_annot, package = \"labyrinth\")  # Preview the data head(mesh_annot) #>   mesh_id        mesh_term group_id group_name #> 1 D002051 burkitt lymphoma        1 Infections #> 2 D007918          leprosy        1 Infections #> 3 D008288          malaria        1 Infections #> 4 D011014        pneumonia        1 Infections #> 5 D012220         rhinitis        1 Infections #> 6 D012772    shock, septic        1 Infections"},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":null,"dir":"Reference","previous_headings":"","what":"MeSH hierarchy network — mesh_hierarchy","title":"MeSH hierarchy network — mesh_hierarchy","text":"edgelist network representing hierarchical structure Medical   Subject Headings (MeSH) terms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"MeSH hierarchy network — mesh_hierarchy","text":"","code":"mesh_hierarchy"},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"MeSH hierarchy network — mesh_hierarchy","text":"data.table following columns: from_id character vector representing MeSH identifier   parent term. to_id character vector representing MeSH identifier   child term.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"MeSH hierarchy network — mesh_hierarchy","text":"MeSH hierarchy data obtained National Library Medicine's   Medical Subject Headings (MeSH) database (https://www.nlm.nih.gov/mesh/).","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"MeSH hierarchy network — mesh_hierarchy","text":"`mesh_hierarchy` data frame edgelist network, row   represents directed edge parent MeSH term (`from_id`) child   MeSH term (`to_id`). structure captures hierarchical relationships   MeSH terms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"MeSH hierarchy network — mesh_hierarchy","text":"","code":"library(labyrinth)  data(mesh_hierarchy, package = \"labyrinth\")  # Preview the data head(mesh_hierarchy) #>    from_id   to_id #>     <char>  <char> #> 1: D001829 D059925 #> 2: D001829 D001940 #> 3: D001940 D042361 #> 4: D001940 D009558 #> 5: D001829 D005121 #> 6: D005121 D000672  # Draw the network plots using igraph # \\donttest{ library(igraph) mesh_network <- graph_from_edgelist(as.matrix(mesh_hierarchy)) plot(mesh_network)  # }"},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":null,"dir":"Reference","previous_headings":"","what":"MeSH synonyms — mesh_synonyms","title":"MeSH synonyms — mesh_synonyms","text":"data frame containing Medical Subject Headings (MeSH) identifiers   associated synonyms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"MeSH synonyms — mesh_synonyms","text":"","code":"mesh_synonyms"},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"MeSH synonyms — mesh_synonyms","text":"data.table following columns: mesh_id character vector representing MeSH identifier. mesh_synonyms character vector representing synonyms   associated MeSH identifier.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"MeSH synonyms — mesh_synonyms","text":"MeSH IDs obtained National Library Medicine's Medical   Subject Headings (MeSH) database (https://www.nlm.nih.gov/mesh/).   synonyms obtained Cochrane Library   (https://www.cochranelibrary.com/advanced-search/mesh).","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"MeSH synonyms — mesh_synonyms","text":"`mesh_synonyms` data frame provides mapping MeSH identifiers   corresponding synonyms. row represents MeSH identifier   (`mesh_id`) associated synonyms (`mesh_synonyms`), can   used text mining, information retrieval, applications   involving MeSH terms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"MeSH synonyms — mesh_synonyms","text":"","code":"library(labyrinth)  data(mesh_synonyms, package = \"labyrinth\") #> Warning: strings not representable in native encoding will be translated to UTF-8  # Preview the data head(mesh_synonyms) #>    mesh_id                             synonyms #>     <char>                               <char> #> 1: D045743                 scleroderma, diffuse #> 2: D045743          scleroses, diffuse systemic #> 3: D045743 diffuse cutaneous systemic sclerosis #> 4: D045743            sudden onset sclerodermas #> 5: D045743            scleroderma, sudden onset #> 6: D045743      systemic sclerosis, progressive"},{"path":"https://labyrinth.yinchun.su/reference/model.html","id":null,"dir":"Reference","previous_headings":"","what":"Pre-trained model of labyrinth — model","title":"Pre-trained model of labyrinth — model","text":"pre-trained dgCMatrix object.","code":""},{"path":"https://labyrinth.yinchun.su/reference/model.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Pre-trained model of labyrinth — model","text":"","code":"model"},{"path":"https://labyrinth.yinchun.su/reference/model.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Pre-trained model of labyrinth — model","text":"pre-trained dgCMatrix object   7,686 columns 7,686 rows.","code":""},{"path":"https://labyrinth.yinchun.su/reference/model.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Pre-trained model of labyrinth — model","text":"labyrinth model trained integrating knowledge two major   sources: text-based information medical corpora biological   knowledge functional interaction networks. text-based component, labyrinth: 1. Extracts drug information (nomenclature, targets, indications)   databases like DrugBank, CTD, ChEMBL. 2. Obtains clinical trial data Cochrane Library. 3. Mines co-occurrence patterns published literature Web   Science corpus. 4. Preprocesses text data 10 million publications, including   stop word removal term vectorization using Skip-gram models. 5. Quantifies structured drug-disease relationships based clinical trial   phases, citation analysis, network proximity gene sets. biological component, labyrinth evaluates network proximity   drug target modules disease gene modules within functional   interactome network. text-based biological knowledge matrices integrated   probabilistic computations, simulating process storing relevant   knowledge long-term memory decision-making.","code":""},{"path":"https://labyrinth.yinchun.su/reference/model.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Pre-trained model of labyrinth — model","text":"","code":"library(labyrinth)  data(model, package = \"labyrinth\")  dim(model) #> [1] 7686 7686"},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":null,"dir":"Reference","previous_headings":"","what":"NCBI Gene Information — ncbi_info","title":"NCBI Gene Information — ncbi_info","text":"data table containing meta information NCBI genes, including   type associated species.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"NCBI Gene Information — ncbi_info","text":"","code":"ncbi_info"},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"NCBI Gene Information — ncbi_info","text":"data.table following   columns: type character vector representing type species. species character vector representing species NCBI   supports online platform.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"NCBI Gene Information — ncbi_info","text":"data obtained National Center Biotechnology Information   (NCBI) Gene database (https://www.ncbi.nlm.nih.gov/gene).","code":""},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"NCBI Gene Information — ncbi_info","text":"","code":"library(labyrinth)  data(\"ncbi_info\", package = \"labyrinth\")  # Preview the data head(ncbi_info) #>                type                                   species #>              <char>                                    <char> #> 1: Archaea_Bacteria                                   Archaea #> 2: Archaea_Bacteria                                  Bacteria #> 3: Archaea_Bacteria Escherichia_coli_str._K-12_substr._MG1655 #> 4: Archaea_Bacteria               Pseudomonas_aeruginosa_PAO1 #> 5:            Fungi                                Ascomycota #> 6:            Fungi                             Microsporidia"},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":null,"dir":"Reference","previous_headings":"","what":"Protein-protein interaction network — ppi","title":"Protein-protein interaction network — ppi","text":"sparse adjacency matrix representing protein-protein interaction network   compiled multiple databases.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Protein-protein interaction network — ppi","text":"","code":"ppi"},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Protein-protein interaction network — ppi","text":"dgCMatrix object   Matrix package, value 1 indicates   connection (interaction) two genes, 0 indicates   connection. rows columns matrix correspond gene   identifiers, matrix square symmetric, representing   undirected adjacency network.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Protein-protein interaction network — ppi","text":"protein-protein interaction data compiled following   databases: KEGG (https://www.genome.jp/kegg/) Reactome (https://reactome.org/) Biocarta (https://cgap.nci.nih.gov/Pathways/BioCarta_Pathways) NCI (https://cancer.gov/-nci) SPIKE (https://www.cbrc.kaust.edu.sa/spike/) HumanCyc (https://humancyc.org/) Panther (http://www.pantherdb.org/)","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Protein-protein interaction network — ppi","text":"`ppi` data set sparse adjacency matrix representing   protein-protein interaction network. network compiled seven   databases: KEGG, Reactome, Biocarta, NCI, SPIKE, HumanCyc, Panther. rows columns matrix correspond gene identifiers,   order genes consistent across rows columns. matrix   square symmetric, representing undirected network,   interaction gene gene B represented 1   (, B) (B, ) positions matrix.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Protein-protein interaction network — ppi","text":"","code":"library(Matrix) data(ppi, package = \"labyrinth\")  # Access a specific element of the matrix dim(ppi) #> [1] 12836 12836 ppi[1, 2]  # Interaction between gene 1 and gene 2 #> [1] 0  # Convert to a dense matrix for visualization or analysis dense_ppi <- as.matrix(ppi) #> Warning: sparse->dense coercion: allocating vector of size 1.2 GiB"},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":null,"dir":"Reference","previous_headings":"","what":"Predict drug response scores based on disease weights — predict_drug","title":"Predict drug response scores based on disease weights — predict_drug","text":"function predict drug response scores based disease weights using   multiple methods. takes vector disease weights sparse matrix   representing drug-disease interaction model, calculates predicted   drug scores using one three methods: random walk restart (`rwr`),   original spreading activation (`sa`), Spread-gram (`sg`).","code":""},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predict drug response scores based on disease weights — predict_drug","text":"","code":"predict_drug(   disease_weights,   model,   method = c(\"rwr\", \"wrwr\", \"sg\", \"sa\"),   restart_prob = 0.7,   threshold = 1e-06,   max_iter = 1e+06,   loose = 1,   print_weight_only = FALSE )"},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Predict drug response scores based on disease weights — predict_drug","text":"disease_weights named vector disease weights. names vector correspond disease IDs `disease_ids` dataset. model square matrix (dgCMatrix pre-trained model. method character string specifying prediction method use.   drug scores can predicted using one three methods: random   walk restart (either `rwr` `wrwr`), original spreading activation   (`sa`), Spread-gram (`sg`). `rwr` method provides quick   calculation estimating drug scores. criterion using   quick calculation met, falls back `wrwr` method. Default option `rwr`. restart_prob restart probability random walk restart method. Default 0.7. threshold convergence threshold iteration. Recommended value 1e-6 `rwr` 1 `sg`. max_iter maximum number iterations. Default value 1e6. loose loose parameter original spreading activation method. Default 1.0. print_weight_only logical value indicating whether print predicted drug weights. TRUE, one numeric vector drug weights returned. FALSE, data frame drug IDs, drug names, drug weights returned. Default value FALSE.","code":""},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Predict drug response scores based on disease weights — predict_drug","text":"return value based `print_weight_only`. TRUE, one  numeric vector drug weights   returned. FALSE, data frame drug   IDs, drug names, drug weights returned.","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predict drug response scores based on disease weights — predict_drug","text":"","code":"# Load example data to the environment data(\"drug_annot\", package = \"labyrinth\") data(\"disease_ids\", package = \"labyrinth\")  # \\donttest{ # Load models to the environment model <- load_data(\"model\") #> Warning: No cached model file. Downloading... #> Warning: downloaded length 0 != reported length 14 #> Warning: cannot open URL 'https://raw.githubusercontent.com/randef1ned/labyrinth/master/extdata/model.rda': HTTP status was '404 Not Found' #> Error in download.file(paste0(\"https://raw.githubusercontent.com/\", \"randef1ned/labyrinth/master/extdata/\",     required_file), model_path): cannot open URL 'https://raw.githubusercontent.com/randef1ned/labyrinth/master/extdata/model.rda'  # Construct disease weights disease_weights <- sample(c(rep(0, 50), rep(1, 2)), 1098, replace = TRUE) names(disease_weights) <- disease_ids  # Predict drug scores based on disease weights drug_weights <- predict_drug(disease_weights, model, method = \"rwr\") #> normalizing column vectors! #> normalizing column vectors! #>  #> Attaching package: 'diffusr' #> The following objects are masked from 'package:labyrinth': #>  #>     assert_dgCMatrix, is.dgCMatrix # }"},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":null,"dir":"Reference","previous_headings":"","what":"Graph diffusion using a Markov random walk — random_walk","title":"Graph diffusion using a Markov random walk — random_walk","text":"Markov Random Walk takes inital distribution \\(p_0\\) calculates stationary distribution . diffusion process regulated restart probability \\(r\\) controls often MRW jumps back initial values. source code brought diffusr v0.2.1.","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Graph diffusion using a Markov random walk — random_walk","text":"","code":"random_walk(   p0,   graph,   r = 0.5,   niter = 10000,   thresh = 1e-04,   do.analytical = FALSE,   correct.for.hubs = FALSE,   allow.ergodic = FALSE,   return.pt.only = FALSE )"},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Graph diffusion using a Markov random walk — random_walk","text":"p0 \\(n \\times p\\)-dimensional numeric non-negative vector/matrix representing starting distribution Markov chain (need sum one). graph \\(n \\times p\\)-dimensional numeric non-negative adjacence matrix (dgCMatrix) representing graph r scalar \\((0, 1)\\). restart probability Markov random walk restart desired niter maximal number iterations computation Markov chain. thresh reached, niter used stop criterion. thresh threshold breaking iterative computation stationary distribution. absolute difference distribution time point \\(t-1\\) \\(t\\) less thresh, algorithm stops. thresh reached niter, algorithm stops well. .analytical boolean stationary distribution shall computed solving analytical solution rather iteratively correct..hubs TRUE multiplies correction factor nodes, random walk gets biased nodes high degree. case original input matrix normalized : $$ P(j | ) = \\dfrac{1}{\\text{degree}()} \\times   \\min \\left(1, \\dfrac{\\text{degree}()}{\\text{degree}(j)}\\right)$$ Note consider edge weights. allow.ergodic Allow multiple components graph. return.pt.Return pt .","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Graph diffusion using a Markov random walk — random_walk","text":"returns list following elements  p.inf  stationary distribution numeric vector transition.matrix column normalized transition matrix used         random walk","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Graph diffusion using a Markov random walk — random_walk","text":"Tong, H., Faloutsos, C., & Pan, J. Y. (2006), Fast random walk restart applications. Koehler, S., Bauer, S., Horn, D., & Robinson, P. N. (2008), Walking interactome prioritization candidate disease genes. American Journal Human Genetics","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Graph diffusion using a Markov random walk — random_walk","text":"","code":"# count of nodes n <- 5 # starting distribution (has to sum to one) p0    <- as.vector(rmultinom(1, 1, prob = rep(.2, n))) # adjacency matrix (either normalized or not) graph <- matrix(abs(rnorm(n * n)), n, n)  # computation of stationary distribution pt    <- random_walk(p0, graph) #> Error in ensure_igraph(graph): Must provide a graph object (provided wrong object type).  # draw histogram of the `pt` vector hist(pt$p.inf) #> Error in pt$p.inf: object of type 'closure' is not subsettable"},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":null,"dir":"Reference","previous_headings":"","what":"Sigmoid function of Spreadgram — sigmoid","title":"Sigmoid function of Spreadgram — sigmoid","text":"Sigmoid function one basic function Spreadgram algorithm.   function calculates Sigmoid function specific node given   activation rates nodes neighbors.   equation writes   $$\\sigma (x,y) = {{\\exp [(x)(y)]} \\{1 + \\exp [(x)(y)]}}$$.","code":""},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Sigmoid function of Spreadgram — sigmoid","text":"","code":"sigmoid(ax, ay, u = 1)"},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Sigmoid function of Spreadgram — sigmoid","text":"ax vector activation rates x, set y's neighbors ay scalar activation rate node y. u \\(x \\N(y)\\) \\(u = 1\\), otherwise \\(u = 0\\).","code":""},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Sigmoid function of Spreadgram — sigmoid","text":"vector sigma","code":""},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Sigmoid function of Spreadgram — sigmoid","text":"","code":"sigmoid(1:4, 2) #> [1] 0.8807971 0.9820138 0.9975274 0.9996646"},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":null,"dir":"Reference","previous_headings":"","what":"A portion of the phonological network examined in Vitevitch (2008). — speech","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"Depicted word *speech*, phonological neighbors *speech*,   phonological neighbors neighbors. taken   spreadr package.","code":""},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"","code":"speech"},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"Adjacency dgCMatrix   representing unweighted directed graph. Inside adjacency matrix,   row column matrix represents node graph.   34 nodes 96 edges, self-loops inside graph.","code":""},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"Chan, K. Y., & Vitevitch, M. S. (2009). influence phonological   neighborhood clustering coefficient spoken word recognition. *Journal   experimental psychology: Human perception performance*, *35*(6),   1934–1949. https://doi.org/10.1037/a0016902. Siew C. S. Q. (2019). spreadr: R package simulate spreading activation   network. *Behavior research methods*, *51*(2), 910–929.   https://doi.org/10.3758/s13428-018-1186-5.","code":""},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"","code":"library(labyrinth)  data(\"speech\", package = \"labyrinth\")  dim(speech) #> [1] 34 34  # Draw the network plots using igraph # \\donttest{ library(igraph) speech_graph <- graph_from_adjacency_matrix(as.matrix(speech)) plot(speech_graph)  # }"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":null,"dir":"Reference","previous_headings":"","what":"Simulate spreading activation in a network (Main function) — spread_gram","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"ACT spreading activation formula represented Equation 1:   $$(y) = \\sum_x {f(x,y) \\cdot (x)} + c(y)$$,   \\(c(y)\\) represents baseline activation y.   \\(\\alpha\\) represents proximity node network.   \\(t\\) represents iteration number","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"","code":"spread_gram(   graph,   last_activation,   loose = 1,   max_iter = 1e+05,   threshold = 1,   threads = 0,   verbose = TRUE )"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. last_activation vector containing last time activation rates nodes. sequence matrix. loose scalar numeric 0 1 determines loose (weight) calculation process. max_iter Max iteration times. threshold End threshold threads scalar numeric indicating parallel threads. Default 0 (auto-detected). verbose Show verbose message","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"numeric vector contains new activation","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"","code":"# The graph G data(\"graph\", package = \"labyrinth\")  # The initial info last_activation = c(2, 4, 3, 2, 2, 1, 5)  results <- spread_gram(graph, last_activation) #> Iterated #0 times. Current loss: 0.0373226717617934 #> Convergent at #0 times. Current loss: 0.0373226717617934"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":null,"dir":"Reference","previous_headings":"","what":"Simulate spreading activation in a network (Only once) — spread_gram_1","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"ACT spreading activation formula represented Equation 1:   $$(y) = \\sum_x {f(x,y) \\cdot (x)} + c(y)$$,   \\(c(y)\\) represents baseline activation y.   \\(\\alpha\\) represents proximity node network.   \\(t\\) represents iteration number","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"","code":"spread_gram_1(graph, last_activation, loose = 1)"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. last_activation vector containing last time activation rates nodes. sequence matrix. loose scalar numeric 0 1 determines loose (weight) calculation process.","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"numeric vector contains new activation","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"","code":"library(labyrinth)  # The graph G data(\"graph\", package = \"labyrinth\")  # The initial info last_activation = c(2, 4, 3, 2, 2, 1, 5)  results <- spread_gram_1(graph, last_activation)"},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the received activation in Spreading Activation (f) — transfer_activation","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"function calculates next-time received activation value node y node x specific graph, based activation values neighbors (including y) weighting factor loose. formula \\(f^{t+1}(x,j) = l \\cdot s_k / \\sum_{j} {s_j}\\).","code":""},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"","code":"transfer_activation(graph, y, x, activation, loose = 1)"},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. y index ID node specify attention place x index ID neighbor node y activation vector containing last-time activation values nodes. sequence matrix. loose scalar numeric 0 1 determines loose (weight) calculation process.","code":""},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"scalar value representing activation value node x","code":""},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"","code":"# make an adjacency matrix and randomly fill some cells with 1s mat <- matrix(sample(c(0,1), 100, replace = TRUE), 10, 10) diag(mat) <- 0 # remove self-loops  transfer_activation(mat, 3, 2, 1:10) #> [1] 0.1034483"},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":null,"dir":"Reference","previous_headings":"","what":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"Update gene symbol using NCBI Entrez Gene","code":""},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"","code":"update_gene_symbol(   gene.pool,   ncbi = NA,   species = c(\"Homo_sapiens\", \"Mus_musculus\", \"Pan_troglodytes\", \"Rattus_norvegicus\",     \"Archaea\", \"Bacteria\", \"Escherichia_coli_str._K-12_substr._MG1655\",     \"Pseudomonas_aeruginosa_PAO1\", \"Ascomycota\", \"Microsporidia\", \"Bos_taurus\",     \"Penicillium_rubens\", \"Saccharomyces_cerevisiae\", \"Anopheles_gambiae\",     \"Caenorhabditis_elegans\", \"Drosophila_melanogaster\", \"Canis_familiaris\",     \"Sus_scrofa\", \"Danio_rerio\", \"Gallus_gallus\", \"Xenopus_laevis\", \"Xenopus_tropicalis\",     \"Arabidopsis_thaliana\", \"Chlamydomonas_reinhardtii\",       \"Oryza_sativa\",     \"Zea_mays\", \"Plasmodium_falciparum\", \"Retroviridae\") )"},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"gene.pool gene list ncbi location NCBI file directory species Optional. specify species","code":""},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"vector latest gene symbols","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"","code":"library(labyrinth) update_gene_symbol('IL-6') #> [1] \"IL6\""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v023","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.3","title":"labyrinth v0.2.3","text":"Added predict() function Added several data Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v022","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.2","title":"labyrinth v0.2.2","text":"Added multi-threaded capability Added random walk Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v021","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.1","title":"labyrinth v0.2.1","text":"Performance improvement Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v020","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.0","title":"labyrinth v0.2.0","text":"Rewrote codes Dropped robust z score Applied Eigen MKL","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v011","dir":"Changelog","previous_headings":"","what":"labyrinth v0.1.1","title":"labyrinth v0.1.1","text":"Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v010","dir":"Changelog","previous_headings":"","what":"labyrinth v0.1.0","title":"labyrinth v0.1.0","text":"Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v005","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.5","title":"labyrinth v0.0.5","text":"Bug fix Remove stopword file","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v004","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.4","title":"labyrinth v0.0.4","text":"Bug fix Added stopword file","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v003","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.3","title":"labyrinth v0.0.3","text":"Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v002","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.2","title":"labyrinth v0.0.2","text":"Added Windows support Added C++ version robust z score","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v001","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.1","title":"labyrinth v0.0.1","text":"Initial submission","code":""}]
+[{"path":[]},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"our-pledge","dir":"","previous_headings":"","what":"Our Pledge","title":"Contributor Covenant Code of Conduct","text":"members, contributors, leaders pledge make participation community harassment-free experience everyone, regardless age, body size, visible invisible disability, ethnicity, sex characteristics, gender identity expression, level experience, education, socio-economic status, nationality, personal appearance, race, caste, color, religion, sexual identity orientation. pledge act interact ways contribute open, welcoming, diverse, inclusive, healthy community.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"our-standards","dir":"","previous_headings":"","what":"Our Standards","title":"Contributor Covenant Code of Conduct","text":"Examples behavior contributes positive environment community include: Demonstrating empathy kindness toward people respectful differing opinions, viewpoints, experiences Giving gracefully accepting constructive feedback Accepting responsibility apologizing affected mistakes, learning experience Focusing best just us individuals, overall community Examples unacceptable behavior include: use sexualized language imagery, sexual attention advances kind Trolling, insulting derogatory comments, personal political attacks Public private harassment Publishing others’ private information, physical email address, without explicit permission conduct reasonably considered inappropriate professional setting","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"enforcement-responsibilities","dir":"","previous_headings":"","what":"Enforcement Responsibilities","title":"Contributor Covenant Code of Conduct","text":"Community leaders responsible clarifying enforcing standards acceptable behavior take appropriate fair corrective action response behavior deem inappropriate, threatening, offensive, harmful. Community leaders right responsibility remove, edit, reject comments, commits, code, wiki edits, issues, contributions aligned Code Conduct, communicate reasons moderation decisions appropriate.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"scope","dir":"","previous_headings":"","what":"Scope","title":"Contributor Covenant Code of Conduct","text":"Code Conduct applies within community spaces, also applies individual officially representing community public spaces. Examples representing community include using official email address, posting via official social media account, acting appointed representative online offline event.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"enforcement","dir":"","previous_headings":"","what":"Enforcement","title":"Contributor Covenant Code of Conduct","text":"Instances abusive, harassing, otherwise unacceptable behavior may reported community leaders responsible enforcement [INSERT CONTACT METHOD]. complaints reviewed investigated promptly fairly. community leaders obligated respect privacy security reporter incident.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"enforcement-guidelines","dir":"","previous_headings":"","what":"Enforcement Guidelines","title":"Contributor Covenant Code of Conduct","text":"Community leaders follow Community Impact Guidelines determining consequences action deem violation Code Conduct:","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_1-correction","dir":"","previous_headings":"Enforcement Guidelines","what":"1. Correction","title":"Contributor Covenant Code of Conduct","text":"Community Impact: Use inappropriate language behavior deemed unprofessional unwelcome community. Consequence: private, written warning community leaders, providing clarity around nature violation explanation behavior inappropriate. public apology may requested.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_2-warning","dir":"","previous_headings":"Enforcement Guidelines","what":"2. Warning","title":"Contributor Covenant Code of Conduct","text":"Community Impact: violation single incident series actions. Consequence: warning consequences continued behavior. interaction people involved, including unsolicited interaction enforcing Code Conduct, specified period time. includes avoiding interactions community spaces well external channels like social media. Violating terms may lead temporary permanent ban.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_3-temporary-ban","dir":"","previous_headings":"Enforcement Guidelines","what":"3. Temporary Ban","title":"Contributor Covenant Code of Conduct","text":"Community Impact: serious violation community standards, including sustained inappropriate behavior. Consequence: temporary ban sort interaction public communication community specified period time. public private interaction people involved, including unsolicited interaction enforcing Code Conduct, allowed period. Violating terms may lead permanent ban.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"id_4-permanent-ban","dir":"","previous_headings":"Enforcement Guidelines","what":"4. Permanent Ban","title":"Contributor Covenant Code of Conduct","text":"Community Impact: Demonstrating pattern violation community standards, including sustained inappropriate behavior, harassment individual, aggression toward disparagement classes individuals. Consequence: permanent ban sort public interaction within community.","code":""},{"path":"https://labyrinth.yinchun.su/CODE_OF_CONDUCT.html","id":"attribution","dir":"","previous_headings":"","what":"Attribution","title":"Contributor Covenant Code of Conduct","text":"Code Conduct adapted Contributor Covenant, version 2.1, available https://www.contributor-covenant.org/version/2/1/code_of_conduct.html. Community Impact Guidelines inspired Mozilla’s code conduct enforcement ladder. answers common questions code conduct, see FAQ https://www.contributor-covenant.org/faq. Translations available https://www.contributor-covenant.org/translations.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":null,"dir":"","previous_headings":"","what":"GNU General Public License","title":"GNU General Public License","text":"Version 2, June 1991Copyright © 1989, 1991 Free Software Foundation, Inc.,51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA Everyone permitted copy distribute verbatim copies license document, changing allowed.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"preamble","dir":"","previous_headings":"","what":"Preamble","title":"GNU General Public License","text":"licenses software designed take away freedom share change . contrast, GNU General Public License intended guarantee freedom share change free software–make sure software free users. General Public License applies Free Software Foundation’s software program whose authors commit using . (Free Software Foundation software covered GNU Lesser General Public License instead.) can apply programs, . speak free software, referring freedom, price. General Public Licenses designed make sure freedom distribute copies free software (charge service wish), receive source code can get want , can change software use pieces new free programs; know can things. protect rights, need make restrictions forbid anyone deny rights ask surrender rights. restrictions translate certain responsibilities distribute copies software, modify . example, distribute copies program, whether gratis fee, must give recipients rights . must make sure , , receive can get source code. must show terms know rights. protect rights two steps: (1) copyright software, (2) offer license gives legal permission copy, distribute /modify software. Also, author’s protection , want make certain everyone understands warranty free software. software modified someone else passed , want recipients know original, problems introduced others reflect original authors’ reputations. Finally, free program threatened constantly software patents. wish avoid danger redistributors free program individually obtain patent licenses, effect making program proprietary. prevent , made clear patent must licensed everyone’s free use licensed . precise terms conditions copying, distribution modification follow.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"terms-and-conditions-for-copying-distribution-and-modification","dir":"","previous_headings":"","what":"TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION","title":"GNU General Public License","text":"0. License applies program work contains notice placed copyright holder saying may distributed terms General Public License. “Program”, , refers program work, “work based Program” means either Program derivative work copyright law: say, work containing Program portion , either verbatim modifications /translated another language. (Hereinafter, translation included without limitation term “modification”.) licensee addressed “”. Activities copying, distribution modification covered License; outside scope. act running Program restricted, output Program covered contents constitute work based Program (independent made running Program). Whether true depends Program . 1. may copy distribute verbatim copies Program’s source code receive , medium, provided conspicuously appropriately publish copy appropriate copyright notice disclaimer warranty; keep intact notices refer License absence warranty; give recipients Program copy License along Program. may charge fee physical act transferring copy, may option offer warranty protection exchange fee. 2. may modify copy copies Program portion , thus forming work based Program, copy distribute modifications work terms Section 1 , provided also meet conditions: ) must cause modified files carry prominent notices stating changed files date change. b) must cause work distribute publish, whole part contains derived Program part thereof, licensed whole charge third parties terms License. c) modified program normally reads commands interactively run, must cause , started running interactive use ordinary way, print display announcement including appropriate copyright notice notice warranty (else, saying provide warranty) users may redistribute program conditions, telling user view copy License. (Exception: Program interactive normally print announcement, work based Program required print announcement.) requirements apply modified work whole. identifiable sections work derived Program, can reasonably considered independent separate works , License, terms, apply sections distribute separate works. distribute sections part whole work based Program, distribution whole must terms License, whose permissions licensees extend entire whole, thus every part regardless wrote . Thus, intent section claim rights contest rights work written entirely ; rather, intent exercise right control distribution derivative collective works based Program. addition, mere aggregation another work based Program Program (work based Program) volume storage distribution medium bring work scope License. 3. may copy distribute Program (work based , Section 2) object code executable form terms Sections 1 2 provided also one following: ) Accompany complete corresponding machine-readable source code, must distributed terms Sections 1 2 medium customarily used software interchange; , b) Accompany written offer, valid least three years, give third party, charge cost physically performing source distribution, complete machine-readable copy corresponding source code, distributed terms Sections 1 2 medium customarily used software interchange; , c) Accompany information received offer distribute corresponding source code. (alternative allowed noncommercial distribution received program object code executable form offer, accord Subsection b .) source code work means preferred form work making modifications . executable work, complete source code means source code modules contains, plus associated interface definition files, plus scripts used control compilation installation executable. However, special exception, source code distributed need include anything normally distributed (either source binary form) major components (compiler, kernel, ) operating system executable runs, unless component accompanies executable. distribution executable object code made offering access copy designated place, offering equivalent access copy source code place counts distribution source code, even though third parties compelled copy source along object code. 4. may copy, modify, sublicense, distribute Program except expressly provided License. attempt otherwise copy, modify, sublicense distribute Program void, automatically terminate rights License. However, parties received copies, rights, License licenses terminated long parties remain full compliance. 5. required accept License, since signed . However, nothing else grants permission modify distribute Program derivative works. actions prohibited law accept License. Therefore, modifying distributing Program (work based Program), indicate acceptance License , terms conditions copying, distributing modifying Program works based . 6. time redistribute Program (work based Program), recipient automatically receives license original licensor copy, distribute modify Program subject terms conditions. may impose restrictions recipients’ exercise rights granted herein. responsible enforcing compliance third parties License. 7. , consequence court judgment allegation patent infringement reason (limited patent issues), conditions imposed (whether court order, agreement otherwise) contradict conditions License, excuse conditions License. distribute satisfy simultaneously obligations License pertinent obligations, consequence may distribute Program . example, patent license permit royalty-free redistribution Program receive copies directly indirectly , way satisfy License refrain entirely distribution Program. portion section held invalid unenforceable particular circumstance, balance section intended apply section whole intended apply circumstances. purpose section induce infringe patents property right claims contest validity claims; section sole purpose protecting integrity free software distribution system, implemented public license practices. Many people made generous contributions wide range software distributed system reliance consistent application system; author/donor decide willing distribute software system licensee impose choice. section intended make thoroughly clear believed consequence rest License. 8. distribution /use Program restricted certain countries either patents copyrighted interfaces, original copyright holder places Program License may add explicit geographical distribution limitation excluding countries, distribution permitted among countries thus excluded. case, License incorporates limitation written body License. 9. Free Software Foundation may publish revised /new versions General Public License time time. new versions similar spirit present version, may differ detail address new problems concerns. version given distinguishing version number. Program specifies version number License applies “later version”, option following terms conditions either version later version published Free Software Foundation. Program specify version number License, may choose version ever published Free Software Foundation. 10. wish incorporate parts Program free programs whose distribution conditions different, write author ask permission. software copyrighted Free Software Foundation, write Free Software Foundation; sometimes make exceptions . decision guided two goals preserving free status derivatives free software promoting sharing reuse software generally.","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"no-warranty","dir":"","previous_headings":"","what":"NO WARRANTY","title":"GNU General Public License","text":"11. PROGRAM LICENSED FREE CHARGE, WARRANTY PROGRAM, EXTENT PERMITTED APPLICABLE LAW. EXCEPT OTHERWISE STATED WRITING COPYRIGHT HOLDERS /PARTIES PROVIDE PROGRAM “” WITHOUT WARRANTY KIND, EITHER EXPRESSED IMPLIED, INCLUDING, LIMITED , IMPLIED WARRANTIES MERCHANTABILITY FITNESS PARTICULAR PURPOSE. ENTIRE RISK QUALITY PERFORMANCE PROGRAM . PROGRAM PROVE DEFECTIVE, ASSUME COST NECESSARY SERVICING, REPAIR CORRECTION. 12. EVENT UNLESS REQUIRED APPLICABLE LAW AGREED WRITING COPYRIGHT HOLDER, PARTY MAY MODIFY /REDISTRIBUTE PROGRAM PERMITTED , LIABLE DAMAGES, INCLUDING GENERAL, SPECIAL, INCIDENTAL CONSEQUENTIAL DAMAGES ARISING USE INABILITY USE PROGRAM (INCLUDING LIMITED LOSS DATA DATA RENDERED INACCURATE LOSSES SUSTAINED THIRD PARTIES FAILURE PROGRAM OPERATE PROGRAMS), EVEN HOLDER PARTY ADVISED POSSIBILITY DAMAGES. END TERMS CONDITIONS","code":""},{"path":"https://labyrinth.yinchun.su/LICENSE.html","id":"how-to-apply-these-terms-to-your-new-programs","dir":"","previous_headings":"","what":"How to Apply These Terms to Your New Programs","title":"GNU General Public License","text":"develop new program, want greatest possible use public, best way achieve make free software everyone can redistribute change terms. , attach following notices program. safest attach start source file effectively convey exclusion warranty; file least “copyright” line pointer full notice found. Also add information contact electronic paper mail. program interactive, make output short notice like starts interactive mode: hypothetical commands show w show c show appropriate parts General Public License. course, commands use may called something show w show c; even mouse-clicks menu items–whatever suits program. also get employer (work programmer) school, , sign “copyright disclaimer” program, necessary. sample; alter names: General Public License permit incorporating program proprietary programs. program subroutine library, may consider useful permit linking proprietary applications library. want , use GNU Lesser General Public License instead License.","code":"<one line to give the program's name and a brief idea of what it does.> Copyright (C) <year>  <name of author>  This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.  This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more details.  You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. Gnomovision version 69, Copyright (C) year name of author Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. This is free software, and you are welcome to redistribute it under certain conditions; type `show c' for details. Yoyodyne, Inc., hereby disclaims all copyright interest in the program `Gnomovision' (which makes passes at compilers) written by James Hacker.  <signature of Ty Coon>, 1 April 1989 Ty Coon, President of Vice"},{"path":"https://labyrinth.yinchun.su/articles/labyrinth.html","id":"setup","dir":"Articles","previous_headings":"","what":"Setup","title":"The introduction to labyrinth","text":"labyrinth framework initially developed tested Fedora 39 R 4.3, vignette aims demonstrate ease use Windows Subsystem Linux (WSL) environment. Specifically, presenting results using Fedora Remix WSL. get started, need prepare following dependencies. Fedora Red Hat Enterprise Linux without WSL. R 4.3. Required libraries. installed loaded, consistent environment running labyrinth.","code":"# pROC ggthemes  library(tidyverse, igraph) ## -- Attaching core tidyverse packages ------------------------ tidyverse 2.0.0 -- ## v dplyr     1.1.4     v readr     2.1.5 ## v forcats   1.0.0     v stringr   1.5.1 ## v ggplot2   3.5.1     v tibble    3.2.1 ## v lubridate 1.9.3     v tidyr     1.3.1 ## v purrr     1.0.2      ## -- Conflicts ------------------------------------------ tidyverse_conflicts() -- ## x dplyr::filter() masks stats::filter() ## x dplyr::lag()    masks stats::lag() ## i Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors library(patchwork) library(labyrinth)"},{"path":"https://labyrinth.yinchun.su/articles/labyrinth.html","id":"data-preparation","dir":"Articles","previous_headings":"","what":"Data preparation","title":"The introduction to labyrinth","text":"Load pre-trained model. labyrinth model trained integrating knowledge two major sources: text-based information medical corpora biological knowledge functional interaction networks. text-based component, labyrinth: 1. Extracts drug information (nomenclature, targets, indications) databases like DrugBank, CTD, ChEMBL. 2. Obtains clinical trial data Cochrane Library. 3. Mines co-occurrence patterns published literature Web Science corpus. 4. Preprocesses text data 10 million publications, including stop word removal term vectorization using Skip-gram models. 5. Quantifies structured drug-disease relationships based clinical trial phases, citation analysis, network proximity gene sets. biological component, labyrinth evaluates network proximity drug target modules disease gene modules within functional interactome network. text-based biological knowledge matrices integrated probabilistic computations, simulating process storing relevant knowledge long-term memory decision-making.","code":"drug_disease_weight <- load_data('drug_disease_weight') model <- load_data('model') dim(model) ## [1] 7686 7686"},{"path":[]},{"path":"https://labyrinth.yinchun.su/articles/labyrinth.html","id":"labyrinth-learns-mechanisms-that-mediate-specific-drug-response","dir":"Articles","previous_headings":"Reproducibility example","what":"Labyrinth learns mechanisms that mediate specific drug response","title":"The introduction to labyrinth","text":"validate approach, initially evaluated predictive accuracy across various diseases. involved assessing Spearman correlations priority scores assigned labyrinth established weights clinical trials, alongside proximity metrics drug-disease pair. Labyrinth exhibited moderate high correlations, coefficients 0.60 clinical trials 0.80 proximity, respectively. Subsequently, extended analysis encompass human diseases, aiming assess predictive performance across five clinical trial phases, including pre-clinical, phases 1 3, approved treatments. evaluated prediction performance Receiver Operating Characteristic Area Curve (ROC-AUC). illustrated Figure 3A, ROC-AUC values stages surpassed 0.90, indicating predictive success rate 90% distinguishing drugs classified clinical trial non-clinical trial. Notably, labyrinth exhibited high predictive accuracy determining drug usability Stage 3 across disease categories except occupational stomatognathic diseases (Figure 3B). Also, cardiovascular, endocrine system diseases, neoplasms garnered significant benefits labyrinth. Detailed ROC-AUC predictions diseases provided Additional File 1. ### Reproducibility statement results article can reproduced using standard R code.","code":"roc0 <- mutate(drug_disease_weight, weight = if_else(weight > 1, 1, 0)) %>%   pROC::roc(weight, score) ## Setting levels: control = 0, case = 1 ## Setting direction: controls < cases roc1 <- mutate(drug_disease_weight, weight = if_else(weight > 1.8, 1, 0)) %>%   pROC::roc(weight, score) ## Setting levels: control = 0, case = 1 ## Setting direction: controls < cases roc2 <- mutate(drug_disease_weight, weight = if_else(weight > 2, 1, 0)) %>%   pROC::roc(weight, score) ## Setting levels: control = 0, case = 1 ## Setting direction: controls < cases roc3 <- mutate(drug_disease_weight, weight = if_else(weight > 3, 1, 0)) %>%   pROC::roc(weight, score) ## Setting levels: control = 0, case = 1 ## Setting direction: controls < cases roc4 <- mutate(drug_disease_weight, weight = if_else(weight > 4, 1, 0)) %>%   pROC::roc(weight, score) ## Setting levels: control = 0, case = 1 ## Setting direction: controls < cases p1 <- pROC::ggroc(list(   roc0, roc1, roc2, roc3, roc4 )) +   scale_color_manual('Drug trials', labels = c(     paste('Stage 0:', round(roc0$auc, 3)),     paste('Stage 1:', round(roc1$auc, 3)),     paste('Stage 2:', round(roc2$auc, 3)),     paste('Stage 3:', round(roc3$auc, 3)),     paste('Stage 4:', round(roc4$auc, 3))   ), values = c('#cf4e9c', '#835921', '#8d59a3', '#368db9', '#302c2d')) +   ggthemes::theme_clean() +   theme(legend.position = 'left') rm(list = c('roc0', 'roc1', 'roc2', 'roc3', 'roc4')) # We perform ROC analysis in every single disease disease_pred <- split(select(drug_disease_weight, weight, score), drug_disease_weight$mesh_id) %>%   map(~ arrange(.x, desc(weight))) disease_pred <- pbapply::pblapply(names(disease_pred), function(n) {   x <- disease_pred[[n]]   if (nth(x$weight, 2) > 1.5) {     ci <- suppressMessages({       x %>%         mutate(weight = if_else(weight > 1.5, 1, 0)) %>%         pROC::roc(weight, score, ci = TRUE) %>%         {as.numeric(.$ci)}     })     ret <- data.frame(mesh_id = n, ci_lower = ci[1], ci_upper = ci[3], auc = ci[2])   } else {     ret <- data.frame()   }   return(ret) }) %>% bind_rows()  # Next, we divide the ROC and average them based on MeSH structures  mesh_ids <- unique(drug_disease_weight$mesh_id)  data(\"mesh_annot\") disease_cluster <- distinct(mesh_annot) %>%   mutate(parent_group = str_to_sentence(group_name)) %>%   select(!group_name) %>%   arrange(group_id) %>%   left_join(disease_pred, by = 'mesh_id') %>%   drop_na()  disease_color <- group_by(disease_cluster, group_id) %>%   summarize(color = median(auc) > 0.911) p2 <- mutate(disease_cluster, parent_group = fct_reorder(parent_group, -group_id)) %>%   left_join(disease_color, by = 'group_id') %>% ggplot(aes(x = parent_group, y = auc, fill = color)) +   geom_boxplot(alpha = 0.4) +   geom_hline(aes(yintercept = 0.911), linetype = 'dashed') +   labs(x = 'Disease categories',       y = 'Prediction AUCs') +   coord_flip() +   scale_fill_manual(values = c('#9fcdc9', '#d4bfe0')) +   ggthemes::theme_clean() +   theme(legend.position = 'none',         axis.text.y = element_text('Arial Narrow')) design <- \"   11#222   222222   222222 \" p1 + p2 + plot_layout(guides = 'collect') +   plot_annotation(tag_levels = 'A') &   theme(legend.position = 'bottom')"},{"path":"https://labyrinth.yinchun.su/articles/labyrinth.html","id":"session-info","dir":"Articles","previous_headings":"","what":"Session info","title":"The introduction to labyrinth","text":"","code":"devtools::session_info() ## - Session info --------------------------------------------------------------- ##  setting  value ##  version  R version 4.4.0 (2024-04-24) ##  os       Fedora Linux 40 (Container Image) ##  system   x86_64, linux-gnu ##  ui       X11 ##  language en ##  collate  C ##  ctype    C ##  tz       Etc/UTC ##  date     2024-05-04 ##  pandoc   3.1.3 @ /usr/bin/ (via rmarkdown) ##  ## - Packages ------------------------------------------------------------------- ##  package           * version   date (UTC) lib source ##  backports           1.4.1     2021-12-13 [2] CRAN (R 4.4.0) ##  bslib               0.7.0     2024-03-29 [2] CRAN (R 4.4.0) ##  cachem              1.0.8     2023-05-01 [2] CRAN (R 4.4.0) ##  checkmate           2.3.1     2023-12-04 [2] CRAN (R 4.4.0) ##  cli                 3.6.2     2023-12-11 [2] CRAN (R 4.4.0) ##  codetools           0.2-20    2024-03-31 [2] CRAN (R 4.4.0) ##  colorspace          2.1-0     2023-01-23 [2] CRAN (R 4.4.0) ##  desc                1.4.3     2023-12-10 [2] CRAN (R 4.4.0) ##  devtools            2.4.5     2022-10-11 [2] CRAN (R 4.4.0) ##  diffusr             0.2.3     2024-05-04 [2] Bioconductor ##  digest              0.6.35    2024-03-11 [2] CRAN (R 4.4.0) ##  dplyr             * 1.1.4     2023-11-17 [2] CRAN (R 4.4.0) ##  ellipsis            0.3.2     2021-04-29 [2] CRAN (R 4.4.0) ##  evaluate            0.23      2023-11-01 [2] CRAN (R 4.4.0) ##  fansi               1.0.6     2023-12-08 [2] CRAN (R 4.4.0) ##  farver              2.1.1     2022-07-06 [2] CRAN (R 4.4.0) ##  fastmap             1.1.1     2023-02-24 [2] CRAN (R 4.4.0) ##  fastmatch           1.1-4     2023-08-18 [2] CRAN (R 4.4.0) ##  forcats           * 1.0.0     2023-01-29 [2] CRAN (R 4.4.0) ##  fs                  1.6.4     2024-04-25 [2] CRAN (R 4.4.0) ##  generics            0.1.3     2022-07-05 [2] CRAN (R 4.4.0) ##  ggplot2           * 3.5.1     2024-04-23 [2] CRAN (R 4.4.0) ##  ggthemes            5.1.0     2024-02-10 [2] CRAN (R 4.4.0) ##  glue                1.7.0     2024-01-09 [2] CRAN (R 4.4.0) ##  gtable              0.3.5     2024-04-22 [2] CRAN (R 4.4.0) ##  highr               0.10      2022-12-22 [2] CRAN (R 4.4.0) ##  hms                 1.1.3     2023-03-21 [2] CRAN (R 4.4.0) ##  htmltools           0.5.8.1   2024-04-04 [2] CRAN (R 4.4.0) ##  htmlwidgets         1.6.4     2023-12-06 [2] CRAN (R 4.4.0) ##  httpuv              1.6.15    2024-03-26 [2] CRAN (R 4.4.0) ##  igraph              2.0.3     2024-03-13 [2] CRAN (R 4.4.0) ##  jquerylib           0.1.4     2021-04-26 [2] CRAN (R 4.4.0) ##  jsonlite            1.8.8     2023-12-04 [2] CRAN (R 4.4.0) ##  knitr               1.46      2024-04-06 [2] CRAN (R 4.4.0) ##  labeling            0.4.3     2023-08-29 [2] CRAN (R 4.4.0) ##  labyrinth         * 0.3.0     2024-05-04 [1] local ##  later               1.3.2     2023-12-06 [2] CRAN (R 4.4.0) ##  lattice             0.22-6    2024-03-20 [2] CRAN (R 4.4.0) ##  lifecycle           1.0.4     2023-11-07 [2] CRAN (R 4.4.0) ##  lubridate         * 1.9.3     2023-09-27 [2] CRAN (R 4.4.0) ##  magrittr            2.0.3     2022-03-30 [2] CRAN (R 4.4.0) ##  Matrix              1.7-0     2024-04-26 [2] CRAN (R 4.4.0) ##  MatrixGenerics      1.16.0    2024-04-30 [2] Bioconductor 3.19 (R 4.4.0) ##  matrixStats         1.3.0     2024-04-11 [2] CRAN (R 4.4.0) ##  memoise             2.0.1     2021-11-26 [2] CRAN (R 4.4.0) ##  memuse              4.2-3     2023-01-24 [2] CRAN (R 4.4.0) ##  mime                0.12      2021-09-28 [2] CRAN (R 4.4.0) ##  miniUI              0.1.1.1   2018-05-18 [2] CRAN (R 4.4.0) ##  munsell             0.5.1     2024-04-01 [2] CRAN (R 4.4.0) ##  patchwork         * 1.2.0     2024-01-08 [2] CRAN (R 4.4.0) ##  pbapply             1.7-2     2023-06-27 [2] CRAN (R 4.4.0) ##  pillar              1.9.0     2023-03-22 [2] CRAN (R 4.4.0) ##  pkgbuild            1.4.4     2024-03-17 [2] CRAN (R 4.4.0) ##  pkgconfig           2.0.3     2019-09-22 [2] CRAN (R 4.4.0) ##  pkgdown             2.0.9     2024-04-18 [2] CRAN (R 4.4.0) ##  pkgload             1.3.4     2024-01-16 [2] CRAN (R 4.4.0) ##  plyr                1.8.9     2023-10-02 [2] CRAN (R 4.4.0) ##  pROC                1.18.5    2023-11-01 [2] CRAN (R 4.4.0) ##  profvis             0.3.8     2023-05-02 [2] CRAN (R 4.4.0) ##  promises            1.3.0     2024-04-05 [2] CRAN (R 4.4.0) ##  pryr                0.1.6     2023-01-17 [2] CRAN (R 4.4.0) ##  purrr             * 1.0.2     2023-08-10 [2] CRAN (R 4.4.0) ##  R6                  2.5.1     2021-08-19 [2] CRAN (R 4.4.0) ##  ragg                1.3.0     2024-03-13 [2] CRAN (R 4.4.0) ##  Rcpp                1.0.12    2024-01-09 [2] CRAN (R 4.4.0) ##  RcppEigen           0.3.4.0.0 2024-02-28 [2] CRAN (R 4.4.0) ##  RcppProgress        0.4.2     2020-02-06 [2] CRAN (R 4.4.0) ##  readr             * 2.1.5     2024-01-10 [2] CRAN (R 4.4.0) ##  remotes             2.5.0     2024-03-17 [2] CRAN (R 4.4.0) ##  rlang               1.1.3     2024-01-10 [2] CRAN (R 4.4.0) ##  rmarkdown           2.26      2024-03-05 [2] CRAN (R 4.4.0) ##  rpca                0.2.3     2015-07-31 [2] CRAN (R 4.4.0) ##  sass                0.4.9     2024-03-15 [2] CRAN (R 4.4.0) ##  scales              1.3.0     2023-11-28 [2] CRAN (R 4.4.0) ##  sessioninfo         1.2.2     2021-12-06 [2] CRAN (R 4.4.0) ##  shiny               1.8.1.1   2024-04-02 [2] CRAN (R 4.4.0) ##  sparseMatrixStats   1.16.0    2024-04-30 [2] Bioconductor 3.19 (R 4.4.0) ##  stringi             1.8.3     2023-12-11 [2] CRAN (R 4.4.0) ##  stringr           * 1.5.1     2023-11-14 [2] CRAN (R 4.4.0) ##  systemfonts         1.0.6     2024-03-07 [2] CRAN (R 4.4.0) ##  textshaping         0.3.7     2023-10-09 [2] CRAN (R 4.4.0) ##  tibble            * 3.2.1     2023-03-20 [2] CRAN (R 4.4.0) ##  tidyr             * 1.3.1     2024-01-24 [2] CRAN (R 4.4.0) ##  tidyselect          1.2.1     2024-03-11 [2] CRAN (R 4.4.0) ##  tidyverse         * 2.0.0     2023-02-22 [2] CRAN (R 4.4.0) ##  timechange          0.3.0     2024-01-18 [2] CRAN (R 4.4.0) ##  tzdb                0.4.0     2023-05-12 [2] CRAN (R 4.4.0) ##  urlchecker          1.0.1     2021-11-30 [2] CRAN (R 4.4.0) ##  usethis             2.2.3     2024-02-19 [2] CRAN (R 4.4.0) ##  utf8                1.2.4     2023-10-22 [2] CRAN (R 4.4.0) ##  vctrs               0.6.5     2023-12-01 [2] CRAN (R 4.4.0) ##  withr               3.0.0     2024-01-16 [2] CRAN (R 4.4.0) ##  xfun                0.43      2024-03-25 [2] CRAN (R 4.4.0) ##  xtable              1.8-4     2019-04-21 [2] CRAN (R 4.4.0) ##  yaml                2.3.8     2023-12-11 [2] CRAN (R 4.4.0) ##  ##  [1] /tmp/Rtmp8Pmrh8/temp_libpathd601a4ba9c ##  [2] /usr/local/lib/R/library ##  [3] /usr/lib64/R/library ##  [4] /usr/share/R/library ##  ## ------------------------------------------------------------------------------"},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"from-human-memory-to-activation-diffusion-network","dir":"Articles","previous_headings":"","what":"1. From human memory to activation diffusion network","title":"Introduction and inspirations of labyrinth","text":"Memory fundamental cognitive process allows brain store, acquire, recall information. serves temporary storage system sensory cues disappear (Benjamin 2007). Memory plays crucial role encoding, storing, retaining, recalling everything simple sensory data complex knowledge experiences. Additionally, memory basis learning, planning, decision-making (Benjamin 2007; Nussenbaum, Prentis, Hartley 2020). Specifically, enables us learn past experiences simulate potential future outcomes, thereby influencing current behavior future actions (Schacter et al. 2012). formation memories, recall, reasoning based involve combination systems physiological processes allow humans adapt well environment (Schacter et al. 2012; Camina Güell 2017; Nairne Pandeirada 2016).Memory formation comprises three stages: information perception, encoding, storage (Atkinson Shiffrin 1968). stages correspond three types memory: (1) sensory memory (Atkinson Shiffrin 1968) briefly stores raw physical stimuli primary receptors vision hearing; (2) short-term memory (STM) (Baddeley 2000) involves transient storage manipulation information, allowing individuals temporarily memorize small amounts data perform current future tasks; (3) long-term memory (LTM) (Camina Güell 2017) long-term storage information, divided episodic implicit memory. Episodic memory consists knowledge processed recalled conscious level, personal experiences specialized knowledge, implicit memory encompasses skills habits expressed without conscious awareness, fear, riding bicycle, heart rate regulation, conditioned reflexes (Smith Grossman 2008). contrast limited-capacity sensory memory STM, LTM complex cognitive system unlimited capacity long-term storage retrieval wide range information, including factual knowledge personal experiences. Memory can also categorized contextual memory, referring individual’s personal experiences, semantic memory, referring textual knowledge concepts (Renoult et al. 2019). Storing contextual semantic knowledge allows individuals construct new knowledge based past experiences, facilitating survival (Kazanas Altarriba 2015). addition storing information memories, LTM plays important role learning reasoning. can automatically relate relationships attributes objects (Nairne Pandeirada 2016), giving individuals ability use stored skills concepts make rational decisions evaluating different choices various environments predicting possible outcomes (Camina Güell 2017). formation consolidation LTM involve several brain regions, including prefrontal lobe, associated working memory situational memory LTM (Blumenfeld Ranganath 2019), temporal lobe, associated semantic memory LTM (Simmons Martin 2009). hippocampus acts relay station information (Squire Zola-Morgan 1991) can integrate situational memory semantic memory knowledge network stored LTM (Renoult et al. 2019). Consequently, even concept knowledge, knowledge network formed different individuals can vary. individual unique experiences backgrounds, leading different understandings reactions interpreting information. LTM stored vast complex semantic network, includes various types interconnected nodes, concepts, memories, experiences (Collins Loftus 1975). kinds memories experiences also integrated network; example, individual’s representation abstract concepts (e.g., time) may based physical sensations (Casasanto Boroditsky 2008). cases, individuals associate time experiences, forming knowledge network. form organization named semantic networks knowledge networks, emphasizing information interconnected organized according meaning (Lehmann 1992). article, use term semantic network represent form memory storage organization LTM, knowledge network refers artificially built knowledge network. semantic network, concepts represented nodes, concept--concept relationships represented edges nodes, edge weights indicating strength association. higher edge weight implies closer relationship two nodes, typically resulting higher recall rate receiving stimulus (Anderson 1983). Learning memorizing new knowledge experiences involve building new edges reinforcing existing ones. organization facilitates information retrieval enabling individuals jump one concept another network simultaneously activate neighboring nodes form knowledge connections, even direct correlation (Lehmann 1992). interesting example semantic network police officers, black people produce strong association weapons, association even stronger police officer sleep-deprived (James 2018). Another example individual’s preference Coca-Cola McDonald’s determined attitude reflected semantic network (Lee Kim 2013; Karpinski, Steinman, Hilton 2005). Homogeneous networks consist nodes similar properties characteristics (Mhatre Rosenberg 2004). Nodes represented kind elements edges connected nodes high correlations, jointly creating homogeneous network. two examples mentioned illustrate different individuals form different LTMs, memory contents stored LTMs satisfy homogeneity. Moving one concept unrelated concept impossible homogeneous network. process individuals store memories semantic network retrieve information LTM often described spreading activation, network also called spreading activation network (Sharifian Samani 1997). network model, initial node activated, activation state spreads nodes along connected edges. diffusion process can quickly span multiple network layers, extensively activating concepts memories associated initial node. node receives activation certain threshold, fully activated like neurons. Otherwise, activated. may lead recall specific memories, formation decisions, generation problem-solving strategies. mentioned earlier, unrelated concepts semantic network may relatively strong associations. implicit association test (IAT) paradigm proposed Greenwald can effectively test edge connections nodes individual semantic network (Greenwald, McGhee, Schwartz 1998; Greenwald et al. 2009). paradigm tests strength association human brain two nodes, .e., edge weights. mechanisms association activation activation diffusion networks depend strength association nodes. strength high, probability activation relatively high; strength low, higher probability non-activation. theory partly explains forgetting phenomenon occurs human memory. Additionally, activation diffusion networks enable individuals retrieve necessary information, reorganize memories, apply knowledge different situations. summary, activation diffusion networks effectively account dynamic nature memory retrieval use.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"unique-advantages-of-human-cognitive-abilities","dir":"Articles","previous_headings":"","what":"2. Unique advantages of human cognitive abilities","title":"Introduction and inspirations of labyrinth","text":"Compared computer programs, humans possess ability think problems different perspectives exhibit greater flexibility knowledge association (Lehmann 1992). Therefore, humans advantage applying knowledge one domain another seemingly unrelated domain. example, concepts biology can transferred economics (Lawlor et al. 2008), economic models field electronic information (Han et al. 2019), linguistic concepts neuroscience (Mayberry et al. 2018) computer science (H. Zhang et al. 2023). characteristic led humans create many cross-disciplinary fields, artificial intelligence, computational biology, neuropolitics, bioinformatics. Humans can use intuition creative thinking solve problems, ability think across domains allows make new connections different areas, thereby building new disciplinary knowledge. human brain contains approximately 100 billion neurons roughly number glial cells (Bartheld, Bahney, Herculano-Houzel 2016), connected thousands others via synapses (Herculano-Houzel 2009). Neurons glial cells form extremely complex networks. Neurons communicate via --none principle (Pareti 2007), glial cells play crucial roles central nervous system formation, neuronal differentiation, synapse formation (Allen Lyons 2018), regulation neuroinflammatory immunity (Yang Zhou 2019), neurological diseases like dementia (Kim, Choi, Yoon 2020), addition traditional supportive functions (Wolosker et al. 2008). complexity lays foundation individual’s ability process information, experience emotions, maintain awareness, exhibit creativity. Drawing fundamentals human cognition, artificial neural networks simulated using computers mimic brain’s information processing. emulate human cognitive abilities extent, excelling tasks like learning, decision-making, pattern recognition humans naturally proficient (Agatonovic-Kustrin Beresford 2000; Parisi 1997). simulation human cognitive abilities shown great potential (Parisi 1997; Zahedi 1991). However, neurons used deep learning artificial neural networks highly abstract, architecture unable account neurons (Cichy Kaiser 2019). Therefore, field focused attention fitting data rather interpreting (Pichler Hartig 2023). Currently, field deep learning concerned fitting data, effect fitting used guiding criterion field, rather integrating cognitive mechanisms discovered neuroscience (Chavlis Poirazi 2021). Much progress deep learning recent decades can attributed application backpropagation, often used optimization methods update weights minimize loss function. However, neural networks deep learning biologically inspired approaches, biological rationality backpropagation remains questionable, activated neurons acquire features backpropagation (Whittington Bogacz 2019; Lillicrap et al. 2020; Aru, Suzuki, Larkum 2020). Currently, two mainstream learning mechanisms identified human brain using electrophysiological methods: Hebbian learning (Munakata Pfaffly 2004) reinforcement learning (Botvinick et al. 2019). Additionally, synaptic pruning may related learning (Halassa Haydon 2010), epigenetic mechanisms also play important role (Changeux, Courrège, Danchin 1973). Although Hebbian learning, reinforcement learning, attempts migrate human cognitive mechanisms applied deep learning years, still perfectly reproduce human learning features (Volzhenin, Changeux, Dumas 2022). Comparatively, energy consumption using neural networks reasoning huge (Desislavov, Martínez-Plumed, Hernández-Orallo 2023), contrast human brain’s lower energy usage training reasoning (Attwell Laughlin 2001). Another example attention mechanism neural networks, inspired human attention (Vaswani et al. 2023). Attention cognitive ability selectively receives information limited resources (Nelson Cowan et al. 2005). ’s complex biological process involving multiple brain regions, encompassing selective attention also coordinated consciousness, memory, cognition. Selective attention mechanisms associated short-term memory, 3-5 chunks original stimuli can enter single session (N. Cowan 2001), attention lasting just seconds minutes (Polti, Martin, Van Wassenhove 2018). selective mechanism allows humans focus targets limited resources, reducing cognitive resource consumption (Buschman Kastner 2015), refining elements deposited memory (Chun Turk-Browne 2007), delimiting problem space, narrowing memory retrieval problem-solving situations (Wiley Jarosz 2012). Human intuition numbers may also relate attention (Kutter et al. 2023). Thus, selective attention crucial cognitive activities like perception, memory, decision-making. Attention mechanisms deep learning, inspired human selective attention, successfully integrated various frameworks (Niu, Zhong, Yu 2021), greatly improving performance tasks like natural language processing (NLP), computer vision, speech recognition (B. Zhang, Xiong, Su 2020; Guo et al. 2022; Ding et al. 2021). recent years, Transformer model, relying self-attention mechanisms process data, demonstrated superior performance across various tasks (Vaswani et al. 2023; Khan et al. 2022). multi-head attention mechanism performs multiple parallel self-attention computations different parameters, allowing model capture information different subspaces improving fitting efficiency accuracy (Liu, Liu, Han 2021). Practically, Transformer, neural networks made significant progress areas like NLP vision tasks. Attention mechanisms deep learning implemented mathematical functions assign weights different elements input data (Niu, Zhong, Yu 2021; De Santana Correia Colombini 2022). However, subset studies found attention mechanism deep learning fully simulate human attention lacks cognitive emotional context human attention encompasses (Lai et al. 2021). Despite differences, artificial neural networks successfully applied several fields, including image speech recognition, natural language processing, robot control, gaming, decision support systems. applications demonstrate power artificial neural networks dealing complex problems simulating certain human cognitive processes highlighting unique advantages models simulate human cognitive abilities.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"simulating-human-cognitive-abilities-the-way-forward","dir":"Articles","previous_headings":"","what":"3. Simulating Human Cognitive Abilities: The Way Forward","title":"Introduction and inspirations of labyrinth","text":"recent years, Transformer model excelled various tasks rely self-attentive mechanisms data processing (Vaswani et al. 2023; Khan et al. 2022). departs traditional recurrent neural networks (RNNs) convolutional neural networks (CNNs), favoring comprehensive utilization attentional mechanisms process sequential data. Transformer’s attention model primarily applied self-attention multi-head attention mechanisms. self-attention mechanism considers elements sequence processing input element, enabling model capture long-range dependencies within sequence (Vig Belinkov 2019). element transformed query (q), key (k), value (v) vectors, representing current lexical element, lexical elements, information contained lexical element, respectively. attention score computed calculating similarity scores q k, weighted summing v. Recently, LLMs employed Transformer’s framework, demonstrating improved simulation human cognition. LLMs large-scale simulations human cognitive functions (Binz Schulz 2023), emergence mark significant advancement computers’ ability simulate human cognition. LLMs possess enhanced reasoning capabilities, Claude 3, released month Anthropic, exhibits self-awareness contextual understanding needle---haystack task (Anthropic 2024; Kuratov et al. 2024). zero-shot problem scenarios, LLMs’ reasoning abilities without prior knowledge surpass humans, rely analogies reasoning (Webb, Holyoak, Lu 2023). Furthermore, LLMs can comprehend others’ beliefs, goals, mental states accuracy 80%. Notably, GPT-4, considered advanced LLM, can achieve 100% accuracy theory mind (ToM) tasks suitable prompting, indicating human-like level ToM (Thaler 1988). LLMs can also simulate human behavior observed experiments, ultimatum game (Thaler 1988), garden-path sentences (Ferreira, Christianson, Hollingworth 2001), loss aversion (Kimball 1993), reactions Milgram electric shock experiment (Blass 1999; Aher, Arriaga, Kalai 2022). Additionally, LLMs exhibit cognitive biases errors humans typically demonstrate, additive bias (Winter et al. 2023), individuals default adding modifying existing content rather deleting pushing back problem-solving (Adams et al. 2021). LLMs produce various human cognitive effects, including priming effects biases (Koo et al. 2023; Shaki, Kraus, Wooldridge 2023), suggesting LLMs mimicking human cognitive processes may possess cognitive abilities approaching human level. specific domains, LLMs closely mimic human-specific abilities. instance, ChatGPT’s accuracy medical diagnosis providing feasible medical advice complex situations comparable human physicians (Hopkins et al. 2023). performance metrics show diagnoses 93.3% common clinical cases correctly (Hirosawa et al. 2023). Furthermore, standardized clinical decision-making tasks, ChatGPT achieves accuracy rate close 70% (Rao et al. 2023), similar expected level third-year medical students United States (Gilson et al. 2023). Due GPT-4’s superior ToM, correctly answered 90% soft skill questions (Brin et al. 2023), demonstrating excellent clinical skills. However, ChatGPT’s ability handle complex questions remains unsatisfactory compared widely used technologies like Google search (Hopkins et al. 2023). fully replicate professional clinicians’ decision-making abilities faced complex problems, primarily due text-based training data, resulting less satisfactory performance non-text-based tasks (Y. Zhang et al. 2024). Furthermore, patient care medically related domains, sometimes generates false misleading information, potentially causing doctors, nurses, caregivers make erroneous decisions, endangering patients’ lives (Z. Ji et al. 2023). LLMs often contain billions hundreds billions parameters, making difficult implement debugging understand decision-making processes (Z. Ji et al. 2023; Khullar, Wang, Wang 2024). Therefore, developing relatively interpretable models viable alternative moment. models trained specific areas expertise, possessing prior knowledge learning exclusively samples. Recently, life2vec model successfully predicted relationship early mortality aspects individual’s personality traits, demonstrating relatively good predictive efficacy (Savcisens et al. 2023). model provides clinicians family physicians insights assistance can help patients better manage lifespan, showcasing potential specialized models.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"simulating-human-knowledge-representation-using-large-scale-medical-knowledge-networks","dir":"Articles","previous_headings":"","what":"4. Simulating human knowledge representation using large-scale medical knowledge networks","title":"Introduction and inspirations of labyrinth","text":"summary, found computer models simulating human cognitive abilities tend achieve good model fitting results, Transformer-based neural network models like LLMs. LLMs perform satisfactorily wide range contexts, multiple aspects addressed adequately addition aforementioned medical issues. LLMs require enormous number parameters vast amount training data, consuming substantial computational resources training process (S. Zhang et al. 2022). Even training, reasoning LLMs consumes significant computational resources (Samsi et al. 2023). Furthermore, LLMs produce large carbon footprint (S. Zhang et al. 2022; Faiz et al. 2023) require considerable water consumption cooling (George, .S.Hovan George, .S.Gabrio Martin 2023), exacerbating environmental concerns extreme climate events. computational power concentrated labs, LLM also exacerbates inequality issues prevents labs gaining LLMs (S. Zhang et al. 2022). LLMs often considered black boxes, making difficult understand explain operating mechanisms. Recently, OpenAI demonstrated early forms artificial intelligence LLMs increasing parameters training sample size (OpenAI et al. 2023; Bubeck et al. 2023; Schaeffer, Miranda, Koyejo 2023; Wei et al. 2022), challenging many scholars’ perceptions. argued resembles Chinese room problem, LLMs emerge intelligence rather acquire deeper features language, consciousness may special form language (Hamid 2023). Others contend emergent intelligence LLMs merely wishful thinking researchers (Schaeffer, Miranda, Koyejo 2023). Alternatively, proposed LLMs resemble human societies, large number individuals collectively exhibit abilities individuals possess, emergent capabilities resulting complex relationships numerous data points, akin ant colony algorithms. suspect possible reasons emergence regarding capabilities demonstrated LLMs. noted earlier, developing relatively interpretable specialized models maximize usability transparency, making safer clinical applications. past decades, humans accumulated substantial historical experience fighting diseases large number low-practice-value papers monographs (Hanson et al. 2023). Translating experience clinical resources bulk become important issue modern medical research. observed clinicians tend treat patients automatically based diseases considering comorbidities patients may , e.g., heart disease, high cholesterol, bacterial infections. aimed develop model simulate ability maintaining model interpretability. Therefore, adapted original spreading activation model replacing LTM knowledge network substituting memory search inference random walk approach simulate human abilities. LLMs often trained using knowledge publications, promising life2vec model uses medical information Danish citizens. , use medical texts build knowledge networks train models. knowledge network large-scale, graph-structured database abstracts core concepts relationships reality, allowing AI systems understand complex relationships reason . can integrate various data sources types represent relationships elements properties. Knowledge networks abstract real world AI systems (Martin Baggio 2020), enabling solve complex tasks reason world (S. Ji et al. 2022). Biomedical knowledge characterized using formalism, abstraction process human brain model systems formally mathematically (Phillips 2020). Although biomedical knowledge use formulas describe biological processes like mathematics, physics, chemistry, knowledge networks can establish mechanisms involved biological processes (Martin Baggio 2020). example, biologists usually use nodes represent genes edges represent regulatory relationships genes. knowledge network constructed, can simulate humans utilize LTM choosing random walk approach. Numerous studies shown random walk can effectively simulate human semantic cognition (S. Ji et al. 2022; Kumar, Steyvers, Balota 2021) consistent human memory retrieval process. Compared outputs spreading activation, computer-simulated random walks showed higher correlation spreading activation model’s results (Abbott, Austerweil, Griffiths 2015; Zemla Austerweil 2018; Siew 2019). Furthermore, brain scientists used random walk algorithms explore theoretical concepts (Abbott, Austerweil, Griffiths 2015; Siew 2019) simulate specific human cognitive behaviors reduce experimental errors introduced external environments (Abbott, Austerweil, Griffiths 2015; Zemla Austerweil 2018). Similar repeated, random selection various possible solutions human brain, random walk simulates random events exists individual problem-solving decision-making processes. diffusion model, applicable wide range situations, even computer-simulated human societies (Park et al. 2023), demonstrating broad applicability computer models many different biological scenarios.","code":""},{"path":"https://labyrinth.yinchun.su/articles/preface.html","id":"conclusion","dir":"Articles","previous_headings":"","what":"5. Conclusion","title":"Introduction and inspirations of labyrinth","text":"Humans excel employing existing problem-solving strategies (Kaula 1995). rapid advancement computer technology, surge research articles drug repositioning aided computational biology bioinformatics. Figure \\(\\ref{fig:banana}\\) demonstrates relative number articles drug repositioning included PubMed, shows increasing trend years significant rise recent years. calculation method also exhibits increasing trend. banana metric proven effective quantifying analyzing research interest trends across various fields, defined number articles retrieved using banana keyword per year (Dalmaijer et al. 2021). Bibliometric analysis drug repurposing. Drug repurposing gains significant attention since 2010. adopted banana scale depict trend. observed clinicians tend treat patients symptomatically based diseases considering comorbidities patients may , example, heart disease, hyperlipidemia, bacterial infections. aimed develop model simulate ability maintaining interpretability. Therefore, adapted original spreading activation theory replacing LTM knowledge network substituting memory search inference random walk approach simulate human abilities.","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"setting-up-environment","dir":"Articles","previous_headings":"","what":"Setting up environment","title":"Training the model","text":"developed tested labyrinth Fedora Linux version 38 39. package contain operating system-specific code, tested operating systems. theory, labyrinth work Unix-like operating systems long required dependencies installed. need rebuild model, strongly recommended use Fedora Linux downstream operating system (.e., Red Hat Enterprise Linux, CentOS Stream, Rocky Linux, Alma Linux) follow instructions . Using different environment may lead unexpected issues differences results.","code":""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"red-hat-enterprise-linux-rocky-linux-alma-linux-and-centos-stream","dir":"Articles","previous_headings":"Setting up environment","what":"Red Hat Enterprise Linux, Rocky Linux, Alma Linux, and CentOS Stream","title":"Training the model","text":"Fedora RPMs R ported CentOS/RHEL project Extra Packages Enterprise Linux (EPEL). RPMs also compatible distributions derived CentOS/RHEL. use EPEL repository, sufficient download install appropriate epel-release RPM. R can installed described Fedora section. CentOS Stream 9 Rocky Linux 9, AlmaLinux 9, RHEL 9 CentOS Stream 8 Rocky Linux 8, AlmaLinux 8, RHEL 8 CentOS 7 RHEL 7","code":"sudo dnf config-manager --set-enabled crb sudo dnf install epel-release epel-next-release -y sudo dnf config-manager --set-enabled crb sudo dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-9.noarch.rpm -y sudo dnf config-manager --set-enabled powertools sudo dnf install epel-release epel-next-release -y sudo dnf config-manager --set-enabled powertools sudo dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-8.noarch.rpm -y sudo yum-config-manager --enable extras sudo yum install dnf -y sudo dnf install https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm -y  # Use puias computational repo curl -o /tmp/RPM-GPG-KEY-springdale http://springdale.math.ias.edu/data/puias/7/x86_64/os/RPM-GPG-KEY-springdale sudo mv /tmp/RPM-GPG-KEY-springdale /etc/pki/rpm-gpg/RPM-GPG-KEY-springdale  tee > /tmp/puias_computational.repo << EOF [puias_computational] name=PUIAS computational Base $releasever - $basearch mirrorlist=http://puias.math.ias.edu/data/puias/computational/$releasever/$basearch/mirrorlist # baseurl=http://puias.math.ias.edu/data/puias/computational/$releasever/$basearch enabled=1 gpgcheck=1 gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-puias EOF sudo mv /tmp/puias_computational.repo /etc/yum.repos.d  sudo dnf update -y"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"fedora","dir":"Articles","previous_headings":"Setting up environment","what":"Fedora","title":"Training the model","text":"","code":"sudo dnf install R @development-tools automake libxml2-devel libcurl-devel fontconfig-devel openssl-devel harfbuzz-devel fribidi-devel libgit2-devel freetype-devel libpng-devel libtiff-devel libjpeg-devel python3-pip ghostscript pandoc R-flexiblas flexiblas-*"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"debian-based-system-and-ubuntu","dir":"Articles","previous_headings":"Setting up environment","what":"Debian-based system and Ubuntu","title":"Training the model","text":"","code":"sudo apt install r-base-dev build-essential make automake curl openssl libxml2-dev libcurl4-openssl-dev libfontconfig1-dev libssl-dev libharfbuzz-dev libfribidi-dev libgit2-dev libfreetype6-dev libpng-dev libtiff5-dev libjpeg-dev python3-pip ghostscript pandoc libopenblas-dev libopenblas-openmp-dev"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"python-3","dir":"Articles","previous_headings":"Setting up environment","what":"Python 3","title":"Training the model","text":"R package relies several Python packages. Users can install packages using one following methods: System package manager Linux Python pip Anaconda / Miniconda","code":"# On Fedora, you can run: sudo dnf install python3-numpy python3-sklearn python3-scipy python3-seaborn python3-tqdm python3-beautifulsoup4 python3-selenium python3-lxml   # On Debian, you can run: sudo apt install python3-numpy python3-sklearn-pandas python3-scipy python3-seaborn python3-tqdm python3-bs4 python3-selenium python3-lxml pip3 install numpy scikit-learn scipy seaborn tqdm beautifulsoup4 selenium lxml conda create -n labyrinth numpy scikit-learn scipy seaborn tqdm beautifulsoup4 selenium lxml r-tidyverse r-devtools r-rcppeigen r-rcppprogress r-fastmatch r-rpca r-future.apply r-pbapply r-dbi r-rsqlite r-bursts r-patchwork r-furrr r-datapreparation r-tokenizers r-reticulate r-knitr r-progressr r-future.callr r-hrbrthemes r-proc r-ggthemes r-meta r-ggally r-matrixtests r-corrplot r-statix bioconductor-tcgabiolinks bioconductor-clusterprofiler bioconductor-fgsea bioconductor-deseq2 bioconductor-m3c -c bioconda -c conda-forge conda activate labyrinth Rscript -e \"devtools::install_version('dbparser', version = '1.2.0')\" Rscript -e \"remotes::install_github('randef1ned/word2vec')\""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"r","dir":"Articles","previous_headings":"Setting up environment","what":"R","title":"Training the model","text":"project uses custom fork diffusr package, maintained @randef1ned. fork optimizes computational execution package, providing improved performance compared original version. View documentation online.","code":"install.packages(c('tidyverse', 'devtools', 'igraph', 'future.apply', 'pbapply', 'DBI', 'RSQLite', 'bursts', 'patchwork', 'furrr', 'M3C', 'dataPreparation', 'tokenizers', 'reticulate', 'magrittr', 'knitr', 'progressr', 'future.callr', 'hrbrthemes', 'pROC', 'ggthemes', 'meta', 'GGally', 'matrixTests', 'rstatix', 'corrplot', 'BiocManager')) devtools::install_version('dbparser', version = '1.2.0') remotes::install_github('randef1ned/word2vec') remotes::install_github('randef1ned/diffusr')"},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"folder-structure","dir":"Articles","previous_headings":"","what":"Folder structure","title":"Training the model","text":"tools/ folder GitHub repository contains code scripts required pre-processing processing steps labyrinth.","code":""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"preprocessing-steps","dir":"Articles","previous_headings":"Folder structure","what":"Preprocessing steps","title":"Training the model","text":"Parse drug names: extract_names.Rmd: R Markdown file used parsing extracting names data. Download read data: parse_wos.py: Python script reads parses Web Science (WoS) data. export_edge_tsv.py: Python script exports edge data TSV (tab-separated values) file. download_mesh_synonyms.py: Python script downloads processes MeSH (Medical Subject Headings) synonyms. clinical_trials_parser.py: Python script parses clinical trials data. clean_wos.py: Python script cleans preprocesses WoS data. Identify stopwords run word2vec: identify_stopwords.Rmd: R Markdown file used identify handle stopwords data. run_word2vec.R: R script runs word2vec algorithm tokenized data. compute_distance.R: R script computes distances word2vec embeddings drug. Construct large knowledge network: build_network.R: R script builds large network linking clinical trials IDs paper IDs. Extract stages drugs: drug_status.R: R script extracts stages drugs data.","code":""},{"path":"https://labyrinth.yinchun.su/articles/training.html","id":"construct-the-model-of-labyrinth-and-validate","dir":"Articles","previous_headings":"Folder structure","what":"Construct the model of labyrinth and validate","title":"Training the model","text":"Construct main knowledge network: main_network.R: R script constructs main network project. Compute details citation network: citation_metric.R: R script computes various metrics details citation network. Download MeSH structures validation: download_mesh_structure.py: Python script downloads MeSH structure data validation purposes. Download TCGA process data validation: download_tcga.R: R script downloads processes TCGA (Cancer Genome Atlas) data validation.","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/authors.html","id":null,"dir":"","previous_headings":"","what":"Authors","title":"Authors and Citation","text":"Junwei Han. Author, maintainer, contributor. Yinchun Su. Author.","code":""},{"path":"https://labyrinth.yinchun.su/authors.html","id":"citation","dir":"","previous_headings":"","what":"Citation","title":"Authors and Citation","text":"Han J, Su Y (2024). labyrinth: Labyrinth: Simulation Knowledge Networks. R package version 0.3.0, https://github.com/randef1ned/labyrinth.","code":"@Manual{,   title = {labyrinth: Labyrinth: A Simulation of Knowledge Networks},   author = {Junwei Han and Yinchun Su},   year = {2024},   note = {R package version 0.3.0},   url = {https://github.com/randef1ned/labyrinth}, }"},{"path":"https://labyrinth.yinchun.su/index.html","id":"labyrinth","dir":"","previous_headings":"","what":"Labyrinth: A Simulation of Knowledge Networks","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Labyrinth human-inspired computational framework personalized drug prioritization.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"introduction","dir":"","previous_headings":"","what":"Introduction","title":"Labyrinth: A Simulation of Knowledge Networks","text":"labyrinth computational framework designed drug prioritization population individual levels. integrates multiple sources prior medical knowledge, including: Clinical trial information. Co-occurrence networks literature citations. Drug-target interactions. Drug-indication importance. Disease-drug distance. labyrinth emphasizes importance aligning computational models intuitive human reasoning. Thus, employs human-like knowledge retrieval methodology identify potential drug candidates various diseases. labyrinth aims strike balance predictive accuracy model interpretability, demonstrated robust performance across diverse diseases, evaluated using ROC-AUC metrics.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"a-notice-on-operating-system-compatibility","dir":"","previous_headings":"","what":"A notice on operating system compatibility","title":"Labyrinth: A Simulation of Knowledge Networks","text":"developed tested labyrinth Fedora Linux version 38 39. package contain operating system-specific code, tested operating systems. theory, labyrinth work Unix-like operating systems long required dependencies installed. recommended dependencies installed: R (≥ 4.3.0): developed R package using R version 4.3.x. Python: Python required drawing plots demos. recommended Python seaborn installed, reticulate package use system’s Python installation. OpenMP: package uses OpenMP parallelization multithreading OpenMP exists. OpenMP installed can significantly improve performance. Intel oneAPI Math Kernel Library (oneMKL): library can enhance mathematical performance, especially Intel processors. oneMKL required highly recommended. takes less ten minutes install package. encounter issues running package operating system, please open issue.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"before-installation","dir":"","previous_headings":"","what":"Before installation","title":"Labyrinth: A Simulation of Knowledge Networks","text":"installation, recommended install Intel oneAPI Math Kernel Library (oneMKL) optimize computational performance linear algebra. Windows users can download oneMKL Intel’s website install default directory. default directory : C:\\Program Files (x86)\\Intel\\oneAPI. Debian Ubuntu users can download oneMKL using apt non-free repo: using Intel repo: Fedora users can download oneMKL using dnf:","code":"# Install oneMKL version 2020.4.304-4 sudo apt install intel-mkl-full # Set up the repository and signed the entry wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB \\ | gpg --dearmor | sudo tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null echo \"deb [signed-by=/usr/share/keyrings/oneapi-archive-keyring.gpg] https://apt.repos.intel.com/oneapi all main\" | sudo tee /etc/apt/sources.list.d/oneAPI.list # Update the package list sudo apt update # Install the latest oneMKL (version 2024.1) sudo apt install intel-oneapi-mkl # Create dnf repository file tee > /tmp/oneAPI.repo << EOF [oneAPI] name=Intel® oneAPI repository baseurl=https://yum.repos.intel.com/oneapi enabled=1 gpgcheck=1 repo_gpgcheck=1 gpgkey=https://yum.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB EOF sudo mv /tmp/oneAPI.repo /etc/yum.repos.d # Install the latest oneMKL (version 2024.0) sudo dnf install intel-oneapi-mkl"},{"path":"https://labyrinth.yinchun.su/index.html","id":"installation","dir":"","previous_headings":"","what":"Installation","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Install labyrinth using: can download pre-built binary packages Releases.","code":"install.packages(c('devtools', 'BiocManager')) remotes::install_github(\"randef1ned/labyrinth\")"},{"path":"https://labyrinth.yinchun.su/index.html","id":"usage","dir":"","previous_headings":"","what":"Usage","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Load package using library(labyrinth). provide vignette package can called using: vignette(\"labyrinth\"). Alternatively, can view online version GitHub, pkgdown documentation. examples provided take several minutes run normal desktop computer. Basically know. documentation contains information contents necessary information training model used project. tools/ folder contains code scripts required constructing model, can understand technical details. Besides, can refer documentation background inspirations behind overall workflow `labyrinth.","code":""},{"path":"https://labyrinth.yinchun.su/index.html","id":"changelog","dir":"","previous_headings":"","what":"Changelog","title":"Labyrinth: A Simulation of Knowledge Networks","text":"Changelog: see ","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the next-time ACT activation rate — activation_rate","title":"Calculate the next-time ACT activation rate — activation_rate","text":"function calculates activation rate node graph based connectivity nodes, *relative strength* connections, global loose factor. ACT spreading activation formula represented Equation 1:   $$(y) = \\sum_x {f(x,y) \\cdot (x)} + c(y)$$,  c(y) represents baseline activation y.   \\(\\alpha\\) represents proximity node network.   \\(t\\) represents iteration number","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the next-time ACT activation rate — activation_rate","text":"","code":"activation_rate(   graph,   strength,   stm,   loose = 1,   threads = 0,   remove_first = FALSE,   display_progress = TRUE )"},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the next-time ACT activation rate — activation_rate","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. strength vector containing *relative strength* connections node graph, last time activation rates nodes. sequence matrix. stm binary vector indicating whether node activated, short-term memory loose scalar numeric 0 1 determines loose (weight) calculation process. threads scalar numeric indicating parallel threads. Default 0 (auto-detected). remove_first logical value indicating whether exclude first node calculation. display_progress logical value indicating whether show progress.","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the next-time ACT activation rate — activation_rate","text":"vector containing activation rate node graph","code":""},{"path":"https://labyrinth.yinchun.su/reference/activation_rate.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the next-time ACT activation rate — activation_rate","text":"","code":"library(magrittr) #>  #> Attaching package: 'magrittr' #> The following objects are masked from 'package:testthat': #>  #>     equals, is_less_than, not  graph <- matrix(nrow = 7, ncol=7, data=c(   0, 0, 0, 0, 0, 0, 0,   1, 0, 0, 0, 0, 0, 0,   1, 0, 0, 0, 0, 0, 0,   0, 1, 0, 0, 0, 0, 0,   0, 1, 1, 0, 0, 0, 0,   0, 0, 1, 0, 0, 0, 0,   0, 0, 0, 1, 1, 1, 0)) diag(graph) <- 0 colnames(graph) <- rownames(graph) <- seq_len(nrow(graph)) %>%   subtract(1) %>% as.character()  initial_info <- data.frame(node = colnames(graph),                            strength = c(2, 4, 3, 2, 2, 1, 5),                            in_stm = c(rep(1, 3), rep(0, 4)))  activation_rate(graph, initial_info$strength, initial_info$in_stm,   loose = 0.8, remove_first = TRUE) #> Number of threads: 4, max threads: 4.  #> Solving activation patterns... #> [1] 7.485100 5.012514 4.741765 6.746771 2.210533 5.461641"},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":null,"dir":"Reference","previous_headings":"","what":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"Maps gene alias names official gene symbols.","code":""},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"","code":"alias2SymbolUsingNCBI(   alias,   gene.info.file,   required.columns = c(\"GeneID\", \"Symbol\", \"description\") )"},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"alias character vector gene aliases gene.info.file name gene information file downloaded NCBI. required.columns character vector columns gene information file required output.","code":""},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"data frame latest gene symbols","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/alias2SymbolUsingNCBI.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Convert Gene Aliases to Official Gene Symbols — alias2SymbolUsingNCBI","text":"","code":"library(labyrinth) update_gene_symbol('IL-6') #> [1] \"IL6\""},{"path":"https://labyrinth.yinchun.su/reference/assert_dgCMatrix.html","id":null,"dir":"Reference","previous_headings":"","what":"Check if the sparse matrix valid — assert_dgCMatrix","title":"Check if the sparse matrix valid — assert_dgCMatrix","text":"Check sparse matrix valid","code":""},{"path":"https://labyrinth.yinchun.su/reference/assert_dgCMatrix.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Check if the sparse matrix valid — assert_dgCMatrix","text":"","code":"assert_dgCMatrix(adj_matrix)"},{"path":"https://labyrinth.yinchun.su/reference/assert_dgCMatrix.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Check if the sparse matrix valid — assert_dgCMatrix","text":"adj_matrix sparse matrix check","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":null,"dir":"Reference","previous_headings":"","what":"Disease MeSH identifiers — disease_ids","title":"Disease MeSH identifiers — disease_ids","text":"character vector containing Medical Subject Headings (MeSH) identifiers   human diseases.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Disease MeSH identifiers — disease_ids","text":"","code":"disease_ids"},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Disease MeSH identifiers — disease_ids","text":"character vector.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Disease MeSH identifiers — disease_ids","text":"disease MeSH identifiers obtained National Library   Medicine's Medical Subject Headings (MeSH) database   (https://www.nlm.nih.gov/mesh/).","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Disease MeSH identifiers — disease_ids","text":"`disease_ids` object character vector contains MeSH identifiers   (MeSH IDs) various diseases. element vector represents   unique MeSH identifier, can used identify reference   specific diseases MeSH taxonomy. MeSH identifiers hierarchically organized provide standardized way   representing categorizing diseases medical concepts.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_ids.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Disease MeSH identifiers — disease_ids","text":"","code":"library(labyrinth)  data(disease_ids, package = \"labyrinth\")  length(disease_ids) #> [1] 1098  # Check if a specific MeSH identifier is present \"D000544\" %in% disease_ids  # Alzheimer's disease #> [1] TRUE"},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":null,"dir":"Reference","previous_headings":"","what":"Get disease impact score — disease_impact_score","title":"Get disease impact score — disease_impact_score","text":"Get disease impact score","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Get disease impact score — disease_impact_score","text":"","code":"disease_impact_score(expr, control, max.iter = 10000)"},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Get disease impact score — disease_impact_score","text":"expr expression matrix control vector scalar index normal samples. max.iter Maximal number iterations augumented Lagrange multiplier algorithm, used robust PCA algorithm.","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Get disease impact score — disease_impact_score","text":"vector disease impact score","code":""},{"path":"https://labyrinth.yinchun.su/reference/disease_impact_score.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Get disease impact score — disease_impact_score","text":"","code":"# no examples # TODO: fill the examples"},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":null,"dir":"Reference","previous_headings":"","what":"Drug annotation dataset — drug_annot","title":"Drug annotation dataset — drug_annot","text":"data.table containing drug identifiers corresponding names.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Drug annotation dataset — drug_annot","text":"","code":"drug_annot"},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Drug annotation dataset — drug_annot","text":"data.table following columns: drug_id character vector representing drug identifier. drug_name character vector representing name drug.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Drug annotation dataset — drug_annot","text":"drug annotations obtained ChEMBL, DrugBank, CTD.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Drug annotation dataset — drug_annot","text":"`drug_annot` data set provides mapping drug identifiers   corresponding names. row represents unique drug identifier   (`drug_id`) associated name (`drug_name`). drug identifiers names data set sourced ChEMBL,   DrugBank, CTD extract synonyms drug.","code":""},{"path":"https://labyrinth.yinchun.su/reference/drug_annot.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Drug annotation dataset — drug_annot","text":"","code":"library(labyrinth)  data(drug_annot, package = \"labyrinth\")  # Access drug names for specific identifiers drug_annot[drug_annot$drug_id == \"2838\", ]$drug_name  # Imatinib #>  [1] \"ALPHA-(4-METHYL-1-PIPERAZINYL)-3'-((4-(3-PYRIDYL)-2-PYRIMIDINYL)AMINO)-P-TOLU-P-TOLUIDIDE\" #>  [2] \"CELONIB\"                                                                                   #>  [3] \"CGP 57148\"                                                                                 #>  [4] \"CGP-57148\"                                                                                 #>  [5] \"CGP57148\"                                                                                  #>  [6] \"CGP57148B\"                                                                                 #>  [7] \"D01441\"                                                                                    #>  [8] \"ENLIVEN\"                                                                                   #>  [9] \"GLAMOX\"                                                                                    #> [10] \"GLEEVEC\"                                                                                   #> [11] \"GLIVEC\"                                                                                    #> [12] \"IMATIB\"                                                                                    #> [13] \"IMATINIB\"                                                                                  #> [14] \"IMATINIB MESILATE\"                                                                         #> [15] \"IMATINIB MESYLATE\"                                                                         #> [16] \"IMATINIB METHANESULFONATE\"                                                                 #> [17] \"MESYLATE, IMATINIB\"                                                                        #> [18] \"MESYLONIB\"                                                                                 #> [19] \"METHANESULFONATE, IMATINIB\"                                                                #> [20] \"MITINAB\"                                                                                   #> [21] \"NSC-716051\"                                                                                #> [22] \"PLIVATINIB\"                                                                                #> [23] \"QTI-571\"                                                                                   #> [24] \"QTI571\"                                                                                    #> [25] \"SHANTINIB\"                                                                                 #> [26] \"ST 1571\"                                                                                   #> [27] \"ST1571\"                                                                                    #> [28] \"STI 571\"                                                                                   #> [29] \"STI-571\"                                                                                   #> [30] \"STI571\"                                                                                    #> [31] \"CHEMBL1642\"                                                                                #> [32] \"CHEMBL941\"                                                                                 #> [33] \"DB00619\"                                                                                    # Count the number of unique drugs length(unique(drug_annot$drug_id)) #> [1] 16716  # Visualize the data drug_id <- drug_annot$drug_id hist(table(drug_id))"},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":null,"dir":"Reference","previous_headings":"","what":"Gene-disease associations — gene_disease","title":"Gene-disease associations — gene_disease","text":"list containing gene-disease associations obtained DisGeNET   confidence level 60","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Gene-disease associations — gene_disease","text":"","code":"gene_disease"},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Gene-disease associations — gene_disease","text":"named list, element   character vector representing genes associated specific MeSH   disease identifier (MeSH ID). names list elements correspond   MeSH IDs.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Gene-disease associations — gene_disease","text":"gene-disease associations obtained DisGeNET database   (https://www.disgenet.org/), integrates information various   sources provide gene-disease associations.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Gene-disease associations — gene_disease","text":"`gene_disease` data set contains gene-disease associations extracted   DisGeNET database confidence level 60   list represents MeSH disease identifier, corresponding   character vector contains gene identifiers associated   disease.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gene_disease.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Gene-disease associations — gene_disease","text":"","code":"library(labyrinth)  data(gene_disease, package = \"labyrinth\")  # Preview the data gene_disease[[\"D000168\"]] #> [1] \"TWIST1\" \"FGFR2\"  \"FGFR1\"   # Visualize the pattern # \\donttest{ disease_gene_count <- purrr::map_int(gene_disease, ~length(.x)) hist(disease_gene_count)  # }"},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":null,"dir":"Reference","previous_headings":"","what":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"returns neighboring nodes specified node graph. neighboring nodes can returned various formats, depending `return_type` parameter.","code":""},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"","code":"get_neighbors(   adj_matrix,   node_id,   neighbor_type = c(\"both\", \"forward\", \"backward\"),   return_type = c(\"binary\", \"name\", \"id\") )"},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"adj_matrix square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. node_id ID node whose neighbors retrieved. Node name also acceptable. neighbor_type `neighbor_type` parameter specifies type neighbors include output: - ****: return neighbors specific node, regarding graph   undirected. - **forward**: return neighbors reachable   specified node, specifically means downstream neighbors directed   networks. - **backward**: return upstream neighbors specific nodes.   Specifically, reverse direction edges graph   return downstream neighbors. return_type `return_type` paramater specifies type return: - **binary**: return binary vector length equal total number   nodes graph. elements vector set 1 nodes   neighbors specific nodes, 0 nodes. - **name**: return character vector containing names neighbor   nodes. - **id**: return integer vector containing IDs neighbor nodes.","code":""},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"vector based `return_type`.","code":""},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"https://www.ibm.com/docs/en/cognos-analytics/11.1.0? topic=terms-modified-z-score","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/get_neighbors.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Get neighboring nodes in a graph (adjacency matrix) — get_neighbors","text":"","code":"# make an adjacency matrix and randomly fill some cells with 1s mat <- matrix(sample(c(0,1), 100, replace = TRUE), 10, 10) diag(mat) <- 0 # remove self-loops  library(labyrinth) # Get neighbors get_neighbors(mat, 2) #>  [1] 1 0 0 0 0 1 0 1 1 1"},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":null,"dir":"Reference","previous_headings":"","what":"Compute gradient of Spreadgram — gradient","title":"Compute gradient of Spreadgram — gradient","text":"function calculates gradient computed activation rates   given background graph. gradient computed based Sigmoid   function log likelihood, iteratively computes error   permutation. function requires background graph computed activation rates   input.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Compute gradient of Spreadgram — gradient","text":"","code":"gradient(graph, activation, threads = 0, verbose = TRUE)"},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Compute gradient of Spreadgram — gradient","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. activation numeric vector representing computed activation rates node graph. length vector equal number nodes graph. vector contain activation rate node. threads scalar numeric indicating parallel threads. Default 0 (auto-detected). verbose Show verbose message","code":""},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Compute gradient of Spreadgram — gradient","text":"scalar representing gradient computed activation rates.   gradient represents rate change activation rate.","code":""},{"path":"https://labyrinth.yinchun.su/reference/gradient.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Compute gradient of Spreadgram — gradient","text":"","code":"# The graph G data(\"graph\", package = \"labyrinth\")  # The initial info last_activation = c(2, 4, 3, 2, 2, 1, 5)  gradient(graph, last_activation) #> [1] 0.03732267"},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":null,"dir":"Reference","previous_headings":"","what":"A small example implemented from Anderson (1983) — graph","title":"A small example implemented from Anderson (1983) — graph","text":"small example implemented Anderson (1983)","code":""},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"A small example implemented from Anderson (1983) — graph","text":"","code":"graph"},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"A small example implemented from Anderson (1983) — graph","text":"Adjacency dgCMatrix   representing unweighted directed graph. Inside adjacency matrix,   row column matrix represents node graph.   7 nodes 9 edges, self-loops inside graph.","code":""},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"A small example implemented from Anderson (1983) — graph","text":"Anderson, J. R. (1983). spreading activation theory memory. *Journal verbal learning verbal behavior*, *22*(3), 261-295. https://doi.org/10.1016/S0022-5371(83)90201-3.","code":""},{"path":"https://labyrinth.yinchun.su/reference/graph.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"A small example implemented from Anderson (1983) — graph","text":"","code":"library(labyrinth)  data(\"graph\", package = \"labyrinth\")  dim(graph) #> [1] 7 7  # Draw the network plots using igraph # \\donttest{ library(igraph) #>  #> Attaching package: 'igraph' #> The following object is masked from 'package:labyrinth': #>  #>     random_walk #> The following object is masked from 'package:testthat': #>  #>     compare #> The following objects are masked from 'package:stats': #>  #>     decompose, spectrum #> The following object is masked from 'package:base': #>  #>     union graph <- graph_from_adjacency_matrix(as.matrix(graph)) plot(graph)  # }"},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":null,"dir":"Reference","previous_headings":"","what":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"Check matrix `dgCMatrix` class","code":""},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"","code":"is.dgCMatrix(mat)"},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"mat matrix check","code":""},{"path":"https://labyrinth.yinchun.su/reference/is.dgCMatrix.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Check if a matrix is `dgCMatrix` class — is.dgCMatrix","text":"boolean variable","code":""},{"path":"https://labyrinth.yinchun.su/reference/labyrinth-package.html","id":null,"dir":"Reference","previous_headings":"","what":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","title":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","text":"One paragraph description package   one full sentences. can use Config/testthat/parallel   true boost test efficiency.","code":""},{"path":"https://labyrinth.yinchun.su/reference/labyrinth-package.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","text":"labyrinth","code":""},{"path":"https://labyrinth.yinchun.su/reference/labyrinth-package.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Labyrinth: A Simulation of Knowledge Networks — labyrinth-package","text":"","code":"# if (!\"devtools\" %in% as.data.frame(installed.packages())$Package) #   install.packages(\"devtools\") # devtools::install_github(\"randef1ned/labyrinth\")"},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":null,"dir":"Reference","previous_headings":"","what":"Load External Data — load_data","title":"Load External Data — load_data","text":"function loads external data R environment. required   file found locally, attempts download file author's   GitHub. package, used download external data     folder, pre-trained model file, drug annotation file, etc.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Load External Data — load_data","text":"","code":"load_data(required_file)"},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Load External Data — load_data","text":"required_file Required file name without extensions GitHub repo.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Load External Data — load_data","text":"content required data.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Load External Data — load_data","text":"function first checks required file exists labyrinth   directory (defined tools::R_user_dir('labyrinth')).   required file found, attempts download file GitHub   repo five times. download fails five attempts, error   thrown. file successfully downloaded already exists   locally, function return file content.","code":""},{"path":"https://labyrinth.yinchun.su/reference/load_data.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Load External Data — load_data","text":"","code":"if (FALSE) { # Load the trained model model <- load_data(\"model\")  # Load the drug annotation data drug_annot <- load_data(\"drug_annot\") }"},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":null,"dir":"Reference","previous_headings":"","what":"MeSH annotation data — mesh_annot","title":"MeSH annotation data — mesh_annot","text":"data frame containing annotated Medical Subject Headings (MeSH) terms   associated group information.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"MeSH annotation data — mesh_annot","text":"","code":"mesh_annot"},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"MeSH annotation data — mesh_annot","text":"data.table following   columns: mesh_id character vector representing MeSH identifier. mesh_term character vector representing MeSH term. group_id character vector representing group identifier   obtained NCBI. group_name character vector representing group name obtained   NCBI.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"MeSH annotation data — mesh_annot","text":"MeSH terms obtained National Library Medicine's Medical   Subject Headings (MeSH) database (https://www.nlm.nih.gov/mesh/),   group information obtained National Center Biotechnology   Information (NCBI).","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_annot.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"MeSH annotation data — mesh_annot","text":"","code":"library(labyrinth)  data(mesh_annot, package = \"labyrinth\")  # Preview the data head(mesh_annot) #>   mesh_id        mesh_term group_id group_name #> 1 D002051 burkitt lymphoma        1 Infections #> 2 D007918          leprosy        1 Infections #> 3 D008288          malaria        1 Infections #> 4 D011014        pneumonia        1 Infections #> 5 D012220         rhinitis        1 Infections #> 6 D012772    shock, septic        1 Infections"},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":null,"dir":"Reference","previous_headings":"","what":"MeSH hierarchy network — mesh_hierarchy","title":"MeSH hierarchy network — mesh_hierarchy","text":"edgelist network representing hierarchical structure Medical   Subject Headings (MeSH) terms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"MeSH hierarchy network — mesh_hierarchy","text":"","code":"mesh_hierarchy"},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"MeSH hierarchy network — mesh_hierarchy","text":"data.table following columns: from_id character vector representing MeSH identifier   parent term. to_id character vector representing MeSH identifier   child term.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"MeSH hierarchy network — mesh_hierarchy","text":"MeSH hierarchy data obtained National Library Medicine's   Medical Subject Headings (MeSH) database (https://www.nlm.nih.gov/mesh/).","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"MeSH hierarchy network — mesh_hierarchy","text":"`mesh_hierarchy` data frame edgelist network, row   represents directed edge parent MeSH term (`from_id`) child   MeSH term (`to_id`). structure captures hierarchical relationships   MeSH terms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_hierarchy.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"MeSH hierarchy network — mesh_hierarchy","text":"","code":"library(labyrinth)  data(mesh_hierarchy, package = \"labyrinth\")  # Preview the data head(mesh_hierarchy) #>    from_id   to_id #>     <char>  <char> #> 1: D001829 D059925 #> 2: D001829 D001940 #> 3: D001940 D042361 #> 4: D001940 D009558 #> 5: D001829 D005121 #> 6: D005121 D000672  # Draw the network plots using igraph # \\donttest{ library(igraph) mesh_network <- graph_from_edgelist(as.matrix(mesh_hierarchy)) plot(mesh_network)  # }"},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":null,"dir":"Reference","previous_headings":"","what":"MeSH synonyms — mesh_synonyms","title":"MeSH synonyms — mesh_synonyms","text":"data frame containing Medical Subject Headings (MeSH) identifiers   associated synonyms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"MeSH synonyms — mesh_synonyms","text":"","code":"mesh_synonyms"},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"MeSH synonyms — mesh_synonyms","text":"data.table following columns: mesh_id character vector representing MeSH identifier. mesh_synonyms character vector representing synonyms   associated MeSH identifier.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"MeSH synonyms — mesh_synonyms","text":"MeSH IDs obtained National Library Medicine's Medical   Subject Headings (MeSH) database (https://www.nlm.nih.gov/mesh/).   synonyms obtained Cochrane Library   (https://www.cochranelibrary.com/advanced-search/mesh).","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"MeSH synonyms — mesh_synonyms","text":"`mesh_synonyms` data frame provides mapping MeSH identifiers   corresponding synonyms. row represents MeSH identifier   (`mesh_id`) associated synonyms (`mesh_synonyms`), can   used text mining, information retrieval, applications   involving MeSH terms.","code":""},{"path":"https://labyrinth.yinchun.su/reference/mesh_synonyms.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"MeSH synonyms — mesh_synonyms","text":"","code":"library(labyrinth)  data(mesh_synonyms, package = \"labyrinth\") #> Warning: strings not representable in native encoding will be translated to UTF-8  # Preview the data head(mesh_synonyms) #>    mesh_id                             synonyms #>     <char>                               <char> #> 1: D045743                 scleroderma, diffuse #> 2: D045743          scleroses, diffuse systemic #> 3: D045743 diffuse cutaneous systemic sclerosis #> 4: D045743            sudden onset sclerodermas #> 5: D045743            scleroderma, sudden onset #> 6: D045743      systemic sclerosis, progressive"},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":null,"dir":"Reference","previous_headings":"","what":"NCBI Gene Information — ncbi_info","title":"NCBI Gene Information — ncbi_info","text":"data table containing meta information NCBI genes, including   type associated species.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"NCBI Gene Information — ncbi_info","text":"","code":"ncbi_info"},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"NCBI Gene Information — ncbi_info","text":"data.table following   columns: type character vector representing type species. species character vector representing species NCBI   supports online platform.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"NCBI Gene Information — ncbi_info","text":"data obtained National Center Biotechnology Information   (NCBI) Gene database (https://www.ncbi.nlm.nih.gov/gene).","code":""},{"path":"https://labyrinth.yinchun.su/reference/ncbi_info.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"NCBI Gene Information — ncbi_info","text":"","code":"library(labyrinth)  data(\"ncbi_info\", package = \"labyrinth\")  # Preview the data head(ncbi_info) #>                type                                   species #>              <char>                                    <char> #> 1: Archaea_Bacteria                                   Archaea #> 2: Archaea_Bacteria                                  Bacteria #> 3: Archaea_Bacteria Escherichia_coli_str._K-12_substr._MG1655 #> 4: Archaea_Bacteria               Pseudomonas_aeruginosa_PAO1 #> 5:            Fungi                                Ascomycota #> 6:            Fungi                             Microsporidia"},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":null,"dir":"Reference","previous_headings":"","what":"Protein-protein interaction network — ppi","title":"Protein-protein interaction network — ppi","text":"sparse adjacency matrix representing protein-protein interaction network   compiled multiple databases.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Protein-protein interaction network — ppi","text":"","code":"ppi"},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Protein-protein interaction network — ppi","text":"dgCMatrix object   Matrix package, value 1 indicates   connection (interaction) two genes, 0 indicates   connection. rows columns matrix correspond gene   identifiers, matrix square symmetric, representing   undirected adjacency network.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Protein-protein interaction network — ppi","text":"protein-protein interaction data compiled following   databases: KEGG (https://www.genome.jp/kegg/) Reactome (https://reactome.org/) Biocarta (https://cgap.nci.nih.gov/Pathways/BioCarta_Pathways) NCI (https://cancer.gov/-nci) SPIKE (https://www.cbrc.kaust.edu.sa/spike/) HumanCyc (https://humancyc.org/) Panther (http://www.pantherdb.org/)","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Protein-protein interaction network — ppi","text":"`ppi` data set sparse adjacency matrix representing   protein-protein interaction network. network compiled seven   databases: KEGG, Reactome, Biocarta, NCI, SPIKE, HumanCyc, Panther. rows columns matrix correspond gene identifiers,   order genes consistent across rows columns. matrix   square symmetric, representing undirected network,   interaction gene gene B represented 1   (, B) (B, ) positions matrix.","code":""},{"path":"https://labyrinth.yinchun.su/reference/ppi.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Protein-protein interaction network — ppi","text":"","code":"library(Matrix) data(ppi, package = \"labyrinth\")  # Access a specific element of the matrix dim(ppi) #> [1] 12836 12836 ppi[1, 2]  # Interaction between gene 1 and gene 2 #> [1] 0  # Convert to a dense matrix for visualization or analysis dense_ppi <- as.matrix(ppi) #> Warning: sparse->dense coercion: allocating vector of size 1.2 GiB"},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":null,"dir":"Reference","previous_headings":"","what":"Predict drug response scores based on disease weights — predict_drug","title":"Predict drug response scores based on disease weights — predict_drug","text":"function predict drug response scores based disease weights using   multiple methods. takes vector disease weights sparse matrix   representing drug-disease interaction model, calculates predicted   drug scores using one three methods: random walk restart (`rwr`),   original spreading activation (`sa`), Spread-gram (`sg`).","code":""},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predict drug response scores based on disease weights — predict_drug","text":"","code":"predict_drug(   disease_weights,   model,   method = c(\"rwr\", \"wrwr\", \"sg\", \"sa\"),   restart_prob = 0.7,   threshold = 1e-06,   max_iter = 1e+06,   loose = 1,   print_weight_only = FALSE )"},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Predict drug response scores based on disease weights — predict_drug","text":"disease_weights named vector disease weights. names vector correspond disease IDs `disease_ids` dataset. model square matrix (dgCMatrix pre-trained model. method character string specifying prediction method use.   drug scores can predicted using one three methods: random   walk restart (either `rwr` `wrwr`), original spreading activation   (`sa`), Spread-gram (`sg`). `rwr` method provides quick   calculation estimating drug scores. criterion using   quick calculation met, falls back `wrwr` method. Default option `rwr`. restart_prob restart probability random walk restart method. Default 0.7. threshold convergence threshold iteration. Recommended value 1e-6 `rwr` 1 `sg`. max_iter maximum number iterations. Default value 1e6. loose loose parameter original spreading activation method. Default 1.0. print_weight_only logical value indicating whether print predicted drug weights. TRUE, one numeric vector drug weights returned. FALSE, data frame drug IDs, drug names, drug weights returned. Default value FALSE.","code":""},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Predict drug response scores based on disease weights — predict_drug","text":"return value based `print_weight_only`. TRUE, one  numeric vector drug weights   returned. FALSE, data frame drug   IDs, drug names, drug weights returned.","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/predict_drug.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predict drug response scores based on disease weights — predict_drug","text":"","code":"# Load example data to the environment data(\"drug_annot\", package = \"labyrinth\") data(\"disease_ids\", package = \"labyrinth\")  # \\donttest{ # Load models to the environment model <- load_data(\"model\")  # Construct disease weights disease_weights <- sample(c(rep(0, 50), rep(1, 2)), 1098, replace = TRUE) names(disease_weights) <- disease_ids  # Predict drug scores based on disease weights drug_weights <- predict_drug(disease_weights, model, method = \"rwr\") #> normalizing column vectors! #> normalizing column vectors! #>  #> Attaching package: 'diffusr' #> The following objects are masked from 'package:labyrinth': #>  #>     assert_dgCMatrix, is.dgCMatrix # }"},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":null,"dir":"Reference","previous_headings":"","what":"Graph diffusion using a Markov random walk — random_walk","title":"Graph diffusion using a Markov random walk — random_walk","text":"Markov Random Walk takes inital distribution \\(p_0\\) calculates stationary distribution . diffusion process regulated restart probability \\(r\\) controls often MRW jumps back initial values. source code brought diffusr v0.2.1.","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Graph diffusion using a Markov random walk — random_walk","text":"","code":"random_walk(   p0,   graph,   r = 0.5,   niter = 10000,   thresh = 1e-04,   do.analytical = FALSE,   correct.for.hubs = FALSE,   allow.ergodic = FALSE,   return.pt.only = FALSE )"},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Graph diffusion using a Markov random walk — random_walk","text":"p0 \\(n \\times p\\)-dimensional numeric non-negative vector/matrix representing starting distribution Markov chain (need sum one). graph \\(n \\times p\\)-dimensional numeric non-negative adjacence matrix (dgCMatrix) representing graph r scalar \\((0, 1)\\). restart probability Markov random walk restart desired niter maximal number iterations computation Markov chain. thresh reached, niter used stop criterion. thresh threshold breaking iterative computation stationary distribution. absolute difference distribution time point \\(t-1\\) \\(t\\) less thresh, algorithm stops. thresh reached niter, algorithm stops well. .analytical boolean stationary distribution shall computed solving analytical solution rather iteratively correct..hubs TRUE multiplies correction factor nodes, random walk gets biased nodes high degree. case original input matrix normalized : $$ P(j | ) = \\dfrac{1}{\\text{degree}()} \\times   \\min \\left(1, \\dfrac{\\text{degree}()}{\\text{degree}(j)}\\right)$$ Note consider edge weights. allow.ergodic Allow multiple components graph. return.pt.Return pt .","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Graph diffusion using a Markov random walk — random_walk","text":"returns list following elements  p.inf  stationary distribution numeric vector transition.matrix column normalized transition matrix used         random walk","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Graph diffusion using a Markov random walk — random_walk","text":"Tong, H., Faloutsos, C., & Pan, J. Y. (2006), Fast random walk restart applications. Koehler, S., Bauer, S., Horn, D., & Robinson, P. N. (2008), Walking interactome prioritization candidate disease genes. American Journal Human Genetics","code":""},{"path":"https://labyrinth.yinchun.su/reference/random_walk.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Graph diffusion using a Markov random walk — random_walk","text":"","code":"# count of nodes n <- 5 # starting distribution (has to sum to one) p0    <- as.vector(rmultinom(1, 1, prob = rep(.2, n))) # adjacency matrix (either normalized or not) graph <- matrix(abs(rnorm(n * n)), n, n)  # computation of stationary distribution pt    <- random_walk(p0, graph) #> Error in ensure_igraph(graph): Must provide a graph object (provided wrong object type).  # draw histogram of the `pt` vector hist(pt$p.inf) #> Error in pt$p.inf: object of type 'closure' is not subsettable"},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":null,"dir":"Reference","previous_headings":"","what":"Sigmoid function of Spreadgram — sigmoid","title":"Sigmoid function of Spreadgram — sigmoid","text":"Sigmoid function one basic function Spreadgram algorithm.   function calculates Sigmoid function specific node given   activation rates nodes neighbors.   equation writes   $$\\sigma (x,y) = {{\\exp [(x)(y)]} \\{1 + \\exp [(x)(y)]}}$$.","code":""},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Sigmoid function of Spreadgram — sigmoid","text":"","code":"sigmoid(ax, ay, u = 1)"},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Sigmoid function of Spreadgram — sigmoid","text":"ax vector activation rates x, set y's neighbors ay scalar activation rate node y. u \\(x \\N(y)\\) \\(u = 1\\), otherwise \\(u = 0\\).","code":""},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Sigmoid function of Spreadgram — sigmoid","text":"vector sigma","code":""},{"path":"https://labyrinth.yinchun.su/reference/sigmoid.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Sigmoid function of Spreadgram — sigmoid","text":"","code":"sigmoid(1:4, 2) #> [1] 0.8807971 0.9820138 0.9975274 0.9996646"},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":null,"dir":"Reference","previous_headings":"","what":"A portion of the phonological network examined in Vitevitch (2008). — speech","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"Depicted word *speech*, phonological neighbors *speech*,   phonological neighbors neighbors. taken   spreadr package.","code":""},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"","code":"speech"},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"Adjacency dgCMatrix   representing unweighted directed graph. Inside adjacency matrix,   row column matrix represents node graph.   34 nodes 96 edges, self-loops inside graph.","code":""},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"Chan, K. Y., & Vitevitch, M. S. (2009). influence phonological   neighborhood clustering coefficient spoken word recognition. *Journal   experimental psychology: Human perception performance*, *35*(6),   1934–1949. https://doi.org/10.1037/a0016902. Siew C. S. Q. (2019). spreadr: R package simulate spreading activation   network. *Behavior research methods*, *51*(2), 910–929.   https://doi.org/10.3758/s13428-018-1186-5.","code":""},{"path":"https://labyrinth.yinchun.su/reference/speech.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"A portion of the phonological network examined in Vitevitch (2008). — speech","text":"","code":"library(labyrinth)  data(\"speech\", package = \"labyrinth\")  dim(speech) #> [1] 34 34  # Draw the network plots using igraph # \\donttest{ library(igraph) speech_graph <- graph_from_adjacency_matrix(as.matrix(speech)) plot(speech_graph)  # }"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":null,"dir":"Reference","previous_headings":"","what":"Simulate spreading activation in a network (Main function) — spread_gram","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"ACT spreading activation formula represented Equation 1:   $$(y) = \\sum_x {f(x,y) \\cdot (x)} + c(y)$$,   \\(c(y)\\) represents baseline activation y.   \\(\\alpha\\) represents proximity node network.   \\(t\\) represents iteration number","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"","code":"spread_gram(   graph,   last_activation,   loose = 1,   max_iter = 1e+05,   threshold = 1,   threads = 0,   verbose = TRUE )"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. last_activation vector containing last time activation rates nodes. sequence matrix. loose scalar numeric 0 1 determines loose (weight) calculation process. max_iter Max iteration times. threshold End threshold threads scalar numeric indicating parallel threads. Default 0 (auto-detected). verbose Show verbose message","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"numeric vector contains new activation","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Simulate spreading activation in a network (Main function) — spread_gram","text":"","code":"# The graph G data(\"graph\", package = \"labyrinth\")  # The initial info last_activation = c(2, 4, 3, 2, 2, 1, 5)  results <- spread_gram(graph, last_activation) #> Iterated #0 times. Current loss: 0.0373226717617934 #> Convergent at #0 times. Current loss: 0.0373226717617934"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":null,"dir":"Reference","previous_headings":"","what":"Simulate spreading activation in a network (Only once) — spread_gram_1","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"ACT spreading activation formula represented Equation 1:   $$(y) = \\sum_x {f(x,y) \\cdot (x)} + c(y)$$,   \\(c(y)\\) represents baseline activation y.   \\(\\alpha\\) represents proximity node network.   \\(t\\) represents iteration number","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"","code":"spread_gram_1(graph, last_activation, loose = 1)"},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. last_activation vector containing last time activation rates nodes. sequence matrix. loose scalar numeric 0 1 determines loose (weight) calculation process.","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"numeric vector contains new activation","code":""},{"path":"https://labyrinth.yinchun.su/reference/spread_gram_1.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Simulate spreading activation in a network (Only once) — spread_gram_1","text":"","code":"library(labyrinth)  # The graph G data(\"graph\", package = \"labyrinth\")  # The initial info last_activation = c(2, 4, 3, 2, 2, 1, 5)  results <- spread_gram_1(graph, last_activation)"},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":null,"dir":"Reference","previous_headings":"","what":"Calculate the received activation in Spreading Activation (f) — transfer_activation","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"function calculates next-time received activation value node y node x specific graph, based activation values neighbors (including y) weighting factor loose. formula \\(f^{t+1}(x,j) = l \\cdot s_k / \\sum_{j} {s_j}\\).","code":""},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"","code":"transfer_activation(graph, y, x, activation, loose = 1)"},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"graph square matrix (dgCMatrix representing background graph. Inside adjacency matrix, row column matrix represents node graph. values matrix either 0 1 (either 0 larger 0), value 0 indicates relations two nodes. diagonal matrix 0, self-edges graph. y index ID node specify attention place x index ID neighbor node y activation vector containing last-time activation values nodes. sequence matrix. loose scalar numeric 0 1 determines loose (weight) calculation process.","code":""},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"scalar value representing activation value node x","code":""},{"path":"https://labyrinth.yinchun.su/reference/transfer_activation.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Calculate the received activation in Spreading Activation (f) — transfer_activation","text":"","code":"# make an adjacency matrix and randomly fill some cells with 1s mat <- matrix(sample(c(0,1), 100, replace = TRUE), 10, 10) diag(mat) <- 0 # remove self-loops  transfer_activation(mat, 3, 2, 1:10) #> [1] 0.1034483"},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":null,"dir":"Reference","previous_headings":"","what":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"Update gene symbol using NCBI Entrez Gene","code":""},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"","code":"update_gene_symbol(   gene.pool,   ncbi = NA,   species = c(\"Homo_sapiens\", \"Mus_musculus\", \"Pan_troglodytes\", \"Rattus_norvegicus\",     \"Archaea\", \"Bacteria\", \"Escherichia_coli_str._K-12_substr._MG1655\",     \"Pseudomonas_aeruginosa_PAO1\", \"Ascomycota\", \"Microsporidia\", \"Bos_taurus\",     \"Penicillium_rubens\", \"Saccharomyces_cerevisiae\", \"Anopheles_gambiae\",     \"Caenorhabditis_elegans\", \"Drosophila_melanogaster\", \"Canis_familiaris\",     \"Sus_scrofa\", \"Danio_rerio\", \"Gallus_gallus\", \"Xenopus_laevis\", \"Xenopus_tropicalis\",     \"Arabidopsis_thaliana\", \"Chlamydomonas_reinhardtii\",       \"Oryza_sativa\",     \"Zea_mays\", \"Plasmodium_falciparum\", \"Retroviridae\") )"},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"gene.pool gene list ncbi location NCBI file directory species Optional. specify species","code":""},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"vector latest gene symbols","code":""},{"path":[]},{"path":"https://labyrinth.yinchun.su/reference/update_gene_symbol.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Update gene symbol using NCBI Entrez Gene — update_gene_symbol","text":"","code":"library(labyrinth) update_gene_symbol('IL-6') #> [1] \"IL6\""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v030","dir":"Changelog","previous_headings":"","what":"labyrinth v0.3.0","title":"labyrinth v0.3.0","text":"Updated docs","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v023","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.3","title":"labyrinth v0.2.3","text":"Added predict() function Added several data Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v022","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.2","title":"labyrinth v0.2.2","text":"Added multi-threaded capability Added random walk Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v021","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.1","title":"labyrinth v0.2.1","text":"Performance improvement Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v020","dir":"Changelog","previous_headings":"","what":"labyrinth v0.2.0","title":"labyrinth v0.2.0","text":"Rewrote codes Dropped robust z score Applied Eigen MKL","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v011","dir":"Changelog","previous_headings":"","what":"labyrinth v0.1.1","title":"labyrinth v0.1.1","text":"Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v010","dir":"Changelog","previous_headings":"","what":"labyrinth v0.1.0","title":"labyrinth v0.1.0","text":"Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v005","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.5","title":"labyrinth v0.0.5","text":"Bug fix Remove stopword file","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v004","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.4","title":"labyrinth v0.0.4","text":"Bug fix Added stopword file","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v003","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.3","title":"labyrinth v0.0.3","text":"Bug fix","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v002","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.2","title":"labyrinth v0.0.2","text":"Added Windows support Added C++ version robust z score","code":""},{"path":"https://labyrinth.yinchun.su/news/index.html","id":"labyrinth-v001","dir":"Changelog","previous_headings":"","what":"labyrinth v0.0.1","title":"labyrinth v0.0.1","text":"Initial submission","code":""}]
diff --git a/sitemap.xml b/sitemap.xml
index 5992352..bfa4394 100644
--- a/sitemap.xml
+++ b/sitemap.xml
@@ -81,9 +81,6 @@
   <url>
     <loc>https://labyrinth.yinchun.su/reference/mesh_synonyms.html</loc>
   </url>
-  <url>
-    <loc>https://labyrinth.yinchun.su/reference/model.html</loc>
-  </url>
   <url>
     <loc>https://labyrinth.yinchun.su/reference/ncbi_info.html</loc>
   </url>