From 9125c1d251661072d17486fe23cabc7d0da713d7 Mon Sep 17 00:00:00 2001
From: luisgruber <luisgruber@users.noreply.github.com>
Date: Wed, 17 Jan 2024 23:03:48 +0000
Subject: [PATCH] =?UTF-8?q?Deploying=20to=20gh-pages=20from=20@=20luisgrub?=
 =?UTF-8?q?er/bayesianVARs@15dc677e127513df78de574aa1ab8635f654de9b=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                                      |   2 +-
 LICENSE.html                                  |   2 +-
 articles/bayesianVARs-vignette.pdf            | Bin 450075 -> 450118 bytes
 articles/index.html                           |   2 +-
 authors.html                                  |   6 +++---
 index.html                                    |   2 +-
 news/index.html                               |   8 +++-----
 pkgdown.yml                                   |   2 +-
 reference/bvar.html                           |   2 +-
 reference/coef.html                           |   2 +-
 reference/fitted.bayesianVARs_bvar.html       |   2 +-
 reference/index.html                          |   2 +-
 reference/my_gig.html                         |   2 +-
 reference/pairs_predict.html                  |   2 +-
 reference/plot.bayesianVARs_bvar.html         |   2 +-
 reference/plot.bayesianVARs_fitted.html       |   2 +-
 reference/plot.bayesianVARs_predict.html      |   2 +-
 reference/posterior_heatmap.html              |   2 +-
 reference/predict.bayesianVARs_bvar.html      |   2 +-
 reference/print.bayesianVARs_bvar.html        |   2 +-
 reference/print.bayesianVARs_predict.html     |   2 +-
 .../print.summary.bayesianVARs_bvar.html      |   2 +-
 .../print.summary.bayesianVARs_predict.html   |   2 +-
 reference/specify_prior_phi.html              |   2 +-
 reference/specify_prior_sigma.html            |   2 +-
 reference/stable_bvar.html                    |   2 +-
 reference/sub-.bayesianVARs_coef.html         |   2 +-
 reference/sub-.bayesianVARs_draws.html        |   2 +-
 reference/summary.bayesianVARs_bvar.html      |   2 +-
 reference/summary.bayesianVARs_draws.html     |   2 +-
 reference/summary.bayesianVARs_predict.html   |   2 +-
 reference/usmacro_growth.html                 |   2 +-
 reference/vcov.bayesianVARs_bvar.html         |   2 +-
 search.json                                   |   2 +-
 34 files changed, 37 insertions(+), 39 deletions(-)

diff --git a/404.html b/404.html
index 71fbc5d..3add036 100644
--- a/404.html
+++ b/404.html
@@ -31,7 +31,7 @@
     
     <a class="navbar-brand me-2" href="https://luisgruber.github.io/bayesianVARs/index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/LICENSE.html b/LICENSE.html
index ba7edc0..bd31220 100644
--- a/LICENSE.html
+++ b/LICENSE.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/articles/bayesianVARs-vignette.pdf b/articles/bayesianVARs-vignette.pdf
index 04dceaab2d90679f58caa6ded5aaa0b49d136dfa..ac5bbcb0b50c3a1b96513dc9e9f87a36ae9af48d 100644
GIT binary patch
delta 66573
zcmc$_dpOf?{6GHge0n>JN)bZJNQH7*LP*FdnMQIZBxf_P4hpG8L<lwHY+)l~R;g4D
zYffXsavWxcv9a0q+xz#ue!uH;UEk~a{qg(fdtLW+U)$?--}m$Rd_14`Yr7wNte{Qd
zN1MV&lr!+<W&asT9OtywwpXW9>GeLHv!Az_Nd-zAc=qYIcvw~psIT;d=AQN<o5j-e
z+m&mtpMcHA&-Og$%i6z8&Wb(VKD^D%{=q+o{H?1l*75omc>6gMC&E<Ef4_2@CD=J_
z*T_w!MrH3T@>NeLMCp6vNp;o8<xlT<weQu{f6qRe-et3Sxmgew)D^9g{bBijVd<@Y
zLc7Nf+^65dC|vl$J@V17o-sIZOS{%1I=wWqJo4bhUhCcHWnKspMf&=u{Vq`R8d<k+
zC}*2_R;-H>uv?X@CwxG9ko&pD$-6BRa^JjlkK45v55>;e6U`4p{{aDV&rN1Gpa&=X
zSlgnbOP|f>ES+jrQo-+P3*;u9lE2cTIkbp<C^c37?(4et&0Ou<Ig0uEmsWn)2kHMC
z|5~as@3Uhr40ZxA;;s$XG;f9<8OVwFN9DEu=;wFnZ4SV>w3UjfZ1)grk5@aMu`=y`
zNLlToKY3KPYqz`0a(vv>%8f&v+OvMsc(V+-sp9cl%G+Z(vASMC|7M#j9lb0Nd$d^q
zI!?sonT1^l&421y)b@2xe;$$_`39uxalfhWyOE%EwfL&$`@mOcdz?XjQ3t0~4n8G>
z{lF0Qv4?krDt`7AnqNr=MkcLPPn`J<vN<tK_n9yFT2bk>*01@zvB9-z+jjl4S&lO*
z00Y-mr)P7&%z#k7k<#~0Ty=D6wtrSE*i7uIYx2&$M=%4ylCEOt3mJ-D6YcFY@)XyR
z*EUFX;~j$UKaT~3OGx0~z`ME=g)2;85iFSR1o-oSV8shqoC%=-rM%6nJcMd4s$|QW
zMB;C400NYfkd6@cg^}Bv7*>W|fEzjKD9IUBH$@-$088Rl0|GCBfXd&XQw1*J<4N5@
z0UIUk;CV4wokCu}PylQVCv6pj8DIgQ2ME{t$J;jnZrTlF9%+4RW5<i%9=sP+;l>ID
zq3{wFwRyi~V#^RQ|1_p$^m_aKys*GrmY5n)Hm?xXGI|eSP@(|y*0*6lDQwr?q#e0k
zVvLp_VpF=pG3H#SaDxDpDhW5ig4|-k0<(|DPT~}=*^{d`?0C&k=KN`VD3+>UriABp
zAcRW>^LXJ(E5Mx(2T1K)!X{<4X*LX?2%`?G0vq!eicyiQbtW8IT+#-=XyVtpMPNqa
z#7?yUK!;K9s@oEY@vGxld$(7oB^4U#!ZzZre^M)?k@Vm;c*W-FsSU#q;q#=a9883>
zPbVnS>!tWn99OUuf$yK&O6VPW4QuJ8F3uZJ1f8CpO1x+M(3|%3*ik}w6J0*hD#4>Y
zB_^AZi%~6K!+;pplpMg_=w)m^n8Q8B0|;xKM$)-D#Aq^A2R0rN4L2!+cTt5s;H_xk
zBXEc=uNp55HDoFYIbcLklaLscQ;ed;<)+X>jOZEE_PPOWLgUK~7q=H{N4)~-a;@mI
zoK;+9Z;Igrog9fSV4_+D3?A46rxm(tOdz>x5L$q;@>+Ao>^UG5l0yat45`1~P^d$$
z3+1dTz(vQIUllu!@i$(l0m6O^oZ;5my<T0%uRh9Ht?-KW)0^UtReKiIMlQ@a>rhc`
z(UClgQ*o=O>H6yU=<`19ei8=LRok?XWVKLo1Je}}wENs9f2}68y^B=q_X)EJ{o-+M
zyo>x(BgD0xemY49AfR3}Ul+3ZRIBibO2l<aUD0T8b1vgFHP{u`Qi_6xzt@|Bc6yw%
zkA_~PQv$sPwV6%(KT&KQ&lNSH@vQ?!)^TDj&~j7&$*VkuFP_|I8jfqmo1^t+T)dbR
zo~2c5dvOf3{VF`8Mli~O*8cLMjn_qbfEONMDeII4pb^?Uw8X1rx0#HQS-sD#Oft!w
zG6p?yjmHp_mho|6xVCb0Mtn*nPq6lK7-}*?E?7d@5+a*uww8?e;${q(8SPwC?1k+O
zeG#YqV%|h4m|(-Z&wH=$MP|<E>yx8-yx0UriwRJ0iC153QfdD+p^z>yNHBmW4V8TY
zk^tBs&*=LruE9KSm`<&vCJ7U3C5p;TsCffLH2ToG(W7<aG7PgYptZ<~$x8(PVU5<B
zpwyx&fns(W)5XXhhZt&)ty3F!fGIJ3Ll_t1z;~g{zv1*wLB0!Za0yJN;#r_`(UhlL
z`%gFpgLOmFCXw7<Bs8U8s-yw{noVscPl;_7MrqD-C{!@a<2qF^z>7D^XZymF2BPac
ztHFa(45kM#m{yNC7zWUKSZG3I2rxqbjAA-4IfaRx#3W`QRoL4;R8+X)R7d{9jSpk!
zFw==!L8pJWPF0w=K?4_}gJC7*2?;c6d$i9I&UCN{xtY}z3%o1B4$+u#z1TX219Qt3
zl1<*a%6#gKKJ|LE4q<qmy!xj-9KVoWK?*e<D6GVkR|xx!g!+8`ps53lA{;+h)q~Hz
zA3}f?k~w%56TY&Z&`%3}$gI9|mEcBX&26%V$j*3_iGzLgaSXee3Zr?A(+D~k+6Vq-
z-RZy(z$=(k`<}t0(3o3BYsLYhWC<7vMNO|uV>?b8JA?=$T<J+`)Rj*rp^2rj0>7bh
zHk`RdLHnRLbH|}fVgh9dObLyK2NQU`c=IVc=2RijS`-k%fXB8Y@gx?svi(gH97BZ_
zzCLG%7xv<71ff&9uDP>AtR^bZj0da;{X^AJ%wUHgxxWrQvc7HeT<Dx)4{2yT_CTpg
zD7<3#uO4c%AgPIK?vsd<QL6+ZTZM(zXxb31ye1R5UIuCl3ijgP!Ibwf*Lv|#Q9?N@
zOHZEAf*7#SEm%@fte}v~YZ@z$6?kVGuXCyoGzPU{_{w;mGNXkV0~Gq@W@EdD=#7Wf
z!_Scua;ckEuusCwAg2NFCPso8v_%t5RDE3nRcf%m3VlkCKG(It59N0=lQ!n!$#DeK
zvA7^GjLgKbbigG-mXR>#h0u<aVprBwJQI9wiA6)yWZN=N<5@aEU@e9qRuD=Cvqiyy
z6K@tr!O6hwt61R;*vD=E&CK%%HJ*(HYaJ1m!3E@k=AwXbKB9kh^lb8~k?@{BWuuF^
zRn1u8zCsKUYw=_niWuL}K5y9N8NcwcmQ+r5-oX7gJ3LYF(^GSeX(YUBieZZ^_W&Z`
z{ZY5=XX)AeZwWbGbk+bhycXw{i39rn6>zwsoc*vpM_6rD>MQcMm?dq;`F6`ey&BvU
ze{}LIOqzedl4xFyZZ{oPTcz}1Gk>_b)yg*tzpD7?cBR&kLQT5XEon^vgS(Z9yZ>wa
zcequ5QIkLf+_403kKpG_2JUM@K6SG>2p-IKP3j@GUSsH3viB2!HWr{0(srJ=7)PVl
ztqidnOR(e5n`evL*mGqoP4i5p=JFoWLa`dIxT)yq;KZFa6NVur2^$Un?HMnMRWK8T
zGhrjy#z!PNTa`>jWm|L%Ws)%F3>)8!ecS(EPJ&Dgcr(`b1B#nMNpuxhSz!BrBRqWs
z6rbh;_otl!_A^R>j>rsC72M$VQR$_~>CQuhP9veyk*R(XlY3(|O32G^j^wuYg(=`U
z;_2+9u_6wWk~GC`C-sc4u?Fzea#}4F%;|*HavT8H8`1WB_n2Gu+*Z%xeqNkV@VlQ^
zMacTVaZd>?8jp>xTnaKCV2Aa`RZ@l5rhrP$o+(CXF*kSRX3U7ZBFe8nVRB`W-!X+t
zqU!aeaJVBMHk~MdJ*%?Gz>ZaG$~5iiT;b%S)au9271I-1dRY8BI3JUWq<Ots`Dh&c
z!+Pmmbjtb%pf_Qy$&FRGUgcRHWKvCV{;*yLmY{ndQ(`5!^XU0Q{T!CRLmYssB>?vH
z2&Tgjoy#Yni#Q=eJhE31S)$6l-i~L90}9w#4UIrORd%bnUCFO@4;=ryhOijG<X~17
z_;rg79GudPbF@5mSHsFhUeGbo)bOreo<Dernak(mhM0}TB=33$9vp|Sr48>`XAPMy
zJ|Tk}lzbhszj0pxIAXC9tkkzBmU0@Nc!#hiz~CZ(j$=$*mUHjw406VjsBD)Zt`DqQ
zki;`xtS5Vp_zZ`n%$74+BJhhi4t7&@OtE`IXE<Fg#CW8rh7r!zWPj)^W)Kr<d9lq_
z?33&ok^`kxEeODv+>#bZu^T$m8S`8Y*K~xU76b<xO*{byQUi$|_Zp`IYkyhHh7I%9
zlb~HnCqoa31MizqBcu%(PcKIJ2A`}{8Frs%x^<8c&N~-vQphiiZNsd$@ywslzI&ab
zEI7pToA9nNMe;yR@1D)bRnkISAT<T-MlIhm+rm8u9)X9c@`5`+@kW-~gwjO#kRTu6
z6AgiB#3agpOMYu40l($E1x4s=O$d@!a(TQ)a3NX{1yH621#vm>U;tM-Sj=R~0QeKN
zowM8E#vKP1xPwU&Ma!NcTb~sc5=iYzijFi^_vzmcg?kAVq+0m$OVgsi%4)xg^>j1~
z7FoT{%K=8=@qbO04>6l(^`iMe$ky(nsr7GdH2%32i@58$R9&4@t0Y5*N6l{x1~*Tb
zy50pBPC7R;UI)y&M}zyLn{Y_>=Z>q7%D;4tciv?;#Y1#akD3Q*A^1IrMiU=Tw_*p<
zWc)gu(1sdEg{*uw1P6GYoHmY|<iQae@8hN@iYwZ2zQ1ORzcZwYlLLUEO|n37++BZ!
zRG9Kr-cSd=#aVXQg0qIO0ERG8A1z2DJR{wJ)!~4+I$BD6_(fk~{vOi$rNo<UpL~1V
zSkq)7gEP@ngr6ny`5o2tGj~bLivrGa7{P#bk5WB5gTa$I1?BP^74fk5@Z}9Gqq}XK
z-4dS!SE>aj_&{{xf?$jhL|3?P#;CKoR?*mOD&}Mky+tmUR#_OX#Xq^`s8j|&CDoCm
zAZ^9?L|)l@e0^5^!CQUje1%7Gldk)NW<sx=gGhQ-wNbRyx`vZ4p;w+ImpU6H$7p^j
zKQY4jX{J_UWTRD8f`g8aU8n>n|8eQO0>9*2VYo)=X$3%8LkiY6)q-xT-Wws92^u^e
zU+_PDc)rHhOez_}mCKh<9jQ$s->Pc$u<qHB>FXx{*5*$&{mGV{-j31aL&LS0&L7s1
zLjo2;h!&KCzkyS|<dd|{tDTR6KVQzMKX7@^lO+!$gD7~S?gZvZ?h(&T!{cA?r9IaC
z7js6k;|vto6LDxVVtewulsdr>^YvSL|JiHgGi4pgUb&AyK@#iKBQR;zj<;&R2Ftm%
zA_mXY<wWR-Rfpbs0lq>1l;1gMy0qK7IkojyXP8WE$tRB|V_HiwO0C|bpsBDUv6PPr
zonu@d14Jtvy&H*<tVs(DO}wc}grp5!`xt}NTk{41mLaDg9>gODh9Kx!(;8f$*p(1n
zNY)qf^P#yE%#qrWKt|orJ#-GZc0MrFD#RE<|B{aw5~LVN;-1Xe?F?5<P*obgy5^5;
zmchOM3$hGf=%73szqvMqG>~t@&50$&r4tLDkNYoQDQV_T^jrxyNa!GXh>u6DHG$h!
z=4t_}aMJ|d!`&8;WiuI-(A4uPf1X*NGYzyVRnE+@$d?YZE!9;*tM|RNhTS+K8${Hw
zdxqBRtFW%#-t!DifBe)EA{PWH4So2-62dWKsW8&dpZW9ES_T~GQ<|B1<4XSFz<|>C
zO6c=_1J)j@FT~LLeWMa&g`j(--yf=5k+;l%64lz5=Z*T7E_JAagzx4vo2-r~1;R@=
zmC)d-ZH}PS;LFz)kVhPm5UIiE6)_fizOJ^&{ME`P3qAEJJ+q#)U_#0f*+FE5x+O-l
z!qKsod%Y1!J~;TXB4?K#s47$`_?8PDnJ+)kUa>o=;>j%~GZSMsy#`Mzi|HM{#vTS5
zqhS)&#m0on%|FGQfy70A#8Uxwv8b=Oe9(i-%Qoq!c0Lg<aH7weZkalgcsx@_z(i4;
zOPHD&;Es0oYr;wp21qzJ&RBHaQ=+HXw204FWR@@Xn<{BD0^se-Jh-XUMKX=@D{(!s
z&6xTIr9o6HBJ81yv2ZW(J^(~wH#m@PIVLw~iqbG3%6Nv0k4V5xux;)2o>F2Xq{ghk
z{ixVAH!lf?7sc(KiXrE(z>6tu<3;%hxQV(U3PmbMqPPi3ePB<f16!WCM}wZnNq1(I
zF*WuO&Bnuz)=uz$(#tug@ea%YdG-@dtLQL)*220gDK7}c=?6M^=Xi`SNq_klSg>(a
z<B;gMq_pb5O#=Q1oA!)So8+}$aKaoU3n>gSE13~R<$M(7bS_e&)g&?qx4`bLU2mkA
z=a!&8=GwCjD9-XsDtqXA9NHwN9z5cyify7q$`j8uakMFs!;cezW_g`j{8&-jTHDz8
z333yqqz2F!^He*6p)@8Z#KJfyxIOIU`9Q<P8Rswju_Dv7A~X)Zdwn@uIL26)5&mIs
zP0W2+WY6&!5`T)vyp~zR#TQn*7Q^2n6ULM-`Z`1dPu9zM=fd#(LUVc^|Gi<7hmh=b
zpT}^3T3C^(3)JQola0uY5P51iu(V&GIMoR79&7M@fXum#B?m5#5neQTdQ7Qj^1zsv
zS^}2WxQtk0p7m@6cPI4oXLbNX0yM=3;5||rR>J$@qkuyGdmc-JSkC6?Yz49Yv1-hh
zzh%M9*BITP&59Wiv9nT$z!2I&i;n0bM|4v3xY^omnsRge{<h>>ds*3J;DtQ!#bY6^
zproa1hu3+Oa{m4z=kgZh#CO>0Q{g^|;P{AV>hd~O)N<bUtWnQ6XAk{vsK1VuQy5m2
z70;IDZqP<(OAqYr?NDsc{ni!C^I!j{?ZF*wdodRncPqq!uCszo0G!SPQ}K+*l}mBd
zl9j>uzhVeIowj*^Wr1v*cmb^QWDCO;-MbAYUP~m77_)M4L-FN}%75QCX>{9$I^>s=
zO&8nN9~tC#dit8bA&pyb3TSd0MRDOF_y)(nUy@8&caTyOzt)o)-2e7`CoRvV2n<l|
z6R63Nbtdzo#|Oz5<oFHqnB@4)j*aAU`2g5w3a&jy5y*Vu?MX=Z1H=pLg!!Cv_*TVx
zT}e&1mVdr&d~+R5_6Uv}$dgmq52RmP`MK7V8WD2BWAe6&`#;LTBdb3(o5v>i>Arb&
zgIk)l?4<Mhdvxnz&?OR5Y~^!K)SEwYiGigv@2n^fzay<0zc}@ke7W;Jt+V-G%!f~d
z`lRrt<mj%Cjh`(Cnr2XdZU4aJaM=xaL)bLh<L2o*8#!b9w^cV65VbDZ#vCe@Thrbe
z8z@lbSPrH7JkH#&1k!dj*%EOw3{xjrVZq_S-<fyM)Eu5%7xJduhc>gl`;#Ao+g?6U
z+*GeB$nI2V&Wu@VIBWI0<n5)R2-3~_61Oj70c|Nijv{}m{kv%tKqz-b8&B7AYeN@;
z@y_&&0>@jCg~p0unjl5c15HGk2Ypl$L^-<<5f^-L{iK?2((?&@ag(nRE5?GfPR`<S
zoqnWR$#<k$SHC1kFCg#rh4iAvk1^%Oa(Syo%!~`D-8VGdPlxGsoZjdEf|MuO@$py1
zV!Et-gx5DMY<%VpAb2gilcJ)oi9k4?XbseK{F#0iVtzsKow-+O>(5^*?OBul#IwID
zmI^LWeTPqI>zjqEAl^RHhWPi)weKq&Pkv^w>#&p@sL?#oI=*cr3b)TcJ-^g=J71mX
z#ApnRPjJ41ZZl6sIA6l0X;|oluO7-A+xt0)ab<f}YkJP5Iv~vVxM4T4%;M_v?7OPF
zRA(X1iu4mXnc2r0fS)HXqeq9Q{VWEN{~WSQ&AFI-FVku41=`N&nnS#nvW1qe=iQxQ
z<ips^Pwu`aW+DQNJ<y4BV1@S=dZ1|Eo&`;Wyk@<~wHJ4s6dkzXrN+UD!>Qv^a1dl1
zQjqZ5$AGNc2k8D*S5++kM^*TvvR>xiRPu<;CrRR>#ap>!A~ebg_Iv5}u}+df>cmxX
zdvO(Cc~HL!1R~oYgPDq={8~^hgrc)$z-c**gw!0`<g_JmRd#L3esU?>=Izf8dlmcH
zmsx=t*@s>(l}g9<7Y>)cOJ15*ITfW5m-XdO;q;;`aOLfsm`1-Q!rm<UcS)gjdQP*j
z{iekr>tEto9psFE>OUuLC&r{^Jd}L;uluZLisU7}xaPe(hTn#zE~W(L<w!2<a3p4Y
zo9dLDIg`At;ge9z`$6EbzPoa|4=<$pX9cG19kJ{RwU4i12JAC4a5;>ek-0p)>!Da>
zO!~{#_l}Xke#7arF6VceuQ4*Tp8nM+KfirMyj&}0Z|qY^Vd%p+o7v!$#66lpPal*k
zt7K%X3@J~lX*84@8>;n7-se!kbrNP*RV5?6lmo_4v)P>z#Iq~L=4jcS`o3dx5~3<(
z09k=~^29#mdu1JFq%>%ko>-bxt)WHJ$-6)AN{vVY`R|g&E6=HvD91R|&m+s6Hvavx
z@7a9D)}KVW<S~t;tMM+sk;=v@ZYgUD$U0?#Js?3`m0M#XXO1~34tmK+jVjs4`{F+a
zZr+u<=#U$Dsni0Uct~iD9*Fz6_mzzq#=4=uSk7_pvjU5*w$w958N-^0OCPi+kN2^&
zt;*-zft_Dv=2RT-ch63-t<N|7w744S5Pw3gCEaCF$7H7d<^Wd)siv}+z4{5HLiTnP
z&jf`4ODdyFcqQbF$|PgpP5n=+>$gPZNEWIS@7u&{DIC13bry0Svo|O+CN-zWh8N1Y
z3Z~Dv1fcUWac!j7E&Bg#2!IN7;*<WjCfMMMhT&H>NYvgk2*>}#*4?pr+z}P9b1XNv
zvJC72B!uTw(H%5E#*qwK(&Jz-{eJrdRH3}D_m;N7c^$ve<K=!>GcY;)4f(q5^bi+A
zaKf{_uB#`6KMk)fZRPi!B38B}0yV%@^!QwbG|p!?m;3Idax?FKm2{86+Sjhc4g|sI
z(`ac;uPYLsf-3j8tF$TS<<K0Lz6@$F3UgO5IAS!m8gL`6ZfOONzPh$`4Bzzs^cyd|
zL?54mmtM|lkf9ag4Fk6{!5DZ;0w}d#EA~uG3M!n#xxQqShFFyWkE3w89NF{rGJAK-
zlc5;9Qaa99y>G@r8M@x^9+X+fsjky};|KWmnPJBZ3ZCqmCu{!scRnF}y!?1RzmF|3
zpP*B~*)vZreiFsFecRfNUnMu+*m5{><UwJ>ql~)33&z%@r9&zo^c~$r@7|rUCcTu{
zF@N>XHEXxAlO895ENi#RAt_OV?ZFX}D$sQh`s*JRsL@~9zIAdNPc~`YQ_=a41$kZW
zICMSs;Pmmn*UJ{yuzg_-YJIO$^L6L1#=TO9{>prnQRhok?0YSV5#7Avt1613eHMF|
z$?v@7HpPbznmmnS3`C!wCm$QxH-9xXQ5~A}ULCqFXaTf;zO3-=d#!JMsPwdVT7Hz;
z&xbZ{UmHa$r)H}|=kG}-TazgN?s|9$`^pk-O)|n*_8o^}@Bw|VKX%(;$Eh;=9!05<
z%tY&do6U<-%Q|oEc6Kv43P<85iC!H}j#4w;mQ^=>Tt*)iY}p+h`ig$b9Q%s?$Q&!3
zXQ~bbn6xT$@%aRv`Lv8Eq5=K+1SX#LE{(&J&E7U&1%UV1oghi2M5%2rzJ}%U9*AuH
zdK$IPOo=M<QG@=9{-P}EK8rfDe?<^DZJ)oI;gnTp;ah3rMm}Z(>U*taXpZe$!V*-V
z`y9JUPfu*ViBeN3w!=yneXwzRv<T!y4e~78EwC{^n`}t1+IQmCB>Q8oS#=9%&>|y@
z<dnQ9M(T5QXxj0^S#{`^FwzV=wapA03|r@j(zPkV;pp17`31hj{3_R5l$`-PESF=B
zjUn|sjT*!kKUSA6!ho7Bt|M&R;8d)37@^Kk<jXI!of&n>w4grV^{o_zzV#-Jp1i}+
zg7oKUQH(m6*gUziIz+Ux?DqL8K9wuc_c|ftiKw>Ppyt?xxPz6_^9zi}QG*1_uNGMD
z#-pe~rn9<j1nj)mai}LjgcHjQ%bx63$kg73G{d&p6#O^zzZe&g1yPKhxHnM@38a2$
zNG#MB5QXd8Y(sik5!{E7@eS<5IMW78Lp08gmxkzwSDg$)oW&f68buTAL`gd-vW|Il
z(gM3BxP~<b?66zy*Z#YrVu59MnjI_$z4qxfrtEvFKTlR7^~j2%cpTC<v)Mm(C|ftH
zjvt|3D6)3G#l|gMr?ksIN0g!=09zbXg3HiGX^6_$>|};E!Xc{;YualjGbQL}j0E(#
z3Oz|GP*}t<*UY-YxO-W30^^LjaV61S!ZOrFh4{Kfv(L2_w#y_nit#qxyKk^lp)>@4
zJT0S+q~$!TE(#^GyEMe$@Z+eFS%dj1d^yh&%Z+;w#qiyp7d2uEoD=yv;-fAKk!gor
z;H&tkKzWRAh{zarOC35^{4k2~@|9?RZdSb8pIJx7t9YqEZ)e$K*`D$c5rW48taGU;
zHn}*(EbUekbtrGf@Rw)>n_<)c;JwRRZqpG(zD5&WGvg^kjq+~(XX?8hHs&lMk+C@i
z@ML@ksJX|fd`_w=lVSEN8iq^a#rIvAMHh42e^GlcN-v=df&PKwXkfHlU@v}Z0v;z^
z=UqK%%ZJNlhU3F2@`@7dE@Mwokxv#-A_-HV&=5(2{q>s8YKAU?wm>-AALnqgxJV)z
zp7;3q0#|pB@*um{?5Yv%>DuN^VDDL0;*52fX>oW~#H`W9^Kg@sbw;RlBpJ*~xC0xY
z+j%*-BSxv7s9?aeKf#|@jfxeQ_BC0g>(Tv0PIx*TL(i@2*GBVXTK2(l>VXX?+DSef
z$Kk-h5B}xw?{EZ}4B-Mi4sWx_97kN=jq@7hv^enGRib|KpyZ2y#smHtDVG`kS*h$L
zK6*(fGUE<F9Y?^IcJeTtu~Y)*BbE8K@qx+Tm2t+CLK?=T9u<w`ZG2X&2fSfrU>)|g
z``6zy3cn`#z)}slO}Shaxd>U8)ZF~+_0<udv`Ejps1Zesi5o=&7TV0OJbq>I%aX%w
zodjb_(IVTB4Sb;C8N>)&tBKpkfEnuWxLQmNrGrhqtHbmah{GUx15M2rN$)9B0vLt<
z93MttS{M$GYuxV?T;k}g$hVD?i<dV(;)%R2!OtaZ-^S_g<|V>~y<-CVuSytj_S;`C
z@YxKV35R3ikWT9FonNL9@$E4{&+lgbMU`kQo&-FeUD<O^faZViph-_S7)pN?sMD=D
zR#|+9Emts#@7^o{H1l#dI&k=vTp1f>cDYIMVmN0d89dIn<<VRD@U8f-lURK5vaox{
zQ5$d1zt#s?JFS}D7_5cFC959pK7G{PrzkVAa8%I}@+Rfz36s3sx2EL>iuzC!AdO60
zU{=?_xWDabq`M&@LLJnmQF5zbkKd!~J?X2nkw3p>HDn>X3*OF$D1L9TFLcWc2rhcP
zB4Nn<>bYoKQ(UrJCsO)NeEvB?4<!r`N3<SI+l^BFRmw76?Cj-TnlIerx$MZgyj`tF
z##z0xr;SK!%Cj1uE&R?PUW}g2bJ>4}33MB6i79gJ3qoK1+x;Bh>wGbwiZ#@kD5nTR
z(YR^>kOG?1z6#d(<CPr}bzApeK3_}Jcc`O(e_NW{tGC*~a*vBfzPd^qF{S-!<4wl*
zbVw|e7Uar&r^IvTuZ6}Ly}oORD1T*7TX&J3Hx}N52F2kCBaMoe%EY^NZMA*h3EVKZ
z6?502UbYw_Bd%rsBwYTsZ2#n&w|VFGy(=v_R>N{yMHz-^=vRA%m*Z^S_M~X}<Q^sd
zNS3wfe6-^cf`54WxK6WN`o}97ueN&j+Ks1!%rAlxoBdvu8xqqbuOGPOS1E&&kxajk
zW36Uk*huvDe$hCf)#}~gk?mvI;A;iss6O%x{UJL%t|PU>u2Zq&BireNA(86<qymz7
z?X8=ER|Ds`hbEML<densk2S}H&2cyGd<{9Eqt~_?eJOcDdXMv=VDIBky|LnxXTDAC
z+}p2uttdM9Nt%4Q-y5{|)s%gB?L85gjazn;!7^?opCC^Lwd!J&s=aT4rh<>$04Sf1
z05wAqdWdTHsMyyqsn{&UnEKL{i01vR?xPITn2t{#KgP~4HF+DnY+dgRQ;g*Widlpx
z-;^PeH$GlLVx(*HZZj-G<~ktHhFs81NIkh4OkfZECX9IG`OqD-`eV%DTB1aV9;A^-
zhSp95f&fTAk^FKf4n6c(PecLpCBly<93U3#9lx@6#kB49T#rS#MuPblk7wfpDa{9Q
z^>ekB;o5O9qQcg)T#3P9T+iPZSHjmj3SNwdtvxMiKH4@v7Yf8p5j_y&iEHc~VyWoA
zkJT;6l0g>l4rcl%{&{I_t^!Isul6U$+Q&4|uJmFiyXO4KKBPmwY>;y);Fx)1SN_2u
z(jKvNv{oP5nzH>xI(q5xtR-1K==QrOnd(V@epsuj);>Rf?oY3^IXEz=v@)~l$`OUY
z(9#wqv_aor>+~IVFVUNiuUJ8j1;vz(J$z~f5t#L>h9b_J{^44)z=43v4kSc!F!`On
zh2ErjZEElp7bE16hAK5jJ7Ckr19?PdF#p}G1%_rW_AJ=ZB@0QG9jK_d0k3d(WIPML
z>C*KmUv98TLQlIY+Og+Au$RjY@W9vUISo@<(%BduCo^}FYCElqHTf04JuF;Z;tRnX
zomWhvXYy3nodf1|kfa1ax^R%E0Q-QKU{eL{uLiww1oWPDN!|wlu?IQDjr&#6q3G~y
zL4WE++eJj@G?t=>x)1<2rn!Pp)MDD^Ut<s<S9YMo0{suGmd!C%;&Fc@GKf;$G_HUX
z)Q(s;r<<vSH=^YCM8k1K8GA@1h1`Ev60)J3%?99g&iZp~IUHx?4&W61bJ4eP#y6R5
zMT!-*p5YPz+XvM-oylJ3-W!*ux7fzS;RvR$-|`=v;I(`CR;Mp&l#X!t>K-61PvJ#N
zetj`2@B}kpdSs3NvaP>eNu7yU%uRFtg)3ePNz1ASiq-eT_l7j$#yOE&Decr&aSUS{
z6Wk;}YD$qwo71Ey@Q=Wf+h}2^k4dqxr>~@&fY^nElRVl-ekLujIt+i?VK7J=#)ya-
zz*MC0jxb*TRAsQ=a-1nJZCxK&hP3gR=Ltj`CCkDs78V&Bb%cE~-eb&j9m8-3@C@kH
znrPUtVEALMYwR!^U2-fmyj-#;0QkaA5FUk@)|Pi0LRxt{+Js9w7bJ!RJJn*acpx)X
z4h_B|{2{oi4j~Gj+gglJFD|<+N*rS&mUHcSRvXR(EsKI2x;910$7D`Gt}i-U&?^bL
z`s>S*#8(6sKW8|>)_^|h7O1P`VZkDX5ucwHh6WdCUi1e@f+oH<Z6C#pz3~xjH}1=6
z=C>~*0<370sDQXf{LRPM9)Qh(=8}#7?rr-Vjqj1Upn+sP?N^)pE0t6{d$8QL`^&jo
z4P9-ICshhluAi-9Nwf4=*$PR4<9EjW6@A8cqvBgP^XQ7c0p@QwlE+B#f9A1$%PLlc
z0*x;Rx(mih2?XL|p79M<Rd>>G{0+}|x0sln##&i!L7qeL$BjX)CjhlJQ24;g8ARkU
zUlz7AT04f-2-B|L5a8n8#vlKSV_klgZs^bk4cIZhaly#U_^b>x$TJTbt5B3TVN@9F
zX=b7&q6KrEq`Qq>{b&Ey<^?9QhOa4B#E05zr)d`PnyjMrx?xH`9>;&QJ&`XT8Fq>G
zt1eP3yw`#hqX;)L0(N;a7I(51k^B}`*j8C@M>yHXGpU2r)1OWs{Y>ex?U@P>*z<*+
zHneD5EPbnx<F$N?zRd!FXn*o8a=-TOM_t)dTf^s1Qf)(C`|bbm%JO+lc+<B`OuffH
z6UP5^sh)U!)A-_zb;*+f6Kc-f&8A-i?ITD#WIoY*M_KfPT}r?rw99Pt`OWN;5BxRv
z`E*{EX)-*UEpt$^Qu^QyEpNwn|2*u9ky>>_o_+sa_vF*EbH|mxLXdKM8a&9AS7Rub
zo#tu%yMNW)a~x4+$7b+8E2KUQ(zvlvA2E<Lr{kz$h#465)WBNxfPdVl=wq_pEZU$G
zf1Fzyuv0|NG*wlH0@>okvv$AllE1~iUA6v*byiHj(j^H}&{{Vd%P|OZxkG#<`&k!4
zzM^Mg`=X3;^J5-HG3EL*Wk@fkw<l4=N(pq+b7)$SB}epK-mf?AJ)<jaB~?EqL1j*!
z2M!!X&gfblxc;!adYgBT=V|eg2u~Yvy$<>{&#J-lbjz`HSs>borxNo@W!7@3Ayq^b
zp|xIVsN~B#ip%6VDKcB^qvTC+)?TGY6Qp9Hembk#_A_?>Wcrg&tg1j>EniH}2^vop
zClzXkA;@&QMfQiMf$l?gAJfC+K<N_Bw`HR1%j<uccBv;O^!sP+4(`!JL`I;LAE`jb
z@7|XKN$7(V04e>dyS|^qKfRD#NQzONVbfA`^snBX7X8?~7fYJ-tkO7>mSZ}4Cp9M+
zZ)xjvNJO$+Tr5#|ofW95Do$Me5{MCze^K93eb2qIusy8YvDXzCv1X7=W>(xSY}4P#
zfm(^rGrK3wipCR1udtsGb)6L5RLAT)B|-hKR;_;ncVo9p+ZPn>d^fBGxf->IEPGp_
z45?Z|7iQlrFD*BowM6^U8TNnVENpZ7HD`36<=o%SNOshSk_JUef}%HG78=WU9zm*c
zB#2J-$L)R9E#fcaiE*Pe(CKeP)S$DU2K)H;0sA^?O5yEyw8ay)rQgy_-%5`-D@OD=
zm~S%y@P$#&e%T>s@>c_iWXaFSN$YqQ-oH{>XOEn{d!cs6W$P0nrm6nYzv+KeXB_Q?
zk|3Lhi<$_(hbINNZSHHZ9bXqC3+F6WiZo|d#FZx#z*&JKf+T5>d}^}w*|Z$#vXsD<
zqyDqZVvDW-Igrn9pZizI*73K*Jrq>@Qtm!_2H3}IT^dw|C^iP>S;rSy*rwgyvAt^J
zwetIP6EWgyD$>u|dw8c>YhY#eS7pfko1XWdy%#r#`t1|3Ql|}3JYW(1dqxsOxHv<c
zij@X6w|pXb{u^yH?W$X)3=v^hI{ozo<m&lzz8{w2Q~uOwWCg}+Ng5q~6#EmNESn#x
zy#&0Lp86sp`Vyo?`S{vcn{tsXSdgSL^rks+$3%3XY)W97^@ycy8CpDZ;*F(k`Jou%
zs(U)j_1s-4(8aMpX|tj1&PgmG#rBHzH_-$O-}E!7FQKL0?XglC7oXIp=Um-(R1@JQ
znKu1dj3`Z%%G_t&XIass`2vc*78#)fq}OHrv@D#TWS6Ra?^Stcd0t+%@uxrWUg>|S
zKe0DVdS&{d;Y|HrLlIB@=1!`~*|b!|lEw^?;3M_w-76dPUG@U1N``+-$Q3mnbA$B1
zHzmCACBXlGn}LGRkd!u#xFNkrqXQPF|4iy!*T6!sOzVikIlk%T*R%-p)05pZ_d9Ih
zcfu~f(tN@`9m!xA%-wH!zs~+-=6HxDR*RhfWVXw`Tw!1_sbfu6NPhi+PKTu|p(@Wq
zi?@#H)oO{pJm>pgUvuieC2aSiYN#9{bj_U9V}##Z_Xqx?_7W>g%04*ti~LG53H$PG
zp?QrSaiH}YzRU~ZG4#Xk#j)nC__7bxDQ^T%%0A>WH(vigeK~Vu@c({&u<S##|FRqS
zOy$5DmCvAVje!Lz_)SJY?R>HO=OpjidA^?NIDZUX<7vMq%B|cZ;?Rp6hQ^mFq~H4L
z;fM0Pc~!9^0H``w22|}esIN_InC(SGaT7M5-^(LE_jgUp!E|L-P>+oKb2G%A7j<Hj
z)aCE}OmtB$h!{AsfhIO2X+nk)N}&3XA=A+R7I&S_xboMs^=^c9ez(T-6?6@>=|MTP
z=eRvo<b&zsoWX-e2&V(<4Z8}{at8g+7nq@I0F&?zpPs1dTti<__2COqNWa=lMz{i~
znp*P@YO?S$F<Rv9(pgIchT__VBsTEjN}`Vszjf}LV2=TM%TUylGo7+yH<Vx4hsuCL
z#?I{W9Qcv`AvI@E_XPGhq^USA_s`km*USB)56!#i6C3nTlz&cV^N|5Bat1^8@sWN2
z<B0>h#uUg8&Kd}@q>B8afkb!^jI4BqXs+)>&(kN@DHdI1YWhn($k306NM!|3NdM(x
zYA{0LlPULy1KPjD5B~4vE;4H#8||VL^QR%kWS1wJa>Py=q(Iex=>JFwbXQP%cEw~+
z5$_d)Vj*!N_aP4@MSZ|UJ)oA@z-@dMa}X55*Bi`NoU#)^6?w{|#1E6l+BQ}@@=RbC
zhWJqu=<e|)3Mlbgua_NGhBSFuNL=n(pb=!_9KqnIu|Rj2;;tKSgAoqLWI@%85mAsa
z_(0MJ_UO(lsX2DV29P1oEM%j|^{IgWz<V3$&QuC?XE+=K-e^H?prUWr2fFj%#KtRK
z<gLPBXUqzSJlMU)ZV7Y~nS`uai<}A<;mh(y`bBSbV=zBxqGclIZkG79Hl^5ik=I=)
z)t;yCS)yy;#epK{L$~aRLJWL=w<M^VkFgusfdK}TkbXS(Ya~SsO_vdjuHobAL}>=7
zs7nWl4QzpIg>pZml&yiTX=3JW*WtGR0}7!NW#foLOi>|jR^5GW%UIbaGRsIn`h}j(
z3v|cb00j^mn#Qk<s4)bZywdRQKMhTSQy)cD!ft7Rk}awjk6#`OmKX{6)4)#LkN)KM
za^pcfv0=s3QUu*OPzY)Nw+q04B2XOYPA3c{iRxd|?x8#bD|$>d#NbT{4Yu|HGKFWa
z-AZ>8xhu%Vov*X(`thGTOnEUQ-@>D&evDV~U$@czH27Q22`1SYW0<bX5lw}G?!{gr
zUlh}epl6-5gB8+qFf>t25ST^lt3}xWoo6|Eyn9ulEyd!`0ZC8@54EJz;Cl3@GGwUE
zvh@|Yk(Ygb6I9;#T~v28Q8AK`ZlbU`2d(S7Jaze56W39c4bf(zhlp|z&#oMbItO`O
z`>E0-@6t|Uga2Jo1@+)7*+DQbF3*Ggd!L=Cwx-M;9&BAah}a{-#`NILh@YOycNInT
z+zE&fo^3`}8H}8M`|JRy+W&H1uwAntkRh_`|AcFft{KNWjFhb>)ai{~Pu{NiYVjb6
zEhqAcls!i7mW5Iq+}`3&i5!1z*V>iz=!0k@aMFv8HMqT_igrRR9vD?5o)Td$+G-5Q
znihwB^3P;8?(%U_o}r?R#|Mh$hSgzDXisupu%G3;ph^9&O#}DzMXl{<;fxotFWSR|
zaM22|{|Yy+q18j!RZJ1xrHg7cx(Kx|yI?7ty|4zqzn2dae}0(Jg23Ra)=i31oM(Tq
za{eee%%P0dXpL;ntq?2%H4k5D@8tc<EPzbIjVr)EfVc^kY0)gCzl85&EPZI8ldiW*
z`UU&tN|M9cC#VDF1CN(K+n@N%JDhr636&OxU@EcP>HIg$Zec$|s)wC0A^K&tXv5_y
z${=rjke*2bquey821Dm|oqf4(PycdLBHCog)`WS}p-d2mTTi+$<TFH_W`8v0tlz>j
z*NcF_XF=fWKmvY_XHIpHJ#>++P@Ccdp2RZUBAeoHH=4GFnHN0XGgE#)WhVt;zVU+g
z^Fj&g0bq7z?l|fky^#}n+>*+4Xdr}2+!CMu=K1N^+z2D>tY5kV+p^i}f^7-qbb?Y-
zfc+li9cBjKV8mQ)*mj_9iS3fN9ZFfx20-IUUTlwM(i*3Uragf6W<s5tHhdjoV439p
zv7)y-;N>BUjz~Bb_uGAYmI+LY9fn_Gf80(8)hYJnhzP=REeV*1-~}UMyQKyU8XzwQ
z^DLK^fC=9Q+ad5ug6#nm<0fjQ$qRghR=IKP_?gYqmcd{__jmUz?AVW1tpGTg+|MuT
zEO(l67@=bD-$x{~z%d{JB;qe_tK~p6MKJyZ*U$fQ2ap6ZAJehmcp0efmR+HW=$pI6
zyCoDmWXqjLVaMuhHBLyJHtZ0rTAu&1Ts{b9H!E$a@}>~{kgfHUM)*@Op>lqM_x50a
z{O8BOew|Nm3&kBiU2g<>6to{H8X=ZPoR^=4M4p#4F8K@2kG)srNmG_c{Lz+-Zs~Ov
z-zbdlGDFH5&)%PW8Q)`vM8v;r4CsL!h5IWe?;eOJnnxWqo>$r3!1^F#>Z=%iFGrzm
zK?U-~n66S!Y@>foIw#SpPnFP#1m2X7E!(s7dsXI*#ft{w3rhfgH>>+x5AOn{lyxb<
zr$F9Ns@GTw_{91l(h&a4S`23WP8XaZ82@V0XGTSp#h+x2w5_Qe`C`0|>vcwYOzAP0
z6ZP^BX|otNf?^8Iw9ZwGiOaQUV<%4bh(wcZ%f7`g0><AK&h0_yd+DLtw-cUPMU68W
zx!oNO8_4)gz>?S9AI|M>D_pN+ZAiz#-Gl8&be}4{u^;J3AL5RRAKonwW<9QbLg#w~
z|Ml}oy@1w-e7KOIDXE)!0R1gu?{oKm9`1lBdKWaseU)&BT-5uSvW#3C*6unlb2nAL
zcF)t&emU6#rYUTi@*d4)Y`SG=^dTpwbLvn3xLo<e{sb(lxWcl`zZ%|nNWVFQifR0=
zTkUNfa5ON;8$>W$=s9szd9mkCp_<XZaR(AyQtC%@kJYGzOy46veJ*vi`8lM2>;H%7
z@PB9yf&UYt!~dZ<1pZHm4*!Sd5coeKI{Y7+L*V~Ibhz^VhJ~JL)iFmg_rX^c9}=rT
zjxW-J?XE{6$*lLgEiosnY#c-P2fMgbn-0`e#8~QSRr#6Or3X_|j>r#UD`t0LbgCXI
z1$*Or)5*gjHY~V>xg9?DGAz5`47r-z)6Yt#sR6O?IVnW}Youdb=LOmT)vU)gbYZ1^
z0GNU&wJ|k0+Zz*(>2M5-vI}&7C`>gYtoVXMs01$`<4BVqyvE*?fK!oYL@b}nqlAjl
zVqK+!g}SrV!K#SOTTAin%-LNvdtf(pDi~uV8Du&)A@ZYVg2`9G5`EBeWDhFfrRPcl
z&?y)rnIoTY6Kp`ggG@lqdC!~LJy{8(C3@ThxLu304_JZD$1U#3;|DFu4;1->-^a|L
zsn`8WnN|scdarqY0@&8HvXK80@)-`kh7MoM&9<B8RW9b%Q%^yH7CSj6Oi~?&!5<}_
z8fWVKNM}6cb$G_f!5+;u{~AB(3>?n$M}dg;Nz{%aV$#{gW0q6Hib(y?=G;{AMYgP8
z0~Tiq`WuFTcG+O2I4LH%=IA%G9Uu4Drs|(oKJo%GFgC6wl5{>wI0=IdH~#i(!lj!G
zF(%SSC%ko>J&wSRBoP^*!PO4zxP8Q=-($zvTeZAaoba08TW;c;ttVChF=edydep+Q
z!N~(p!1pH|$X*dq)5`&C#_ZUmp}{~Z`O}!!seHs(@u`hscJs4A>#h39(S;?{nDbK9
zI<O97b7FY0t9wvl!a>0uqF|hF!-KI6Dsg}Kjy!ya*Wkt>K@HglkRKo7dm}S7fWtyw
z)LQjGD36kq#GCpK5OEvy@0e00)~**`Jf2!b%6AQlI_tpi_AfHhK=}AyGJQF<t7UNZ
zAT(5#@agBV#kP_&o)gM~(Y9&wF`Sq~eCiE@a9!{@gsxXIY4jIH_FH6FZ|@mq%Zfzz
z<FKoD^5vnuyU#d0vU+3STBBy_I;sc{^PC~e&AMB7#f)15BbB`!F-XFMAE{{|KaA;&
zIrBFZvFSmJm~@<temP`7X#I<L%$euxmT(b?_$HntgcQZ)DScIdo+cYNR6kk%Z6#78
zb{7cFFx&VoadR_^QY@!j%kvIuhNQN2m0T=qcp9R)42gRjzw~|NXdy=~*sov@Rr|nx
zofw8B%NHQY>4yJMq8pPjHr>@qoA|<j0z`gnhpM3j{&B*_=c%6l15?yC+;aQG3S97+
zuCoYbZ5<Gnumqic9>g7SN;wC@CXT1?a}hKR{O#-CEA?@2m+a23;!-8u`<%)RllC5G
zTzqQW*QA`Tk((52eY)QU(=!`lcrDkD{L09pKp#-intuy9lkzvR`@ZC+w_JK+yu#3V
zpRbp*z?M%{2Jfc2RILoC)xZvZ%7<Q@MzJl-j~dN>YppTUPsRic<S*@WG0)CQ=~pv?
zKYN$E!s+XKuDtvAk;J=4F)#TCM1bl{&S5@SxxnQ?<@v+4=InV*X?qWO5CJ~FEv*&@
zOanyrc*ubv>v%0Qv{n}^sFigB5(ZAPM;qs?EZ1lvoF9IS*whg5a?krhT|XoAX6k!J
zY|uwv55o+<N`kuHuBMLaTXg9wUis~lM_a7K1f}LU*UlHw3@o}r<Uqfk9X@TNi1|SA
zlLR%#XI}1qD_4IyH3t%v5_pl6A9m%>IY0+Fj!p@z+_AH5FKZN@X_~WRM}meuQmwNv
z+~mpMB7?LYnLP!2_e6Cg{q}la%(dHJULAhNG~-LbTZ!Rib?Z=4(dx~d?cP25dy#4d
z+L{RQiH5-V`2Hmk6>~X!HTUf&&6&)MGHuAI@Ap!3+$2G+XF^kRf=)>e4c$8o8cPKn
zw>5iD&%wosPE4OeHn6>4uyVxvAW8k_$sz*jIK(Oah$=!hCv0Z}Hdd9WDI)ZAt6DSX
zA%pstp`W{uWiQUA=Wy=7yNa8AAZO9lz0r`FxKs5=dDzwXY@)73*SX}4gM&FsSK6G0
zp9Wr3y3k*jE24|=BKcV3g~O(wgGvG8s}>3((ob}#{qT3XRBa}cb1^LkW*uMp?xs9w
z-~6KeFy=y#{H~6tGR?a*!cFg0Ky9ZNk!qS*V`<qkdT+a1E)B<xD?_fXiO8q7pEFnA
zlyCQDz<&k(2SA&MF&neAHG85Ask2YG7JtGoP7btdA2v`^BQ`bX(a{|$r9j4N1x=()
z%UJyxopR_>hWN$af6PR*-m6>NkkiK@RruY>#JEElSC}GpZO8P;<G>c3jJdUK{^n?x
z(0TE^rQ;A!<^Nd8M;WR!^;(kMf1*-zM!_N)FWVgbc7HO4qaoE`ZWwGM66Tfet@vk1
z5~Q|Y%(_ChX`9q8&6Ge*z%g2+)>MZ5_HlPQE-t?z5=~aw89OPeo^X9O{-8pXD1B2U
zO~j*QbM(#yapL`lZAL2d4?Yjn3Hn9s90VKa{wOeBbnMg8L}cF<QLs4bwl4@#OoC&Q
zCVNB&*~avo)C{0WL?3o!u2!s^cT!x3%Sqk3eK$2HvUXnG3Al#&kdk`A!q#j>#HA6=
zSx8#^)_ygryUSM7LqvG0%oNLk)K;HocUE2+CX{A~5uG$M%q?C8is<mP4D;kZnt5p-
z?Z1Jd$X;)EA7WqoFAd7b@7Y#RQOv1Psc(SxKuxK8sma(SO@vyhanA9}3#IRsozxHR
z<;`G#5BMoBQ3JLFMuPSKy>&AG*Zu$13wwUxdodr1Z+Nc;*xFTpTJw6>w`Qb;sQk0`
zp^sMbG*Z|$`0wE0T7*sH;b3TwQnUF5UZ>%!$WxLyJ=F&J)S;RC<{^Yn-shQKt&i6F
zcVDdq>Ujlg>9*GRta*%~%WN!2xi)s}v^7{(K6=~ym{k@%0)BD7HH)MDt6TK9`i)y|
z&A~qQ-%I!xbNj_^UTc5YHNU&*U+U+XsE)o5<KO0=f>D8MW5(|R$DX;((@B26{_p<#
z+{!^b_V(J=K|IoPZCqS)lyl|vLh}|w$OaI2ey4Cn$kUSNmy_Ok-2`FL_uGF!XNn}|
z>)dw4yoQ;&F>t?2u16_9zgF4>X#8<am6E0{6zg@nhTWf4m^t-!<@8TiVc40vq(RAC
zymN9jF!Ya`ux6$9QH*-q3{7V2RJP&4(NoDqX`!3U<fhSJyZ*24Kel@aj3PilIq)%u
zcg>~x@6iLn>qbgRCNJhv?*}Q!5EJ0@O=<fB586MHC7e2Auntu|Yj&QFbDD>KshNx{
z{?@Ry+syq57M)QO5xK*C-8rJ}NBu&%Gw~VM90_@G=iu|;<2P_`zbZc)j8U%PoOSpj
zIdd5%|M}GH!AZ9=d1m0o0(>I`>E;K3__zKdK33n!R*iDqyY}eT>$@KZKfC^Vsc?6+
z$~VD$xaz`j^MR`CXAcItHgqT-X*`vD>v7Fhh&VZkuir}y`@F!<zW#qP_SI2se9zms
zyB951oZ{{jD9{FXcPZ|^xVsh&THKwYr9kmg+}(mx-1-Zj?|J`!&soA|_wL-8XJ($6
zB%9pG7GmNRcqffTzhvE`lH~M_u`Sy2fr>R*$O*az;YhD^g+lr^Tp9D5({KLY^7Dhs
zN4jUKm=N6I%cB7nrx+E&1MbL*+i(;OXS)~}jYWSuk8QO9xg%7iwBv&%*-{0*F<Yk!
z<I4VTXbdgE-_T0iB)*MU;>?O%y5RhvW!GS%k`duy)$rTOEmMbmfG=578VLPa{M7ok
zTxy2uL&Y5jH*C2n<y%S({ZCo~6*$`Pam?Sv?KvtEzaYc_KIHV&1R9Z%As?>!&_56k
zS;R;~Z^#TSQ1FP`mo12%>?R6jR`&{%oUNI*kXEJ_{LzJ5_M^QArtD!{O~h}}Cj4AY
zQXou>5UU$rEX|H+c0ui*UO_3GO}Gf5)ESIYUb0i^p9GFk9_e+~qVoI;m|PKQ6VTas
zG$oD;v$?t?kONUxsgi;R(f8N>%!VwenM?P74m*BpoiapMcStdKtKMGSq7$}Gy6uHr
zAtYB^-t9o|!1G%j6hXo>p&siVnu%t}n_AA^{?H`(ZR;GK^P7|#X=FJLE22pSOPwh$
zsWG!C)n5J^ZbRnZZU@QlgA~t1g>L2T%5%u4>6_G10NY?zLXaQ~D`7=<e_C{XTAeEO
z2113pW|?}9+PiP+fYuZmCKKU&2$4!f*HpkqiD6MxC1wP+kd)}g;SVX%DdPw!57<%k
zgv7~23d}@(GRY&Tceo#!Mc-HI=r*^86_Cn83>c!-F}e_v&=|NjU<u^RP^bxaI8jt{
z(V6F002ozg#)!KlWa_)JwW`q+OU4O_^Rlmm(G;^v%lN1igejs)MJ^A1M|^MAM_os@
zS&A`(;?OF{XUeL6mkB&ckCg>}kd*B*lZk~K5h>gDmO0FA3C9u73Nn(<IH?>qr;)Mw
zU<GK^9e5vL))(IH@eu0Xd<r5UG$n$S*?n6K#lH+^65|~ym``En9-Z!TCJ^^NLgEAB
zA%T=S2)-C8Qbc`28u1=hL|qHXVu&w$x)DQQrv!#iJt7nBFCWj&CgvOFK($_F>I8gt
z^+bw3DMSr?X7$oEhX4+|#2qqx=HlOhp(3MBR_LhlFw}aq=lQ`0SXh_w@R${3l3yB)
z0p*~R$IHE+b`I~R2aK+*x9%d?sWPf9R~oukevyp~(u(+vz!z@bUAzbHR8O?e*L;`%
z_|435U%GNxp8rCzop<SZ<f*k&*@gY7!*P%>WYc5yIL>lKZA{_9c^iY<HwHt>lS2Ax
zZsHWyPk2SP|Iwo<9try6ER1UKJ;V6O{=n!o1WAtj@PYa&D>;ntD{2hwklQJG(K}!2
z=?O*HC?7*czZ}s~l+{z2L7fIxNq!pwR^?XLLQj3R67$BkRk1D=MXCr)Uy~11e;d3*
zN_Es7p<gidim)+EeB*9dBgohvbAO`7z)e}MJ0N>rC)W5mad7wZ&b!_OwmnNdb9jJe
zuWxz!<bdDa5E~eslVz>p7Yht(m7!_uvUzxJD4~h>=|5>7O%Rm)o;m(tiEY7Rb?#ea
z^>YcMEvl~n<2lYi=sCZc@g19Sz}*EW=n3U3JHIb~`1r{*v-!;(ENAqJ``1smE*w<`
zUfFu&AvR`X3OwJI?A1Pp=O_jBjeQ5~=TGwUMY1C?9LW+}U9*(3>Pf(pE=x6X-!bRO
zy2u*m4y#DIS|_=Z<n5f_gk;7K<goAPEiPDYEf4^#2&2k988|yaC>DQ5e|h!_u=j>B
z#)oW>+pi!~eGPHlj5X6w$aWm$bUqMt82xPi9)#)on;qMm&?Ax^VD>E6j?MtM!~WQ4
z#o&ICi*258&{ZkU$`xoJJ=qeZq(rX6S&Q}c`r9gykdt9Cf7ut)cYzSJ$YwELHw{^P
z+p$H__DJx{sQl|C%nr76?lGG%8kJF2rkN61u$6cx!2EtSZZLwOVt}$cf#k<OD}pSn
z6%Z$pa2p}4X@kVsS7DL27-a~6_h(C@5By)PzpnQQJD85=Q%{a4lyw>>OBsg=q8nJD
z{UsTRoD#_~LF_XT<X$%s=X>`4<q?Ew(1hl4W&MU%%V#58nZe^Tclq*^4vl8QU!R|i
zE0)&8d(Ao|k~i-x_!Th7<7_NP_N0FpJJFec{cYlyxn7TydAcgL@%i{H2xyQ*qWF!-
ze{~ejyryeDPM&*$o^(YPs@0H{ZE0B8ZJC>Ip7J|_oI!x8vYm;UDEv(&e!$zhuLOII
ztk2(CPdg?&<ny@ickd)zO|E522Ng=sKHcj^{T1Ckyr0IZtVK0%oPuo^G{b@c`QSU{
zIAaU!uez75bvMG_@^uQu&I49R*C%>HKOMCw&RKF7&U-{niIJAL-KHeulysG?d*%ni
zEwkLI3!e>$nRF|^2Nq1SI_T53%sUpv(9AOm9y35pM|f4kS0U^_nj8<|#zil~9@D}o
zyq$^s-H755shj-&`MBz`(izZ-&ubNp)xC^T{aD@1o-@+Xt7~zvxd$H9H>2j)w0&(4
z2bv?K&uSel{l6TM8LZtvZ^dNPk^3#wF_+XDJL3K`e5y14%AJUn+dNm)#Im~4AvfvB
z&cM>;>mF5%gf`6&IK1=>_X|`-HT||94Map;?(!JuminN6Q?m4}I_26?FYTReA{@TR
zI{&I{#RZ?OL4ua*x^f_l&2bf{-{TGcblq%toPgc*(L2hPJ8Kw(zi&yfD{XKw60Fk^
zis--c^lj*j7Z|h3s4<i1_F<sQF4_J-gnyp}^a)^m+cwokYW&;Ji23ddhjmE%chdeo
z_4sf%j>ZsNO1mrLA9C>GU;p(9qSW+5v=Lfqh3NY%P_aHcd1_h#jBN^2NoWq5PCqn>
z*#w-Xy3i+#oG?SS+d9(v4rAGEoU}uP$#Pa_39;f`+=@a{DD=-~C9zL^-A+S}sCp%O
zkyzAGeo@0k1+u`6X)DXYP*Z)SlSmEJh6~iW_!P*GdxCI|U#A=>%c6?b%d}MnQwcYs
zE2-6MjGK%YN(7?^^g6QSpdYXYcLhGf5n%jsLX3vxmnC_lYyp$WvWJ1gL-Pkl8Qu`{
z++C&-c15;|EZ<S$2MY(5fZ&h4z(V*kEE3<a71$s-3CeTtz$4hM_r_wzkJx2Yi2x3Q
zX8YSnR25qksSk-v!IAX}<dQH^mmhJIf}8L{Gc?js3qPm20l^#iJNX){XsW6NSp+Ga
z&QdjI>8<0cDW+MiJ728_L>xtBbvdW>lxj7`u6)@x<G(`UaIl8-9aG7i`3a~JD-us`
zw|TBjzIVm4uQeaUC!X7mkI{QQK6my=g&D80Bcf3x8lH)j7N`(x?*^$gQTl4Docs<s
zH7Vwx8@9g;1lT72%sK6GiY|=$jGfG(AKHE+AmZOsp=Jx>35;vvDMhvW5Fik@;cubO
zGHyHy5>_Vt(V|zbn)PE2`&xsv-;z3yCnAn(02H#|_lqk=iP|oxWLOaaop@Kjv+90W
z=W4{qaDNX2l}qHQqbr}Nwp-T$J)14Y+1Ub~ZPr5P7jWdcE1zsTZ?S~5!j-J&^FHb6
z@}^?WBiW3PitzL0*j`b*e9fYO?ZPm>yZM=!?-xLv;mnQ>S)jdC>>>3@Z3J&>N>_Y#
zi7m`i%2wsMStpOVrn6Es_owOE-GE-_aZrOcx_|n@S4O@M=C%a$M~T!@_r%#u>f80W
zY8<>A&Ome=!yAdzhk9>y6y83sAHB^s8gqeEF8sM)H04wqM}}pV`Hod%JGOq&X8P4U
zMN4Y?VLZj>eiJGoNg1$kwXh`$zU26Tj%)a}d|=nW{3lWA#_&%IiDejeg5m)Y_8M^I
zQ|~yg(GIr5aLY|R@>{FutNyqV5|Gjg)@S0<_j5pr!C|a|<BEP>#?{ccic_(igRle6
zVG7ydEU0#+p@gNx$N`i=C+bRTJ#(0yJxw8hmddh}Ti_vjuW)@kU93I(<(mGmge<5D
zB)<AF!PN(nKHo3U+UezeeYP8r_|kE0{GnvXq2@{Y;dsyGSMj`qjkCn^StVEMAJ^7n
z1r<PHy&8NJP;z_oLig-B+GgSruQX)M?BCW+RYJzn<yDl^Hl5Jgls>&$_I~qfH_oLM
z<XCgmffMux-1S3H)JO47y;#XH)z#C--GH;ifa%$1*<YaaPh;!qu8UGO<vxMT@*wy;
z{7+A79>{Pdk1On;<)V_rpd`Z)EBxKFUlAbEK0SSeCKjb@4^EBMSjiI^)~yBzqpPm&
z;w}$B;Ik9zt^n6N&coa5Gnq=tY^l_Fjfa;#0jEEYqCn2eIymAIrc0cON|_Ba%(}9(
zch99Im8xvL6vi=)VS#-5(|zzb>{Kd$*;%WB+KVx*>B259<nilQV=8R+TIKbg7r=0r
zxMjQ^!aP}b?b4oblE{<cxquSxejm{hdZfEc4e`QVTUpH$kvV8(bFMJA9_^oH;v@c3
z<TKEMn+|eEcXSPJ3p0Yh$RdUdq>?)7>&IS%?cKH-w{}`J@BP~^0He5lOR6g#sGZ37
zQW|vIFAN{y%3YPS9#lm9*^}gv8>lp_K10rNG+6%nD2tyJgD@L$>$X)p(xD0O##)*d
z+!EwQSa7YaMKZ#;uUe))l@`@9<vCl(8XYdyxKtq2swtJH;y37Eq^0auI!ib!-F*26
z$wkW|%YKHC%s^+{M%YQqXX&py;eif8VZ(u8bw)=6UF%fEG;IPp0X_DBEs$1IWnP||
zrBTC|`vQmP!Xn=8>#uLtzJ_sM=&;qb<0bF&w<veiY_z>&pO~xTOZIowKigEq%cMx8
zrpuf?9W7$m^=#v{BERw#<L8?(iZtx975E+-#dyv%?DWE~ntyd3skaf;pFF*(uRd||
zhsYN<E_fs~$b{MP|Kuyw#s!umg)OPu_93k=UDPcn(~zx)9Rs_f98coqIr8^}$Pm+P
zzQ61(24CVc{tR?1gcIz~svlVpFE^cCrD-oy9`S%e23oc*n~}%c0^SXcHiK}NDKGnR
zmi14`>*J;+SzB_uJ0r6#T}@s4VH=k`1~_muT%_z-h}J>KAV;%pHbDKXVacPWk7>T5
z#bgJw`6OqyrJu=||F@K|m~FNZwQV;4yBR*!(2jN~zRq2~Xe+lbR?-Jjj%x3m=9lw!
zt>GR$R%e!QYjIXk&RaG#7yQ0&3{s!qtwdMiQ7{_ZlHcysN;xTh$(xUl3)~olr6yWo
z$B;m1b**<i;%7Ng3<j1D`x*HDjHfPzZ0uE3qByYIMsEyCA`q?kO`r#p4Uy|OjpjTA
zZOC>RpgWKhq)fj}m<Gvgc@Xe-crZGs6a27P%t7%ebZ0U_9v&}|Il)OfoTvDP8G$)<
za_LNer<cl550TxptAu`Vo(_>UHTJwIGq6}3*HBGbW#r{IZv}APWPTHpmSvOnP)s>)
zaRfCUbu4r5`x?}!(5;R~SsXaDy0$pgo2U}4Rk6}^s$f?OpMz{tLX6Zt%t2p)Tlo<Z
zV<%sF3~bEvSox_$jcfn#F^(W2bu#cyp;}FAb**4Ful6(dOVjYylZf&C!sr;x7tDjb
zzljk}>l1o4m<R9|w_5E<Yc}W0VdLPpgbIlj^Xupu_^1^Smk}l*(C)HK;x4k22p98j
za!xfMzeHUP#&X)aZehKRrz0`z7~MBQ-MZ*z>uP_4(;LC_u9zRAin{BfHF@o=E`^xt
zqJ_1iw`oS0kj>oMLSDqpJIgN|9f(dHw1WlR&hZiTmGl7J3_opeP517bLzj)HpxVj5
zaLD*V+%x=#rDcv`9};W2QIp{dsaULq#TI!h9cRk;wr+$G1G)98x(&p3Cm`ACpc#HJ
zwcH%>MA~dChAnT$`aGKCvL%IyG}4S`QBTsL`D!4Abs3t^w}Kw|+!VSCZRr$aNtE8-
z>#0kj3w=PJi8S!t@)@(#?Z9z92lDs4RrKsAb2^kGwrX7)SsGv=8n5bEO%X6>@H=pn
zk20KiWbV-UF@2~e8XXd*J~H&6m&FX5hCYt-eB!j4A7fl##SE{$Zt{ihflP)w;$^Pm
zc|J9TZQVKkf15q;L%i$L|J$yXfLn_FJBs~<jK(v8A?ViqC16PKMK9hp=SwwY9bDbK
z^m31$s_Xx+ORjuYI{s1pXojkC#^!sL7-)RzVyGiaY~u5;Tj+mrhf1e^+OfWVRfBii
z<u?D8G&FCCN*5}sD0`YlY{m3vxrlUTd%Au>PTM{ibSKcNEhQ|~$VLGgFH6^o_eWGz
zVAqn!it?Wo;<vkPeJUy|8>7x4@W}0b8Qb9{pNpclp>OT|v=#mbF;X{ie+W0yG5+t-
zs2S+5Wv><T@2@SdS&A(!IvKiF40yc8KD%2Ii>Z7$l-VaH(i+dkswR9M?rL%8{RXiL
zd7+C>BzPWv`~)-r3F5X_4EXT04=JcRgwOIaH?;>!QZYEs@`zC{oAVk(%~Rk{tWgI_
zYt0mqFXpy}Q7`7f${&DB_J2Ci>tf$MY&+sU50`}$if2mbPJvt5#shyNjOTdKJfxIQ
zvOTC_et$qO6i03UaK%6!8~s8zQBM5qtoH%n8j_28p~C|B37+M(TMEUa0yU<<os&tz
zZ!F-Ux}1i+(80MBh+khT=sE_tFpGtKp8~66gpfYV8yg-d`RI#56;mgANa62>>LL6V
zbT!luy^eo7uZ5=6lnwf;TA%9PT`{QHQa}&t(m$lM?i#<n<-aQBQ*G8<4%eqLYAc-r
zKVSlw^pP)gYN%7-P*03!`2)w_;Lv0E=i%&xWczn3Z8P6u0%4xz2O{SAAEqA_t{7;8
zC4kGlA64TGBk}6^&*d3#LUAwV5m=EgJL5kd<het4n)<Iap&${vLjlj>fO$T&CxVKb
zM2EucSo70o8Uqt)3LJZd{yh8vA`Jk^MI0Xhk;Ve?#k6r5z&6``@j!`2nd%|M52o%w
zi8&T?3cRx&YB71M!T>!c6@Q@gjt2K(dlu#)<yOXRGF!2huXsZc<#|}FQueA;?1iqw
zDq<x1#r#q*Up&*5CtuvVHF9ku0a^T<LD<RQeCP|sIYSCP%+p~t-jlp}6#xT39@EUB
zBZ$#9lnQ0J&lwC9KT{=;ZzusYec;Y}<pm@FQfexR`Qn8G2Q;vxpuKb^<g0li03yVJ
z;+aoxH<X^$Ik(33FG<m!<gvzxpUR(E1E)8XPK4_7kpQkt`Ch78%U-b0UjkOJugM$|
z)JFPm{C6qxS#U2i?ZA7YnO6zL2Ze7Q5r9quSbz)lQ!kYx0`!tbF;Py~38ueq9mUQ{
ze`ADwKzZ>3!@ZanCIpH5pQFEAO(a3@cJlo@f(URW4fIl-8T3;1&~7B`i$RS&kv$(O
zqPSB7JcGs8)S#-O7;h+fD8xfg<DL&yzBwOK!@JuCqTrt{bdjGfxZs}T$yt$~>g{2k
z<R>zq;7mB-J(a_^@k*at`7LcIbr2VbIxldZGr0Crv;hC8q2N3UfT0#{F7{H9@%Mrs
z4iNyzFWeKG^C3O5^U`jk@>9&0g>SGg=3s5+yX_!^r*eKJDEM@6PZ!<TPZzQ<Pvt#;
z@NTpP*ogK{tw)F7hy?Xfosd(fkoc*-8u3X!L*o_+__LSVf&}1YGW3EGgkWA~R7(B6
z&)u!11dCs*w)-N>{@{!{UTjwl7B6QV(CEU014Ir3#l7`D+z7c4So@mLogYebU#fFX
zLv<Qb#Q6^svt`n|Z43On?KD{6ZhIX4F2&P6rWgFu8Z5r8+D(eQ1q+BE2Z=8Zp$;cq
zKO(;De6FIrOZnIVb(mw`hnwRRFQ})=9=MJ$ANGz5YaS{(p~O9LVdfya=WS!I6fsM?
za}C=lnMo-w?5&Y2Ef)2ioh-H7>{;5)*Mq&nFxqi1&Ek)xcfleuCvfPFU$EDgi0W=!
zW^Pqs`<8-Ii`4Q%J_m{`m&+KB^x}7njbr$7o=&E{i49L9Z(D(uuqbK1)KU#gBjZ2!
zm0y#Qdv);Vs!fIFlNu9-^HR)gmsFL3ve+=1$3%8E+GhSenRjf%Rhjrtoe~x_VD0V|
zOCtX8%nY$V-klsbu0Uh6oux5v4D;D9#hQQVh?!&T@8`vc;|E0QFY0Gp*8=(BCLYiE
zlB(ZZ;}0JZZTELJ!|v`h4mOMK%Cw)4^v-F<jrrU8<&mxtrO-|+eOp}5Z<)zSPq}!|
zvJsx3gKN;ab+lx$O7||BoOAZzm%a&)Md)$(gyM)mLQ<0&cMoi~CO9Xjcbs3JC{<{J
z(Z7l!*e^3)!VH>Uo|!7EFfZ+C?vX3Ua$dICRn-4lNOVnqYe{)FFPonr+WTSbYcokA
zR^dRW8yv%6sdNPA=jlRSAC848`pj|JT31}os5d8kU!T8!o$gd7Tb@YyNw5#&Lk1Bu
zgyUXYk4-wm=W+sIwHZ#{o(DKnx%?PPa#eX>>Nk5l&Fl0D?+1%Rj1d~>^E(aV3@cIl
zANX^p$-+cdBld%{BVK+2rd4EZHhis4TMy`IF&X|mh*%pzmTb2i_?edbNkXt5Y|khH
zc|$IH5pq?zbH^*H?M*KR<c4$P-}@=+1x6QfAuc7W^#V)4gu)=^q?d%iVZdtJd2#Gg
zpA~Q8Bj7~YxDf~@cH8K}s8hc6w<vA7&69sYR4^~qK+*9kN4ECL<vBfP;_by@FhEK3
zs^<Q9LA_*51SWzHJsjH1f#y+aFeC_JZ5XtH@Z!{2HHC9V?4&ODUEoiYrQ7&#CL2_)
zWEV%6ss;eNyPFd))fx8K#gK(DR>ex$<M+uA>&%d&AiID^Stn#pa~L5kGvsye7r5JF
zs-PEJN5=4~puPtrLX){V697Fpda;zsrgN6UY6RiZoWI7Loy5}TZIF<EApTpP>==bM
z7d%#+S;4WlntMsGteS}zRSB$kohKp6*8Go6Zb86$j&*UJlVfqd^Nq@()JWb0%)S_3
z?-T}|Hm}W))A1+LC3&*bB(FUa8a_j?{fwqN)sGp$Ifc$umTqmocQSH0Vkx1R{Z?Op
z+&22zFNpWVvrAM?VJzhR#wQ7}#w5Wxz#)*7C^Bx8=`p63Uf$>`@~zt`iZa&`KcGKm
z{i6yrx#(GWBkRAxZEMN;Z6)v#{Un@0YQDbb?f8BjUp(&H9+rzElBcgQIi|l^1L+d^
zZ7U%?Pt+?E-8Ro@&4^!)($i~|2=ly7;+jj-4fTBmu3jE2=UQ93r*`dYX{lqA-4Uvb
zzEF6Nmg&pck&GXGG_vaBlAc?{x119?ntRg@Ow3uli?DgAM+v2;VSQ`jtwRj1;~5BJ
zVf|t<rI+!u&UkQ_hqc9o2@CzYBQ`=t%mLOU1Z8=sqdl~m!Qqq1GFrCo!sQ$$r(K3K
zH>RHbLS#Q*m&2AxOahVw2X^p9T^3Q_d8DO0hi<SuK1SXwIohBqPf2hA;aHt<arHZW
zfaARIo;bGOScpB`f<st5_K$G`QV8GI@0Fb7pLVT3B6ACqY*t25V%%A;A#aPfA@w<H
zuSRdUeaN5_VC~CAotS{2A<8jzsPqk^u`fXK5pUUqJo~QE<4v5%v`V?!a;B&;Pi>tn
zLn^-u{pkA~4{~#=4}tIQ*zfaBsPT7t0^o^IejCf6!3umJ$2FFbc!Pd`Lb3y&j1Y<k
zquBeNB?tB3U2txoKO6ztFKa{vSbiB2LS+pY9hN;b99EhJnBQ=Q=;!t_uCOaIRmAy5
z62Y(}H3FQygON1wWf&x`VL7ltvJ#}{Hi4C}T{Xsh#>*(~@RyhpUIPQL@8o94B>=bH
z8CcWWNMV@s$RT(UY-`cLTf|Xh>wiI4@aNbrz05=iOw?$YlHn*cGKn0TLA1Z9RLD{P
z$m3`RA2&pENj3~i{zi3C!U@qVXGD=#vHY`pXe!XL7(sifk)c;YMWZ!8o?jM0=FP4$
zy(q_U8Av`484}OvoPQa2Tb2JPNC7Cqb$ox;HyV-W_Od8l&o?~S?aQC{w-xf`$eAPO
z&vO1$e{_EdOGRo)7wC=ifJ;({bz%u)6giJ^=Btrgv)iQjubN?E*v(Q`<=u~zct3F^
z`9G|pCgokbhsS<spKpr5>*Xc+chhl}Wos~_D9As2lMeU8ZBaY!sEHV<j7EUVU3@E@
z6;O<_?gI!^X<;H#-ScI!*sQDUGN5ZGuS$#cSLyssjHS~LFwx)X{eL#)x86;52pN^R
zw%HrcKKPr7#*UoDTE6f?i@XbYHM=59$3HP!_(tA%Z_}QQEf-EcTTEZUH<WZREILhU
zxC^2Bm^5}T+%Ar+p4eTbiq!*eBo055p^B*z;}?R_SmC!_no%>UGX7}}4$JeveTz|i
zYrLQz`px(zXNSjxz@R3dIpvUs2r(p@{?}f*D*@-AzCE5=HMpBqx`t|&l{;ZzT=s3m
z-qMFoJEf7p&h*$8zmA-W*vaXZvQN|fM)dt@4y`)As@I4`tyEvwQtozu2#&8xF1sA1
zI;i$`e>?1>!R#wE($AMAV_p@ai@$F9x8ZWWXfMg^Z;0vFFu4AS<@!!ZY={vb^`&jA
zrA~RF+{AWaw}2*+qL5d*(UW2#^*e>uXtS@=Pj(-8{*y$OIDF5xuk7q=T@@g9ln~!G
zBGd|L8JXZ6U9Q^`ubb>$Al>ka8+6oOJheA1*y`f7deYWU-ePZ+sRlk#@U`2p28-Z2
z+{r7v0|`}@8F=x%6}}@tf))tj91W+yV}l}i!{|P^klzbI@>Yz_><kaPqk^+=r0Ox8
zr)`?EhwsWr)Ln3ius)isa##u)EUDw{xlrCkB{_|o^p|dj;*aA1G6*kqAX3cwJreTv
zPo&z1SL1JJCF{H2GGX2G)tN?;THQMa88GyX<=0K4-oRPRWO2!EH+f07q(~vX6_{6p
z`CZNM&;?c{(r`)$W|svB4mnBO_{0`|JK~gmEGMBIN}+=Ls7f%_mb~4B$^GU@HaCDN
zP+2PsKBeIoNnzdwpn!AKJhME?J`R&-fzR356*-7}VyZ&&WH_PkwSvL9gHF5obwt}7
zLiR?W@gJ!i_Xm^N;~l4C#cY~bvLdXP&n+}|IXsqaf{(PHa(%GQJ#IL(MULQe@?_A%
zm6k~S{{8U^eCB4JLKOH(IUdM{B)X~JgjDmJ8pLVBgBC9=2lQ+RPw|1egxwNcYwWU^
z2qFxBEJ!PECc3#1t<jKTpjIDncS<*rWnc5pMS?sN_gqzG)O>LC-s$4%(FrA24<MuG
zn>_SLSWK@Z59u#n<ywspx6ZilO#L*^Xyg@lj_jne2U9+NWy#LkoHF=j>_$Jqi>jHL
zy+i0E22+b13QSb?gZ7&#AtUVWCq<p0iBV<2;F%Ir;q;CP6~iGvr;i5S7|V28;ncZJ
z44f;Hr=Ji7@Tbj<LiMD+XBy6ooT}Acvrk_tpLvv5in2`$mNR^Xs1sPd5op@+EdK&d
zPY4Olu41cszgWUfUtGs-uL4Qt>CM{9$;vY>$;o1O2gcJr+jk{2tM>>O*ZAs2c;4t}
z+CUl;c=cnfGCD4Ej4hl?6%AZ1rgIiH7IgSt(*3PGWoI`n^{4Zi>$Mt++7<XwH3e|j
zFqu>iy6}!Bgv;&WZx(lSMG*aps~gXzmemc1bbAv!)N^uI#Rnn@8tTmjkJaxFRkgB8
zYW(Hv0GF%pZ0YcEsJ?sovAA<sQyVQ|y*<(p-QzO&VN@2DDfb2AInnyxc!(i|&+m>-
zcf9MLlaCIyAwJ(Og}hz=`~@9f@yZDarc^2#EH%_?rK}+6HHzI*=w(<g;I-gSm9X2{
z=}ZlKTz<Bk_9K_HpS1(y6LJ)@tCnH$e53Ch0kU*6I*J|K$`#qoZ<$)kJK|f_9%&h~
zr6E3+<&_z%%`?qPsG71gpgPkZOA!h+0k%+7pWiD(5<7NOw!Z7A^*0JLRz$5^DG}8s
zUN4Dl$4`gKWikbY_O#coF-ArAYvp2>P9mQ|!K8cIL}bW$=V#s`<NG@*DFhfdr&0wn
zE|)rH==pIbMHi1+XPW7>C0EB0a40@ue3au-s$)m8KhGpFfNV9TFI$e}7`x~DN(o|h
zP$0Q*bbg;H2|^IRXoI-zX3+*HHKe@dPOYR^kwtXP7g}ugbjT_wWW=Gykorr5702<p
zL1J1b#mg9KtVCb=f?D_tz2Qs=yC@))(UF?RUCe$_KbrYjwVTqV_OH4@DLbMokC##b
zUh{G}NeXnLY?-H=VVU5xB;uyTMAF6&@rfkGDkU}>i4lUujJrWu@cklX4Dg$uI>XrA
z?Hgy@f0G$N@Jd3fT%0BoJ8A@p1938_-OCYhKJZ_loUaAA7O;m4LE{Qy1$?~RTtRl;
zYa-v#h&cFJ=6i~VAe<d=W^Phhg^8-EJqoJxDO-jY6wj@suZf>GhO8Z>>OzF}5m;#a
z#{Ld-aGi*2v2DyOIh4R?xY22sKpnW>&5II4!ANj5nAmZuv-(j?8?-be$&5dOLiCW>
z!8k*!5f%uhb~pG!iIb5IbnK|Pmn%jY7(kNYW}zVos-uCQVOG`ZhJyVWRkL~AHCW?@
ziID7qqXH$g#oc=6DU12rI&Rqi&jxT6D$9IYAK%PnZ>@^zKq;k~o$<PbnP+@H{D<5T
zKLSRxXe{l@!-VJL=1rl#papUbYEN36!_RJOwKIlg%#-^!hfXetvkHT98{-fdeRhQ;
z;ufs8U`{4$KPC}GWW70TYd`0Eiofq{q+?g05mdane$4QHzI#|+ZXFV@HRk{Leubv{
z?E_o8*13oD--k!+TJbLEZP&*c;sMW_*U#@AEfsy8{*mRs#MK&gpKSt@;#`{v;D07s
zo~Qphq3hdpjva}wA0OR5ii>yaG`{)!(Av6}`zL*qaBFY-g^H|g_3`AofTsV;iynEc
zEvGFbhh7~91m*F7FJreZj7^RcBT`R7Z1P#{6!3pO5W5i}z8;(2Z43HW>SvDZ%?<jr
z^z9k?m8)a@`wRlC0PYn4b}^;5vlP!&ba(KUhuB?+yM*@XpIEKC4ZrgDn3%_ZR4&Z&
z0`odCLXE{Qmn@ff!deN6U6lg^r?XqS;HZ`z_^$8Qr*gAQTJ_UU(UWX9b(9$$+A%Et
zRg3L68Q=cQ-n+fv3+Z5$6f6w;=${JoW#;bN6$JP1HCLo$EI0=MGHWgj!dtI@rVK~U
z?b|=KFQbN2Ky4LjHKs%NgkZf>b2)v8`Lo+4#Sr3xT86N6f?{sLKr=r1$U>^^KI69f
zV}6XTn6;HW!<`zIu2~fH<iN!5ZB;94=Mw}fv+0Z5C)@Xs)Jm^lU|>I8NLoDC&kCk4
zTa8^@l-)fXAR>lJWMQ}y;J9|t40m_=9S#*{X5-14fc>Aap^}N;Tc%bA&W8xCZGM>&
ztW>LQTt5ZsPKv4^S1Lb8LpgWJJYPf|jh2$04l!Mp9BCeaf7d6p?P{bK+v19CD^`FP
z*s(|}!-ovIfJk`P+P}*Yz3Bw;YjFyPfYv5J?_=#P*s$lm7X0s`f4%45CUjAu<@v`1
zN26}7ZlpS!zc*=@>8^-=E8UA{=i@h8x2ejHo!5_34Qc1Q;GW03)t;9#2At!)=h4i%
zlS$;T^Vtj|zw1@1?z=lCl*SV+oK~vk*tCml;&s25m+>AyG#X+AfdzdMlt-pOA4Y26
zB;jKerQ)Of(4RA-47-DRzfWb*T~Yg<?2@5N!9q&KpLGSL|Lu($`?p_b#nYn;WIvyp
zpMdi*n0#ApMZe=w^0&gan$O=nMtFU4s`Cux!G9?5i-l1GJ2QL`22Y|E<o$~J{h-d^
zw`C^eNy{ht&DGQV!K|b8vCak?aNQq$YpS?G!mGz&H)Rx3s%LtJHk~RBFN2zjfvh`&
znyRggzQ-0!vLen;OyMxJed0A{G|Lq6d=|iIET2d3owTE^nh;ESy&RCxYpDW`a@B6g
zhvXnO6jNf_@}wX2kbJhg%K5~>GbEGiI`-K3!iLHwCncPcV<InVD+BHY9uTI6(%yGV
zU9t{y(;3T)LRz2uq+)5vC<L=VG-%HX#^gWzWJ3_lS~$4t9CuuKVu$3&uzOl_4GYEd
z%tWc59_%L7u4FJxJ+Cxu>x|wf%iuKzlBpWY2T^;c&lG7pO#QSrE~4G%jgiT1TzG^P
z=KI5!NpB&gAav*zZzwMf<jQB#H<#5ac~mJG`v>nd8g&S&CY<_*Q=kh^HHzh3z(S=M
zUfOVD$4nMZh_w}ns6O<{?!UoOo5w6u(y{_;JABtkwvsobqlXUEiP?H_8%{`#pwecK
zDFEr@6Uryi%lrq?kMPT{F@|xc`{`72L#gl?v|NHnQ3&Y(EgR%hbi&6Z`lRoOpLStC
zJ{@Pq<^3Fc*r`-nc*GEQpwuz=$e7Hk2vI4Oy$yJq;P@EAOHASIrR`D-$=P1B+PPpu
z;3Zu=*dk`jkfFP!esm?g+}6`e*83PElyZs5)KKh2FVmmw;|4Tj3{4%(Fxur&ktK+*
z3(9iJ)u#LgT#5@O=*;CsNk<5BNaAdH?DKk*L@u>m^=AZ!{G=<^e@3F0SN_E>h5D-l
ztamTQDaecFeiJ*2yGX7wHnpC}r8t6ts-v0^E~LdEQ$+gN&TYqem@`IjkIR)A1V_Rr
zq&)-*pS3DWQEIW!JWo&6R;8gEO7<b31GgNRUg!Wbi&kgImjbCtaalH=$ou$jz;!_)
zmTA0*^iY>Ob(rH0u)3@=Q%~HIh}!Wd%F%owClgbv*Uys9EoH|z00qDYlw}c5fJ$ZP
zL3aoxHDj3S#wxU(U^8#?#&`9b(|C&?5*6t(4$Y^AN%=&qi5UoTZ-8=;Z-VkSVzvVl
zBj{pi0~`&T+?ziGVt@L_<_ixrtn>@f&oK>yqGaOE1NKteC9SUNX8YayG_c^w48@g<
z#Vzt(n+)ZeMOn~o%MijWOoFB3;@%@Pmmv^p&!7yAGnvIJ`Uq5-a99+4#?li#hKw|j
zAn}P%A)k^U?vVr$t0>J6Bio9*jx4}PVgqMZ(r?J#9&T2JcQK)Vc3FU!ltvVcbE2En
zM%)&C^NaBR-LHaOI?hZRIB82VZkcw5tt)zb#bKv;|B^r7-=?$s7$S)+C)!EHySp5r
zgNfW9KIu3v6o}{ygZ_xS_kJaT9|bGb>=hEs-ADXRN?FnW?G;k4O@_Bz0qKk90;EUH
zJ}UcW1aT08+r~a|vazwh@7z?PQlOV%WUF*%OdXlb^qX*x+;YkCi9F1<Rnp9-7bGiX
zx|Qm;psfl~AAQJ^8YN=E;{Eaco8NlMbboL7WGriT>F(vqWDb`b-Aw;uCVqIx;-7{n
z?W~XHDzrIQP-O4Cqcu?h-1ZM}+}oCoCMvXcqiaHf#|!7C-k=qBW_bj+(&sTIGig~q
zKV7(N6z&jxzP$=iz4c}F4Wy~;CzGfYY)mj=S{f&V6`#mgGu=o5%rv{t99b#c9-||&
zFxyv7tHC}Q`|TWUWpqkFh!jy6IhxK4N(39rBig5#{&G9!)c#BW6O4K*Z}Py|Ap~H2
zQ7?Tq(xG(Cakul*$7PSo&p|mT<E!q;4_8p`KDOzU(IJw#jn{@6&<VEi_H*HbZ2p8=
z8aR&MCUQ$j{V86g8u5tH3K|8=4QlFG!h9mE%MT05zm(-^@`dl&Q2Xb{5ayz=jIdtS
zcwWjEtWE<!>hvI$529>Yq-9qKh?&x`6r=iU{g?YtM=E`@$YtF|(DZb9ERjKl1|M6}
z=g@du<y+LzGGq%(_%K%6&=ATs-%yUz@$~f1=8Fy)4$65l-dN4IC;=$;0}|-ADpsld
z&{5Tt9u}VU$-n%o|46{bkeV2^IF3SH8ibhu#17UlChsfD=Rs3qP!xRSWZ7R@VutVh
ziALMQbFBy?hKs;LhM<lyQKd%HR~}3igo(XZkL&DC%Itipq|E2UNXCfPT4wKo=_BBf
zM0J~0VugR9PjiSnQqL%-k|}fT#Bq%EoMNNR&@X^kX@BF{hx+|$Svv60H)Po7Ryhj5
zw|=}*tWVgRxs_5YRig)$**|zvqkMm1a%FQrV|I?_FMF)G4^mi~2q11Cjh9`xUuR9|
zKsB?bk<P`B(p+9SE1d3sYj-A2d0hkbS7eXmy@w%VMI}NQ1nD2il@kFPOozpL^nxh2
zG`i#&|3}WsP4nX1$B11OyT*51GlD>pIXGW+OF1fFHCe3^2Ja);^HD!!c6bEzf`rJ5
zCO{x2<xM2h)6Qh=4$Cf6F0yGHJ|>8nCO}`#70AQI8XeHCI;Q}y&&Ln?B}BcQTCHih
z(U1bcfK*AE4PLMrsc#{udP{sU*Gb(~Hmt;UdP5}$d-MU!>zZhb886%1aSsHOGE2%-
zFZAvf&!V*Dl+XlOqE>THedtgav5wtC7*w!;bD!IRAT*>*tBAF(NhwRyGyYZ}>-DNu
zg~3ojL?6e~!_I#-4zG#g#ki0Xq*@}`*knhOu1MFo3#UjMOO7@SPoqs+0I8Zngi@5n
zKyu&`3QYub)^<5YV--L`rMFOaA?6wp8DE=WBj+4-#0qQouyqKjIvmE1K=BRpw8exC
zG~3EGaJ2|V%2<iskHsa*Ox1RB^D~;O{3@u@bb_6TYQ{pIZhisfb|~aXuv7$!=Fd<B
zN|*O;)-{#8JO7sek{@)4EG*P#&?tN9`TV&2kG5vYD}OQpU&7E}WU=+ijSF?M=H{(w
z4CrO@K$Si)FrvpLy5U0pZ$9Tou|mm!svuE@Ev#}AX;(XSTUmuA1?Hojmxa*Wuu5rF
zVcyd(eL3$U4TZo~Rr$fOrYP5jC?UaemqHLh#4#~}1}a(;9_IPW;BeC!sXK((S*rRY
zydRWUnWF#!)}x!EO9R>{3}}EfEuo+5xN%jb;O74RcuiQH^u_31G=y1Fr#cb-w>&Qa
zawEICQBJr*&c<teKg3DY&ibcPX>CC`LGKzszXTCcHMHR3Wp_tslw%Z>RVwj<F4a_@
zC_Kz`=Z0NC_qZdCj1RkLvK4;?NW*>=Kv_i-sQeqaK`HoVE-FA2?jtX?yFblL)X^Im
zjo?e&YgB|(TPZBdTpY}O_E~6wldMDy!nt-b{UQZYw4;UjTwF<)RF#$?!K3>8(LfH>
zeZta3cV>}kiJ{CjWv37?by`8?ZLoT@3&&Aa{{Ym9!{SoZF}eYLaLpA&kwMjpoc8Vj
z!wmAa(<!#T`W8Vm9&w~Rvxwl=Yuw{49`&tMaQ;M`dyG&~a=RpTj1-ndGbzzWP(?cR
zC*$0M6bZPN{DG$3O69XvD=5MvUz#}+l|6+{|4vjp$0+0kYs$ehK{n|Kdw3Mr2T8L&
zYRS{+NKl|#u-Ix?sq^tOP9xz%p#`AW?jbe2mW+s9XX@;Ltmq0l+}hcF30Xe7r6pql
zq~~3^I#8MbVvDwwJpyaA$2NI$2bDlTxUz@s_Wplr9;N=pw6*iqokiV^YnSvU+7(Q-
zh>7{62ky)#O3Pot9BFcnD(p~D&QBBZ-d@c5mS&^FpqQ&HvpKa+49t|{ssD~H=!~*k
zXxh~i0VNlSXLwfpWI6<8R#%daZ97$gq;qqtyhATpN7PbhE4n(}?YicMuj~<uK?d}o
zw}nuK<QhQ_Gwez-Q3;pc#GhGkZ>&HdbW_J5hqecV%~_m|DayIBnepx}Gfa($=HIG)
z%?zyal%crhcn1TdkCvKg`Bm{1=)=S_%BUkYybf!yWMY{MvYL<?jjf$9pEk4lmuR>*
zVY;F@o|}WiElFuY?r8!t)sz$8>D1SFEe)Or>i=43M5$@!_SQAKemf!U?Klt;WP=<d
z+`NYAQfVJ5uXnYIK~L9fA<LiL@jh3|8H`WXpa(Gmskao_gmpF76O0vj>RwPQqi<{?
zr8tco|Kluq7Bbcz*J`0z3pH{*SxtA3OW8{FtF;D`X%B5<gQK5b9=a8H45frG-q=11
z=!ntd4eUZ&3*YOZKAeHTbzw^(Zq?%>9A)k7yjE;6$+%VGN(E(kM!DlS@)#F#W^n)-
zk%jzY)}c#3=gMj%3}5+;r7%z?DAt9aalM9!U{)hI_>bW+%faILFwcpt<SWH6fAElk
zMh8}Q%yc1MX?_vJhc+G)4=Pxt();S`!uhb9ePrpYcXU|Puli8vlZLvfjI?en87d|~
zePXXOQg8~ih+W$4=m4qv%I7E+Is_-vy9U_IBXx}Z1<H(@A~*O3Dw?v06dLnSI5aYt
zY59ekr-#ybwmibq&0k1fIb@_ljYkS~zrOx(tEsF}tx2eY9d&CVtgZrKx5)=YC0VwA
z{K<}(=|3m_uYfY9hTq7h%jjScQWlWM6t;G9qPYWw)hdy%fs&S&KgbuNL0`sN0C`Tt
zmQZ=kyD3`*{|2dRTQgG!EzQ?h-*IEWfqW3gU%rli4FYf6Fjh1R0yN%|t|SYUbm=gL
zvifpk4OC`GS;wh|BvkEW<w_sBuSapR$VTeHBm7Xt@6B((9WmUO6=Oi!J!}DBz}4{;
zF;4%BJ^BW*5aoHF<LSS@YZu0Cpe>-43T-Su=$ZM`EAGLX^AJ}8G`j?fpl<f6@z>Tx
zo-v6&Zm=c@s<>YWG)0V5cInlXl;dRtqZ_2`?H$dvQ#X}CAKnmIm=mCcwLqiL<*ylj
zZv|fJyn@fvMjVjxg&26nh#*vG#jJMcwMG%Ts~e#FHA@4ahS<<|SvM$0ENAAqU&g9=
zXC?=_SQXWyLTwyr;Qekz+h2$g;gX=-TbQmT5EvcYfzwh5O3UxmT#wm3Q9RZ-T);qc
zryfhdgs*&6A9~c_^KsfY&Lc<~)#x4c7?6CL8uqYero}u%&mgXOujTIHB3nr(YZX?w
zGQm@ovYE4jKcii0q!A2<tMTGlrbx~m4fSwXM%<JvbkdBd4NNp`0w>0o5@SW$U*yp>
zc||dJ6=J&5xnbbf8X(W2wFZ_7Bi)|(>Cq#d+aV+zG-c=;ZY)zO1e$3#5eJCowbBFx
z3n<%pEAXLfRw<*RK?C9P4hj&lp~D;*Yz!P$7+%E$h5oBFNx9iYQE{RZt=D+mkKNr>
zKi)0iu3tS<F0H{UWhG#sdQ7WjBv9Ga&}1HI(PQ|pcAfY;Xd7A%7%5=wand&mL7SJL
z$Obi9er0uf#utFl+KLt$6a~t16iT;d-hNQbt!L*Cij?vFj|;=M7rm-%YMGrCbgy|I
z{I++r8HNN3aBT~1puk&~%CHoUL<MaxI!URd<>x3Hit-27_jXkvAo4R#7K%h%LP_4?
z;J;=}YVoGt`#D%wJs3@>V(rNYn&xGUGZrbbfEnlOl&ksw(LoGj4r$z4cNup@=Uz-t
z&5Uf#Dx(^dp;P@sp}p44O>YnRL{rzD!wT5rwWvENKYvSpWp+b<3j^;65+F+G>yYR?
z1ez!Wbn3a!M{$EE$Ax(6osxn?y8`NXBV&3++aP1slv-v<i|R4*QN<zPPEkGA%58|(
zvr4NLa!gSD(A|09ki&XpF35mm+n-<>vSz(9!Okq{P2sNZs`=`lu%+|0GOS@Ud*4^A
znm0*`4%B-NhXVQ1yP>Y6ujV5+BdYi<{I~6XG)7I!s?<Dm>ul{;1LJ5xRb&I}$OvO;
z5riLoaOtK3j{pVW-S5J!W86?Sl$y=lTC=_yk)datQE4WAWx}$zGYu)S|C?wnKT(D{
zLcwnJyG=7K*2A;692D~gO{s+45~nP+&zD;~GmGV&gz-I^Gl%wC**~<Ap)u>SREb#M
zNgUsk%46B}5>L1frF=uV*~6><Ub*5`yqrN7qil25hq^^7&d+h@zt&1__TGhrWdcg4
ze`3%Qq_S8>hM~D#(Pdf%s)dB24X)%vOCI@%ZS!rPz$Rj}+~@e}logKjRH*TNn9zDN
zRM4qhsze$`zpFw28jHV{g|<bD%AnE9cz+fLDJNGPfy%x>vUsHi;1G7?zma9EPa?f4
z8s<kI{>BzUVwoKwYAV+?&aC24U+Gj^i>LBm*JqR!1q{y(Z%^B3MB%UG$M17D=D&Ks
z?B7`&YfJ+<nvDA4IniSg=oOopTaoa~7Kt0%`ONF9*swLI>GE>3-V3j$r-gy>1eojW
zaw!0JvvDtIL*k8I5Dj&X3*#dSlv0^SP<xD>)W7INqC>MC8qTj7(s}NRRGoB${RM$n
zJy;4?qvr<ewY6kk$%mvlvA%`sDYIFMty?6d;q}VAe=sZ7OyUuDUMc;lrBB7t^5|(T
zRCneLqO9DkydXbOHZ+Tmf&U?vbDn}GcbHW^ZREiE2`UK+WFzBbLGf+CjqEi{Ldh$X
zsmj-dW~%vRGW2CmAr;78tZe^Hc}w^@rQ3gGANgbH7yD}=pHnFi=fIy)h8o8G4#*TC
zowZ4OHQd5c`MNAt0|FH{^x_a=vkQa_4k1FjO*jhoF||0t6X*#b<Ub$?0=75Hba}Qp
zy9=OQ6kN>1&2?)l_L(F2`EsaMytDhCMZ{O1R<9SfgUN1}_m4jRM!)n>*#&J+UaXhs
zv)^ew=eGR!#;QRk?cK@I^HA&ib(UwHzb+g4h+mZ)^dF?eRz|TI=x%QwhnB$<itRd~
z$p-iIUk~C@O<Jjf=fQ^JJ^qh3033O6S|j={0R|j5AM^tQE&#hdjqQyyPJc_@28^U;
zG{%Rz*afb{r>8{n<o?t0Q_tF>uz08MRFyuK0<O_iyhGgqe;aDgi~#?~=cC$Mb8N8D
zg~0Zxm9P0fg+9uu&X=3jqlDU?Cg5)ka2rGQ@BHY!;rh#+W&GC4D7oO|BvnuDp-}+l
zAESSlt-QL<|A76I^?%zIhX3v@je5%e`Tw&VHyJxJT1WDnkSlMx7w~`HKCf}<>FTyl
z?CZXn9K4I=D}>!0Pht)1__OI2!Q!&?P?V!f{Xz~dP9*F?zu=@~GA%pe$@LbFnLNPU
z9r3<<t4?V2xL*}5%g<<pwf%FuwcU`Nwlw$4ziG71c_sjDo++#!g<&bn1Id&e3?C)a
z%CaUzg3wW;(6utZ4`VuoWd5bJUh^b8WG!AvcP92!TUGq&S2?M+Y3cK?6VCc3%9p;5
zCia)p0KvK@v6m^=_a2u(KKn)dSMS4g67LeQ@A`oq3AoYs&%s(rKybX_Wm3P=-_BfZ
z1v!R{$aP?tyThcM;(0aK)!4tgsHX^DU)xlU%DevWMCTvs_2(usZPSP+<aou?PA$X3
z&TZex1f>Uifg)diT;a3t!a0ijxJvENu`|?ihDYXS-6Z%&rCwQBRp1>Pp}L<~f`3&b
z$i1smGTGJiecQ}?;$vi5b=Rm#*ZlPrBHW%Y2>@=f(r<-RU1ZPygo|$v&t*>&x!=40
zN+K5C8Dq(<0+e=k9{2G|SUYAPQrsGX=sg-TGlUj@#pK%UrNuXX*?mAIaC=^{O%nz&
zo`A`wJ0f);!E(2wZ0<Q1we$OCtmplD92Xn`FVFj}wNcQuLysoY9n6~Ei=9YwO#F#3
zHUip(P9LvIwFG6K9u9yGFNE96n~(R`hY|CWFGufzf2WS?0k`YGzx$)<TSub+$Q^0m
zxZKHkJ9&v)U~R(s>{Nl^%kf&w?=x&N$YKM<<|y#EcjWl83y23ACn)#GeipJVbwDxl
zzpt$AzCWq$enI)rRq?V5X=kf9dwS670Umct*Ps8&y;TS~pP*uALwUpv3}9Yn6t?tX
zKZ%&IC~MX9&q;E<cL)t&|AH(IIgy&&eWZS9kL9G`{Ox)l5E;Opf_W|ME}nt2;c%nG
zuPGY*$01k5*F(<B(^anrd1*VMr55?Z0YQrj<%Gl`)SG!7bMpw(aT<%WX8#cv1Q_)!
zd4x`tzWX6p>eV~nD>eY%jeW(GXFQpHwsWGBYCB3WsOCy{6w)k`LWQdWe_)rO6tyb?
z$tXxY#%9p|;k(-l$1HEoB*$gikjF{fWH;f`VhCceGJvja!pEX<Py0)&W~~QFzeCt}
zWWIl}KLp(+fmTd>o-gT-N$>3UMnGOLxUu{_y<X~<<mC|kvGkirjRu|WcN0y^YGX_N
zBO0^EOe>(ryI8Q+KapCEAVjBUF|L>FkKOX5^`=11<+bbP`Q>%xB8^cGsget}+o99{
zo{4Kk;BGwBM`b*&)ExvkK(sj9=BP3k+LIdoyoYGj_%+!l%Zn{#i2}PvTC*03wO~&T
z=`V|PM&f7K3BDTw=(u^=cOv}V358-Y1k|c@A2&aaWJZ6cU|ytIpYKLz=Jci1Y!9%a
z07e&AwnnCey{@px{SRHMz|t-_|3}kThqd)QUqf+sE$;44i@UqKJB8wQDO%iuySo&3
zcP;MF;%-HXzv<`ud!L8=k%Y`{W@gVhGxuh9jGamn3mQtYZ&L3(v)d}U_UTvA<<sfd
z9p495Vo^zkaSX4&jy>p(QlA*x{{YZs$T}Z0F%qYcWmPfA?HDMi6eSm`3Ml%^%fB-i
z;ox9!Xka|Jz<S+668R@!k95%PmS%Bh?3C$Bx=2*g$7X0^OS@(4d?#UueU?G~Bm@;Z
zBktb6j{n^PUhOAHAzEdu>XLVQpg>TW*p$wiMwFVfp|nKF4!1l>Ge)*-7zfz2pHvyE
z<GI5kTU3rLD$CK~T1<*A@kyN3P$rh8SmbO_QsdZ_A^Ifr@Tog81cdCoONQy~K~r0v
zgUz+5-Vk{rtE&E)-nOFDD&ZtuI`R31Qn*{NophU?zKH!XdxSO{Ggmb-1bc4sfG3LC
zDyqasZB}E(Ht(d(9_0$Y{|^AKR7qcyA&D442|FsD6I&W=XJ_?UL&j4}^|SR)jTwS`
z@PBe&!6WkpmC+Y*$|4?bPB8H%7Q|1hXQJ^p6#3Ldyh-CIRZ)m9HmvQ1#A~@pGh~R3
z!!l5XrD5%i?LBHIU~?zeK7OG;<>J<C7|&9;PZXo@?f8qq_XS#1T?XKe!p4+#ix2<w
z?z*s`F~wlVK&z;@udb$kxz+1VN&odTzj){%`l4k`ga#G2C8wB~D%eZqW?*AX8tS~p
zA)15MI5@KDcnnHv>d@?BBpfBxGLab!M(XOtu~U1rwo??`?%L+2F}Mr(-20k|MV;^k
zp`R??+9~M2%Iq&J02*T%85OHo!o(v%cVl)%vt7dl^sM&CAS2|PA4Woj&x5X^YF4FH
zKUDl2Z6#H8ck_zc8J}L---2#3-t26tdihGi6>>OW9An3uVVEuDI!lsu8SE3e6x}tk
zkt)7Q%4Q#$HZ`SSjkARjAelvma#$)h=&DoA^6=og&`@&X0E$hz_Ef7<&1^0u@M#R7
zf^`Rdr~ZLdPv?9+6(z$vyD0Uw`qXfrvyzvWg8>+v1<Cqmwdz~ikNH%)ptbu&nB)cS
zri>-r#glhv7*EwOES&m+qIhKLjjTj3ksoxyb4$P0)Cew=1+Dln)bo`45j8>c^SDc(
zx4o*ON7&+UAk<4mCR9zrczu(qOepA>9Ls-!zm^%Uxv^ls3du1z-rzfHSv(GPsDvX)
zbukXVwPE7zhSD(-!{Gdgwtl<16S=3DB(HURj)IcaxXRmR3`IyJzFfL@LygUcI!8=A
zY_k05I`k*1TdAbL=IZ<3v$wsC@naM;r3k|*Qjspjz-Q36ths)%qd#MCj?Fd2+yqC3
zTyPH;oAn>DjY^2rJ&+@%5j5tk&Z=q$^u0PT6g(+tq)FXz<kGhrgpjcOusB6Z0@ETH
zSnc(aQv2X<64hXJ;M0X0`YLYfOMY}_rNMF#4DM-o6LdCXuFYjwAdALFH&Q^Iay3pv
zP#31F17&DsWa5&dMzK;n<YhS2DzeF`j*Du?%D=JqsPh{UaLllnsVc#?5;{BCi7i~e
zm)kT1H!~YDgd5CzH6^PFfJeKlpB1BTpHqrYg_^e*Y%B;FE5E6)jzWAL69I?VThpj{
zQR(ld44a~q0*!h}nG<QHZKs$ZJCp_DcY459zL6)VqSIlb@Ps`4Is!32u01Uu+whn|
zfk{aLDjnI3p#ob`YTb2dPR7Vp4hkuMrqlw4ULXaA4?7Bzj#g&2a#j~Ju)Pqu!lrbj
zQjH6SPO^o|Xnm8>E_X&?owl44QO%9V6@8q%tVnJ)W2wkSATE#C71xT&RuXkdmJ|?i
zB*uN{s#KkZx06KBz_2)N`gs%LN(kba_a@Yg%+X3L5VV{wi&x0C#6g^Ko0ftUgMr>(
zNS|(=tn`a(_J5QE1cIp~WhEh(WOFU8L)31;$B}~iR7yQYy$avtcF!NInAkrfnyRY7
zZ&l5FB!HL9wJdFH0=NXQg7j2ji-D=g<{{Jl+dTm>?ILmgo1vNdt^&?{0p%yAi4~MD
zyf^6MlysW{@$Q!*_Z8TV_Jb$n&({$rMe51~`Y4k+SAy!MDqM1Tb*!ev?hT?#o3Oeq
z9~Zb1Q$70{9DMLO<xnX-D{DuKI<V^Pl<Oq#Xjg(z5Iy@mK#FMQQV#<*_CRmONIybD
z>hjJ((%86yJ5@Dty`5SeOb0fu<tI3HdJ9Qe@`~a<+{Jm5b+tOzO>mldNY-;t2EEc{
zk_<;D4)k`1F;`D<;4W7GLT+YE`w{zGmL%ypAHv?>ertn0ZX=hc1rpI!gcT?hAlK7N
zw{(-b5sQx$6Lp6@!D3G&0L*wI(HSVp@?-_ADL@|j0`}q2C>WlS9t)DES*mKZ9?EGy
zCSOBt2tgF*S$#%P$t-J=rt*{s%?vkx&}Fk3+m~%MkwP0P!V8L%)5WcSM48DWHGC`l
zrOWd<nt(d|H@EJ-Lty=QH1{nWRum@9BJuAkySOj7&y={unS)@-3{Z!bSs`k_myR+l
zFF7_1j3#SVa`p>QR;Y7m#7j-&3>6OAl?gED=_P1Rn<|GAWTNJTO0=Of%GS+}iqeJW
z5rIrPRYO#6(|7Bd%;oK!)JikhvXTt&3k$mS&r}a169s;t1oMLPPqnOQ3Rb|=&b7Lj
zTe&=`3V%5#y5aHx;7=|8E-jl~fj7Xv82x*#z4{=UTcu_bhd`$k()~$*K?7|yT3kZq
zo0w_gNXn95nNgDkU)*Pc%7X5T_p*QZDw^owe>v*VIQ+&Wmsh}W-f8z>W2?C4JvPYB
z=vt=~elo~{x@6=1#S0`-o)UK-K&rUb)THwwa{&U&o*O{y;>buQo`|9`Az0-$A5$Yr
zM&3MbwCT^m*q$ZLBGpb%s?|BHRiFYSySFTdQ_LL4&nvobh_BrFX0t1S^~IsO9ip}-
z<ghqF%yjyC@u|?9A~a9YHujz4)dclatj_^?|NLGeurpM8Xtz*$LaM?D&Q`QgUJmp~
zAvOvgpxVJ!F;Z$Va%YW@H3h1W0`AnUOo?`EscW{r1UyndOKP%KH6wFO7vr*1#ey~c
zgP08nyC6ebPcOazdS3K@gRPP0hR>rCeAICohewJiLmxXABV_i3kHe&M@LZi6pUjt+
z%HSsg;o_Q!l8=ee+YMVa(1I58lz!g=XJEuGQ2S&GOr|VZ*!S0@O0Jgbk%U&!(MvRq
z-fGvIM#!@<TNH4T(2GTM%q7ZOnfZ+FVI<il*|Mf%<CZ+bwts7@BK{<orlGDV&xkb7
zf<vc|XGxkMjHo-?72gp1Gg;ipzEkHCLjf)QsIk9%%?>>CRwz!flJ9ACFxEhES0h`O
zIg2z*93~Cx*zfM1D+C;iGJ%X`CA8*W8)>*4x`Rb)?i=*0?sQfm3bzC<6&~3sxKtOR
z1WNy)mcA16VbCL+nOSC|JIO_@MjAfqBGkfRd37#mkT{Km!+S8~T!8VMDx$+Y42xZL
zUOZR~)3cbmW4gDK>!Y=vD$ozMuBVq&&5!6oEO%^5es*pso7n14dIS|INdO{qkuHSK
zrw9^2tIZ}ETN7tXb^IockYso6qFaBjpn#OlWTsaAxu#l{(Z5!Q&V}DHB=o1WOy3d-
z=Y3zvsl7>Xf$E1)M-j$qJzqV&G)HQ12*h$RVAcb*1mrM1O<+d>Y;^Sy@Qi+-*vKRK
zvU-lMEA$o=p^w3Gfw9sO9b^I3et$2NrvnnDN4l-Sw#7{ae@KD`!%tsa+o;Prx>5O0
zf2R}rG}sjg`e9mnO^{8EY*F%bC`^Mt#X70Nr77X^iGmPSP&0qDJR;z5(VY{2m#W9Z
zm|XwZTM#1K2;dpUER|FK%SsL31l|>+#$p(L1v+A;kZ!4;7K*68!OY^k|E0gFZvw%`
z`dk&MkouL##<>-vLAJ}yS0=f%DynR0N~nC~AlD4H`=u7VH+x%A0jzNdFjZkOuBAK)
z$N4w#g@YMNZgu{++Cs9Or|2sU3AKXk`upk};N_8KMmQ>wORH7SJSBxH#t&V*mI8*G
z)h#nd%;IQ!Kq&4al+Qt5FXqYV3T`2igN~{}drj=vVL0rCHYq8+Zpc=CH+?UHVE8ys
zfm1H+-*d*f)s|Pt@l${M58xx-69}vWm-q{drdN#qvT~MdQC-&Ij3IOO7NX-)hXPqa
zZY41vLJD}B2bzl)j}-o+!C4;e#zu<lF5YuI-A1mI=>`g6)C8~%-)xfg6fM1a3PaQ|
zTJ`vJ4aUR=qxb=+hRds1`%1w$%#yru(>O<YMq<r~(;r<&HFU9@BL8r_)K2oosENWm
z7iQI`jW#C~zb*bP6N@9s0){ms&BGwI?gJz);M=&nCN>Qoq_149vHEd$ea8p?3k@!{
zLPggo+?Wv;`zzlv$W{sgs~OMHc8FtdU7&kTx*vghx1#LQC}5$jph#(qLxQlH`1dC;
z##q{b{jxfdabt1Akz_}3gP|YhG|9*E(K2H;yS#dru|=g(S_758f+sI!_~K4Cff)Fo
zo0Jwg_6jWs+B-j?{To{}`5U_&$5sECehsMtqg>kD3pEQ4gIL*UK4?{oRMS^geisEM
zX>LI569e>mlousZ{!#YFsOHDK;sgt>frnd*NDB3SuxCecpqmPy<?8tor0m!E<2dLn
z3xkwbcEE|$?4~ZIBO?(Ccxf7yK<rbG+{Q-;#b7q`rJGLqi0Kl@HHklloU8Db1S_Dj
zo1S9Z?;!@62}{idgs8Q}tY`GNeq46Lb=p;0ZBYk<4p?Y3>t0J?h|SXW)9co=q^|$B
zkZD|$9U9vI(_CY+u4!CgMslH3M9p34ZYsW>JS*a=LkEHpRhf0)22-lSj-4l)inIPL
z^z9q#AXhKb`<NxP5*%s;TF|sU;moAkBt-fV`isW7{@dVv^+zxaho{vo+Lx@qLHOAl
zskxfeh;l_;D6Cd+HgIQOY0Hds2B0K!ZmNR8rXLJ{(}D+U*wcj}vQ43n(T=A2U}PW$
zK?1m#D&!b))||*m4YTcEb-{ck{9}G@&B1=Sxh`|AR}7g!qA9+KYpblzJgjLpQyM{i
zmWSc58P&0$={@rlcz&Dd3?R1kwQx9PI!19>56(3@k+R<W#`IMR-BFtV+~=)nwcJS7
zfKc*H;g<a8Jtkd?Rl=DSRD-ig02@$EXPS6sl}dRJIaC0RxIU#**i;Q31!SjaZ`(QI
zX(ly_jiIh(?T6G{;-UbTz7p3|&eLpC7c+%i>d41zUN9R|E=teY-DmP~T~3fLn_zz}
zYO1gZop$W3QdIzBArg~r@dh{giBp0(hF3)8A5%5P)5#tRaiNT<)8SGBDzfWn8*`CF
zs0R+hId7|COiY7wVpJ3BHVY5<!vADcc}k0k#)Q5+66cuU{+h#7p)dK8%v>8<koY!!
zr}!-0xB3h>r{gN3`XL@3B`}uf;8N&rS%FiMDNL>CYh6VQ^9tRRz(e0to`GPIHwYV(
z`q+>AL=gColr;dbWqV;B+aED8;Dm(l`cZA+of-2f<OW>NxT7}PG)(uU517f$S)-7I
z)*Tsve<s(wIGXAbf$)FdjHtQO(bjT_pO1A)ATDyJ)uC~kfZ83FF1y98pP_2T=~H^p
zWYO9Rm<QZRz-yjooa<i({D3L__u8r-I<O`n)E!zA`P98y*Ua#Tu3Y_zx@1E6J%oc0
z#34x3IcpVustd9ThuiP0XpAS=M7mzdXr5C_DquEj9Ed-&YM|hJg0B85C7ZqCWEJ+~
z9UMhuxugQ4rp(ahgSh8;N!$6>Be3M74yB@#)NYS`NnY=-#H~-k-LuS9K=h{tY;`{R
zdxI7t&aF6!HIQ`W?GxE`WI2p*?l$vH&s+CryQcHrIMm@Qn5%?9Aq3!qbs;*@NvK0*
zn(J2SqJ>2Bt-uj#uKT%k-6S4cfsCBEXs=}<m7C{24gM-kKH$Wy_38TYYYf|hX7L0^
z48!7A=&}3<1y2wgjwo!RlznO&93&~Zl&R<}xNNmBx$Q?j>768C2WKUs3Wt*|6BPZw
ztesAMgoFGD@Z8utGsa<-RgF(%4BxndDZ~Q2`A58?EfXs}IMZ)=oSvaT0U1<Ct9Ka)
z+b!7V3S3~OVvt<V5blogoNSs+1}`6CcSd6*+4^`zur~Q*rp9bRE{n?%NkE3xH37Dj
zeti=K2QIS8s2uF@W1hN`|6?D;9hxyGDS#ctqM*g4-=hJX%M1#W@aYt0;3w%qKSC#?
zc|DLPn_g3-Jxk`{8>$-;LRD}@bv03QD>kc&e-M>`NN~l%D@!eSKgK*!yg`A>F6`z^
zM_a1wDKFm9e?^FD{A$tcTFV5urh_lNvD+gDNvl+BGG+byJWRSG9VJr;fEujqz^fng
zBzOJKfCk1rl(93f_kR`to-a?LA1zop&9b|sRM|Fw<vE79Z9}UrI1!@Ht)g8W2liFf
zzNhZ_8h^R!<Des?RuN6U)O|{3iUgPcUCIp>ir|@<ZT&O+-`ZE~K2%S<^tUxvsmZSY
zB!@WG9}QeOfTuO0Qd&UR@WiN_FXO=}M_@{RyTl}wc&T03jd&ZV2Ub4Swv6W?Q<%%X
z7C8#4R<7V_#V~~wczUmUl4Xro{x>oDOlmPzM~8eFo@Uu&sh*<H$$!f8M`&{@#CC>1
z#_^biWr=$7r6cRNFRp@C#-vyFzTmh6G<VA_Vye})%+-^q(&kj@o2hwzx$+P86AW=q
z?mTpoR?_BiRHRZ|iQJBAggW0vLrv@now3O^id0sHo98OHtBYZdH<tVTa&@+l@>SJj
zLgv!{SEab==q5Hc){pQKI+%CR6Eh9wgCXdnZDVu(4P7v%g@-YNAK1&lHwG0eXRyxk
zR9L3fd0n#}FBDAT&Txy>Y8q*3B+G`g3=D<VfNvhTmVy-{K%~|^u1UrIoL>N2EOTfx
zfKf+9@Lyws>QxQTwvi_rY#IzujVm54Y_<Gs|HPbZ{gL6N6qx={<iOeRuQthP)=gSv
zHoRgMW+S!$T))dU9ZAya-g+o#bC|fgY1;;?S$Mdu6j%=<WgFxUt)}UK*!Ue&I*SwB
z4$>ewX7#sB1-Ccp#LB+s)!jMEBXMM!fMwUEI<4+|&4bM6{2*j%-S_`trC6e0v9hey
zWtzP<Jy?~G(eM%8CEWg(4o{30v!<uH1b}mu4+lZIY(2spShLOQcb}J|KfN!GiqpmA
zRnctgKWQk%r3P$9tK<;ih<WBov93Z(7Yd#DiByA~Xtn*k75Z;w(K*m>%*mW=r@3r+
z{S!9o8sRWTuZxW-xJ5$wjdkeqNPSDfoLejVPw<Z6U?p9=2Cc8R%7wXZ1fVeGvy2NF
zxY$S0gY_YFq1<cU+M`+gncwnrx54tRJ~@*dV#1kUV?ww}jyQd$s@YYAU#+cZw=pQS
z(ePR9ANbqpf$7d~`Ls7jLDZ_>qr2l@K8ngN39r$K(fJ(0E<P>KW%!}#CSC@E8&pgT
z67_F}tn`^NOW6|z02o;>9ns(eE)3IBJ$@@NwsT54yqd@KN8}yx%)o~3J<ADdCW5JU
z?Eh!NTl%#?<EqWpK3JP+WSYpEXZ{-GiXS7RalkS-rkeD_nrt!$fiVel@7c!9pWo$Q
zFX|BEm!n@-(K*7t1=v-4|9Wdga;xIRl+6J4RvIb?e=n%+^^5$ju4|<2zm>bbh3V(>
zbV146O@@md1(sMTpJ*KD4Qw5-I_U?zM1G1{fv|&^Lm8{rpv@yy1TwE-KedpcH=q-O
z+k2W;w0EfoW@_ht%rW~{nhA!UCq{?F)b14?-6_Mae|=atlg36BV8(xU%yvH00pJT^
zD%SH#uCt|fhf$@sm2SZ|Cq|wB$VCS1KyYhdB>yV<>29t2zc{mwo~?qlNMPOv&}!A}
z+QwKR@`;Ai6enV$-+#!RWQXNo`>8OX3-yaxO-Gg_LXrhNQH)OlS{aU|icEK{TUcV2
zIXGRN2eI@KFQrJt0B_{t&X__EPtio<TJcV{^R<MrLP7Q6l>Wt|r#%(M@jR}_9AeNV
z<2hQe^P?!ih<Ow)2Uv?Ns^N1sV`p4D4brNE(y5+*J6MbigosUn8HdVSse-E3mI)lO
z?fE;pe~lk&{~R;A4sN{%l7G1>yfGI<Mf$&*5;za<O!%g9yT!Ot(OL{68(?~$t@WQm
z)Joi2UhO2T_D}p5=t?vBU4pv^0&;Z;g~aLvo*VSY*uqFVdk0(5M700P$H+vgp7vg-
zaeEa>zT_^|ZDh-3DH#%Kl?oVePwB(zoL7pU+SBz)!9QGtUZ>V>m$`X8i2|}BBoWlV
z-S~66T7Dt}D-b7@F|Ny49XF0dL)ZU=TGFk20GLJ8M+>3lnzWi5*_c|L^15qntP528
z_liv+ZNd0mssf3Mq*d4rcMRDC)(Qk-RKbpE$X^`^_f-3-ba`sEzpr(Z+k#gfZ1H3_
zZ7_2tFPMR0*)45#01-k-cpQEZ2a}A{6L;UY)d$GWX<Ba9^(%;W!@kPah3=)`<}+p(
zdLbCPP(@9nj#Gvkp72VJ*PYMSmPZOh#%vp&#-0Y~ORd5E1c^%$MqPzk&!UWOdaM5y
zx}HMP5bpfP`RGFq%Zz`rSv5(qzv`{`%vXx!Uljr<*}~8Eh*%Xamg<5;O;1l8_Q*iu
zDi!YBx?}AEY|lo>9Uddsbf*1Fi|HP)sN_!n7}}&lrT<GI0KR+6{tpqjH^PX*EA>wt
z2KDnSGczOevP^b=>#aF=1~)V6rre8vxZe%m?*33rw*s-j*tX90R45upZRgv7L?rou
z4V_~R<;r2~5%WxrDeyQv%gHHKm00t!r2_vZ98z6r)fuxCk|YLpgV;b!GRqvgnF%<K
zCI4#FNUo4p!aJNzu%Nh5*@9=q$%@(e3!A6H2`<~HJ>-!FVrtSAjEOeWZ9r_YDJ!$1
z*}D@VVCekTpR&@S8mIy(4j_mpb<M8z4i-W)^yZIkj*8F)=TTLj-w_(mYcG1x5n8JB
zvxR}HoX9Okqx!Y!e}y$NHL`b&l)vF*R=OD2mOwH2j14xKhKPzT;Y~caV3I`Dyt2K-
zT)DNZL5T!Y|7}!4BUrm=khJ&<-Ng44hiWDfe|6mZ4GxDM;HCa$F1~c)DWZ*w8QfPa
zMfDD{_)VQ=uZ#?4jK%)-Iq>i%Qlk>A-o3b6%rW8Y%b6*(2WJ>y-(nICzr2Czn^mam
zb!SBQ<tb{pMYy~A;Iu*XeTVU(zE&CVS#>htJl%cj=$fS&(grryN=c1`IfViLr?>M7
zW#UIeX2ta(=0X=@=*j!0c5S9}vn+pKSKmX=WD-YJu67^jTH)@uoBQXMcPLo<a@-br
zOvn;cC{8}!B-MQipesW@j;aM^<nopN1fNuDsq(hf$M}0Ne&!?-5!uyPggMhE*u>9N
zq{2lqa-jB7Pw*ewdel3%JO#bnQ*y<|{p&$<+<Ffh0N#5f;OVsEvRV59&c3QC^Q{ZG
zgL>h<TnEsAciV<_VN|DQk~jZPd|WGUeqR=KS{^Z@YuABYDGnX`$jyND=4Z-a8$V2b
z0*&1u4jZ+P3pREnuvE#Ao(!;S;g=+Qqq)M_@Hg`_J*3_Kr{t7qJ&nSSFu=u{^Zaxs
zgyDgM>k0`;i*T=T&nk%!$=0Xuo^AM-%_HI)QLEKQ+Y)d{7h=7e+9^Fjo@}rSG24~b
z9#Q3YSrow2doJK4?9{`XSzd}?5#j!4Y>z^Um2k9f_xs<I+0BuG97WONNmN}SY<X2~
zIC8{~(F5%C!%bx`P8=r|k<!Fb2qaP%LCymD>f-jqe!Wc44bne<^4nb|y5!08@Z_Vy
zp!l!O^8Xxb!hrByKFbd<x%L137NmGYYcx_X0F4IhtLx#c2Kji>EFUNjpFXg%`YEBu
zJ>%LriI48aYPRTnN2Z~v(Vvdgar>-5rNVyflQ#L8Q8{c3=1w>=U=o@x2HvT76V#R5
z)55aB4_IeUemMJWfTe@Z&gyvP$z$*n78ert_kvV=!*LLgSZwuG3C~S-wQ?gQCtw%S
z=m`UyB37b?Ve5N8JiS(K_5Ptb?eImTey&wEx7irWPq%-#4kVYj%J)jrJOAT_mqwwp
zj?u2vn%8Xhfw=3s+t#=-%*Sm_C^N&&H~r%Bwa-9__F%4domKe}ziOAKY|4b(-}IKi
zgur$J={aP=Ec0@QvWixSkyNy!z^|$UVNDwN2};5eY7W6u_#5h-(wf>?<t;c~Irz&K
z<%B(OX*W#XVUhhyEX(Mh@4jOb87+mK0pPM|#tFn-RWoutyx26g&)G4H^-XRh%Di&d
z-sHgtd=ZL3S$VvfvFfb{BFaP`im{JkW6>-<ujC{8Li_^kRt9(jkv$9ac&aebq$E*b
zcnu#`(JuH`blgDTo&^L|BIw`CQTF)JY%a`0-rA+17cj^NM3^J}CO$(#o}4P8b>~)|
z&|fsLp?=AD2L74lIEil?MfhdlsQE+?=-d9{q~?R{`jgG!op?_~3&vFl(*|KwRFexG
zD?a$$L5nQ>qmddJTqxrD5d?oO!2*E4_X$~Diwq7>68)C7j!_IZ`WW<l4Z(kjzy-7C
zR}dq%1apOEuOf1R>%A-CLhKg3BcLKei_7H5ZVM^!1HD>-%7h)GTHzokoOXP#!qy%3
z+$6V|RkEWzhH=)(Zda}`bFKBm`C9C6)Au^4bCRw1Yv_&a?9tRd;>ABJheUu#=vJVM
zo^I1jpPXP{wtWjq%pd6i!(8PJ5}TmU7ii5W+DenGD8!X<b!cAa<}$-li`JvJUucA2
z0=&-FW--^!t0PpObrf^_m`W~`!6%Be>YE?AuzKtM@c?Ph*3}%RiH6h9*urZ3#Nzu*
zJM;O1n$><Q_T~x3uaka*7$yU-YQ-AS8d>h;+W2h;>PFmX+q7%_iXM_V-5Dl2K#`Ip
zItcMEP>^-fD*xiNr%3c=#FZm3i{J8R5l<lVI3rZz|AosL#S?WnId}XLLY<{Z5N-kH
z6wv=WGNrER)^5IeJiVVyl0$8k!nxTuD4$!{!FlH?-u~4Wcve0_rAB3(H|#ip1)q0n
z`_pjV<tOn;S?qB}P4I;Zt)ypJ1F>6pJ4n<MwbPCfa{%Q;DoVk#?Bgr@rm0~EpB7On
zlT){G;6)K?NO=(|?4~<Xg6XHTFCK#hr=}E;QL@Ve78knS912bfo~S%l?_%wl>#eaD
z2Rt=}9c?iaz%eQ!I>W?%^gNut_w&KsIvV))H{s!qKb(&gox4F7g2cF;n$~iTLG%V+
zJ4%^zJq7R>6-8#h0F7(ib1B>Fm2M=IucGfEzy!4Vip!pP(43~tp4#voaGQSoD)&G?
z1AlV$UGsbk6?Zbt-8Y}uGWv~(@)ChGH^Os-eFec6@HJZMB;MCFySmlcF#l2_K@5fM
zk#1i1IQOXo;qHHL-XZ<<lJ6A6DVw4t(kC{Cn)iY`wS8o8(`HV>tA^oW2{(K7DWp()
za^*42`kpYSB6;p!rZ%-rSTP<esA66b6@eIcYS`iVo1QQ%R-Fn?QIx?`Qk{Y}gCalx
z4NzB(yGPNj_%N{dO)VuCo}ieCxD`=uQEqR5D)jK=Bn{bEIi*i^5hIyhs;|qjfL&6w
zM^aQHraA<CrU<o#c~BtYcx0Z5O1968?w)MA6$$>ckc(4<@~5+YrcY;3IeP--2J1AX
zKV@qgG+>EJ@ar^S)l)a2EWeoBj86~ZaRBoRgYKhT+%ClIii!p<L<d)Vliwnt6`bGh
zf4>dnzIjf{ThpFOwo2Ts^TxhWB+5gOm@Wor?=%vp@Tnq9b{5>@$LqS_O}6ofwiUK3
zhnsN3EF5%s9XRY?qIk(S^xJk8ubNc|u=Jn$&aPz?wTLVzwH(^u9KE(3;~V(?RRARP
zD2~Wy5On97Xp$^TN@dumZ-3G&S`y+Nmwl<o&DD=^9~;W3EHv0!>d>c{&G0*Rzqh^i
zim(<qv+FV5m8AI;H>f23OC`;ieB3f%=)2vM8&V;$u@m>O7OB(u(;Wr9d%923M8#z$
zsW47aW2Fp2T>G(BOQ&|xw#(Im1uanAI{yb=x_v_5ZAkX~kkeWT>~0ixdWjcgF*f?E
zXW@dP-nd9)S@NG9QejlBoE7sWzu+zLKLj~DH$f)`6%vg`4ca97#K8~W`loeD70r;~
z<C$RF43Wto`mu_u{W{dli9#2B<*-)Z+Ickzmoms^yLRU{u~S}r@{6c_*aeE6$Vz1$
zexjwe_bn6)=$*jwh80=NArZ*7d?BVd#;u{G+U&;(1g$f^J78YVc6p;YtzsR@9$Fns
z`biX58Vyd>Dz1TuoE0+~c8kX-IJ?%5NjaysU$0zb-k-uq{-h8T-CqqgjIs?+eOoXt
z`{owOa#it$%QSnItUQEic4q!JF}ij<-0H^5?QE5^&1R}RM1a;zc0NPmeq&y`VyBHZ
z4ZN2KY4DH3Z$i;t9!kDtUnx?!s%1N?ty?S#xIg`0IKrDJx^|weU(Rrv6Of-_6bxR3
zCLf6ZTmh|HCUPzM43KlwFq9Pdl~bu4Dv;DbV&6!Y9!dKj7MIBd`GP~yjB)^1{>>^6
zM$$MYT3c8vs==;o!mOm*dECQ@m)5-3C!*jkal3~gz)Jwa#le^EXgto6ywPphpX5+T
zJMLoJn(JW2$>LYk1Qw0PQd$mW$(tc>xT3MKUw{fJsdoH_iBo<JDdtnF#VDc$=SmNt
z40rK5uxzW`8)^97zgnq%r%Fw-jgUce$*@h2gtda#t_<&w9CK8?{YUo&&A>iT0Ta0i
zckK^J;|IC^f`i2$Cv`0hj6&a{{`93*9uysEO-TMA)bO3-S**=VSw2;cs`nXrEv?)9
z@s$8|He(Ne!8>b`C!hu=;b4P0cYSmQ5GL+2^F5S}OOdhOF{;@*LNoN=nqC}0e6xTZ
zOkb0jrT%=Vam-qD2*C1V`YXg*yxF}mI1|tqhf$7yFKhkz+j9D1{_RA()}r;S-spNZ
zQJyP+6;JZbA8*J#GSfr#f^gBpIM(iC%OQcY6B{(!@cWQLWYm}Sv<1O05~6lBfIamd
zg8tAfX<W8&BNyCmq%T40^RJgff#pN|WyzY6A+h05`%rWHo!fQAp816XTfcMsR=HLu
zTf3<RrBj8@Umt$AXCJiL`gg4I8Ybx*X*T4^Q?t?*!e=#mQ_(05K*C}KCCirL9ZCt7
zSe#s4Y`XZn>(nc&DlhireA;>M0lfK<OJ7^$x&mJHP$fBqoiZgiHR>HwRi^Za#or<u
zsj?gAu=5y&<O|!5(A`k4r*Zgym(m9Y?L&YZQ(_sz5-X4T$fZ$#+f($?(!Np?{3R;d
zmV_-eoH$Uc9+&)_>gZWuBs!I8Hd|S^!)zcvON9ORF(rFYw?l`y5~}v38!*;{D<a`e
ziW2ZLL}NwEXx7J#3QjU>V?tHxXg^0ia2IyFn~iL_L_0ait5g#3&~}9Ses817{G6cq
z3u7(VH{5{O1omR4Ffcg78B+{IG7_Xr^{S)yEF703PEZ^weYYl4*~T(B$!_jH3f+;(
zd}}!sGJT5b8@<F9Q&up*4|J6+*cw)r60z|So8sza;-6n==%k6xkyu0O@dS?%86baA
zy?xZcwq&nOai?EjJeKKI-DIi!P4rHbtjgkLNw?nFLoAqTc@Px`S`80EPt+Y@C3VS+
z9uy_35*G>g+pvN&Ux&PIIyG33Hk|EBu055!yWfbzmA3B+&nQck2IM3DOohLyw!q)i
z!;Mot=FAIF#`I{$&}I|&L)LZ;11y(HP`|6{{Ty)%L&!oGjRdkx!*J*>e{DQ<+e;1h
zo#iu*|90j%mDoBrP4;X0{dY-u;Mz#1_*$x6qQm7$GmfsqIWBw_(z(@FOS=|abhj_r
zIuHGdtUK`idh?5V8z6xBJNmN!TOPfgx_eBTWeD%*Ao7F3YH}lEaEtpEwwcFlA~E;s
znNlb%Bc<_`UR@@O8G+U{{!O*ow^2e-cN1%Pj<YwVP~>g!O&!0fjn|SS9lR%u_7A?V
ze5cSo{`-53Dgt?1?7%Yu#%&>^pyfIIUN$J>E5FG?=&f`@Zv;@YSu=dgLZl{37Y_xw
z+~23>*8Jh4e9c#9sE<1Al#Vc`BZ4RACWcCL4oDNloUZzPYiSlMxJ%Guai5#NabIQY
zI)}3w6W0TquqpBGSHWe~CJ3e{+%R!}d`~>1Ncp2c_Y8K>e?(O&j|`&-J=6!&B5_R_
z#Qv7sup63Wl*<6-FB*;T>Ir7;g@>M>Pq*j2(w4GW#2-J;JUwucl)RSkLerddMSlGF
z{tf=q&2zegrOXR1lA~M8sc6CC?BCqqvRsKAUD8MUkET+XqrTLlWQW{@Hf*+@^3<qv
z-_!!J#Iy*gV}%Mi&vSbyi}6Vj>_RMGOqAeczjdP${^dsldJb=et{9=B@WV;Pua7M}
zzavIoDEc?YQYBzPwm_s<INr-r{i@~e(REY%<6DiPN*1ldajA#g!Ez*ttDUqeY<?dS
z0;j9b6OT!5-i(QG?uc_)AUO`3BEp#Ge<c}qFOx$&8Gnj{8*bVE&EKfDH(+eTF3pv7
zB!q7WhRAOaust%O-}1(R2tT^LPq@5drWgioEQ^Ort`Wrx>4$458}E@V4G6rTe)t9(
z(>{X;3W_<=+@9c!$;(@MgO$V6qM32=4lH;5{egTCFSuxm&kdaopD25_9m(nhV|$w#
zzSM#nHTv~6T6BZp@iV)u;EAW(8)JSiEbTJdDhBFDwf3@}l@KoSmw<i9_{JWh?_c6$
z<F1SM0U47jt|dQei8|S<Q-_)bXyEr%o7&gi{W-<*+tzccg8f%-<*T5Vy__GO8#V+U
zZ68-4kKZBn*Ow5&iL@cXj}PDn6!-z%LZl5-y@|=s#)ZhLVCiJ-X2U_o&cVjdmzp2_
z0UL&$laqrj6@wd+7HINv*8!a_v&hKFQj)Q`d3t)vQ*!r0h)<%6d&+ytq!zhRl5LWa
z!OO_YQEtLLymcF&eSQ0khu8V(p~AP?x|;iIS0ji}uPQM`<!d|<_6PHQi(T{&Wn~p&
z<SI(2T9WubNgN=5neG1iC0baV>;VgDYC<Og3lD_=`PY;O!W}?bX(;9lj0zS<kHIBF
z3kuB24NZagsvO7w54&*h2JP|zKM)}1%41>E`kFV<fDioxR$L@{hGtGhTG_g;3FRyf
z%}Hwse2G>^{s$x|5Z6P+6Hky00{#ay-H-Bm$c4O%88ZA{h){@0Q63RrMu-K(Jd`=O
zBb%}A#Sf!Bn#;iKPcsW&kE;EXBY5{WQkFTrHGvx@2m)DdM_mY+R2Z__!s6G%S|0eD
z$?Xi7ypmU@cdSF*>$tXxs^+5RZ(&Td&$&bpSrEMO4M=~LSAZ{tzE-e*d7G@&8Li6(
zj3FVAdDzkl(FHp?Is`MDDv_d5P8_BVY~JyN9Y59~^%4Uo4sc!p0>NjZ0M0Ou*`t9{
z8p@F~9B=`h_n{d=)IbD?-pd1!h;!s09d4rs-Xzc3(CdVS&fm4*S_4JiafV@!Zc?iw
z!@>}IA>qS%MbOCxcXTmeL)zLQCg1vV0MV3N01Fzj$S^zmeNg4qIN?uE{|l=-{n1w|
zsr#*N5#RyX!bjgZ*|{Tadc168&O^tmLz>)ojw1`iSs}R%@0!RJgc5)2*Pvd+-nVFK
z)YZ)Sh}K_$=wE{xf^q=)EdLxQEK87lsI`X&@x{djk}oGy1J{fm#r=zy)p^i;sUOF$
zLxH}$>72872YW&I1>&7opij%>d+}m!yMDVsJ#Z@0t5icKnf|UC5ccyeMmN|50u^Eu
z0pg~v&h#DYtk&oe>g(IL+^Aot=kV^(I*{jvBHbp8KmR~Yt-~C_e(23I6z$%6r~iA2
zMMx{$i|GLA>-6VPRrGnHgJKX75&h+3fYnV;!MiO)SkEx%+FOb_AKh0uka`p`#t;S3
z*9m|Ndly!}kJE9jxxcp)fkPh<W_VKO&VyRn`qP8`o}Ul*B%Umaa&iXI_#TdYMgeyZ
zp=%dtipvAwTs!~0lu1Ol_eBp(P703+&Wt;Cj+oAks0W-j5AUGf8oF}!{a)VVKYJF3
z5<~CpeE5a+A`k6TGbC&yY}`va2jKywi34zRD;@y%Q^tDr=tyPM_+}`-cJ`qlva`SC
z*bvQ3^$txST^;yc*vZF_7;k<su?`k7ouPq{)=p#XMLdbVKa|RiS$UUxLwn)vM=@#n
zAj~`215slQadPs)`io!H<5W14;MO@h9Iw}D_t*w1=w&^Yv=`M|>CPD9p#gPpu@87r
ze$aW>kCR64MSVr%&asw@C9`H*xY_>6Nh|$)n0(E;JyT;K;`B}sb4UQ``*J!t_IP$6
zT)*Qdw8|I!M#&s&JxKg?iT?q@gEjuob}z6UGMsC9yIkb*@@)5JE&uMm6vQunF5+V{
zaj0DOuA&OlpVdcl3EoUP<PQPx_6TghN53;docGYa2v-#s<&kVveG1Bg(7HIZfNu4M
z$Rf(}1OFV&?)sto#a=8|{|J}?Qg9%kZy<E@nEk@_;KNr+xI>d4$8eQWC}$DtiTx<B
z=e6ZT3`Wx&X#JU!9+D8LjIdWQ<7Ks>HC53$?&T3euutd1&{u?O8fmlxJ_9WapcROx
z@}~r|F}c`-llo(5%Zh0@kf$2^o7y-t8{O~pX$N^5O1Puw3t6OQpe^EehFAY%ww$HL
zRx>DW@69Ewfna-;WZ9e~vnbdtpdO0k-Yu9E59KVuz~sqF3tft@BIIo46=pr2zgk59
zy_fG#F@AS^9Foyr+Hi^bX21pgi?BeLiBJgLT6bC7Kup``x2T_AG)%J<4mm!EzjS1+
zJ~Vv3DW@H5k&|N%l{i6JCH5!9j+*EiK5DlhS0a!TCbfv`-$_yK{1AXM@1H^f(9LXJ
zk4FDEj*vB@8BzG66u6-!A}2gy(Wkq8ZzepNGtvyTdEDVQll1JDas!au*V`@I?4(5e
zu=S;!li}p}d1#i1DT6%kc3FiZ)0j~7N^n7eqli<m)MEYUZ|kEcYhxZG<XZuSS9r7-
zr`OS|E{;%W;u9pgR;yC;`7-QJq*sjTl%;3?aJ9`bq(63mUiGVyOKR82wWe^;s=YRN
zL-XUO`#hmv>}2BS`&_^aev6<$*g>*GjY=5f_0G~E_H49s6a#&9g>hc7N#Sq$39Io@
zhCSPaL892O2)wX*^;2g_WEkesKNWIp8+4M0cz@!oW)>Egre{gX5Vi6Niw?O5jO?8@
z8n-S#=nMo=z;(jc!TvxSD2E8gmpVV)7kq0WUrD(N*{!bK3X=q=eP$bPCZAGDGmVnF
z{xa3}e4A@kyTJ`E^+zkNB5x6jg56ID^Z#iAi#Q>%p_$HFaYqOD1*-?eqseuOB$-4`
z)tP6&bZ&oRf?^myG+z$ON@6dx;@Je25kxMh2=tT!H|)mp3Y*_II}G%}EqMlAPC{vm
zA{Cd`sCP-}Fg;)<RaFq#o7O}BQ(yY+E?qvv+RDiaXd|f=E_8?N1JcGw?I`W}Kp5&9
zL_euv;AC=}SuOFX+^LG_jt2#b+_<3(VAd`DESkAIt(9AaHA3m_qBl@)=J2ZqlSDL`
z(Bvi)wp8J60M3@WIC#+mN`8omO8~^M>BOSdh-U?V+ytDQ+VLe~rO|Iis=I!~CIyiq
z%?pRS?H1ys?AM-(xrz+4GG2@KC+%^BV)lqAwqfQ#M_mqawVu2GS#@u3Gj0SGp~NZP
zN8_FfEp<21y~KEZe{Wl+Xk7bF65F?5COz!&;XaI;sj&EF?DXexy-QYi3QjJEov$s;
z$hyJpas;s0(lCT+Me}jy8&XnsacZltc#WL!*9zqD4fqZp2e)?oH7H{hY4le&zk14E
z2Joq($qUw%EMvCz{TSkw9tWI93c~e!7;gsi76CO+Yfj#ig-?^qeHUL{{XK5+?TceM
zH_$WCQl-}%VMSC=KC?6lR3ObU(ogm_Jw#0UjhX_&wnlQg1v_{zZdK=$ogB!XZC&Mj
zgW}yM?<K=$_GyIyzejDWTHZ2#IvkQ==)~($+-VLJzc8dv3GVyOxp^rUy1By}^gkdc
z@T$2K$tsGG^Ddqm8e}+pYcrR4{sIcn3+((2hrv4z|2gSbbi%6|vqV^yznrN<A4y-g
zh94HN?R`gj!YA#!lJ(M?7fk)gWHS2Ko1JF_{i`gWMBow)G}BW@5nuHKmr~GqRM&-#
zadzi^k<@bgvw80Z;iO?R$SeZacEpnTIy<IUbhJQ8;jdr=a+9IH!G_j>m(k`MDu2|*
zeTCYYjKGlX^;-+&cGo1fk)B6f+R!6?l57!Ryc*!@T$_y;&(r$(<L8C^M+#B%=XSSm
ziA&GRv>lbIAY?pctNOW*haHtC@H=IWw!%@Z6lP-6)nbY@xuX$#SqmX2N&ANR7;_U@
zq!k;Vn&<JuT=kWK;tfTDVnv|I8^HkIHK|Z;LNLY4XXZs9m+h0%?g_80vO8FW?Y#oK
z7tieP;k&cCc-Egk(Uc=U`cr`NC*AY!5(R(omzUrB`=oo1EOV9ySZ&iRv4wqRq&9|e
z#q=bY-q7V-ihT-YKKUez@z!}IW~iR>MAA)k4;j0MayDTXNoe$;ra}5cTyt^HY)6G0
zEySZ*h!kY(%SZdI?aT-QvuhOx%)>`O%_Emz!Kxb@NeYSUIvSH|YrJ-To2jL|&2V_!
z%l+IzJx3s{w8wj_^6;H$vpv=M*wqce+uCmcgFyio596gURV#w6QYMX#dj4FBO0Ujx
z`}7yeLmGi49zf`ImU8pkg5JV}zlL0SJ5gOYYznIbVnAkSLRUCQbkV>fz<v>^sl6eL
zL%@piLV6#%QpNW@cbIYs?&1zCstp#vKm7|M`AziDbRy-5+PB4Fj<^}w6guqA{uzU+
z<-!9q{Go{ZKI!9?*3Y=yt@Hq#g29qW-n~c~vU%Cu^XW0A@OimbNXb%5j-D(wOegiF
z*AkG6U{Vm(U#DNoiJd(t;~Zo_c0PGl@Ozo(kCE+}V0dQ_$>?)~g=hJUtzZ^SnF8Ky
zMN0f>#=XU1N3$(UlCawXZ_>`p2@|G=Y%L@P*?G}G-R&d$9B=akbAxTEi(jty>qZ=?
zuZ|K;Cj@2cWgM?O`TOx#sz!1qJ~(S@8B3TQ-q?n@Wz^%}<JuCxqG5^vNpO5x3+a7R
zg*HE%`72spLsY*xH(=lVmAgV<$ND<dc8C4SbY12xd=sx*`S(G~?S6eVpD33v92V<G
z{&q+0=@wx#k1S`G+IFMJXL^oT&Gga7%4_>gk{Yj#Uh>e+?}et>EHkzllr09Mt&|mn
zA|vI83)2iP+O8C<QY%IAz$aQnM!TI)v7I_OkHH^qs077*$ZA4tdoNkBTMj#pIjPMM
zA<T$5ShSO}rMX!2p9DVal|ptuF7xlYwiuh-|E0d^gVL=edNO$x1z`myc(pflqIhnu
z;}CvKecBE)5dLYw_jis3(tyCwRkZv}!CSUWK%OA>$`t~2&c~hcE08{}Jc-r8YM_%`
z&Zyd=N%O#Erk5B<vLSx#`SSQ2X~{1~m(e~da=SexhXQl<p__Me7*4@=cCfD!#U6V}
zKe~b<#Vy`RqAgpSu0hG4RLSs<oD8L?v&t=&LUV)G)n9b>#<QLrq~;O1)9+z2p5w1-
zheQ~rIkOl1$e0C0yTIWTq^s}Ak9r<scL^vr-Ou#xeDl%<<|9c)GNm0PwB3?Zo6ooh
zqL=sM?4m5hq~YJO$y}_5;^N+-a#(pqWL}Z2?e2ro8@ORz9C>g=zFPBsJek*eb)VRx
zY$m7iD%2XY+%Z^TDVm|<Cr=2UFpDZG-7cUWGE3Cr@P<L<X9K>avwe8zEU4OPbfra#
zeZaZ={!$Sc`MN}G)gl0cQz091JN3n;Ri6v$_d&Ta|Kp2|nMOX{mz#y~jZ_5E{7dye
zCx|Jqwpns8XS>DqYUAT-D5@9qEY42zz844Mq^6-Oc72-Nt#Pt8h5@PY4R?tggLoaz
zmJC0PqGA|(Zr6d^Z_8rwA7{AfHohr;-rK&|<H)UWLfVNG#Kf9YCUyUl6tfi(Uaq|9
zP0U`Gpzt9VKA^z~ZT@>b2&ihkJ=vh^e!mjfay)z6L?+ry5NM$d->X@1jv;snYWJc_
zn7mwxg{z%U{JksF;Kog*yNppQ3Pnk(!FKIML~ZXEs-_2^vfUdP`o;iHP(@XRDs)pW
znjhS2u2>>LmbNk%-CWX&HyNzr9xw-fc$d&E(z_045fQ?N6J0j*VV@siUQcrC4q>qB
zXwbZ6lYfsiPy3i>Yj~N57DE~Y)Y^u!FmdrYzqwV%sc2Q-DsxG6DD#UgVdjB03Q8(C
z_u1J&^$)-o^XnNDvA6150cqCSAM1tb6T?**vXg6-gQt!!sI%xLiQi`mwzX#5+30M&
zUwIq_FvbqV!$hgOV=&Bk)=s1XB)URWKk22vp)3aCk)j!<e%IOppRJ?i<*FICq057Y
zd)C1za1FjC3Z7inwvo<~f4SN%%HZbg3vgDMUjqdgFq1k=`Xg`O`kzKHt;q}2_{YY^
z6x5EB{xF;}@Ot12YvYX#m|W5F>An*_R~`9vm9(p*YWVQ3wo&7xv;XVo-&09P1NNz9
z)v<2nK(e+AE}}b4^-MbG!$%(=%uKZ_S3HZ)>1Hr9&~yjVE<?PeRg<Zor*lfohN3&S
zm+vz$YwuZm-qcL#makqLXJaABZ5)>{Xc?>hpe9WeF0J62B~^hEVTEF5R+y0;XUKzd
z^6T1?xh2x%lb%?Ql#NebjZ=A7Y1Ud?CmKz-^Mi6_?Cp_e@d_tWn$GIu&Ay?^Lx=9>
zpr8o+6_q2Gk&U&?UXK&1VaRuTKE69Dj_r3Kuj790yPp5==t7c8+7B&#2~<PD_iPOr
zC#^##6q@`rO61r$w-O5Sa>Fz0dw>&v%}CkEl`W_AZK#Mj!lsak9b<#~B@D$)k)RX-
zTs};HMYT2S3n7O=s0hgtNjz3|sRSbV*?OYSgL31qMFzYZIq0aw8Mqf=wismVwF4f&
z3W2r(#_Gux$)P>+W1J_pt>dZKhF|4O>+GxANy;Cn<OGU!*0Yi4=_dZ>C3!+h5(ZP}
z#0uQtPj@8O*7Q`;&ws-_0)^EjV8V$@F9?SpV*IuDG#wH~z0t<&%ch05&DaY8L_DSh
zs<ioGtb$Tk@JVc#yWtboE5|xr)<S#0sWb1}WJ2bKYlzd+*dB`UZSc=X<T;flE)4yl
zmpzjYJGbrfioG0jmEH9OlrBQli{1xwfr)Xuu*<1mbOuQ*rA6FeuIM#*$XXFJDN-9V
zYdKHwE{5QqDKaq31Vtv$+p^-XJzY3Po=NK2;+3`B4-ywlOe!Fx*9$9NxoWEcy3Y^E
z@Tssdb>=at0BU_h65)JoTG6lmb`QisXY44sOqH)Mp_hUuN_XU(MZ^h=lvCcw9$xyk
zTiRbfH5sE-$-V}&K~JugZ|VqDJ3vEex@!*EMWr-o?&;x4i|A&ER+RY~m)>5_1v|JF
zry5r<d;BDq4ohz~|0+iA4~ZKB(A!;qGpb{a3$l6e&h<x9Ota!m<Wmw_d$$2*LO;Fz
zT2EqjjB>&*V51MkG08EpM_82ByqO(4)L+&yuSFW>^s-r$uUb!`$<*qONpN}+xQiV8
z3en+;6&+PL1i34pW#Gj6_s6bj#Ha^`(My-Sd%xUc2Cd7utCP8GzYt*!(03l~+GK9i
z%@Q>27Ld5(gh-T4q~=e+;&*!vOSHQn7hPBKW(DaSa&x%Mz5TfjVd8O41*0w@xx6YE
z0_DE!3k{t4vi&n#%K4c1@9I^yNVu--M||$5>Vb|3r*;?XlJT9PFD$$`ZBbidhUkuA
zlTofPmS*gcNcT{raRKaJfEShdlh$LA2GO&XdB7IB$M?eXA1UMYebkHN%0?X)h*E@W
zkO`@O!t8V;Nt6S+I;e(U>!Vl9{^++ZVRC|&4pzRS1c`Dr9Ptfas~^tYo2F5_P-S3M
zC#Uf8T1g++N3J8CA3ueLQ|wz@6u#+oXobWPE$lA$@Ir*~ReX710JMv0ntl1wacKVz
z^Kj8-?BXJv%if%k)EcP~_j@<(b-Ne~W}kp4*_e+we|?#`jW?}udE5EZ7ASZzxo+oz
zIy+u&ZGSi5z2{JJ)vR?~X&jU4wVf3AtBB}h?)1ouaQp$I#FnSI_W9&kCq)Atm$6pi
zx(v2_K@ZYby>aT$XrLa2g#OCasmdxgmpDaeF!DU0+;-IDWB^rX5+^2mjQgwn|10dA
znlyoyEZSw;wr$(CZQK01s=933)n(hZZQC|)-<X(~$9XxC5hwq^&fK~7U}PBcdS;!-
zr3&7!d~|)Pxrr;HGbh`g`<&&9yCvwwy?Tzdt-KgU%=ht4Fn#YT*v+Ig{3@&9v3aaS
z(HO^)W(`2orJ+4@{{s~yF#&wcw&d)Rakpk;))PC9DFBs=g<-zJt0$P@t_+)b!~Z24
zQT9Tl(xlj7#quiy1;8o1QGf14=g9{fw|pIj3wbNE?z=+MEI4$0=x^-jBZJ@KZi<7`
zs`vU<{80s+&us#;w{Si>M8lDP_((;yRt`mdap2e^GC5iS?NYF(6fTVnpJ7Y)y0K&N
z7ks+$9ss`j^-s=8jM>{S7vIQX!c%Z^W-jRBv``9cYVaj^C$=c6_C7^f!MN4s6ilI%
z;0)AU7<M+<u8p7`Mk-##H7rh@*#`bVGV-GM45(mm`;#D)$kfWJ=~ogcRk;*DU$MXU
zqZQO8lRUY?EXcYn-<Un)5SC1-$jN2EG@j<U0uVDxae~e4_KSa<>o5Ha&K7r;{8-OW
zp+?c4-Dhyf<n668_-wd3)6)~C5Rpe=`z*aoSdb0~yea{`MCRKC>JffD2C2&bG<FOW
zwsDVLCBkvkMv7)#Z1aT7G&wTvf$9&T(%;>0d19_Fyj}t3w4jlI8fPGaV8_SC9o>*~
zz;qf{AfauYV%uDb(VtqO*L6;BHRc?p!D=X&JlQl#&{geM^Gv1DenQ!f+PiDT%n**<
zZZCgx;>iFFk;#3*t1iRxPRIb%px>49wAC9}L-r9?-V5E{j(>&T76vhW`B)P#IGoql
zdP5=b3<TnbGWhpGl4<_B`f~&j3m^TN11R$|rBjw<c6_zy=xb+at%P)z@^493-QW1A
z><M_I6aoQ8Hl~dC>8E5|>A<V5rFbqUC;%BH_~E4eQJ*0?JHeoA4@M`EXKfV+h%8s~
zV;hvg1oUdo9kh_9+*(%D3{t6gl{mTVuR_^>Um&NoIkC<*dFUx0{Fp#{ugJce04ZbA
z+~P^qb2UY5D$HK#h_t+~+$XQaPcD`9G5Yk7$u-yRUt(4k9ZaRB%WD^1FvVBbFPQZn
z+L$+JwiyHc*~u|=AAuxh55_PA{~Q`C^iQ<H=B@(TJ##kXEyFf&9vw`_rn_V7g#2G<
zOHL~U%!r+(2U6l#K}xrEqAZfe0Ik|!ye4<YjNkO0$8#3DiOd|Z8`k{ctUbY~{(V&0
z3WCJzuR#{vH^n!TO;tZen7?l}S=H;7qMW&ChcQJE>w<<vBL}%tyhz^fP~d6l6W^mk
ztM#l7{#5hM>eM|euXAyVfJY@FfxJ+CoqtduFW^*|oSOoy!E-2yn2kzy0E*9s;L4S{
z9m!92FvWu)OHyN|PS=9mFpWL>=cYX`EO6aJ(oK-`Gf!*M?S*q6%!9z>oaFaxx9vUA
zPpL_lYPSU+r(c-R(rE~9q*ZB}0}}h^lgYD3#3QS5ZivGGRUX4T)`^Yago%W1suRgK
zWcz^_XO`^oh33UC`)8gdKwwLS8BE~QSHI%bFN<a5G(Zugzh6(StIqAbf8wFs4#Gh>
zCvx*f8r-sLV(Q^$tnlK7k_674G8wG#iqzlg1E<@z)^nFYsHX*`;&}-_uT@z=Rh9{!
zEv+Yiso`U_N%eCv!gl>f1xSWEvyo>gku;BGE5Gg%AdMZKfBZaOKzI=!Uez-<$5~T<
z$2@a~76jN_FlDQoRAX#gVFWD2FADnDrZJ6AF$wJDRs@{v;?B}4p}<78s;7b~3VmEH
zOYpBd+Lr7Hr(|F;;AGF{cFcIua6epw8e=5!>dP$J$PDF(ORu*2;b*bp9^HjtXF+sf
zvD*#-8lHi!6EKNa0K2A?b0EiY%uPOppXF>520?Ym)?AzT#zzvIm6D!0vb7a6isrxD
z5Me&e))=0~<0;t)Ijw2?kqt9BHbul)<B&@GnGx_g>J;!N(^_tUheR=w6v57?480l>
zZ3<^Gi&7$LuCvukl)a;#S<<2`ooV5OBuNS#tfoPmxag=70IuGX9CTZ;3YTj_mwJS!
zZYj$o%=sJpx_-IAqT;T~Xde~30Rq@#JOPT?VNudzvLO^1=QEho-e6dR5)9I2OU!$V
z9sK!)Y8j9nxZAWQjn7*|xxgBE-Wt;2W8L32vkNsw(>uir$9%RdyQFfeOuSVC&-mW0
z&I0AzdF2>F0En7lFopL4>puRW1?vxTlBUIWXpfb%;D?)%-I?8fo_C;LNwn({w8!-e
zd)YKTt1K2bR`{cpg##94kexO?Wf$_pvCAGz<5+fbGi$U{zAOr>Vx>EWdd==gBirVP
zP1I*|{>=(>2d3Y;6RY$qIF@YmlgWy=JduE*`~0{&fZhl_!spu(Esa1aOzWj>j!|C~
zP(p{=P9*>t*NxdnB!K1Ui}UrBFyF?Gy(0Z4>sC_Jq65?>K>Ck~lgY$`MTmN!Tg&a5
zYJwCp*-5i1M{_(aH@u%sqKj2cZoJ2waV8ZzorY`+7hA-@*2b)^em8AHeF4TKbwn(u
zxYGm+c<E@#mE<vR*U8w(;r6K~JGcR^U@lDp>BQkm^-A52sK5YehdhZIe#>cA7r@g^
z$k^8v%~SDKgVoTS%jZeXr31pKqRmVM6`nk_Tx_K6Tk!4L#$zoG;?LH|s$s1_yr*Y@
z5kOxOnPiyI0BWyW&sYIfz$d1$3_0Z`@5818ob-txY!nQ~ACvAr^B1dGCq(|LPGK(E
zm>GEk;giT?kpf?L8z_zz()*Xs=@T_@JR}x1JudT)dA)jcm^VA^7itC8QSTW{bJwz@
zLk2WCrxHb7d_|KYCeJJaeRY<iW?p76&~^5u$0Xu62`hbdYLhJ(MBUa_Okc*%c}AQF
zsKTROsh;tOe9?G}5p^*<7_;p&T*yB&%1RP!WV`5)xaClA0L3RIW-y>OJ~lu5u<BB9
z>L-IxX#HcLhOB5thgof@wuGDMTq0}%NK4isjGuDj?&@A#Zp_j9zBP=i-bzaQAfj}R
zoRl6JTfz%yZe?Np9lN8yl~<kwi3Xztkl%%f)nK>QX`^ZK+GHpd2=tN)7p{XO?FFEp
zY9n=jBbB5nwqu(4O5=TTyFJ*JrF}j;7FyR>rQ2RjupRG>#dEq}1^D89cg0YWi#|tC
zfCUiRZW9X1yJrjip|Iw^a1I#8p}~_vUfQY1<1BOFKG9I@GmJ7+l3p8m93nLaC|e9j
zpXi|QClI|=g{N-RD1r_AH9&wmWE?&l<${ZqDyd9|47>c3aJ}oo%IQME>lJs|<gPG?
zRVz+GmOc}VGXY+e%;vYESi^9DCm<701D(iJvPu>&Zq<b9b((-CxR{EH)PMT@t_Jm8
z{J`EkT|i?DjPv}|RhaFrx?FSym`FuXFr9i>b3{XZl_3EId*Xg#@bB$Z0%Ed3=_eWE
zGMp50Z4Fmh-eby*p;TJ^(3!rCO*@43cXSW{J+$53WHdn?D8$o;F}@Ylsy(sZ+X5k1
zp1a{(k{k(#?WY<QekNEm8z-s?G#{yH0Vi0)M|sEGd(O?_AE4b6eNagN++4f6BHiDI
zx05wEV3xR)o9#bP!X}weu<l$j_vq9_)f_>{vN-QLcTZO2Oz~w>l*Z%PEB8c}!2`M6
z_i$gq#eb&xb`?rpE%chB<|pL*44Q|oao_O+wv4PSZI3({#eBVu_o1r(;HKqB63$tu
z4;PlOE-wWKk1|P9g-vz?`uRV;^Hg=dRTKQnw3HG{1pX|i!dJO<WGcMY(^9;oGyzF~
zwh+TWip4u$zz8k$W$$DOUdbW-C_;AHvkEep{0hZ5DkykxfBHzXMS&9sjwNoDTXQBq
zihJvyY>MoteCSYuznr~JVjYM|X%(QILHO#?zw#lM=b!i0_c7=MP&WsczqMj8Ax?bU
zQ*pTZOdC9Ptz5}OvTtzYQc**m@hV-y{HfE4Vi_GID>#s|sPr>PD=mTge&1i6ueHji
zS}`GB2tWRg*Q?vEEc_wTRsPA|=(IYz9=9{k_bpAdw1&;R@ow0sZeZ-3409t2Zndm9
zQWT&yL(g%021URD;E?zw_&4-Rb2i{>kZq<i=O4m*EC%g<O0#urxuMXypvd!!<VH-Y
zYo)DMbe$mZm^z~NW)t~4U3ra5IP;An*Bj2G$$qRY|I-gQVL;6SA}xDUm^eF}ulC9^
z0*l0tVE3^Ha7<`6>xD-yK}Cmb_3!B!R?9kc8~u;Pu=uus5%|6v-QX;=kqF`9RTBGc
zs*k!;Eo{L(`^ZMsOBmrRAcO%cy#&9CnT3w<X*CQmOUftR?ii!KoF}S9vko2GcN{M-
z;FiY$J=}|UWxwvkmO~teog$vHCoA=|)2(6K2K72M8|+P_mMd7UQL5(gStXJ}C4c9U
z`ATPO(iRthf}OQ$37}H1HF0{2Sow)u(dPFb+dnCLF8iyDj?59eHjYkN#j0;YYCt4a
zD>88_1{G3BVWnESp?%0Q!1u;ypp#4DqPxSuaG*{Cb$DgD5ZXcbsn-E%$(T-78!vmT
z0&d4*@bD?Fo4$lWBzhYZheJ(Vs3v<r41kxt+Y4Vnq=>@UcX37z2bIu$bg%lTO?zxF
znef-}Kg6^|SxY!yLcMx~>g_Ut!Ta#m>Wx2cd)p6KtVD51I(MSFvuTBW<f&LR6^l{K
z*FsZ`FC$d!9V#a5=+KdsH2m4#57(264wy;+O?aKPpM?#%l-WNoT<OWiO(-7ZUacm%
zJf#SrOmi)IV)$*@s(U?^a0e-1E|VCIgA~8D*RV-jL^B(zO$xTjn#Ri3hwZEBv03;$
zZD5;5Apc=IYwBvWe+4O6e!!Oqp9Blia$>otR+h<vhFz%7{}4n%Pe|{M=C69FDGDf<
zvuSvW*_B|!)$rT~ZtmQ%JvHnCoK5@`ZZRx?R^w0FUlO$z;w*~Dc>3RJ>2x%KYm#4Y
z$C)^H@7=XL0Tn)GI7=BNnlZ@-mq$x+*<-V*>qoHx3&39tn$2o-xU(Ae_r%HeA~c=`
zsN%k<ge;BchL0S8M9pgddd+=zCO=P8Dxb_%6Fk1~a(Va2B^5NZ^8$fcz2R3JnPUmS
zSZii1%izb0k5pM)-S!pFTyZ(U_Yu1b0oms)3@y3%^diIlFf~~L&VpOyHcusVUACFS
z{<r*MtdNqBj$ZM>V59TQkDZ;oPH);;LzVtmHJ2O~7E$6;m+$Xc5w?7OlW#g)gfeE1
z>KzE{Idw1%HBTg4lpAS?G*rHvn(1f2=?b!lk;T-v%Ss}Dv*V3wEAF9(1);n!qgDUt
zhLYlK74%|#(B3cXtW4&T;n%n?_b#P}QV-yoXQe)7Je_*^?@QJ*w&R}zy|Np<4epvZ
zcUCOB>(SPOu~n)};Udmt%X9~ov$O0<ZF><HMaw8vpE320vaU_yHahNt8BBJ-YZXBA
zw{^h9XRq5SK3lw58iDe+0zx&xhNOb!Wr4=dYYF$GOvZAt>_v}duevw@7u0uA$@2^W
z+gkpdO3`T_!Cs7Ywk0Cu{l`<UKO!2VJtbY)R)Df0?sCfj+T&qhx`X>Ny2;ZXo`~OG
zFUmSTLhD}~v()DV>=N`jQP(O!)>P_GAwi)G@`r6IUmcX!8=H%liX)eS9@M>mWOsE7
zwAXt5l2)u(EYe2=;TCD3-utZ!O(462CU`^S$%wFr^C*S66L~<2lT{_l*a$Q^Q+?v~
z-k<_jN=>KvY1=kcaqhW(srB@l!EA~@B=cbTi?V?&q)C-boq2o)0(=yJ&-NaM<ZK-B
zJ7l_ZT~JKndlb4npjN6f!4YsJY8&0{SCXRv_sp{$*H+KWUWbSq!UpNKLw=}?@%eo&
zw)*UyywPP`V3A^#mQ@H7*cA(g<qPI=0ndNn)R;kH-xx~yvVaWhbph!d=F-@<VH7xa
z3OQ#h;oz-a$zPlGQ+fi3Y{ZSJV049U4kJ>3de$(wa~GPwnb-SCm#dge#$j#|WuCv|
zAv2n`36%ANqZvyR%9XM7{0>YgKXPffMYhQG89VEe8GtxE5{pE-d}U8rn$_SU6@PDM
zF?E{4wTTEf2w^5ap#F19HFS+Tsi{BQZpV`D%2Zl0<Qqb4?vw~P8Q|?)FBn>njFcYd
zDfj4Bej4gc5%CBXozLq4W-uQ{#|&k9beCy6c*eM0(4p^*6WlWZo!iI?SN$H?!EwE#
zxt<B9CkT^9{HWZ|%g<cqr6kg&1#hq-et{+|k;5KITJo3vJDz}^)>vHqK1u(e=lpF}
zGls+3bGwENBYy?}<*uQS0y(GAj`Fa`2Ck%vB~YhsG#j#2>-?(ny@e4*+%jul@()WX
zePclp*+IKtp*K8%#HQw?oEqd@gt56uXeu>o9s~35ZB-d>cY`~l^bVIT5pKye(ctc`
zSPheqpWL&~d$k6_k5G#6#T?*<<T%IwL-H5`-nm)SMT-mYM+tR;3x9<467f;bfyHyU
zEaxch%E^sv9(Ix$M%h=MC6O>>a7!BSI>C%0CNlY4ufR9aqY*j_IYvewHB-xm<Z!Qi
z>-F3r*SAByv$x3S8GxGO4hBz2!XK?3E-0v?7h$ehY&BL1zE|FygMjp`^()b0Lz?d?
zwzZglS2+Mk!;~~|7?gfTZ^_>N6Q>Hws}qGSdXJ?3DhNIpEF@e;!~FeL(?mKe<I~u{
zM<7(T3AG>gdt)3a*pZ1erdlfK4oFyfq?ptxC%S>kj$2^^p5aB@%axG=|6=u=-y3_W
z*=Gp7MlUu_ney>V?ACFMr!9ph#*{RNr-zd;vi^m10Bt>>_Z)?J<15BNo^jZ@(B7^;
z2Zy1x={bOghs_rGz(9$#rVr*gCp0DzIv#77eq3`~A=?_i9agv0^XLIJ^9}DP{xO(-
zvoIhsaA}?Zb$SGnYTNR~LYlYf)K4jEwll$$%86qVxiM$Yy3mo`ZLiC~21`5*X*7S7
z9a9ECEnsg%Bx1@f2;^5b!&5H0qVKm(0LHMNTHEM5)8d(4kze}+Mq~sfbCQfy`!7jV
zwwfibhTzHa4e7)tf00^GpiCd1MG?>Sx=sjjr%Nu6zA6KMeLCi0NPg3ucK}1sPu0JO
zRaPradkJ=y!9@enC?jaGg!(=j85@VqDiHw`RyWAmrv{Ilyt#+qn~u#(W=Sm>(!Pbh
z_it%^_M1MLnoxj(HEI+M>wikdsN1E2yO*rT%mrXyGD}fXI9k}VHV|Na7TLZQk+?zg
zG-*b-QPz|4Tc&|4AymaMfpe*PQ=wWXW1%-<ks-4a7ZOOk#!BD2$;|rgPYar#*YW`l
zJ-$dN$bysUoIA-ERYRduJsCXPz#(!dSy*%IXhd#}5+n5scY`GAk<5X@H+B%|C@L)i
z4hlxu6jac!Xhk55D+@u`FiHistu_2(xBi8e7U8<n_<s9hJE(%qv1p!3T4p{R7{V7!
z$2PDmL!0J8kQ`!p+wV*sZb>F&-mCx~qLj-5S&jlCQ3@>X<XOR>SwC$QpdN+tz+5_f
zC+Z1+w|<&)NEh^5HzqpeH=Uxl;_Haj8odMtC8;ZDnuZ7gLSKs69<eL|Q59Tcwr61w
zrIs5D7VWJ#71^!2`X8@Dq5^&440xe2%RRzrHSD*o{%lV{%m^j7wn0xRMNhyaXS0qS
z<(I<Ei_s_~>ICj<gR!o>RDjeoTXnoy#Ln+OzUPGFepREnb;q5m-Qs~|sAPU$w1v)^
z-Jd5GEpp!&AO@v-JM+VO>*K<nF?*{MnC>R#bNKV5x(DVCE|b<!ip+F65IZBa-eLZ|
zBXQz`QTp$&Dl4Mdd6sZk*EIki7dg1a91UFL1M^((Q$I|c<9q%e<<DtgoKP0d48vQo
z^AGtV#b;VhDq}zOwZlnEB84R6^>s7cwMq5G+Ik0ze}kt2-RbP;52$4xMI6&0WN~hm
zvh}f7?9iK~6|=vu?`=nRNmvh~#w;Q^MnzU;3VSPQPDWDoY(yIioa+F@jKnub*V^qI
zA^K3)(Nd&n$EskhZ+n;z6Ez&uzdug+BUN;MYA2p~c60dnF(%BeDMR{OrG~E$?IMKH
zo35XpSu1W~b81odHGMmoGLD~(x>Iin_4z5DbnfGdky81m!pHnPF}y&)TDDp?Um>(+
zD+V#HWR~5dEcOwFA7lXdO$0B!T+sB@t#oHmmv^pH+OxBy0VhlIMPj54F*s7|RhRe3
z%>}s|6$x9LqbqBtmFcy7DjNC+@^=l}UbQ%?(@LNEd+}f)dv)S*zHm=PF<w>{9VQbb
zNE24;QOw_pR4!v9_y-)bEqGYuWM=<Z{0`{4f-x>S*QwyYO>qEz<Jl8c&-y$Qnk)(w
zb`v&sF3>GeDxM;w&*Lt*9-OJ;--G0<d5!164ou7Qif0t0lFS!Z<Es-THKi$rGtGo)
zujotBTX8ok9YSDI&i>b)mGI0O6#~zcfD_9Od=DbdP2?A1wY-aNlp?~n1aIV%<bk6`
zR{5sv^xj(T?G}J2N5!^HL}Ku@fzCA-smnN|Jr!chcwE1+QyUy>vt1<qyGV{uBf$=t
zJuJNhk05+ft`F=>0W(Wn>^+<<7}>QS9FRT47fh`4%~q_QKDrg58AEn`)Lob;+%FMf
zB+=`X;)Rd|6iAjHvn{PR@~sBQw1t-DR7(65tg1xYydnUlx^BtS+tkY1f>D<_=m5cM
z*=4pYjuLwb#^`(PV;ajsT7yCP_I@XUsBDJX$+w~ynh1xZQJ7@Lg_#sMoM7}KSm~wI
zMHT4G?yrQDG3R=6j(mCHzu@f`(IRGx@|+&=^L~xWJxH_NOxvC&KH0UG#L;HH9*Dth
zvO^fL+O`0~W_gnL0b_XSk%LPUpky!wg=OtRr4=)-9!tPwNVE&I(B&z|9Phm;sh?Q#
zjDZdVBDbHPhslToK37;FmVozO1eQu|duK==ETSHx2#*(Jvsn{0tabz2YK&qBY{UI3
zLgyf)atGV5N7wS&3TY_N)}8+JNEYcydjImp<r@H*-6F<d(I<ZV7tK|0otNG3B6SyD
z_g{Zi(L9^idkFDkFEubCP9I(p@mi+hZO@R}s}{hiV%H6glOo{j&%8uCDsgCP=3*P0
zP*l9{3`~|-ibE>SnzZJ@_0wc5eQ6Q8bb?eg;51pu)gDV3FfLfOt8Ae05ar|y&m#;j
zhj0P={*fgjS>N{}S7cO$-;X$+Y=~wTd`%q9=hMLUKN#b&Q)N<b>Nx#Ua=GWY{*bl+
zZSK{2jV&)x{=(K;s3(SA>72fUv>be@`^Jf?qS`2CR_-@jU6O#*so@?Kb&;>9IGC3E
z45%=+puRx-wAJ68G03^Lp=)lvk7}rAqauJPZA6y>Z%s35P90E)&MI{4<)JQ&6qQF-
z<U!odhin+5qgnWxmle*?E!LJkHj-3{;g=+>6ocpVgX)yuF0yWkAN^cl-#+ocSH}E1
zU|f;ne`gm*&<+GgZ0CVku!4!FUo62^1JJ<m+0eKcG9-ACezy%}FpMQJedbmU6;lEx
zl<H-9;BAEyYFg&CvG9u-rvK%x4=3{=l7!dWd0cWJ|NT|L=8r@AcbJ#-N}rx4H~SaD
zHiFaGKT@RG-WJf<+-=th;GX$fZjW)SG@I;B$L6IgtfwV&!+baQ7x<5$IMrob@BOgU
zc%0Vtv(X*0P;R2<3FQlRP|$v>@E8HWVtZ9Gi|in!*cyCgfB5tnC+8x+f<<0ZR4r@L
z11~G?F#QiV^TTT!8eXj}XQ8QO++q3PwC_%V$~vEg<byO0ot)MT&U2&hMV6@=OP6|l
z`@z}kDf^uR*iCqaZ&G{>d%8xcHE~ZEyy+f`i><SCDN`u;2c3{9fdCm8OW_w_y`G!r
z=l)dKRspr@l@jd@Os}7!fd8qANc*{B6p}U5a8|Uam0xf+v#;Rydy0`Mi*8nhdeRd}
z{34xVUh7bB?6!+fEY>Qtt?|@D5rx*xg>x%HJ4P>Re@A-KTYk>X8qEW0pjM%p<vTxc
z-V&gFW|E+d8nt_1t!=Ef+4~3pF#?q(w%dxk)G@XVYr!8?t(O%|#7v*s5>@kaO1-W!
zHlafb;K0g0V<%gXuh^mI9n1N%%9VJ&99^nXEB#sl=sTMrRXoz<qrY+qi`;r8JME%y
zt&QuM(ggOT@%-Hq5pbQpd&e0zKf-4B0}Z&H$jY>Rhfau!5Tk*^cWDFwz@gP#ZX0>2
z<vZ<%+wqKKV#T`gsp@t|aYZQpwpixRntLM%6W@$sSmP+7k=$J#Kt>2I=)Q7S+v|)b
z)pk$Qyywm}OuLr(6D_BazjO0kSUu?V&o9+K(Acc(X6tIxXi*(P@sQMI<|n?mV^Ax{
zxI}B#As0bhH7(_j-*elC(iHT#65#5d9AnFOqHi*TjSL5$RD_m4*VbvQh7_5&S2Mzi
zrTgc=SCxj$q#x*NF|O6#-o5E>L#pn}Tcx8EXX|KikoOR$v1pH_VgE|HGwr#FeH2~d
z|8e?|KR`hc7nXp)My~z%`PKda5}v9w1%lFIV*x%O0>qWd<Oxibs#yV!3>f}Zk2eyN
zI-1Zjcnd<-L`sfIm8fbwQ`57<rA8WABpF{1fuakF7`e36l=Xgoxs~Kz7at8@uf)%~
zo;i%U@Vc62h;1fpqSo7DU8!-pk4>Rh=xUo4EjJr2vv`T)u$jYLcSn@((@(bmx&gW=
z99zmLupBmR4Ch-C#wL=;3#i5ur?kPwBZE_GEo3T@^2x9a)6|uyOhASl%+mIXM)!q9
zpX*~=7e__H;Ge(!12b98N)7c0!Ad@h03DSiS}HQMZ+}}#s*OW8J}20?I?BSTj$J<Z
zFB=iJirpCWSd=UfIm#8=^UPE!++Bf!5S4TbI(md$kw~#tp%NB23Lq|q#4D(MPA*!V
zMy`z&UP=zGX#3l_fQAGpPsw{mvKI}dX;56sHk}f}Sw3I<wS1RR5r~QoC_w3#f{w0<
zNhw(HU$jDC#OMN-dGwAYOrxw9A5e-ITu_{7Bz5lgNTifeCP5XWjnX~Q+Dhk1nsKaM
zn_p$lF~fmYNs+!7B!EuW)jTqKCh}`6T-MiM#}>>zl-54YC`jm1?5uF$nY?30;z3ag
zyFO6p-t#{la|gVx{r)DBzSX<1Qc=IV(*`>DY8$18$A*<77uxI6h`oMGRFs@L#bHvD
zt)GPzfnyU2A9T8bkgu$^+gK7|)k4Y8Nu~2SHmD<K<%fx&)&ZE^OMv16ej{2KZ{+D$
z2u2B_0cUc`fs_mz5ohs=3mrEOlmQvTRY&_E;9rJw7}RWt<HCXujOBQ8;$}A$#v<=i
z351vIr+fZ^$-B2wz-=l?JSJ(aFVw_UlMu$wO3oJb#9W9c&)d118UuBS!076`+28F0
z(HFwalJ_16-~^yM#OBPM5TquKJ-q^gUms=!czU8K29zu~+Yt|>UW#_0k)=?s1b&`s
z_&&FSCLzmzWhHcrU<30m8htC*fy{lxD!sM7l{%r;b6F3D8u*iu{Q^1@rHt*9>-n+L
z6Tk_hl`Op?wM?{o_nx%LXK6XRFg_=}?Nj&5zW&|Nd<L*qITtK#K96ufZnk?D#&Pvs
zwYN}2GaPADGdR72x9wBcBkO*0*HM{vr)}I-_xpQ3*xJ}cG0<1!b(@EzF#5R0dkXa^
zAEhh7;lzoZgWqC!DUQqUKOe?%l`8dMrNjx}fpL}N`Dbmt&ot`l-Z3HU?$+FOJ!b9E
zabe=Qas}YW`8i(;>t%Ts$6B;s%hRu^EZ;^jv4L5U&_}90s(<e>d$f_{u2yZ=mb2|~
z#QpN&s^As!bR$`bML27CShfrmXF>6Gea~w>V&~!Ny@sXmK9ITn+R0}!*4@o@_9EZG
zY8@pHcw83%E3+m_R#eZ5U2piBvG~~fn&H@*D+1s>^<3w=<-@ffxc~TauJ@eh#DrX(
zIWl&hXJ0F5yqWVICtiC>`9h(C7YSKIHopCIIHaY)x54NS!+WWKmxp{3)tFI$cYO?Z
zlx#JiH@byztYk0jC39lI129ZyhR)1huyxjNa-t{-ug^enLc~Htk!}=vp7Kzb_X=(#
zxd0rA^kBHUo6+A*&lG(Z%6pk@6JV>(_hEcVR)hsc6UkqZh*FBuuPX5P;$TH?ct_Vn
zZ(DPz+XS+N(>j)aT_v>b7Fr4+^@{Mksu`PaHM8UKT5SMe&~Wg}170oLeX3JSC2+Sy
zi(j4jQXY=({J4)kK`BRPwyP*7j5fkjnt(&fd@kNB)l}|T63dA7wVU7dZuLQozu+q`
zK`n(Z*ft4&4Aw!?y+HkR-hAA8$`CM!9M^4EFv@o_uHK1jjall{49tENz2*FK(m-f@
z+{|zQ_^<pPIq|p^<jN5?k)43cQLsUe@c#TryP*bO;Em%*T5Bwh5M9o&4K`&nUI2h}
zw8wetyKTZ8WcYTzQtBti%phH)oxAk;G=s81tpwE+t3gWOv7x^OwE2vYqF_U`+j7IJ
z6YrpMp7-{CZ1R+*1{;p=%)qiOVNEpENBP=h!;i~!n2*7rs7ly42($`89|m*Z)P#JN
z)lGz|*5~lYrH&G-;tt*#GqO{eHUR^&5<UnRh;026z;LDAvD1|M2|W`xTEH3=qer7=
z-6%~riUm`1ZN#?=f4{G6j=h}IHOgcp0!+#v9)U)L*(F1OpT_QskVpwt8CgNH$TYk&
zB(Dw`11=g#;9p!hz~mkrbD3FY6whZ_4r`=mQ#7b9oWs*}=vX~H`OJZ$lmRow1Agdv
z`{8zm$JOWJI>=S!3P3z9KFZEuMUpyWe&^q9v1%!Qq{!%}Y?FOC94J@XOx0iz&t~;5
z6^rntU_eK+wL2k`7?S?t9mh8q^ad%#EB5iWB;bRsL)mAm5%g6h5tYfqM(*<#!os&C
z{}CDn0vjnL2?%(j0$hdP<pIED6D#)yh?nwRB^N_KMXVc+y7vi17?Q?4_v7x9H-rgM
z_TjFg<Pjijkt^hViOWpKCK1EnV9N|KNTNablOoAklb227Nv*8@5$6>4G&vx+Q3*)t
zpkhekl1<m?Jf<mfF8JrIX~;q_hw=LesqSjOC}Ky=*i+rkFw5rmQ2;PXrhR*QL2pun
zX&!7>d{W5Ipe}}Qr+UO}S%fLQj8y!B-v#vXO)g=JOe=dx1sx|1v5O?gJSV})5t~_M
zyUamcJ1%8Rd~G9*WXVzuca(0@5lZZt&sfD1=)A2<6N=R+E<RWs7cwt&nR*wkje|@N
zK3HVsdqFG_3Gv1*6M$c)s4VR?ZpPu9*;Qd6O7RmJ(;L%lH$7>x$A{iv=RivPAWB$D
z_PNv4XG&1&Gd4;$x)nZ0TgM3+l@_n?o06)Fy00sHVz_3{v#!r;{fn8p)pjbvIMVp_
z8?|)gNBhsqXWkDf<qx0h7EXKj%hefgVczC|go!k;WEp6+27ur(^UH3P^)suB7E-zf
z%q>R%e-iiYAEV#N@RR42m{j>#uYz?ke@)zU-}gmCAW5t=|K?{)UC6LMxVG}wxhu+2
zZXnMnOQHPc#q=i*c$ALtqFVmUX%(n-t7`F<(GoJJyF0h$65C-EZ8KFTreK@6iupTd
zURG36In5}mI}dPcw9N&gBkGNbNgd(?e#q?o24)d>Y3wC(%z;Tzh>iMdQ88f=*D>4Z
z#mSI|f+NG0&Hu8eOhs`r&|2hsm84(btj-{;&N6>o&#3PRllx9vFeZ8!GzPEsOz)gH
zUhaI5d<Yb-MW+yNp{GI3Z6Be4(W-kwnM|@Q&3qo9RRc(yOiON^UIgV6n(`G$ElR(Z
zvPDl;t(>P8OO{d8PucsGX5&_qL8JE%Qe8uQ!`{E__Xv$<Bn`6`r`UIJs^y!)Zf|C$
z>zj>xT2EZ;KcrCc4E1J*Ji$JbxZ!D=&ZiS-Oy=pIx<{+a*EK220%6;I<c-qFms6P~
zu2v%-s{_E8R-HMal-NV}Il&LPN;Dc!gE#e0&ALP6Od(o23OvpdDxmXOOfdgjqy<PI
z$;XnM!2%ud`70q3t_J#1`1pwuqMuT?ObRYp_c9c6yUbuhSvvMtr1g$(+<Q7L?0!>!
zh3|-9@3q9Zp>H|jQGq!HjNe&3s5X+R#UhHY!x`WRr)!+)u!=USN1BN$p8@(?7KWVt
zV^ZW6Z*)}i*x>%YOnzpDOHy(e1F(H(H({mGph!<_@j&_pTVx&@EsLhb-l#dR6oAE!
zN+K#0Ku0`aPnLg(7HXVSV2W486(@-s9+x|z6dqq`b|y);FBpNvoR%px{bV;@U-FLH
zLICJoT0FA(%J#m}(RykY65DWWGm4vAqx4EpQt$##lgN@vYdaSLheueo*gNxG%%W1M
z48bK3cRx?hM)gVy$X@Neb{k!zmphTAr;d_iGSQSf`NW4Pk6>LkHd<uet>bCiRB7?y
zcE=~2w9y}l$qd%da!?w)7^?-Dc0P^dItJ{LBiASd(d|!zmdy_FX^bb_O0j!zJ&opc
zS@e)HnreCNRa2Z~f*f9EIB%052}TCTB7^%AVb;lD9wE|Q^h61y!$RL^UURy-@H>YY
zt)=I&ka3Z0&76=)HAY2qWeTAR2M5lr5||R8^h=|ha!yQW4oOGgJDA@L);<f;jsv(-
z%4#Iqk}h$ijS+1|DVoMy-V<Vn()-m_TKZ(RnR{5#IQt7ot#>Js+UNv;igaTq)Z-W6
zaNjJaSbVxGINgSIN=8O<Twt?mb(~+?G@1Nf9)H=xq)7_=HTEZo1Pj&S_9?C;=lU*0
zoMKD~fkSOmsNCmjNU`(VH<eDx+5}Aa)IVN>!(<=$bFH9pwV<pyGS`PI5HojlBp47?
z+T>EEx~S5A;JDs%qGUhO%#12Q@X3n$kV@0N9s0(Qf5g$QF2-)_MqM^Z98*m8{kzcH
z;|vs?=0+x(5YO}E512{t_J99yng@I{p>PE*C|tGA5L;>SHiH>B6qX2JzyVMuH*r2|
zIFBo={`9n(@b3s>ogNzWG3ADMa^tvwJEK?9EtGu1qP8TVmH|yGn)wuNU9#YCfHaxz
z&~RQj8+}{b;!!Can3*^G#$k1F&gdR~RT_76&QQ5jD-OtWo(MuGpyjw}HTd#%3%ezw
z>#;aQ&(2dojAKQ!r9TDsS_9G*O8Wim$g2|qH^Oc|`|YL|s*&bp_s;DvoZ0&Ij~jI?
z^sgIqmGAo>9$*vskmT5*Iy7)wZx5PLU&~^Tmb3O3{^GAT_f;96aLx!5Sjq@=INH9S
z>WxiusvJqboKoW`WDF7qE~v>-u(vdPEb_E)9bo}SKuxS_)%ecbJ^(OL)^J_$UMMN+
z!xMv1B_y8zt$pkRP*lBel3<EAF>u_(oI<H*JbXuIzIZ(j;N!EUh3AsJ4=JOLQbGQl
z(!CT0qka7$)^L$>=!pxhwm{%|{3HsMd+<pqRLCBR?Rh})wt2n7vuJo~^stf(A;?pd
zG1f>JnR-9|cr-_-O9$*_32L^cB8|JzQwck#6fX4=HJmh3^Ef8t4q3<apezi>PT9t!
z$s7ji#jD=>S<9HYmPu$89G2zg_#89W(P*|P62iI|sf=-_YEO%~#Qsy5EPm-L&KV8W
z7rU6DUb(#GjSbq<bg0p#XRF@w*CI3hXYnfT7gwDY>X%&Cy8v&e^J9nKa~<Nb8+T`=
z9(-6a_Z<>7?Ha8Q&qo(40^P5<7!0u?8e0Jg9|)meEUu>>48ze=d<6G8lOOEw=DGfm
zg{S$b{W{(p&xXFz4lEx@&x3{6PYcZ+PFnO<LhUUz$mHJaXU{$*5J8dt8>4(qyjC^a
zN{lTO{7FKJVt}-RdY%G~oNprdL6SSw&QtyDEVm&UCrWq*pM6Q($b;=6HH}UbJtt!M
zp>)fkwpPT{{rih2Gaf{~yN=UH8YV;$CpvLsTJn8_yOLMp?g&6j;hw-lan7S4WLm00
zzrLrQJ)fok|NGwBT_~2OH;QIJZ$z9q24<yV0NNF>Ccw1~PLtUV^rFraJSoOrsF(wU
zbij$Eo4&x%LwK*G#lfFA#cM5F_xsjt9)7rssBT)x$7$gv#vVdc#~2?Wb&^{qi+^Nb
z92;IMAi&Zb4zqi>zx3aNPU)rbs^ugxajesSi6_%QwtDFG_}|=di2owG=3|SD_mkd}
z*Z*nn6tG(oC_(;A&lm-cd6}c#e9=TmcNelYi1R@z0EXCx9(RJ6%MnZ^AG*q+c~E(!
z&>k#AFK*{CACso!Z*q2KQP$B8w4fV#;l7BI6K*rQW3f+|(i5Jh+niV(X#0h;6Ztx%
zckA<WyLQlWgK_JEvzG|Q`>ys6RLrp6wA)1+2S8#aVbWSHN$I$JINGaG*x}P(wOE}b
zFSk{_<Ehy(DHL*8w_H2-+WUh=09#Kyb<!QyEC2rRx;2m+3(!nSdXiRc1OBc44gc#i
z<rgIuJEf9xe$9|h>Fl%j<^=)-a3vgX*b|1@pKU3#@5jfp83+Llb@zih9T8l4f#5@e
zGXQm$ArCs7-RAVHy<#fFJ`qkV!63>VNe}ur3hOY;?EuZ43sXnU+HJRy>2@T-KE@r*
z1<HJBPYcR?*}{9+GNC%R-Hi8_zed)xcnOEb$=5T?Y9wKHd7e5WUqKi5MGS)&&!|DJ
zL%ZLcKNr{y(3{Kq1=RNk$ay)#bSOum7{KReV{#wCI|yCfJ`CA9(%*8Z$cYXGTu)qc
z%2oTv9QN<?TQS}P&t<MdNaFKl(ab<Gh8L;95w+Hw(+31P!Ap0TQ|>9bAE&hr)0+$A
zhqueJjl@+eM@|QR#dPTQ)FkiM+0dNPk+1DBksLYs&;7bT1-`S5*Qvzq1QdMZ2w--T
z_S%`l<o5N70U>wF-77S(F@coqUcrRqYJRDMHPf|7=frHtJGFlMNJMggY2sIFyK{Be
z$1;zGIR%N+?cVJ`W_t0eEyPQ9Xrhz=vqpfwc1`2Yv55F>+jEjurKDl+L8Q*+7&Nc<
z_x{~R;TifDo{lL_oMY>DNHR+I3IIJEY5B&{$%M!1O#-3@VdsZ=YgMlj2hI+5BqS`I
z-Mq)k#HcHwCdpB~)-{D@q&1P>`|0!b^wzU|!*j$wo1vrWj0ev|bEZ^*T~zSm3V1E3
z46<QbAJl?^?fG<QLjr+79bCY{2#>RF)kmcK5~sHe7w^QQ{&j7epgtP_0r=QP+oCaK
zzJ{Nc-V`k60g+~J^Y9tfgT19Sf8`ZPy4)o$aHkl0{oY*}Xqkz;@%(%`H(e#z!Ez}!
z&lQmu`!*>x3}*2XfLI@gtNn(gGzd?;mdx7h4gPJj;7~KV(RvP-$6jOqp1AGyQ-Rrt
z@n+F`$ra_J!DnlPl$p(X4w(BGFxdXC4V%)-2V1l*_!-K_`)WY`;>**j@xTABk?5)r
z`?B^}sJV&F5%7}Pz=aR-;GSFE<QKaS_^t?7O}2849HVZDtDM)~841--qPWkM%CPQv
z8%R)f8#$DZ0bp#6;pIFAuK2S5QQ;##Nq>6vx{K7vGKN3cxbyVs0dP)SES^ot({FB&
zjkr64zbk|my89_lU7Jee;QhRReswsPZ=@kXcLW`Iq*!NCEdg)LAEAe0vzLuo(MRLz
zyuMGo+fYQ_PE26b=y&C?+=vIo{bR%XuR8&@s~$uy+Biau&<Nxngwj*fyudj$C^%oY
zhX`Gi(D<iF@-x)b0j>R81=j$~EchAkAehU15C51jE5RS(gTi|p;5zXNt;*4OlR^<D
zDO9|i7sHB@yMw57n)}FKav*(eg6crqRmA^H>f3H9Bfv@lBo{90Z;L(DzZolG86l!z
zHN~_0%%C79%mbzZ`S|{5SwgcpX_Wfwr0G+okf(!f&68@30bCOM>FgIAVAdYRT&DYX
zEDO#1{6TD(WBml|+*YSo!q<j}xC3PJ7P&Ubhm#kt+$ShAU2-5vIgO;!@r?g!l_#!-
zspv+>C*218o~ak)T9Rr^PCQ<4=*+=K2{du~r9*{3upMT>Sg_B<3x8K1F5PKgFjZ0y
zWb*Epi5d5j3UDg6@G=$_&_xC)0?XRADWpf+6<S<_6G!dGvaZW_MGp_##^O+v*h?Q=
z7T6?spe4Q5T~bGApb{lSHsG-p^Y&W#`r9suj#KnSwyO>1rD?W?wQ5ITJEDan;t9s1
zB@Y#a=O{L*{T{&NpfZa)i{^0>a@dK~2Y*;7AI|wQ0R)BM<l_>!kI?-H0<J*n!Ng{M
zlA?byXA-w9d5M<RHCsZbQekj+b3Au=GIe9#a>t!^sLVfWW?PS@Zm;q}#g$WSczfj<
zt9|&JWGi%vmud==zk#;U8#A`7w&IZWh&4jZoRe!tgU8boX%5&A*SE&D;{0^TIp`8`
zRHLh*0cdSjIL=$whxU?%bn!Ne&@8V_lX=ZVD>w4j?@6=FO*EA<N+VI)?P=0lo6sAM
z2*xYLUMb;l8^|(0p-wQy2iJquGxM}49sSLy=iAl2`|8Rm*Fab6jbq47?!h*0nle>8
zf!Yi=LRagauIcM&;*9RuQDe1mP1YXPE5lk7fWhLbDv&RRCAzS#P%<4HIy$5a_19in
zM41$0@k@5*;~S&#W=|sHHQ*Y?KNZRLQD_2{n7c{pW!<r29S2RpW(gXr!gRN_<yQLf
zt)v=O-LYC5pXI&vWco(%smn|f${TUM2n+5bDnI`^Hzli<<7|t61f5yNUKky52o@OR
zQ96y|)HEyBFWEtN21d8iarD)lXeb%dxV0XJB0^;RlaGDU+B=+^CrCTwNH?@!C(%(u
zebPF7B!r9A?bw=Wpt;V}q%DfCe|}}MqHz4rmLYinJP(10g^`7oRhtT$LC(Qc)X3G0
zkW!SJnUR^5k%@_e`M;kn>I>ki;7UxaY@D1zjN%-eLY%B3%q;9e%p&YuteiqDA|k?~
zY|Q+Gy#H^D-v7eZ2)Q_zQUm<JsQ_e)GL~u@m_v78*&nmVZuZD*8JH%rtanH#UHpld
z2Rz|SOv8&Fa;WP8`+}UpRK(H{`I1#?GV$op!S_P>ASW?Vi}S?r+)m2{*^*S83GfkH
zKSzfn>nGPf*KIyE4I+?xcF4lWw6MM?l2c3;Xrw`)%T8B(%JBC{xg$+)umBjv!*$>4
zDHF=avhLy!HDNSW5x647K<&~XZ_QsYMo6r^#+ty<v*tua-pAyc*Z^KWXcR^$iCxgF
zfS2@^C|0<WUD}yR;Gz;KC@b)3mbPHmw8kuj=sgHn9VwL}v?3G71S&4%i9-k}T8nZC
z6Shx8DVthkbs-4Xe(+VuEx@dO{yV}~-%>63vM`1Rm%R3joCKD&+(a!D`vCeX^oAkN
zqcFka06gaJY(~fqLypUUu3CiW0U*pkBg}vxM(9vO*cjp1;FO<7jA0=fE+7d(AJKSa
z_gREQ%fttq>1@9+n+eJFC0^Cv>qT&U;M&nBj>z(TR-7Q~^$m=0CxG8yZunDF?JBM-
z-DTV5S4qLwnHdEo1qWlWR%sNJzz@EYwxlIq2fBr8`c_8P@+;-EpZ6NRecxpZwd<8{
zImkyZ&@d!T*ylEj8)qRZ;3cQ}^^9IhD(Egg-<c2XoHddE_i$)>zN+rz8Eu>HjXA+>
zl$Cz(Dg!LRtL!6#F&Z-JakY)W8q<^U|Ajf8;hD>QVL$(Ba$7%ZT(Bys38Giz>6qf)
z;e#1+<vv@SD=S<OEeV%L$tLF}bCo`QoS{lsz+lKs-uXf}KV6)oYF}_FS(A^<zUFRn
z+g2<glo!eh<OFkS-&^dr`g1`!OQi_2o-Wl8umWf=wx4BNSy!Pdk*j9TvWU>S8&Q|5
z4h5@)Y&7E*mo`^vVf~AAZb-!1i#!|OCzW{aJ2y8EeyLMFU_w@EUXcgFZ3L@6fIXXE
zi_Y1lf`@gHkK=6!V*7{LHgKwT=u$N|$F*XukVATd*6K^ws)I}~Hrbml-LqiPV(c+g
z`U{Yvo36TUG-ffA3{Jc9sAyGkM4yxlPWN(=HBUdW5>0j0{An39k3X^FSJt?4|D&(6
zZV`N^A^lOs_>edcM5EhT#cCRi!I2bFrCF_kUAHgc5roD7D^CA~*Y>G)2t5#MAXdE{
zu6q;fi@h_ue~A&F=@+j9d*h4M=D~CC=HPYsog439L9&xfva=KWdCPupNq_GncX*N-
zzjOcTss3v^tP;6A58IkB7+JA3irJKvqL2%#X|z;9oP(J=Nhz*!E@Ec=z=8LN^b-N;
va7_@}%-+<+)!EF*?!T3Tu{E@dtC6#-r?Z(kG%G77Cl?bm8JU=ZIP`x2H{$M^

delta 66505
zcmce;c{E$!-}w9KNGD@UTNP7PwZf+pHALE?rlM$zn5w9tsMZw0=|D{-HI$;FHB{A*
z)?5<lgq9j&3PDmeiwGhKBFR19=eg@ycilgpb^p1)wf5Ra&VH}G-_t%P=XKuEhL_93
z%YDSz16G*+e<cW92gmKl9VEz;xF5$ZDxTcV@VG1~`{QdsUzJI2=JDw-I3Fo}^^{9y
zrw{(ZqOhoQp76K~m5HhGB(1P7FQpa#IntSM_86J(%Z`rY-Ir>R%)O)X5d~~tediSY
zBS_C;yUkf?LbpR+<+VibAyK2%9p<}irJkO>y!ymE>a8<?ubtMDq+E~qY<*B?t%f;t
zNvmoV?#j&5d#xS+G24}CQ>B&sXH#!}Y~IBr^<4Tk=$0UNj$Zc@>0S>XLfAWm?(e*G
zH6#3n0<c$Q^MddZ`BC<liXjht8e;W%n0VEz#d?F(F7v=0qyHcT8i&!pSDH_K>zvuK
zuwZtmA3A<B5N;f^FM_njzoGg3fqlhD=nj=Oo6B)@*N>X6Uo>Fn10MZ->Dhnp@aid*
zyY`QZG`SozLD<^-C#N75?L1Or@*iWr?-9G7_U(27N~9)j+moE~(&~?k9Gq3jd#;kY
zS9{+s%YFO%-v41Ow@p_5o9g<z^C$3B5!wbPe7>@CX^DBKMf>o{c*BdaTXH*DI?k60
z{m(reyN@WjqGjVY@igZ~zd-kQ)+?y4v)@XpuxhK+4uQ-)d7}8%7ay|Y>5`<XyNJDk
zULgl+o;X9gh4Ig2gw?H>t7f-zozu-elH1Sf#M#;t|5*r8M`f>=Z<DTRy;-zv`#Ih0
zPZ_(FfY%IyC63dPR-ywZ{*|NbU?`tMb^4MI1b$DcF1aH2uYq6l_{n1zdaEAuSzW)`
z75E>g%ZiJtPt=Ko-q`w?t#3Y~`qxP^jfVk*p28IyxM0*q$ifIYo?7ms0+sO|2MTcP
zbCmqFX;aJ6Y`c&T2-Y!hfqyfSN2yns*8tpodBH9=u%=X;PhQo=JWj9yZ14c?Bo5%)
z6C!wLgj_PfSD*rf9tB~Kjc`;b@Z_xkTeZS<E>p<ww?Q|vgliejwk@kGlMUxj1+|l-
z`Rg0)7<a(sepE?kJZqQD<|%IDrTETWWOALAZ0+}rC)Lp<2AaJS4=xKLfp_q?->0AO
z=&uE_Vx{|5>jB23&HYA~UPzS7ztrzU-T|#mIG@FgAPKh$1aupK9X`*;6E}J`kql@4
z2Y!P*bN)10Hhe&9D~~2vXU4n~@)U?0r3C_x4r2)T9n>iZm1i_G15F`?G(pH~n`p08
zfD;^@YDb-1$z&`8M0`&u&0UqNkC#6KOvj924O}L^Z-q7{yHt*HPH(O>9?zd8Ul=b|
zz*w#ZN6GiQf~UL!#s-Fie{E^D?>iR6<~O(VQ)^gsUG}NYP=TKfk;0*|bz{g-PJ@ST
zCAYD_J%C(rAG1~;CC)fGS}>IQx{q96s}kTW=%)k&c-EJPIPS3U3oiow_kpy0DW5RH
z;*`_r>I_Xls3Tk^^LGOS_qYy#kRC#X3vdAB`#oX)cMZ6E&q}ufmmEgcBu2ZI$!Fa_
zZ|UEl>ceNr`Sb~r8w<UTG>CdxBVR87#@xgC3|8np_Q%?<tXkp5M(IPLg-t<xj2^%u
zzw`C=k7u5LJsQM}oMq*AbHZK~M6SO-Qqq!L05sL(_<lZ}w&L7PB70H)7e3ypD?%7Z
z80104_?)V)#wA>cjV!UjVAP|mt5HX1(l@@W`vd!9-9~-@b!Ol@7mAP1&b4jfP)=Q4
zMRWA%9%Yz5$j6pnK0#d(;#>+aqvlVUX<(inTRZe1U`js^9+}}B{fV6wnGtX_l9?97
ze;f9yg=0?)02gkgKVODm*`J;B=&Y8E^{)=d_|`5nHgk!2ux0*8YO{RlW(mCuTO+UF
z?n5eQp{$g=bJ@Jd;|#Qyw_{@5{oo&9UO`-1c-_qPNe--GV^Wt51atUX7WNnl*O~FT
zrH_<a(1d6R?29R=$FWE#d4E3u<7_(8|CTcr5E_M{$(6g$qx#TSdLGu!nwv0C*-c?Y
zJQ+T=>Cr66_t8FpQK_3WV(8KpKeUs{WwRrp(L4{BA!Q>xg7Fy!(Am6C<*?7%MG&H}
zr-;SiGDQJ>&2eHKQ~pjRth9+9b9?_-=siOFtUq=RW#>=J8NfE=&9A?Ib`>oc!iK+Q
z_zL)=@j6b@l`vdc$UNT=oZ`BEcUT3Jm0DNa9X7Xq18wb&D=FuM<pZ+UqZR6e@<C2U
z4JkQ%`TP29R%mLtdSu5y!BTjBKaW7+)6M|lz3vDe?Q-@2?Ew$nLp?D-p@qB#mvwHS
z)1-a)e77KH46c5Ai!iEzNA@V<OmSMM&kQk3v>5k*gmM_IYSINB!Y6Ff^*0gd^tPF1
zy&+i#Tu?vVjhb4q(tkHIyww)H4w;t+oM@$WaKdB|`P%Z_`aK3J{$bvOejto2JCceC
z+v3cQrH=BhG1o0~GnY7A>aJ4cEkAp6N?zR1JCS^V$2wC0AkYiJJT!BCv?GLA+rVZ<
zLjxks)yrYccdOP?iJ<F;g#G#gYGNrLmJck%SG4>11JnVm>!i~34UxMr+RQ8{hHqe@
zd3=3=He1-t=L1pX$Ng304ZDCA4i|LDO8ByYUJSe#TiW$~-gP+%Yf77K{IEf*3Ic=?
zV~}WE4gX_TX(a(v0{iJiarhwmJPlF_8xCM{32a_7_I-PEl~yGzR1iQWo^~g7kY}ET
z5coMO{S;y3xzZs+z5sLtGrWJJRFlQnSmOb9YTt(3DSRfOt`uv-xJXVNyj2+ns>kuE
zz|OmFPsYyrF^Ho-#;SsdKz)-ahRGI-z0tPXe?qTrHjpuBOv#!YBKo(JxxU)Q=-{>N
zunO^WOWdkBlb(n5M0c?UwFaUvm1s{4nuHN1#fyr&v^L}cN7aC+o7pqEZG2!gE6>#q
zg9{rwkxO}A-7q=I@TI8)bs77T$+Mlf=!nuY*j0VO$Fz%vxuIvf-Mb9RCVl+d!)t67
zxC8v8#p1#Z9pa=G9Yu-Cq^&e~de`v?bftHFX`?INDkagltyH2Bzmgz)t>a{A!&;<H
z9@n2P!<9jN<W7Ai5Tn9?r)7M8_?9<2+j2Jf-pfMv1tMKZ+rBT@kMXI*v$U1t7}KuM
zU4doF19?;&g;y1inR!~252SS4^DCW_KwF&X9=;6S(JL^}>Ua}cX;;t21%n`VcD>tF
z_?CxZodDD;dyQ8)&eZe5^#ZKW53szU$t^^I8qIVz8CZO&T>dW+%t}`6N*;~<f>XP3
zk`|yO$B1sHy`B9W%2(n2qD_u?;aqaztIfSIgm@l5gBJI`ezr}}8S2p=IoxJnC2bo>
zCs)$o)eeE1;f(i($&AhLE%{Bu1xyJvHX7CeX3KfFFP<<nS4nHwI8Wrvp~Q<eBHG<=
z0e~-uypl?U9Y8bQS2b)p^C0d17`Fv2E_f-TV#8u+Ubur@fN5wWXTtFq4wl3p*rcye
zib`8-hE@X5(?h|w=%f~5n)b-)=k1eqrdlZlL&DVmS{yT0ez%ds-Y|RML%=HFV3F`5
zxIB1RK;^>jU<fTl4ljHQVDQ!<rY&Pw?E;0*N0t}@1`5OR*o=zAPimErgqy+VR7&X?
zGo>-aV!;&A;bZ9-mW%&qFQZb^GPgi&u;(JSfhx?QnGS9C(;Rr$CZmUV3jMsQ@#UPB
zc678ehEG^vz3%5m_2`&((wYVQv+Vfv+L$Kd0I#$RPH0~RFypl9As=r0CZ0RsPUEVG
z<O*EK7y~x<@5;=|3-6vt`~b`MsEukFk-L(`Hr$Mu+JL#yStY&(Wbd`t?YyqYnT<bL
zu$hf-oUB>toUfiOC`=xKzVZI}NHP{@aYw+7$~r8>GSjK<1>*MnbDk}0+&Px-I~X?T
z9P0=RSXqyLfWCOw&}Q&?&J0@e<0r_P=TLI7V3z%3@-0p$XM;p1DWRKi4o-Xm&L@YR
z&R&WJxaTV4RecTDBAsYL?aET-hBiEs5Gv$Gx-jA&B5<_Hhj~M+O+u<K0XOS|ep6M-
z*a*<_YSG~sM+>XfKI)vMKW3y|1HQS<1GIL|HJS2xj#$*2QKKC8QFK`<dB|<l{<p|D
zwWGf0WBY@TWU<X`#^-B5Eg^dYTQ72s)4*tbP6$M&bHdT*CVPe<UF*{I4jEg6fjuX-
zbVr-lJL}`=sFD8R&LhHZP6$zf6V?V)!?5)8dW)eLV3c*mEQ8P!HO$}E&<+T$+ydx=
zoh*B<&K7zrjRMajq!9ZGM1Xe!ei%$(&A{ko!3X#vTP4-WvLgkj)@huEg~&Io)%FEe
z2l2amC!(Mkp3Uf)V0w48k1PDH4sRyT%{IX2x&FYpNrg$w(1tDeRCAYy`^E4zc*|=J
z$j#rH7)m(-=mx`i{eXbVX{Z^BU&8o|;Ktp34Cw;Au!YZmF7$$Y*(9O1ZGB+HKIH~c
z@81Vlq7fFeht!zT*ohS+mBbS__`!LEGI!OI>v>u_Y`BCk`3X6XDPxtUw)&s4kzh*p
ztZeyH?ZsG-l{XY3g+F29{S^K+!%sNC{%ci7Ob$9UY6JX@U;fY}NDNB!8JzWv+DpB8
z)<7C^iwHEs@r%)m8Z9TVw`k$b5bCX?@8k@sT5RN#$%FJ!))&?mp^KNaaoFHvmD=CO
z*$lPyw-I-A*U`&uo3KzqgqLmvV44jHU%9^W+fTfXD9)@XFmGS({l?YSf-awn_`q;t
z4Q-<3f#mi`!!x)_*3<6f*`XC`iwK?@z*KkBfVi~V`)lDFi9rz$biHg|>2SGwFiZSx
zZmt*~tgH3L?98~TFX4<hj&s#t0gvy_-sA@gZ{Ve<fo>iU_w*Sx)}|h?u;qa^VLauM
zna~(5*q2Gk&;OJ)Chv)lRI_^u*Sr!tdlm3Xilm$WShh(0F%c)btvHppIrdXUJ6XcU
zz|t9^o8~RyW*x2Kb;QOgK|A#72SC=mwB^y}6^W<?u73mN9=0}FGWkPjpIDHb8e>xN
z{9p?yr1V~Y@qqy2Q_iQ%9JL1|8!fVEDO=t(M}jL_%ANEk#hn~Ie=9Yk5>z@+CYONf
z@*Tyxic)JGsql>FHB(E0^;f-Mu<^rXM7|bWz}mDSroa;#_5m8_RnLb+KVN>aawt~z
zpc`%T!S`~E!ilK*W!vv+?)t`anxD?aRmtexnS1B2Cgy4&_UK=OZ-1QPZyk;Zc6@eO
z?pes;WqOE&Pjb)5@ow#4?WCHUue$!;1Gs18+D6WHzw@Y1saa{iC-1|$INCm_SRoU9
zVZ$-tuirt3@{hT#gG-(lo|k*OSWWuL-I{Df02MRAb(o<`wK7jGNLHqZ`yWS~CgtuK
zvaWq*SW$wsGVwo+@cfEM9vX|Uldb%UgkJSO>zGG6%Kvz)=z?73_~6mhq1Dn0@|6Gs
znPTb>b7YYy!sXqC4HAUJKLN%;Q{SBvsn^!R3Uv<ROXjSiBb-Tc$yCJJ%##LL!dLO2
zGm#mj(fw59+ATPKWR7t)=wf8&*Su6JCte3aV9#}#2EltDne8(~UMW71c6BF0(zi{7
zk#PQaU%Hw7_JG9mGf#(%bESQIEP$uV(9L4=G;NckhkS3oaY_q|?)_-isce;83+-(*
zqZ=aKilftZ??xQ<{VNji{QTLz4zqJR7|+kIJ<YnBd(_vz_|xN|tGT?(Xq5x0wHJG*
z%<|yKkmA9#yQUO*WJK|_!g*L9$82&(CFZ>FDaZ7J;)h+fR-%4JH&hG<lHT71jO$cB
zD3v3m1~YByR6m@tt=!{heFJSc@aDaY$p!TfR+m$f{08H54-Hnme`Zpr@!r=qfOmbe
z5FtB=f5S1kuxtdKS}%sRKzPkGrf~|Jraw>+iGD?lM2k1X&+2ZVW2n+~!_3LQ%#CQ*
zTXLv^(oHl#lVyIVP4-1FXq^C;#l>&&H>k1;WxVfYEB(CAGWKU~X)<?G5a!Hy2gk07
zbIJBB2WC$Tk*81ciZ*b0&xdH04w|&}|BB#5G$-6$jyRrMjd6J0XNVOi0GRMm9#;HA
z^u;b&6f0_KW6A@z8fA1IS%4Xf>2WaTvNslwH+^G`s{@5S@oE5<8<hy}pS5rCDG1tt
zL>n6BXu;eIDon9G8-~#>#8(&s#HAX;i@Ypcr0*dWMiV}thipcMNJIa!7nCXA=_7{*
zzCY#8vX^_K2FaE<rkjmT3ywUR!S122Tn<Sk(^cNb$1hH|P!){9JH^GsTh)a7qtq|h
z#yPEeoh1pN%XlrY!9jB=2tAX|A7bRI4IAl@^y;&>VghVjjYUSNO^(ES7}r&JGa*bm
z#S6Vu2p^`ZMAnZ!U84nRxvl%ut7)yI1qQhppQAt+?Zf42f)@;@P4|ZIBe-#hp*K7X
zz0nsj&pAI?IXWsUNMo$nbBr6vHWYuf;0vu9AlL|*f`9lmynX!jc?Ri>pv|PVlrPy9
z#C5Qa7B2O?4jsNG=zc6j0&|LygMvQu^Obz{O?z&Gbqj`nMhN5i8#jjoOW)jUiQ#X>
zMimx`?e1RD^MWb!wIaS15#SZY7kR8WXtgx(N2G-`K@~Fyc(qZ@>*R15Has*iLTgD7
z)WN$qhxnm0?MprGt0GNAV1@g8M*@|J{}UE+#$aLyyb*m@fefNzi+s9HC$%}x9@!-v
z!%vbT7zR^Ho8$PYTfm&b^b?1#&I_!Qk#Z3GpECdA)R(PRURQWGu&_-n>bb54@3HV%
zB<o6_>8!bUR3AQZ?niggUSMSR%*|abURbv1y>O~Yo7JpG+~SUvYQ2E21&@G&LR!~o
z(%=vKjexu&o}5`Zxl~>54c4LlDazeeIA`$DN9D^t1+1yTC##IV@m=>T+B;hQ#c1RP
z>En(!;aI;%qt;v&)mYcLK`gsb)BI95gt&fCXnitj(JCwn#n=JJ06moqLg~G<NDYW4
zWQ>YgzdzZ^*f@b5t8(D%!GC^X<q6?08~!d=sB~Xm+=SKZI={Pv@+N=zo;;J>-Ko%0
zirNGdFj?RY0Wz#S-Dow@e>TLo`tSuS<LrN0Fw;*=$2;#_A<capxHOX!C%Rehz4zL-
z<iibjw_6@NuIuy<@Y$T87oe^F`Bd|ffJ_TSnS#=RiIV6d^|JCuw^b>%5NFyORrzlw
z&~<|(vASbF-nJGN-gUH4cJ98?U-T*H$O5Uf#tgjRA-*-$Ej#^bB6Voysgg4hKINT#
zBSj|d-;O@~;H@KdSfeA&#^cx0Es;-e+Vi1Y*T|j8k6S;*odvEQ-nt3fDklorhIR*q
zt(pGQX4)5r_pfXU`Ag>)$@i@uxY%3yMUihpoNNb*4OY*NcYA4()G+}h_Bl$tVL<xy
z27e53W{VTJ&6Mc8xP!QH?aFc5uAm{1`phls#MSYV6`ELc;sJMqgT6|JyDR(ufH8V&
zXJgE)bI|I^!0_DnfWuAk3Rm{aKyyeiJqRq##B%eaNvG?~UkT`i{Wjkl&X*YEPDdIC
zUGXuhy%+Q#A*<^Q*Cn{Kn|<v;nK&Bz>CZ2XTK%}-h77Jo#^D#$HV-a+z^=|^UM@9A
zch^hI;_BK|zyAW8CKV38Dz=n>a?hUF5PQp0{$ox{{sBy9UX_uw^Jzx@DyJ7t$B}N6
zs(=1^Wo!vew?;)Fak7`+1lgdxtidx^V$4wZfu~K#4aZpHT7zh;iNVHJz{y|@C1bUo
zAe9NF*6u>@y<od4Dp4q>F5?^UV@#t^2C5~4O3o^N5lZSStbG0+aqyCIy@_S=lngYu
zsW}a}+F`5nSK}1rr)2_Odwclv8*+&|oz^l<<A3W;e^Nw@inh)(3a1IrzOWBECTEE#
zm)S$c0Sg;cW}1ge<SgnnXsEs1gq(G1N=VYl-&gU^BcDgv1vX3Hr=vsECQj`7ZN1YP
z9e!3rnX<=&nq%DQY;7S6#hBC@=mx1JFwX#qSr5*dc1)_MeMrtC|BY2O#+%n>iY7{j
z|G8mR=-o`azkSe9Kk=9Pi;HRTCnlD>PS^<+W3d;bjcPR>_%$OFe<o(+MYWU~bj&wo
zM8!&s$*T=oYhT_6o0cqUE;YC_6{cn=w>^aPAkQ8GJsN6uNw_F6aeuFuEOghI{Yql-
zfH60--LzNP`S@?W-47PQXDZR^cnN>DxKU=m`U}_EOfx^xhs&@7z7oImevFuZJwNp;
zmv(|=$N%*r;Z6J=<zMrV!V;=b0@}H~>=8%$hlORw<hGFWJ3A91jB2A~5}wUbmGS_%
z;_Eld63|_e2Uaw;A6Zp8dQIZ`-4dmVR5QRYK{R&$MOx;xn3A-_h&rYfe9Q1#Q;(Br
zIv$jeB~`NDyhdqW0vZZeqz9hd7d!rF*{F7GHnUyxsPzj6bydt;`lp1f;#W3>)Ag5T
zH1BMGyT@Yg#pa)AxjWLagG%AMrNl{R_Y{VkwP+Pir<xhqLD+)*6VPu*`<T|ji-je?
zn@tI*c*(&~lAU^-%Cs%NQ~3(Q#Bx+i2CA_5SQBA)<xO(dw2H}si*ctB$5hr_iiu9h
zda@+rlwkGRxHDP?inajviRw|r)Ys`n<z5d-(X^ADZi~uN^dQ-sM^dh_%5j9`8zbte
zrS|(@UePtS+=w!&%|tF_wu_|wR;B=7X*EB?=Fp*%fu=X|jV%*jyVl;YC^czur4f@K
zWgY|bz79yp+G<J8+T=?>bzat**E>+QHt0()V}?4*x{=Odk$fC;oeVY)3eZlB`LD()
z`KvEr|5pp$1o;O2|JjI%>ty$y5=42WKnpndj_NHcpne+wauJ&%`)&~3|3Y?su+NX+
zlyDvrFVz)<UUXmXPT}4-h5!+I<-5toORA94eQTz?UG{Q*iDL*N$7z8qIO1LCZ+fjV
zeTh?gD={XZg1q})NtDq`WHiZE710joPCiw3cUk&gW^q1zE-w4?cTCUei(@y!LxCzO
z#_1>n*E_tQ+3OKaKP+RlR#rlOQ5s3Z^Xc$vi?yK-7X_%Bg|q0bzEdO?EXMccTrkGn
zO*OKmU8_tAN?*?pskc?#1VJ@?D^-s~1Xeff>BbJyrw;VpJbkmF<WeymW}wzrCym@W
zPq}%CTI?VGs@co|re9LSyJJQMta~Mj{Q)m!qb81=*oi(iWJ!8HQv8s;N*@~$*K_dv
zC+X<D^Av?6c`vHCA1k|FR7IA{$kcv55}+Cz!a%+TKM!llfC4$&Ae7XP67!S^P$oa=
z#fvKA$JS<K^!s7ev?}8bZ8Ne!<FcK8$I)VcEd)c!HL)-8MO7^w2)NWFH6Yjb^3nPU
z*FBbBi|GWH<2(9p%2RY7Xn5>-6+)Q3*&y5Za&g0`iR!65V%!vFy8TL%lr7?f>ttkF
z)s&*R8JWi`R1_VG6G`7W|3S-BEO%HwXU{yP%-^?<O=TV|lbW}<6q->Li}dY#c`ZiK
zb&t(>@oC+&(tR%@<pCt9oo!G?)mlZUJ@swnPdhV*vm=i~R5LD_IecTCn~$Vr|5PmY
zFRdy!cQ}iR4N=uVWmN6eIB0&V@8z<KRg+X!gE<*3cxCSJB;-|y>d)uq4i8-KnmL62
zG_E&pN~b3oHT9YF9w;t9`dryv+4a})1AQ-DO6B@)0$SCoeQW@7dQPN|9fLV*A{b&S
zt?r$-$o`cD`b}t%QAMGC4zy(|fUdm$$lRg&=f3$5e6(1u8SMR);?n^q!0CL|R*&{g
zqke!vCiWDULwbr&=Tuy3T8PM{UsZkLc&%ySNAVkxw5r<AOEqR>Yp)OHWNUf5j4FBC
zo??Gfue0;O2Vzh=7_g>7(EWLx#iu2m?)BZogc8gg^6$<zx$&dH*`jpFyDW{v@<Q0$
zUGo+%oHMGXnmS)p5iZVM_DAvguXC=?ntGr3`M|Q7Zjl^Pa=;Fh;vGGo<|2EnZ;&Ta
zTpmLlDo}LQ54~p96t<9*8^Rdy)tLuZJ}wIM-6Z19n*PUA_l%mte&Wp?3R5f19Vm+N
zFRJ9LgZggbWRFPCTU6PLf!*kVd5Zy<_IxC>^o<f|PcPOydsaWAickTjbv^=2YutaV
z&DX0t;i{j;JK@SRlK7{%b9Xvuc7Aw9m0v(tapzEiIl0^L04Skv{|X2hTsiAwJ;(%S
zp|t2CmDzf|%B}(|I6CU}Dum(X26h7_I`a>N|ISlJsNakI)tW}Z?$Ky*d9J4NP3MN3
z%sts-kOi~Jwu5GJ;E4Sf&|rmeo8t0%2QSOwa!$eLDj$j>@IsTT1A0#H&x<O(g6^Eh
zOPCX`+CcraCaKV$R!s)*+>jDC9Yk>_zaKQ_om_DzC(YaefWOcj)dIE2(LUkYd~Q_H
zl^3w83U232pFVb2vp;C!{I%<#$-i%8RB;`1!oXOrM`u(;8f0h&#GV6pUBzYTdZJZR
zlW^Po2l}sL;D$~Dy(%Af72F9ih5#rEPt!=NA|PWffCh@R7MHiU^WFZF@kDcn@B~Fy
zv}o=yfpiemr_9_T6h8ANaa^BysfqLS_Z_qB-+wPRagwk3fSVA#E9FvCnCyY4A%n~_
zK7q_d8_&L*1wkj=6R#<`7YsAnxR!Y{{Q_>_lWSv;^c|7tWV!_%K;f}#v(cg!E4<q=
zES|+q4_@@UuFytyi*!d5(Ug+^+4UaN%~ybJrLh(nBku&a(H;C&`X`>fvS=5-i(|^f
zEbf~S_@Te>b>Dh56UldbCckpB+bxyLS|`h5Flb0n>t(CwrXnyD-tYwy<Abw-Fd+q4
zaXj^md+~TXkZXgFLAQK@O>O)XWXzTu-+eh-Go`3I(>$f9KVt+0^<)or<-L45_2z6(
zgBWEs>YnNsR-JkC?8DQiu}o=tHu<UX^;hI%dDU-pbHvM+#xJtSMaI49<nm3}>^0_Q
z%Z%bz3K8oIyQWXqr+ZT>G!Eh1I~KW-bS8t{SaBIhg#@|vj0^G9Qa4!y=hBs54TKlR
z27J&@6srP*8iCuGf!Tb<%a)isJ!|Y<PAU(_P{Pd)Th<GkV6Bn84CsO&9R2+G{!xE^
zv`0@U`X4l|JDk;`#E!OfdP$fQ+(V!JB>tFTH7Z_j+6;d}p!A5NZ(Z)F4-XeP;wjFZ
z1Ol}zQSDf3YN^lg=*|nC7^awC%Oc#M=0Sko0E0iy;Zyn28`7(?zwH-o1_Vkau7az8
z`?^Y-)wsJkwP82V*72DypGjB}*ZoNmNTq3w7ZggFa2t;8c?_Rfq44WsR|5et)g~T#
zM4%*iqoTT~1;N|!GIBVhG%v$B0f1DV8x{C{qkV!mOM9gf^Aa>_;bSx{!2_<(N+}E&
zrY#8!1QGlgy*77RNku4NSeZjj{CR@)539L*-yf@dwTowe;FbE_SfW=rdbs<&&8TOz
zY*f68+Ha$?25tDMfz3fw$9(ajh|c-rha#S5#QgRjo_)tUwz!XlUHqvJL|ms%JYFt*
z4)?}pI$`vKly$0ECs#N0aHo!|f2~=a`5hEN-oZR4{;)3_b)>P0bzaRpr@<Ak#cJG#
z^>YiaO7mx;R$wecX6IdqnRiO}KKYT|HrGaSXv=NiL*GrLHK2Tg(bza2>aux+Pngs}
zgYzP8xm^xj_Jp3-sM8!EZ1z(F-R!%6Py`mWRKhN2QJ<P!jJQ9Wz5Lg&;G~_Mfv#>}
zFMeaKCH5C*FIO68J;GD`R%33#!1Idj9--Dht*#n_kF@h@G%`NF<63@ufq(v-@pMWz
zYtI@3UvRvnO5`K%#q@RozBSgUysYHe?^aRC-<3ZPCa%XA?{@~Y&5nf?Lm$0qbcS4h
z(-4uKrFh3kKQ`O%>9uV?6DuPm0;KeVmF5pCDz6V0To$Y5d43WNI;O7vN&WB@d*v_G
zvzawN!ktefbyLTl%bfLkob0u&)BX0bWX}Q2CB1VyYd_vqC^{TG?Ubo<U?4r8^+kMC
z1(nWw9uD1|Yi;n%qY|jG4lQK={|p)b7sd$upMZ@23u6TSPe8{1g)svECm`eh!We=7
zkC5?NYK%oEd$^n*GqNx}J?jA3#weq+Un^s)M|)TowVupcO@*#U53D@wVRiyN%nbZn
zK}&}H@CGFF#~kf6O4cK~XDTxs2%2?jgdZQNArN1P4DqB<F#v&>7CtGM+bH3ZpQ9Oq
z*G9msAlhKLg%=T=D1uSTzIGeW-pE;TqzPY>|1#+-ueHgKha#A)J*>+G1qSZNJPhYU
zOLQ7$bBd0nTe?NGe8VoVa_EZa7JVlL0|lv+;n~jb%`vn<aaE1(CEOA#YBBv77AM}*
z^LsXDPN4_jP7SWH6`NDf(|4iJ(Xg-Xi*&1cgrk8F-WfT_`r+6URshGh%<)jLYi<fo
zTH(y??)aFr_IK<dRt|a6_*ZU=n>2rv%vyB)>%?@G`p$VoK0Uh6cwadB8_K#~2X9%;
zki?bNJA`Gw`HCG$9QsBrV@=)@FUfbE4bVanfNxj}&5<XBm)hj$ELc{U=|d0PHt!}_
zys(!+fDJ#LIueQ|p&IMe!WYq_d8wMUunroX^8pjkvt&#hp;~pHV{q;}EqZTSHk6JV
zj*dpr0x@c)LJu4>RqO}0f#`uQkTMuoFsoucC*Brz2zH|Gj?+E@NZ<$RE63rr4iEse
zp$0kO@13dqw>tomKuYkCTdG-(S$D1~4*(YLn?tuQSU^I7RH^WBT2Wv$SH#9N<F8iJ
zQUl*SJF}W-z9)AH^R2XrxRCD2@nHx@XjYoG&*0a_pDbyql_#$E>xcek1cA371!nU=
zU?9e#aHN3C`$Po(;BrKctpN(G?K7j@(oqD|_X#92;>UlwR&`zl>-keoZupwsKY2yX
zHyeiVQ_p&GN-u6w@zZas9;2?km*%(JUC>LO_43hw#eKc3#(I;nD8eedmhZ|szPi?x
zcUJHSb%QF;*K~X7>hsm0lQg|t^*ib|Br7NAcjsAD<&1C6Mmrt~G*^8>xhPSn_$8&t
zaRb~%^gmRGKV>tI`3Csj(6}yXW{Gu3ay=Qf^7E^f`M<I7S3bAST_2HNw%!{X2C-N}
z&6dmrEDzkH?hoq9{Bai3HH?|5%VGaC&<VPwFhPi$JtO{ZEqJz-j0jgvoA#o<UJ262
zwETjc!JRPwz`92UBz2f!zwB7cc}<!#(=)`G`8ZEj*w&}+$ugZ$6rslZH;WmXJ<^>~
zDe_{~`}3o-12WRX7rxqkFUfOCcyqVed+)P!ll>M;h`);V{o6%Lq@bAVkoLcO;-JvV
z+r>^1j}sAD=7@AC<>M2lKBZ5?PsHij(~9ZY63KW9!s6R7@ToU5OOK@ODp7V%*YTUh
zKxE9l?AIAvFF7TU;KJUwCFhcMC{^VAQffXjq1ci0N@+wl|K}68p3$&#Nn_IeI><*U
zX+xg{QlW*@uh#TPp^Yv@s`|q;#@L!wJTgZuF2skqEjepy%c(XhVoDQHziXTs)*lhl
z;hBayZ`?Wb9f(Jcyfy{w*DjSBNI3?P`b^r;K{<3|XR*)@Pp2D;#@Bcf(Ed}18ASvH
zewS1oZI2OaT9bgX^r|Tl3WeS^C!Qfu-ATfCVob>t`kA)oSkt>tjZ#ZylZ;q+AE{Hp
zk=V<1s~TRmy%`6sLkoTkN0rz-$Tm9Fge)`h1^Eu(Ymk^8qgP_!_i(Sxx6)+xIdgYx
zZ8JF9nqd$s!c)F<Cf_&B>|ro&bEXQF_AB9rTdd1sM{=f#s@9zb`=gRja}vl%2=L^r
zxP#ZV3iEWr=aMUAB(=Vk$_B{zDl#%xPp0)0PEU5pOF&1?fG|fUI>5M5vp(f{#=*;S
zN<c%({^tj;{Are)?2^1Uwa%l6uHIGAPRA!W*B)<0N<q1*AO_48BiSi0|NHlNq!=kZ
z)7;ono4j!Kj@D3|<Yy6;hn}{Ot4KHF&X_TK$Qg@fWR46ey>=<^xO>8YQSI?v)yY}@
zq6PE?ikS^{t0Xz=LC3Q+EnAZHBf7?@IdK0>g?a5g>)zvtQ&W;^KU!NCiyKov=D(6x
zBUy{JE;LE+ys>26H^YKWpGnGkaEcMD>-{hxD-x@Ope7?X^sku7s-@it98WBcunj-!
z8b50Np6(@1vb+A(x&oa47n}Q0`W`9l;xAuGsH@0`b$z~_{=`9%MQgoxW5*GJ$v}p9
z?V`jN^_S0fCM?S<J9?e>QY^t6iI(2BF7)=N9Y^f-17Yap_-&V`LNJ-_`o#td%mXC5
zUaw1sOG2y*r?&??uB(CUtRMsV`I0hu0#Z1wIX7Tpxog{gl3ih^1QheRF*_woecHCQ
z?)9pf#Zu;~SrSMe`!x_l9x9S&0rr=w&y5Z?pN>f|UnQeU4v_98WNq@E#a4Yc0BKlM
z2C^)3x8m`_RZx>cFbZ5bb<C4j)~MrgVN%dt0S9cM=?6%wq7vHBLe)@mX=Zy#W46g#
zvBx=DAf&ubmeAFWJEKpvlo*z1BEMYv4<DYcFE(h>TnCw*pIrJmpuoCNRXxqM_TY~T
zu`~(w-Ng<a#YzHe;M`O=AYWLb$d!Pa`yQW8%bYUzy3PYh?KxR4CCD9vobM8aA`BC@
z%Ruwb>Kk|dXWeQLdvhFP=UcxuArnPwozL4Q6_{W({Xm#`$n}V_Z7$t-^u>~L|I$G-
z<CAu7Y9N&b=*If3waG3!okAq<g(pFjM`VmGXKKuAUF;#BW}A>Gl@A9bF<p1U8`Qd5
z31QmIg~DluUZ1P+4T5*6!6xb%vTXeMGvvr(n)?m4&HrY8v;U0)|6jcV{D0^m$Or{_
zDL_8mm<cc)n*}9$q_UUeoBIY51GAQfKze5MsY<aZfz|ltQ1HzC&wnat4_tc1viBhl
z{%pK-S*d@Q!^5YWyYyT%k)pvaiiH<l-`wTydsn(nt~&1ur|I9^O1J3e4JI9`ul_NM
z@XNI(Sz0~ZdI)c^3Hv>P(mhu4E?UGPh$9~p<u(Br?JBh{dB;1%k8pXCIH)z>J}Gr8
zC&nBO$*xGrJF2g_F==7N+{@Ugv9QwOuE&+U%SLZ~v+M7N#RQklPeX`r2G)l53l>C9
zuzQKTQQ}53uq6(#=sk#j;P!T&w7(V8XH>1NwgQr5tc|#pzvQ{oZLze-{3z5vJ|Hl#
z7N~&u%_G5KtjhZms>RBgE|2nf#vM|1<a-zy6zW21Fl^5ahuhRv6dZ#3`#2Yy+&ZTS
zDfDLCKPZZfXEdZ_v6D?YDWk~vl&qARl&qUmFRqR}33Ci;y=|yu3Cew(_~hDkarh+E
zf48CW8&LH5_#dYtZ}vysn}yzWz!9q|DQJ0oSFtqM+rPMlj(ye+uKTJuKj2s?)oGhs
z>*=_{N#oAQnm68zRrTdN@NLZA>PU^ZZnYIxWRcF`OpYY~LhnO;Q|ilw0@&a!8EAP5
zy!1;lCEhs5xN`wDcdN*oF@MnU?K~l@2$bPoMmq5bRY7;%d0&zM&S^i)4lsW=!yji2
zdd`i!<w&&dgqHV&+dP)#hZ=P*d@JWd5w`$%e|03^aL~=S*OHXhK*^AYjmXBhY9VgX
zMCwX1D4v^Yh+D#0>~<~6%E1UyQ*Dg<00fUlI?)vh-MhLJ_~5IfDEC{qj01shy&3##
zwG{#<#*-5WV2FzGGu1&1QE`R7a!`MM(1#1Im@?Qn(kWmDe>ogReMh;(;HgU#k*X>1
zH@+!Z4B3IBpguj%s3}>4?QOfk;9=EX6q2erQH8#xlp5$`S@t@%uwazEH*<IZ>pya{
zvPEG#^eD8vy<yG~bk_#1Y$!wyqenwJq2Yji!28V@)<M22#>4^i0vGfGu>rK5y>)Id
zH*gZY7yrI}_6F~RoBIh+N#cJd@}Kno&snJ%FqrJU#mGP;L7}b$RK9)hK=0oT$mu39
zHsd&00k}X*Vpbh@WJwJ%G=oz|ziCk*9VJx{!AmLLY<{5q9}|E_m;<wc5uFg@T_+OM
z|8CSAdJBw?pwQQZp_?B~`e?5o2)Nf)q?it-2Z9F&d3QhYkz2jh+Qg)m3p(Xc)x8zN
zgAkF0gsi%atB++;vg%60KDd%+Rgg~kVI5#BWCM{-&Ie6;#Oh`r1%m1b-rUP^p84sU
zdusrq03>G(PUd9Qc@FM27{F#+f`FO-``I|=d3yh2nY;?A88C!F;<XiB;r1L`{UIik
ziUo7<SaMbh5wvM^2h=lnskVYpkeGGh^cA93AjWTBa#kJBYQ&krJpJ$Fb+qPOEV%sc
zHJ<6Mqylgk)tNq)F%L_s5gK<=&K6w)QtY#mcBq3NHp?$PS9UHT{~hzHxQMo@z2}0V
zJ;<E_&0<MI{Q;YH6Hd|&D^PF)v|%Hr&>M`hpA~ubmpy|NXmhC(0Th7Ax2d<wt(N_;
z!|fiuyo%_Jdjq)wj$HrqLlZq6a@)iXZH)z$81e6D!{38{ddvm-8ET{|D^ACfD=BQ`
zbgwVc38odPj=0s8^q(M;%ZUk~gTOzL?F^K0=etQGPREREzh+%2xsTwDi}?P4boTcv
z>rTq2?PsbNUv|lbDO5XXGn>MGxhY-Wi_nhILJSE6|Ir$Bn;y{NHa{CQayt6^(qGOT
z6Z90|xeIZ8<a8hoIfBF>M;Z%|BOxt6-Bvz<M~zZ~5C-No7l8&djT`8}1}`TYxjR6|
zr<N`*hedU@FLw3fm?MA2g0iEBh14a)00GuSLoy#$^b{c|lPD9@c3v;{71bf4;5C!U
zYx%}KIe~VibDz_k89AHZ>o~GNH%1VG;~!|!-<*`+q~FkwuJu{t`V_YF(N>ctkHb6-
zUM5;|x%w%lW!pUQh8_lR`}zoG4_3T8Wf9RM>stoy))LK#7A4PjJ*FZH8MCKxI8Q@+
z&hjf}V@prG9(<+~ZKnWQ-&@Hu>W}*Ew(S?gBfHbEz&%5+Z-(mwxQZ`<+Yx4Mujp<B
zX>X#sfb*Lw_eYj2&V9np^D%5|zcEK~GQ02b6E_=qvsw0#Uxy>+m!=P)yHm4loJgm`
z(5w*jp9hLQ4YF8uXT1FB`_kyiJJIr-gckBghhgU(=x6sv1DUVIt}!?-HiaLv2ZZr3
z_O+3}<M4t-;FHhZ%z$>gbIq{$rV$3=^zc)=o_Iop)nk06Q2j;2Xx^fCAHz0`&)}Tm
zoZ)DYbjGQK>ks2$Th|sIG$@=vBLu%cHXP#LMvpbSiKATweII8rx0#lF1;HcAxvgEU
zLBO|>9~_<6hyUG5L)&qA+?=f=+es_onp>xNMmhrEBfmt1H!c{M{1UcR+1`9>5(6^;
ze{_NU(_;c)zC>+(je!aNP7+=$s|6K0Ee25v8sL%A>jA)^pq4)}N`mC^nYxk18x8h_
zSLEDRaVw>XZVQ5Xz720uJW-TGR`7|3&+PL8CYkVc|Cw##q+`oZItLH@*^^xa{VkVN
z0}v)8;{<oyE}y?NdjtcBGb5JcyNrIW(eXb>HZ$rXiDze^@tg6e-*x(<$J{38psd$&
zF+S9Xl6)V?6oLLb84{slF#nA>hyR7n%Q5{Vc;(>AONYAA3L@n2bme1%28(x?W!;?+
z)AFzV>oNHa<J~9X+dpwFW<#@EOlMI*>4y7kbjJxD@sN5vYGR%;u>2!qF@klK<;nV)
z!S!}S>8>nlyASftF^`{l$GWao|J73KgVU>0H)HWoss*rMQ2%N+iR`wsEgVbd`clEK
zQ>;JISRc12yOb{m8?g?#IzB#_??o6@+m9QxB(XT3JgvNN{alj4)*5?xH5m{DVq<E!
zg{bECWE7qjkAlrt<<Dl!kx)X$4~w-&$r{#E+gCd@PTW4===bGtjl|a)=XaM3jwoOC
z&rucovL<oq9eyBexchFoOu(&%+6R8=^4~0Lem&FZM_e$LE6EBs_mPZx&p9*oX2wwb
z(|f0VMizc=#f^}DHs`(G56vh6cJeBJn&q_#<@;Qc#F(oNd0`r<wl$r7M^vLVDkRVN
zj&$w$lklK;#;Ka6{Z};dhD!BV_Q;K6XC%OfP^J5R1GzH=F>-)JkgUKFX@Q*EJUpQ^
zd0JHKB))|`{melDer^5UMA0*wmav3g*E>FYU7;aAdacOT$C@IJ{o@f43w*5+tNeHQ
z@5lX5wfE2e`-eQQ*ez2~WBxke@4bUQ<^Sfjjx2d<Kd<qAZT0=0oX;d?M6Y`$xE@n}
z*%A8JZ}k)9$Kc7TTOJpZ%g5XO4m)%<=C%!LSF{D!2^_m7{S+!6BcVI}l@W3ZLuR$N
z3NJ`k<RcG6J6e)*lZI~9wmqo>WGX+4gKtEik`T{_ygq6dUD))fyy~xwXfD5Sta94M
zUk@=#I*J*J{5VroC!9E79HbN}{dF{f`ZK;kijXtMFb+ByVM)r{T&^$Fks;L0b?pk$
zihQa-y|Z?rsNo2nG*=lDQ9~L{q26E1da5H&V2B49MAAul7;4O$ZD}$<RBx=AoN|h2
zt#aRUaf+m`S+Q!`yPXs%-`^sYiRab&urGI;AlA$?lmn8^pX&Q?sYaQx`NktH%OqFE
z*ZECw+E8>a-i)VwV1MnueLuu0hkd<@U#8vNO*!H_Cjw15f1z*4%yUO&%6aZn&efyG
ze1qToO49~<=gi7gx>9TPdx2%MPB=2Mn32{df>7`kh%hjImu~DU%#|2;_Wo+r`$M)}
z34W$GjBeE(|DayrXL-Z35FtJA`aLJ={bgHdlAqm;yeCHw4So>6aQcI*?SW)Jr|bQN
z6wcc=lRC{0sLPCGzd!MaVN4=~K`wm*d50Ge@b7cFy$>^>mth(!r~+Wwu!<e6-f%mz
zfV<gJP#Yue3U86+dqnjt69%y)g@X$E?Cffe8mwo4hOL7y@SY1gg55^XZ8i@G&Okac
z^ZEw=7-ADr^V*zq9=J#HHY3Knb{y`Rs%Eb&&F$?GN!`5?m6n4uD7cvWjny?V_pDto
zR`)(W`(fOx9?hAeGXU+aA!p+{^E;~qr@$&Jlx&N#4Dg%r#1(k93~kMAjMbI3M`3b*
zM*fraYA6})WaCmEQK92~N~JHA)#TQtKUmN9DIn|9dgva?xn337Nmq~1)xB?RMwYYN
z>A^%Sd-FD7Epmzdiw+}2j6)h-EsvrnZq*_i)DmP~FCH)fo(@AkJGuwp)_LOSt8~`K
z@NU)Iq=PAeS`$HG=)G<v2GrDO1k(J`;U3O)EqwH0b|)tUZR2h740aM7GMsFT3q`N)
zSx{;pF3R?|Wd>q8#15eAHAoN0>1aJy8pHdgRW+=udu5zzc#jy4cHzo#BKL6c51z&o
zgt5pg(?FXTFp<-7=>pnRV9wQ<avxFH_AnG^C4J)fXp*y*1gU}+{##QHlAE2-vLv{n
z6dtkeysEQ^kxUumUoh?b+v7UAz;yjB99KBrP{m8m%%Ka4`7mByCMj==r1m~O{JMA<
zEQN>`0oL1u5W!jRK5J8Aq4r%<@lNno;sz|k<B0-|g^BOH2BLd{%mwk2RaOVIem+)v
zq3=h%uS<qNf+Yj)y4e<e_Q85!`%$J&MP)SBDLGNC&+p?#P!_Ximb2Q)-#x7(j^Sst
zaZhK*))i)8I-m1qE;snwen~!MBc<>}O|okYWgnb#IpQFV$T~ID-d*#vm4WXhEdVnf
z1kQPUg+|@!t0DY90p1^?Nkwn<aSyfQQB)HrYgUW2p-#k~n&ofF$Y_>vVZMG4-i5vc
zGEX@+^D>pA(-Q)g;FtQJ`)-7D<4CzhN{t6$7oHnN(5IPuFa*0lbu(qlWBYjZEVu0O
z^WtGn7OcN{2)UVGza8>*!m+RK0XP@S`hECP)*`wL+CS4-_I&<N>CAL@_Zb_lWdw`!
zjrDC(i+tiS*?kI<H|stiGlv(x)ox-xa%OyTx$l3`@tUIau-y)Y(JMPcx1jc~BX1-h
zzVWi*$K$uiZLwTG)D*%VrwL=fxJ5aBcOp)dr_@=i>Ytz;u={YjwcNi(%izy%;y@ng
zt05zwD-TtbKbr+TZ>#q6?0}MkZ^Dg$VzE8t&<dkJev?r@-StLW&rDB*&M!K~UD!D8
z8fLFrrP)z--ywT(>W|`_ZL;^zXz~Y0M=tu_vAVB(+4mR54S!JmtuQA!nurfa5jH4;
z)wP4MQDCW59e!J4C7*wOUnmfXt~qO|{Ud^fMzmNOX5ofzcQ9q4Jtgp3(wW389$2aN
z8ciuhe%2y{I3HTZCuN;d7~&Dd;*p<EF})8#`wf;HLDD!P)1ZV{@;CT$rqrM?YCox3
z_o!#p@NA~(*fS->J6r!qDfMRLhDM=k;V+OnU9KH_x81`i_9$3?bz(4W^`zwQjDLH^
z)#g<avhY95Yj1R5Ww(`Xq0;7deRnh#)4JkXJC<$o^J;=@**@FAgskaYS>3%FAeRa{
zc?N_;x802M?Pg76)at>C!zV)~ot^5b8h>Su5PrHRG=eo<SGy#jH?MbIjrDf|$yhR2
zh}O>(3tqVTT!!QUOetOoBdv<5V^m{r%UdSM6b^gEx+bR$`J0E6(kTP3wJs@N7N2+$
zO>4cL=3rM7e(K1*J*y`P9obe0Ru!&U6;5l*ea-f_A2qvMonvenjY(U5&xHR6T^ZFD
zb3$aH)5j6D#dd<@2#C^;>(<!(ga9j4?N9v@gPF2Rj^}_=Mxs+7kqS$z*1zl<oGvjA
zQXzLc$YZe#=@XJ5A^mU&<X|x2{nY_1=$O4;HmInqOw;FofBhOOLXJ?*y}nTg*2!%`
zz}4P3%0h!bfc)_(@>ej3*rdME4|b|gpNaj)UZNs7>xmdxIcL;SWW2OZGBf?69fT?I
zEef!%dsjLsc4mOKd+MT5$E2ijr@uMsxN5W%)Kj#<&tKfEMkJw#LfNq@dS$<8ZJY$_
zbW)a*F+X>2ET1aD-E+r0^GKnp65FWu;7+;1k}3rdnT>;$e$!VwCT}QK$c;(=NPdB#
z^c#V?dxAxW(IUn-_D)DZnTm+&x2`rQKqDE6`Uh&-;kid0Bbh$)k1Z=5EUEmq%|j<K
z)u?vAx}pt8>h#1!x6<yMJtUfE+#!}zx9!iP-X~RH1?1&ggDV{>5(J}VFmpCrb`;v*
z2JxXq+G64K5k-DHvTRZl0kMDlQRx8&gqYJxh=q-KBuZ1+_@KImGKg{~l>mgMvZ!2Y
zSE2W{uXY_nW~lwjzvGeWINxG}Q?0=uppJ_nBFpIU3B?Cnl@J!O$CL+dDBjb98+Xo3
zTvNw@a|atFW?eQ;R*!TMwkKyXn8IzO2d5;4jLQ_2m5+*%s=@HQXjZ=6Y&J5~>2Bx$
ze>1xF9@+{HY`}jX{-=k>2vPz5|1=#rp&mImhH^|GVrf<0xaR(*Z>^t$n2&qse!5hX
zR`l#4oED`lZ&gZj+eD}uRu*ln*~^y2G|6}EDr`7$aBVy3Ka`GfME3a%-{0Xy3<9XS
z5C*&yOiA@Q(=|@9nt`krLk&xZiWRbgWM5F_HQ%<gCr&DMF<sMhXu|c2xFRx=R#g&F
z+w<_E>!a4GeHUMA-u|_`>zK8yM!<sAYt8k^q^q<p>CGCyt33_tajHeL&RgG-NRu+e
z#@lOA$?n&+B7^kJ7t#H^1y=zF;J3RACxz?6n5<+0|6y-|BLHnWA28Zj4pD#If_TV(
z|F!Bt&&$Hju9(mRp@T)*mD*sT7<u^S!>jvtJ!}2ES1kQshs!zA)2DXQ_dFP+)vYJA
zHE!A(x$N3?Ew<bscVJiJ=I6GCEpl8#hYI&{NT<#5#%ZEv{Y_DCDawQL3)lFI+nuK+
zQx1|>VR3a^CyeXB|0M7P5z~AJ!48KOzt!Dp4Ze8*bI@X7FY;ydjME?va<N<j3(dV5
z_+)x#O$Xt|6V0oquQ;)<Ptgl^Jy_gT-8RzpX>Xv$;qo!+wdT)8({4|O_8zkJaAdxf
zz)0L$S-P$4$eh44!<3eH#N50t_0<7qTk>T$sr<qE-x;>7MIi9KDyhLqy6Myje`SZb
zH|k$*Dpef`ym?Yttp1ePoJ8HJXAhp-id&X@byL^n&u69b^sGPee;xNa88Rzp{CeUl
zZ-2Sj`uK3+O;vk6r;mOvF32|BL(j_c-(N5){s2$AbkoDbGhdxr>H6wJn662Qw#Chp
zQDViBT0_bwuimWg(RK%_8{O(e2YJrh{x7cH0xFKCixv&;?iM^~aF;-E9~^?ayGx*P
zcMZYaCAe#FcXtTx8uU%R|E>4#x@+|e%v5z9seN{x>h7sxg(yqV8E?5NH6o@GP&3$>
zDupbARY<O{DXrz_Xv!_+ixbNuOrs|e<781Ym~|+9P)AJ1Nui%>4wmT%y@9m|j+kL7
z-8ricDmc!>L7om%9X@sbOs3Y3QtzR9rLNzrbQfN9c-%W6(p^%^{G63+`&Fr*q5+Un
z?Wz-2b;L1nEJ;YAnCMq}%~NckJhBN?G<6^IV&51w+?zw6^S%-kug9~mqW+-voGe1L
z@?E)|q4`pUd==(NwY|7@ONE>Q^<H&}75`SXSVDJgNCm<MDl4L#Fp~dnk?C|UojiY-
z68lh9yRdH{!|FXP4==(U4$HBuPX<6hLE-V#>`9Ms@4>aBZ-XKww53mKh2auJO~kM+
zn+;QDCA6Z)eHH&H01F{sc;L@5p*mqI%RZT>SY&iaJ4A&&N1)Q7X-;HRh8Ky>*{jO*
zA^C$!%a;3&Z2zP3!zrNQQhIBW4x5T-XjN0_^i-2ey-8533jY_15ogDLTL$RYl+-MZ
zQ-f5aPnG51E>UIDFtC#=MHfL;zTOp0I?f%T;mVntjG*zvfS|#F`x5pk`Qzf0fiiE$
z#gYVjk{3;qs{Q;qW5S*WoD}^ZtsSX;@!a6}h|)}2<&L_93CwW=>ytka*p!Uq1+6m2
zP2&S}(2<aFh`G*}`A_h@q9_1qvhi|h1^zk0N6%j&Bsh5=ze--@2GEMdL>P-e$E^8p
zOYCbk(lP`AqS0Pmm>l#$tKCr}4cN4wt5QE3sVd?^NbEb%e%%%CZ<cL7B!TqI6Q6`w
z!(j9pWS*Xm#_50(vydq%;i;1=DB<}fpHqUG9ga0GE{1`Cf|cC&Eh8Gh#2M^@$HZ}3
zF%UcFCCd|GAq*-(PlqNe3souk3gI3QI<hZcMnpoN%1t=RK^acX$)!qLflBD~5Q04a
zAV9r?d+!j8lsqe*NQ6tTXjY>5)ddO>r`Ot0WM4Bb4vys9=DpHSak#4LyJ2}reD7Di
z%ABvkJfsGShDffYMyt<&Q4-E(1WMLxsNK;AyMYda@_^$%k{%hi6>#;={kxl}c8jB)
z;I(q}*7KQ;@t!Z?+wTXD#-3<Pyx&Y>A6ORQOpBp_^TALAu7WMt)(AjWQslvr+nMqx
z$d{s(mLi*<<!4$OH_=S<4IZhLMcjv7e`ndQ>l!0=c(m5RhADBtzG*yt!TR?XIby=e
z{^T{zm9u#dCXesU)2Q-2W0Yur5ysdiYsHNW;S47gCzaVc^~X-{YrI$=1W(2e%~<Te
z{JB9`JjrtKY3Z0PFf$QizPU556t3e{&N=W}<En<X$=`PE)elw+`W(|SAW(r-A#BHf
zf)!JvQN4qLfH*z{m<*0Qc-%~473=1PR{WgAJkrQ^SGFIiqhbd?MA)W&LR1>k@N&OX
znfx(()6GvGRUHy)i3^(k*XJmrQP}5Q@(hd<=<wynNPei4@$tFX^_v_a0~|pG`w6=N
z8zoO01A9S>CQW07yWi@@J2lki<V5_Fd_3w~9oT|Q>fk%8fFV;OvyXYj<luBet^BjS
zX$za-a+PZdt9tdU@vrcX2NYa(W;pNGxJnc2RiX}q$U3~7&v&co_@goAXgb<|QQ5oU
znjb%Z>>bCYendl^9B8XCiMjY{-(=qVr|n+*rQL&5ZtC9CE)U@{EaNn?OZAj_IK4JG
zwtAn;QGnIV1bDXW_Z+(3k#eq&J^0`#APo0yi*&<c$s@L#>^a5v%p*&Y>$y<w>%A7K
zj(jge<M^PHXIkJX`X)5d7SpB&5zp+#5S+$pCZ^;Ixe_t=x+5&lRew6iaDSI06U2w6
zrXb9Z4qoV&TU@f}S`I;ITRjMc2H}w&pK@R%g5n3x4gnGfG@6?L(e_#Vnmuw3lq6*X
ze7_VTM3|5xbI=%c1flsDG^R4N2QfDE8=WklT#KJW4y7D&U?H3tl$((od(!p7Em^q;
z1(I;&x!u~GAFRJ&Dz}L24}=I4$O2CyewI}f4X9qj75wRV7<<3Lke&hWrr9*;_aSvw
zLJk5LVqn9Mqb3hPa=9gilzu{jWc)T3#h!_Te0sYG7yOh6Awx>@$zBt|IU}doo>G(M
zH-rUgD&qu)=+T#M-7b_VW~}<2%Q|S`g~;MZDdrc6iPxsK=;Rbq9Tv=Mv074jdPTdN
zOJqR9&%9Syin4kNi&iyLVLIIV87Xt%BtzPV5ST9Y)z7><2!9j0Ns0FjsUqBq_dGHT
zTDcr+>9aG}pmQ;S$zQK4C*m{Tr;Hxb&pwHe_Rgy+DiF7{>^h+z!;Uz;*ql$luczgI
zL7hM7+2?1GD9RLnU2IFB3Y?->;U{zrxnu4^(Fo<Q1WavvR$2@9PgH_w^fm{cic*@I
zfzIvA`$gtD--f$LCVlTSgxu5$L5nk8>q%{|{S{+Tdp++bKExD^+L$bs+6ASim?77C
zCQ4r9n^<WwISfNcgxOKPzd@yR3Oa^bNQoPK)FBA$3dM%CC@PLWN){~z<gOhN;E}{M
zzby1qGYUEyB95wS&h;iVF@+hnL`7O^0#<z@#3@Iv7-+KJIX(n<qDO_!*TaWMa>xcu
zW6J$BxUBf2$dZyrhl5;c5CTJK%efukk2l_BKny$U$jTVF{MRrEHjxv*zGqE0dfR|r
zQk3t<at}Njmlw-+zCY>M`xY_uqM<8ez$!*&!d5MG=%;4>a&~N;!b2$-ZU!F1av&0y
zxcP7c=7705q{mTbePqi;B9()uq9+q&;`|s9PLj8Ex@Q=b>G3!R?m>|6q2~_GfXpBe
zhaP$l-CwL52f~c8gwih#O%9VxMwk(zyTapBcY2r`<RgM6t#CdL9n2QCqpF`8glT0C
zgFy}aFr)~Q-{-9+oCz2gydJ%7UjPCT)}$G--;ang3b};MZw$v77M>8p@<RzkHSA++
z-tq@K$_11uZnAA0e7Fo^LucV>2vxW!5?Xh^EsS_LzmJz*eiukT(YD^8PT&qnsr=Pg
z;GIOUi2|@Of+EO_D<~x6g!6<k_Y%pfppQ&(%mQq%#52kMLVvYLjtuyL4W56BQ``HD
z;)#E@r)uXb^K@SRl$XVQALXigE1~Y6U(~d&G!s#_k)ZxrQrBKJ-c|M9yW$zEp@Ww^
zwd8zUEin9P!wY9~{A}8r%IbT*Rb~E2Q&vvJg*I_eRnBeYfhdE8kFl&aT!TH)N6N1k
zcODA@qnPaa5p}<Cl;8E%VH*I^^OGwo$uqWrr?`^n)GB$Y_XX}DwV`(j#TpgJ#*f$4
zN&e$>Bp<7)6of~5{@C~^V{XLew`3&+SK42eCxy5#xBkIwPghZp3SY_)1dWK<qtW5K
z{*`LMACws`A$=!sHt^nP;p=X!Ci-O0#49Hw_4>;7XVa^Fl8*Y$2UG{Rx4mPlO`Uf|
zw{(%D{^F?L6f)ko2sDkbD7jFX$C650hKF^M8DHQoP#=*O>>SKn;5*R?aHk9!rxSej
zEqL*`<e|23Zh%k@v3j2TMcj167dzjIQbZR~(9zWbZSS)o-DGv4a*%wj+O>(@GqaL^
zOe~YOQc|^>RA#$-VOR(#ew`xtc6Tk^Q8qY|L!LhxI1Y>1V4i`#04Ve>HMWr1yIIe9
z*Nh|nW!cN0Pvtx)Z(--j9?zUN%SEx7+2$G{T`6*LRjNH|v|TRmeDh+{ShXZ5ky&t_
zbYa_U<gJn!_Px{WcTPrbR;pMX4tRZ_!dl3po7EZ)xwzs@JK6=j5^$PEhofc8DlCx`
z(G-D6PioPHzY5b9GAV0e=&o%0DVO9@jcYpQp(bNz4vxCi)^-YwRB6tXZwtd&>Be(X
zu<IED(zO<^dn#i3M5W4YNj6%Z@iEC!t?@dCCi4&Ny+=}J>k5*eHaY~Ibqmj2z)g%U
z9HgsOdA+j!O$Y9tyqzSqpJ{T-eVD0F=+oStiOGjMiTbB9k#*j)CK$eX@Drbh)A+P4
zRV73}KT{HT<_82MCT~9P(|Of*fck&ZFMG=V^vpo+xZ7!$N!g%fIyX6vZe>tttU7QN
zqBgdlGjy!Im-yvK<J>j)fUUY3b711|0HTR}*i4ku!vnTj9;=?y6Q^I>71TF%ila>2
z05qC5`;KP2BeO)m_=NqYzTq30q6ZJh+SfmiRtZ<!0T!7R7Z(R6Z_&-)XrAri#3GD7
zto~ZaTxA=6#l|YRFh<<6d$2z%#`o{aj>NIS?r~uajJ`C)<6lBuk!1~vX3@u!T_R>t
zU=8*HFaS{;?IkHYH`b8PN};hvC>MJSKg3x>qjz?g?x!~4`pJ&(oY#A7i|xqzsT%bv
z+}*veu_rgjb@0kdtb-#oGVCU~=uCV}U0I`|$FX!6^WQqxNhx%~j6?D8HokByht<^m
zF~=FsN>&jnmnsT~&2-cU|5bdM7~~w`6SV)M2S}dS_G7x)gIG#U#BWf7L0nd^D!8VU
zibON|a}RBZ4$&i-MaH^<7WZE@aEqe(M%hz6t^!>e@(e?&OfvGvs-&efgk4Nr_wZpX
zPmf*I#FK3+n(N;-=6{%RyA^(7rdO!jZtHc9Vyo<w)d;DNELO{Jq;L^G(wQr|$7#cy
z2lzl5r7_|?e0Z6sNu<cP34$hSr!PoqbCpbY>qy5diKIJgyxgqJ<CCi|7OU4w&lDsh
zvUewZc#ce;t37cJenRm5ZgGBdIjcr<WSYzt7%DngQ=>Y5wR5bJPZZ72PUu}>M+TSZ
z@_nwYIX?8Ft#G>TL?)MZ{iuf`jO=(0{0KFD_SVf<{p5IVz$2T{$8ofyqy5;%lbS+6
zRY~|=ub<ac9L3R20nu9%{i(i)>gSgu)=WOSsIsOkmV*(AeaCT~NM~8)B~cjt-z^lP
zwm*mF5342$zh&}E?Tci#eg8Grl4n`U%vMDkE8wnzj^KH&<Ce`-94M%{T_RxS1H@P{
zh-SAH-_6zr-0`N4NM<`@Xx{Q78hp!c`?lhx_amvRYSQu7Eiaerkz3y%3MrUtt_ys<
zkHx;Gvx(-vwA}0!nZhUfJ#24J!>V-iz5h(*?#wj(Q1ysh-;=S{oH0d0i9|uO8gkhS
zq;@W~MhP2nZceffQ82fL%$!J0m;#Ovn?#D+Y|CVdDd*Pt*7r|Yk6iAtrjAZwb7px@
zF*8v=&-0P|dRaJ96sVllWs;=IjA1)iI6_qpm^#w2Hfu&RMJT%QY$J68wJE#H$><)4
zq#U0Y9#*omwb}=rE%_At;G7IWoK<N!g=?3)%CL!4GSnkCFjb33*l=rAIRl=$dwfpt
zp*l3hh%M*$A#9HYIT9zX6Fr1XwpncbA(630?$xD6;f(y)ZQe?!D~j;|)9{2AZGvvk
zb3ONFwcW;>^WfHkHIv*P-E>~&L-<kNdY!~PJ#S5<n+@L_?*Lc18-A~7gDg8$Ph*3X
zyR)OlS$#HoS?1=V;@KZ`6M!KaYYl^;)~0EEdVI+x<qST##zrGYYsif|4wwFp2gW8H
z%9C-Gbv)`js}=6I#>ijb1sQ5}9VLF4@Joyj;iWf?I<c+CosqexoSlDbY@_`gXcPLz
zZ^GY!V#0qh8f>GXGf=`l;;ewrFaeQhr;!0=wr<CyzuMp511iK%s|N0^>iIqQE$Q{N
zE3)&}a7blZZqg-N+T!7r&X@ie)4F?k>$Vt|H%(+terwIg%q?uIovUzNtaBOUl1@2)
zQEGFc35akx5FCm-hxhp9asbWlG+_6Ienz)Z9q#nP;r*AaEX9w8Hd_VzD?2X>ccv_k
zB>4Dk@{Ms%PMb{At15u!J7{O!mbu^3*mV3r2~i#IqKXrbR{lxOC#{TiZgyaW^mu2z
z;}2Qx9d<dZOshx-Ls_-y9ct|2@%q{ESLF)@Yj5yd_U9PR1o(w{erg-jfbaLxYz(2y
zo#+_U26yehILh6rbWrcA<c+%uh;9F(CcgY12k>%r<Njmszr*DJDti$$OmP0C{`!E1
zYi;}f7irE){<GU~gBIF+98tgW^AWCiKKOc{Zm5IMdQ8#myp`{hCCPa4s1BUE<g6v<
zKyh)=t97d7`s#vSrQFF?af+oNS8Bh2NV58ClZNH`@??Bp-zxC=Sc#XAQ|oU7QqKAA
ze#HT=a@?$H{``Rt4-*ULA4t*|Bx2(sm@mZO>wSds#CR``oLZ|@$AEe`I0m|EL{3~n
z^)+)EHn;H>%`?w0Wk>3`%-<|hbUTKtyWY0X^jFPY7dC~CUc%?Um0N#zeMwx-(0ZhK
z(2YG_9T)RNd<WV;Mc((>J|1)^dsJ;_%V9`wlfJ_2X1n(?LA|$2E~4+8=$9n6Urs)^
zHIl)<n^1%Y0c;q9y+Ub^7Tehm7Aroq`NIhBCK11o-c8&<m~X>k%ZTqLkSoE!^_Rg+
zlR5HjQlYj>&?j7tM;a_khL^<fbL_X_pbBtxiv1t}*zo3hB70G};JpNO$U=#uyg!!1
zzt3?`*C_P~rRk1#3wi3h5tcoq1wF1k9N9h6*lRiT2rWZaUxMgIKZy=|U&6odp=^R}
zx~Rf-n2j>W-03&tdmBz~Vzwt`@}~T@ojvx&V=MbKAfs34uI<of7wgW(o8mS6&97JJ
zB9|V>1ZykB1-3u%gWn|x(k2i{cxc9ZlQt9Y6)OFp1-6zGzeniRGZ<jofA0~xyP_`O
zeo^tx-8mtA8&*#W2A-h6Eg%{P0Bm%aTiH=dh;PG5J#g<PA}ikvug!Soke}MywvHv(
zw*+nR?<YTld*Bk{WwtB+5+q^w;T<pug#=r$e)R+8-J|xyyNUfP)r(3xdF7iU`|WB7
z;8hORz<+Uz6pi8%<d%_r32HoMww~%Bc}eu1vg#AcIs<#CX>H>&)|~Y+c986`_{&8Q
z@Z!@W^l==lO#$rN6*BVM5nklmV}mFlfCFA|38^aL?Uq0C%><7O=M@-c`bqz!!km(@
zaR|z!{T)L1D!pMRc{?Zy^^{l|kMugszkQG=bmvgUwO0lCI&8xTe&bh^xGfgyJdEYX
zyrr@Kz&M?K%Gbn}%GO3B3SbgIOp#V|PGvtGQ{{0#&60k-9aKTT9kj*(pAvrv5-dM_
zy|?g`9L9aU(t&#gj?ieI*nL+IK|%~qyEuqXiPD#TZ+kwa6Dw^choJY)JfXBo;as6*
z7&CFeDjg;e^{3LOr`c?ciFOTvL(u&MTb|H)&{X!RkDM65=4zKKv}=<u<oV!RO$6(&
z<29WvhmUwmbD_18BeYC36ZZD_ECLAJeFcuav*z3YwnG5)={x?DiU#p(qY~5u!(r?J
z@gazd82y#Noal8}K?(D9I9;}d>nZW0{l&^5h=L3MDKYa4_3Gxy7qC6mhoDb8ginbb
ztFec%Z3MS8i40J$rLoMabB7?GDmU|K(pde&Sb2GHvu=l=*S6Nj0s^neJfW;LusVO`
z^Mts7>l~qvxl`G4VLZ1q4{P-jKtfBzAxLB|=&e1wK`z`8fe7I#5k4aDtw15>qfIhL
zPxhNw0JuNEFL{OeZ_b-ghXYcI<<P*(u%FOJm+ox2ILOPv6S7MhZ)4RS(n;QK(rWc?
zQf6ClNu|~n$W5kr3-n1D;W2T!2=OryC>=$4DRn-D2H29D{DofC1BA|hb(20wZ^fFk
z6EDB{1qg8!`U`zT>Lvy0cV~MYl=4}B(isU&W2@20=wObs6A#OJk8^h|B=9Vq_$j@6
zMX5Nof%c;#KmKDjF=<zZoGL^$6V`fn<Q?fwLmh9XPz3Xz#Jjqu9JsHr-|d8$flU7b
zC3hS)-nDqt4<zeR?d0wvGa+l@%QF)sKk&ueF9L@%CN4a1LtrbfyUDjl`XXM4&|3Sx
ztBYP6&`<d#>raK&wIK!y(!`!~IXO?J*<$s3-)9wJ2vyS%=*1M1xe;7K&oeLF_QR%R
zt-;IR&ZUm`aIz+#?^?1mr&V{v19C0Pp5;H|q&sh1d>Flb7b7Pv%N|fS8KL3gl^C(o
z<R>-7YN49ze-p#=G72xJr(1hc&_whzR6j48UT{qG9&pt$Vmcg)v*#k#6e$1NM6xpz
zoN>xnHtC6Mct|s*Gf(o&HFsG>HjrA8G5`?M(NAv;p>}IlT&AMfq*XeE0OMBTFc3ah
z6>iM(uFoiI1S&GJT#?PQ3~cfs#-9a40?@#Zhxp&Bqxf<7PN#ep?6DWtlyRTAY$!ga
zd^8nK(6hT?Tg^(kO8e!+D&KJf6Jo8$MN(3kfWny1&3shxiox>3^A}ffCQ|sX%v9Bd
z(^o>4xsH8kv-TSa*6ZgiB)|#6E2WG$sINgQ{W-3K8FoT^lSZTp4r(DrdO*M1_S=ny
zpwIHY;nXs@6T3rdI|l4T5-F+1Xf>+j*mzv}f*>=m<15<s&M&zo2!WJ>`L|(*j~)7S
zRr}k>YP?xVnN8ZqW?S{ObWyrm@4Cbrf!yKzrynEh<PMq`GOzp`>Vfd=I-&``hQtlD
z{a#Y3N1jLD$>w*O9{B*MHxmiCjXT(yICv!pejZXR+W2UG#PQ~$Q`KcyD0kusK^$b>
zAx#YWVL@ltK`C05pMmD~qWF#M+<C(c-MN$m7ZtpvuM;`BS^i$r$iZLyliSBpriL#N
zsghGY3cb`rd%vAZn*vMy2V96ZZIDTxO(ZOY9qV!9m5@D#8TeBqNS)<)47+B7c6ctB
zPd5&*$bU{ratOIjLC~r$UJOIinGj@Y;e=2ce@x5ih`1y&-z5B|`byXhsn9u^45<p}
zdYQaWs6SMYk_ve9u+q;W-!bH_XAH$+Je~>SGLD77gbWO8djSY^b(%yx=9Q8I(-m|z
zRv+<V#(%@@n#CeNa<4;kO~@c~4{rIF))`{7XS%}UX4h+WAF+Q;H+o!>zVva*FnnKw
z(e={{GZU^B1zk>?rRS6q%#c4)SYT`F9pWJi)uWhcxS+u@{T#~WMePV`!Tua9=YGV<
zrB~Rq6e5;whzF3>6J^2c(oFJSd*+Iw?eOQ|{I%$VHc-)F^jnxvN>_WoUUhJ7o_*3-
zz^$m%*6RGC=CN~UTVZeI8F@m4Y7e}!ooHxx9bPvqW}$lxuLLe^G~BlrYE%hw-TrAP
zuFf|gAU`=optL4&Jqgp1JYh@dV|o7p-HVO!DUPM1A`t-n__^~Nr>KGDs8Yq~2g|Mt
z;ipv=oX~Ku%z-{C(p0}1;h=qEU7g+;EWDx`n_%(JjGiX}_$i~!KSBz*7=Ih4>ET_~
z{Fa0w`vKSEp(!6};2~B=h5xyy00DN=3Khnf?x%9k6r!0X%Zxe}6TZjv3(|lg3khBy
z+mB3A2}>5>Ti{osvGsX;)FLO2g1|JAsLOdVG}!MP7lBu#I)Pg}6?Wnde0j2%#3@@7
z(B|f@!2$W$>m^$zFnBhuQvuJonQdE7FoP`3(d8-xIunO+u<o2Zbv^UCB->k`Bx(4X
zmwF)d*KfAKsni#qHwYwXiyDgH55yEP@HG@UQ7|q5H1aisG^7ZEAN5u$&IGgzR*ypW
z7YIZclX}QPzjTTy#1eYH3LIw`cwCHcA1olMVICXvT0WdpEFeuWll6Q+*2us&Fcyo4
zP=$?Z5t;B4rX)jo<mjIF)6-xfs<91fh4>3g*1ogzA(7Gwm5k5e_`@%a-y|PAe(yqL
z!PS542A%`<q3Yj-pCBILcnxrn0+P^!L&!v-<0+)@%eyi5VrgTbkEq}ozPf!APa|92
zBRhm{p$Yq1u8JL0K+Ap<XQHn!cZa@toJWri9J)}>M~!M@mBsJfKVdLgJ9u@^PL>av
zrz>=2bOL#LB<Z0~{k3-m`p-vJWlywrIN4**0N3@`YU<RctW4^mNGoBNWEH^$lA|4?
zR)W=<R;(;_k$G|E1JZIReWu#7^D{Y5zW4sy{;6j4Cjy?u1dg+=ldta8qSt2OhTWJ-
z7u(;xU79dlZc&6+RWG%`vnF|L?nt~@`01l86Q)$td2AQnejP0ALR3K?UB*Z`I~sNt
zKxK27Z|L`7v}qbw2h!)nvZglE5y7;RA90gvuGgLAa(}*4g6=llQiA<Je@uJ%muka2
zg!EH>jEck5f4GZm5b(4%z-8u6D2yh=(tEViBN^Qhe}stq(_(|L)b$0a>LJaP_X`4~
z;jd0jJk1g2PG+RnmNEr}fk7!A1e@+42Y}!Ma+LKlCX2VEMiiEPa^#VSZ!t>xXfl(@
zy^dIv0rq-Ii3yvoixM)uvZ}4!^k3U3U2L9y)wUs^FRqOxob`t3-L04)LbOlJq^b_W
zlQ&fUPa(q}gYQ{O8$XJQPbx+n3b7N$wKT4N9~R!MQhkx;|84L^!}(|F<iPzAWC?&a
zuBl2BQZVlEhmYtl4J<p;2cn~uug#e~v=9FFCI>sc+Q09bYmOb|XxU7gf~DsbP)zer
zuXSSD5=DQjFS?~_a-Pa%ZVc;I;f^Mfu)2*mrVk8E@i<;cC9wxbx~oKD!^IJkwuP!r
z*WP$|vU|0pRCxyNdNOOzAB-+nI%xsx;|`XuI1V1KD~_R{pZ;}g%qa}x97d9>nGoDB
zE)A-@M>IGQ4%^b4wSkkjrVsD&Aafq!owR8P=a?SI*`i;8qpA?pKmJBDG7XL+TCm&@
zx$ESR8-%g1f2GDVobgK<MK4-XPh;y6v?ZKwe#8H8;6yhs#(wM9O^S#?grW<~xj_;C
z3d3i}OeulRTvA<;I;LvIGw8<ykL?Mpr$#88i2S-joZ}J!tc2{j;QQUVkUjB)R>B8_
z2wQEdZ3u|xx(rXwwO<S_osJ+4`>ab%whGsV;^ntdsH)Xmjc)Av<b=pdZf@`W;UoeA
z26_ro%d_q95lq!Hb$XFMoV)<um!~LeP#by*1WT-#s<b$l-Yo`4qU<Bc(V0Fl%J-`u
zc-oo*^N)CItoTTX(>qaj|158BrZUPeJyIvYa1Rm(l<LoM;n;VT1kmG7Dnk~V&jf7O
z+4FrTMZ&s|9gbphU@F6sMfQSZK-Pk~!QlH{Y9B4pBRI`mLKNJ%y(tYm7_<mu8FGDP
z&gIW7bO5>FB4cy>ki|lJLWUwd8_;gq)TFc=;V=S^4S5>5IGHA1GLn3@jK&be>X~mq
zm5UZ$#SBBHSp@QIRKZ|T_^@<!<=-UIi`Y-GE6N&am=$EK?Mx$OH}r{LK%g2c(ge^4
zl0$F-w5VAplWW6#2_UPxNF`DQx&L+W*IWar_ftl#tnOEI-QUX8O1cH^$<oJs97HVG
z4ygp6+BRfCHLZB1C-`@+R;iu7a>|RG)SBPEH-gVYP?zqj1wq|lLqM23@#ttQBdCv#
zCF=SNBB*Ow{AdVBIG@YPGH|U1i#C@99gP)g_9o78(r8ixLF8sfLH9kxF*iE|UsOz6
zWpzDcq<K<Aip{+AzPSyGXdWAl^Q}ekDhDJ%S6lFpDuQZe`YS?+tl_Np#VNl@r~T}h
z_{28UI#*~_;S)XK9Dt*0E~Tr*sQHRDirxr8JD(C5nQp8>Wv*&!RD3{Dtf29$e@Em_
z_c&pH3ajEA;Dj`<BHVPPIHi@Ut*oeDF*xNhK>B>e74llt2Ng{>*+@$4l+fzZ7`IHt
z4pq)Zo@?aXs?3L5S~$k!YFo$ooCD)h$}owbUd)g(qVN{BzH$Iv-nr|?c)5py#^{=?
zcATKGWc%!s+`32QRx0}juLniz`iQ+k2NAm~M$khF{A!*pWcpjSo;pqGQhJi4LoIaZ
zB1l55sk<SvBjPw)**eF$$XR(IGx4*KsYCjEQ{YOqhZ{|@)YY@X;*O`{+I5D(y!&;W
z<WV$>SHL{`pz!%)RG&ts)z5{+4d2M@)rQZqx|T6xCg1Ae&QNGu=crrCWOc8*F2}>&
z@Rd_P0=crfPi>;d$}K9gx`T4aQbr`oiyZ^a^yY}TJxqw7aEeW9A%%xjJTsq^%wV0q
zOX-#s8<)zyBDOm5Nz0t>7x2iYw$f@^+135VJTN87v|0)~7Wm;uVSkzTJ2O|>hM!e)
z_Sb;WCVTNnm7P`TNfuTGDx9VEG3JL;<;8|inn1koodKZ@yQeHu{4O+uskmz?(*B9N
zj4ss_Z_lxtL~>!3xmoxyjx(xX@mvOz>-fB!?dlK1tc6A~p5>a~FNsGhF1IWs9aBGo
z;K<9vC3<x$ZQM#3#?v6@SUxJJ>glGoHZO7TOe&|oGg1AWI<=jM>ftY_(jES0G?%^r
zg9X@>jc6a|4b)&&F#d`+P+Rzp!fw^vdc{0P^4cqxRcbG#+q-`Gy}0b_^Z39tB<+cD
zAz52S3|7TR#%h4BISS;EaWI8;Rf4Z$1%mo-5|lH_z1>4~-D4T&fcfnLa#`JHM%O#h
zBhJwBpIXNvoxkRSnaoOM5wZD{DD-M%bxX{pW_}Do)I+*nI)%&?V$;j&(xV48x|Tq_
zC$JU$<<GnP@2H79*u)41AAKq>F|_~x@&8o-SF3qb7nhrClpg$Sy<-QKS9&#3P^I)V
z`wmCvy0!Peoo^;PqEVnYx!C{Tt#2wb<tINk*$zHT7W0gn#D{Nit~}w$Pj`n`$A1vl
zsDhdci~fW<<@;iF+l#>qj7+-LUA8nA{G~LHi7CM-ZMnZ|OZD-2e_CB7eFCaY(16aJ
z_}$S1hm7;&0Kq))!eDKS*?K~qulL15zJG^LR<k$89Mw%<$K1ZZ-JYoez!Rq;@OpGn
zUhnjNaJKorYpVDD*G#{w{N3l>Y}|0{WOF0r;zl3>DEuCz^X}xK;nUUe#d7FYONcoZ
zbhK}g)6Sa!Q&NSxOj31dbrPGP))0Fog)#W8y4LsUp5P@)Me2e_cQ_R7JxSKBlX#%?
z>uQ}ZjD!Zi`)h8g3J-ECK+YMaX+_&o?>ENVhdr~O-U8eEe&JQgUmkeoaJ9HJx}cMN
z7x|HqT=(0G$6px7(m?CuhH1{*<4eM)^&bRKmEGGKDcjuNR&xHrELO(UlrDxWP1z4D
zIUEeopqa`nR^Mc^49T=2&wu5*d+L^UT}$svVfbQS#7w9Ay%g#I5BOb)R$H|Q_ieE}
z(B}BgKQX_{*Ps-QkKZVHhFaZq>m}f10bFPuXycpdvDLg6+b*rBQjHRlyXn&-sNE`n
z+5fWdbmW?wSm@J3>O2c8OM)y~jT%zA2-`Px(P^MXU6b7y$MmaA|B|`c&X#4Y8Ve2f
z;%uAi$o#mWhvoJdF!=E_j<;qX)nri<ciTGEvr?(K;qzZR|B2Cn>zlLIznnU;q!>7H
zH_|D|j`{SFPM1ks9&13yI_UXEuQQ7OZ(lmIxvtn4c>UHh>qcb?pZAl(&V}Xd@0zL8
zGqTzefYL$DVrW<U2<zT!{_A`{rE2QE)WJy~TD3m#x)}TZcz&_z`v88_^6sSX!9E+X
zcp>S|lB}t&BQE6VUL4eL4YrTzYJI?5`O%cS@p%<K`~7cvb=Tj;_gDA;K@kt$#F#yz
z?Z`=?=biLT;QnrqY)mPYwQ6G(lW2Ezvh(S-PviZXHOz^;{U%g4sD>2EqG=}k4&!L=
zVa6Aj>p;FcP)6=bz4%n`O>ivbtgtrvK)KQp#Y)0@>HXJNtSxmo#QLaVO%}d++}FA8
z7)M7%hD6=iguAzG+%{s+ve2g^%pEJE8fsey;YPiaXj|TUKo@=Y`MVH#1G5u(6e}kJ
z>E9nQxJIMKd|L5mq_csQvqfhUy3L<3oC|4z_RC0z?Zw%d4~5nt-~B?^^tmCthNol~
z+A|iRJ0Ni2=|PIT2_Nk`B=6GjTBLDXFlPGYU!!=o>=Fej;$5g)DD5cjB^4=RkbcFH
zGU>hsQBtjgMzx595T2yFLAv=27<f|j^rcmXTo{W-L-sEjuz$&qj*E4y1;%69HSWIx
zzAL)EkKjgQ%d>fQ=0n_|3tXt@)k>G+A*o=5uS3<Un7%42OK;a~9T#zZR1a^)0v#Se
z`_43IiWayPf21PUmHx|;9yXcc>|(ntKHZ2q(Q0O@6z3A1$MhNU6O0rhVd<AqY{8vY
z3c;_xRZi$pBpo{~5*@+=WEW{)lK-dyLb<+8t&L`%(fj8XRH27lxTCgN5swJ-e>SP7
zTVT*>N%V$WTaDhiAiPo?(pdTiw4~G!Df<<Ckm7NcKjR__{uyn6P5GzM)=iSU&>EYQ
z$TqKg*}|#;nZ3&WF<;V-TsOaj0M_iv(3s?(yBZ`_ICpGA!nHyOScUR<8fbyMj)fv3
z<$wZcsq>ojFvo=-S{VqmW(2`;RZ746w(Nf{V@nyeG=1SC_#WF9%^oH;lUzSeh+&Qr
zalB1a>T4CUyteBkm}>U|eMlqlT41{%#bd2FHWM3H1umPu&@TH1wb}I5rmcJ!Nl2yj
zsfQpW&nqrKX2mQ&sfvV6O&eezV~Xg6pv@;cA`~BIVPT&nF_#{pM=;O-ORK>im6cZ4
z(fCo-cAOCW(}1=e+JVO?)^dW7MmYlTEmzD5mYGkYnA)HBJkBUe7qhesH6B!QvLRne
zmWL#zBIqsCqn%nBl{n5US73h?sj(xgF?mMmw0^m=XCM2<KK3c%r=c(KCBjAtJFQYw
zfwS-ST{{81wq+@S*^-IWh+k55hv(hO4YfwYYK3FjxBS7YFji~U6ndp|?dPt1@d68o
zQZszPo!W6i81v}T%Crvvr64Lneq0snf}G#QYK>*rZ(<lr8I2h(YwQfpqWkv7T{YWG
z@WWwejsiwe;9t6JsX7n=IAQeg4dX-$&DI*K8qejY`Q@SvDmgqzQFVuob_X8T$5hp8
z@odUymtx9)O4a`O;<}2=4J%dl=EFrFc;-oiL`K~Bo1}YYP~-m4iGm0`%PhZw6c+4m
z0vfE04ktePn+V4JByjVeJMzUFazpK!v%u^!{za>hP|Gt}-d*Gkl&O5_q_qC~c28Ac
zldqL|v$;+xTyj<bsHIMhuRFgDtqM~hQ{@Re#E>=cT^+y+im$8iC1f>KmaDuo#htC$
zlVrV=hAGmxWFSa()3irn>4a2cJ$*utrE$p<6v%8Xe%-qeEoUx2=F7b}htHalj;66x
zUJQRA(Ma1;`{bJpxF!sw1=}GC$0y-<98*QJBQ<QYCdIK~gRAP&9F4Cdxh8C^btuvt
zk$e}r3VHoabo{$5a4{V3K&%z*@FTTL?B#Fkk3PI!O{lq-RD$zYwM!+oCzxvU&%+!$
z!%1@xxCzoE?)!#B4~cqfAg`-6TxVnd`L$97j~U84UkX_O6<1ICtZ9C(K09N^*Ga}w
z`%#WS{R>!48s%LC^RhaBaT#T8O;4&jU!$StaB{dpsiC&8eP%>?173Qy-iIbStTF|j
zwwwnn1+~<55?;f**g03DrS8es-ws-ubCeB%$$YIuY#K;MUa=D!MrB^TiD3pCk&5J|
zS@&2IX$F94?3|}YRo#JaFE6+aRvFNcfHhXXniJiLRLyEpTmTKtaMTnQK>YbRydp@K
zdLjj0*0Ep!YndcI*ij4nz;~u?<;2Y&yYe@zg0x_vb<2hgjg)yKRsp-hQ4>CsiiAH^
z4b(x@CYKJrKw5%UHW^;(gG5zat?*1^Z10{eB@o<$wRbktC)^!V%f321mj=n8%P}hj
zh83S@lsc}l#K$}jtcZ1mnTfJQ7xS5W^cye^hcCRCGf>)`QtL3>^y3IY?G$-!inIlc
z6BTOh*A?58s$wP18rM#Z-Bl5#ZXO~ucWw!86PAS)0z?#JTCjM^3+Map>zEU^G?oXz
zC6zqEHipz3l(bfO8Bw&YuRjHysHtXtf&^~fqugO*VwbbWb+(ngfpp{??n3E#m_XEQ
zAWsM1-)aeR$|MWI8}QCZaGDDgu5(sur`UP$K?RI)V9Z_ZF{ibu#35O!tT?^RS!rK<
zIhUmsRMHHqrKdj(H)6ep$rXFBsA&U~e;_HG<QPcC6wpKpoMDMC(L|sdhRe3DWrk6x
z&a~2Q_+Z8=@|4DTWz&jG(;JNy2K!%1lM!lSv_(UT!35Q6t%~axOaDWxf_lMy&5XqH
zhjQ~K@MVQQs5E6<sh~6Lf&qlEIdBV4stiV-DUK72921iIzq#SCrrt|a1EtcY%`PnH
zaFEid>c0!({b~D@^eVGF!>;+Pvi%Z%`tj9qf06-FnPy>D)FV))qmO(Frw79gfwmOE
zq8XVb&JU7Kt1y~?H_V8iohe<naNIOQ)vm}t3;8+-3%#^sC0fa@$b+~4vuPG&#YnuC
zyzqrD3T&<eB!z2;8t*VNFopF_aG+wlc_@LJM=Y{!!Z}Z)AY9O>1ykL2N<$a5?H}iY
zvggu<n8p7fy|0=m<2~mMi+LxUx6(Ns-TSD=V<kgCe?}j8KGv3k>|O@eDPvH~gG^-v
zlCEZjd%piRII^g*W;{UW{)U?kJ)!o7!3t6<dIBp~;mw)BSAzg>q3#~ZAwVk!fz%(c
z4RNlxuF>}~WaNd%XoXedN^~LVzm@Y0Jm>86VM(VIt5_o$O2Fnl#+GlHaLz4;YbW{H
znFVF-%qDWXX$;>ESlW_re{uIh32mUcN?`+8)e?eH#3*8$n9Pl#E=4^n3d-`9sre}*
z@j6u@7N3;sU<6Qdg8Pk0t8LifXwYECU+#!-$koUlD@-j>ZI#p0>$;UDqhg3QiB$_f
zq5yf{eO@2P{JowztCBeHKrjfu<u6BvWRhO-;@a(oH&ohCgq(^CYotWR^9>>Y2L}Ch
zzBeij7rR^ZM_O`ITBP!MlRb73uw3E_8(B&;rI)4ypiz2?pA!MKC^Spp8r#rR8$UQe
z(f{46N~NVOCvm}THm!>6U+>RRExC<nm6YTwXzHd)VB|f5HgGh<yP_$?#Qm<#7AKrL
zaNwwW(vJo_4zh}pJbMum=X@8RcB}K-+k{VfLwJzDae>Z>5Yq%9lQgfygOI@2hxk5@
zYz%}sgH%QYGTZDvTsUH89Q$Fdo?Rd4M@rGQS*u<5oYB90*)lO5xHshrhob;Dk+D>u
z6h!+*39ocQjQ<~G3Uv?s(0CAJiUPBU9|}H=nI(-NC^jYKj#Q0y>q~}ke?==2jHem>
zsQ+Cx(O{KUff^n=6&x%91@~Isiw3~YEA8EX8Y9+t`_J+GiYKIDa9Uo8t=c5~r*fRj
z@<~?hwNq&4-`86cm;ZXy?&Zfn)FIN8mnukj;{7&g6uASVLj$yUwXHkVH}n<zU7V`P
zn)L>OraYY>w>V4yW?vAi19K|6igZ>Ofljnvm;Wb7+VgDMMC%Y3rT|$P8B0KbLB+8B
zTS9`45?h*}?39o8#r2N|-mL-QVfecMn?iuu9F6j7v!0U9h~jgQCyA1JfZbRl`OkAE
zH87R;R{2brXFy9wgT2QVL*NI;R>&CFev)(fcdav*qCqgdKQWFC#PbUwzfBsvw)R|j
zB3WtWg6(vgEwH$~9bPb&IJ;rd6Gb6*H4k3OpRCl;h&{hGaV5~P^%*KS$upvD0vX8E
z7+7v*U!`377YQ^O(81H|nluH-g`dU@vB~Kv=30D6X1>)=AP?MvhYL(vt6rP^^0|d`
zh6q`(tK^jV&-QUMmlr+Vtuw%?_&!$^2K*BQ7`5q!ur6K-7bU?YB)N6z(I(Q;nU=Ks
zY$_j1>$g)MrLfJlSfkvUrZ4rn26*8~flNlr%bHyfsPNlP#^49&s#^{ldgf?+u9PdZ
zM+7Xxqlv)WMcg27(<4?G?Z7hlLMITP<S8RVM9$u-KM7ViIC0Asn~a_;DTT$`=Nj<G
zR1TpZA>uBIC%D0sls}57HG3F|K0({p^^3C1rMZe>{cPx#zckn;ysjK_-KLOW5I7?}
zGureEkdEp4asiI=)gXe36c)I;P1GZ>CshJ6sb;|3u~mZHrXUraxCI6|?A||w&0v~U
zgM^eM#C2(gleed%k;MptY5|#5O7^@eB05vr&n>9=8R+0BZ;-FbrFRHLH0L`Ub_6L|
z#-F-gEFF&_$w(2%$p0?KxEq(E6D<vICTbA5h7wl};d5VS?7=C_bYk26@vWtpPFK*A
z)3V9>dtQ6Y3XSbW5?-oY5eYlP14rRl;$}xP)niV3rWK3!9k()g>%Sy0E-VxIdM5DL
zTVahDTA8wfk~~LbQjxix(EA7Lg$&a_5i~rV{gcxh?ok$~#Ak({$R|_aa)(_pqcoe2
zV^n@uMayJLTU_trl>b5LyBu6bm^~$u5^k=;%qLe>J}cO|WO>SlC@H!dZ04|YY5Cu0
zgSgt2?0>wcoak<VdeHCfHP_)F#U!S*?OGfX$~!x$BSNX)!^v}2y{BDI>WMl62y{1l
zNlXN0V(WOyBZ_*S(<2;C==<n0@>OHBBG~#YlAJAW`Re}PbEHNTr7dyJ)@eaWi%Q=t
zaWW+gO^{-KDu=AlES7v)?BSr)b^xigdbMs0<Xj8{nIxZsH&isdxugO|z)zhvrE+cw
ztGKxTs$(Z9WDdI+NBl=;5p4>YadUesMaNV0-sLn*U_RU%`4=w!9<M6+Bbn#m_=OZn
zK8ej1{}#4m%vT=nflm;|24DDTCuJdvwmhHs^q4$f{U-8@GdQ|OF7)ECoD{XRX=bAV
z;sHT<x|<-ZtGoeS!H!=?+tN~L#bECu=c{8z%>|PcA}6~u&z5qHEJqASQA#`hW6M(M
zwo3FY+UmS2ZqF_1NtRl4Q;ZRM+nO~n_vl||Sp{7Dk49wVFUv6+2*7B8Dby}$G7H}0
z1!IW*yfwJD^wZfyl&Oh2@CS@pEQOVVFUcv(U;&%|3BW0=D``p>R8kVh$*BKdXU;))
z5h|(pC$S}|@TG((mA45hQkl|LyD;8`v{;Hk>434Ri;N5&R6gcj8u<|gJ3*>$*+Oi4
zL5-e6YGa)L%c}O*BlA~;96|!^BA2$z2R{Bo8qgc*pWFHN%C+{3BaIyiJQ%ckL=2qs
zrQ|6kx@pUYs%vTsI)eR^*vOJs1p#J;occ2bFw7uw^JWnGqlv|M!<BKN`i6v)zS6p~
zPT{_QlBSiu9sMMfsNBOXK@gT9I9_v=7gsS@Js-E}Vm!u(o0(5k!-G1b+y58&wLpz^
zM3fdeqhhkY48h3RivRANTbk}vW%J!Fb{q|7Pm_7Ep_u#!Fzk4GHo-$bvIuoyU&MbL
zliUD=RuTMFhfgfg9#Qnz7Phiwol5`e!w>{xh5t_sdmGp8A>-*!3bS%O82Vk=f88ul
z+EB3GrtHJO3sZIt{szEyBT>Kda*lC&;F}c&!Gc+Wl|7HIBMCf2{WiNDc~ale{Kxsl
zGtvFg0{;NlVJg|uZMR200+qCgDZma+Bt7l~yfU`NEo$#<@qE962bdP!19hJqHNK`#
zb(Jn93e1Jz9=<GCNN~Hjzq7UHQFPRmNh|hPJ}GI;1^$9(KDh}qE&Wj?0!)R-h16%3
z;I=8T6;{kR2!s?xMij-5t_dY<-I*1q^<*^W?$mR}A-OH2z1dcCa<7~JqmT6I>K8F_
zlKteyL(4f>{N%yHq}<-B&wL>z_2Ot|n+fq*QebM9Zfp7iqLc)4MM=dA*7w;V(Im+d
zK*dw#kmVTkp!zWtOvG)eQ~hCyjRd-LokBz^^4c;QAPgE0W&T2M8#a>Ly*Qfc_L>+X
zg0P)Ht{rH5iIb*_$q*#)2m{6*qeK?s8n~aNk2uo*3rEw!SsTKJH7OW7Fu8)U;a2}&
z76kq;dS&O`pW-M&M&byNF_ut828*th(OCJ9RS=nB#}2ErB{)`slXB;RMcJK<bH|jx
z2?gY-LE^vU&!TH%sz4{lR^n?;vuaL~_yEh#NSS{m(;JDbY+kr(%kunF){BDf#)W>}
zuA>qloOV+G7acsvlyFL@Ic$17aZ36KifKAvz?}nn*LZb-BsONmTn>T<${kuVX*BH)
zWSn})YjQ|}-}dO}No-zgdN}g-E8^HxmHC!JlqPvG!DOtAhuT*nKdic%0v3z}r;Jw^
zSedQu;3JYh)09+|rBHWb0DjJUWrC)T`7Pq|AN997;QyD6dKRu$k~VteRm335L;_DN
zR$Yvs23v4wmDHDj{tIBR)vI(_5*+kk57f>Oa7V4s{>QDwGkFtHC5@j8k=2s$C>8SI
zl`wK5S{*E&7bhm~{)#!$iEal&m1cs^`0;rtlqP`<q-SY<QLL)Ney0it_Hl0{gmO88
zFN^kDVMHAmIZ8))0Fi_(|LZ??px&pBcy$=#$P3PoV@qnq|6g`j8tHF0cRwnqmI208
z#LUwM+Khmbf+%>0YpoF&w|p-85<ajx!FW(j;HgXCe^FTNa&X7a+i2_LO-~1oFzlrB
zcFR4o5%Q!~2hHe`0!A>%nTS_1!I#{<#T^0gWN*K+Z??oISydVAO1@-;PD>NA5wfJ#
zA<Zb_<fK9u@o()zEW4`lwnX4<&twTk=yieFl1jqT7jR4aZo0s*Rb-T}?{MSz?rI{~
zs1*uh!n*@@N=20sMWrM3jQ@O2T_{>ZnqKSp<yh`t_NiwQr=`s!XZGv`luR)cj;ScM
zzJ(~QhWg}4wdQ4jKRh8YuDVH*%twK(u9)!^ie?o17BquPutqDQx(8-m`1@I{1#|x$
z<q$-f@;46foQ1WF7H=tP?YpKiZ30_TBiRvLS4AXbKiJsMQbkImL8`)F>qcY_wppyS
zgVDgF(6K5XTCozKZKw?KO@=^5JMQ+ztnuvPGcCEUlAd-Pa(*M-LCF%>M$_bpzW+H;
zg%9m{nsc_Xix{T}tXO^jC<h`s4!VqXyIj6oU}-(LGEOp6zi-Y&MtFV@m{y$X4G0^E
z<O)r}5omkG7PX>2{V$Q2!DW8E%p6wFX-)#bojT1EHS^uTj5eXF#2_hVz;Lcn?c;w&
zc%(_3RVgNGfs<`!y)ho}Mvl1uGL32%4%GZkkT}ji(B`;sDEyag%mpiG`{R2>k^Y<N
zv`_u*JF&OV{_gR71_EBL6BnC=)WR&w`=4(_-tUnfYOB!z&-wiwMYY@5w}%sSW1j*G
z{nJoVy{@NmoRMmM(oW#{MKr%_QXp$_;|PX@)b@LH@Jz=lWwNC{bM==HJrdZK4!>8g
zHjtZbI$x_7k{;mo`gkAO^bWObAM>HkZ5JCEl8xiPTRe%C7Z^ZdUGkkjPne<hWQ{)=
zgE%aOriclFl&`N8>agL1>O<S)Lhqu-x{=x>@k9n;t#ypg>mO77%?{u1&#xzg8nvO!
z)gF6~*Zm*+yOe7rIjXySo}Xq<Lhm=*Pr9DxLBQj2-6FS8N7`=vAJZ7j-N)xyZuZxc
z>aR~H`tOYr4CerNJJ~^?bL^yR<Mf344%qDY=8S*y^w89!JL0Im;$))#?&zV>vZ<HN
zIb@*!ya~R}$(0AOKOSH40f$@_J4eJVlpkVFN*l=dYfiT^=ur@sYhyvf{`ZX=M2p*z
zyOxGVI=7n*99>Oz5hV$s?f%|vD)g<FP~cyD1IZLS-~-!x@zO3;ltt5LSoT<Qf3e$1
zf#KqsJd~jYY2!*h6u1r&z*335)%cb8#=#&Q>q^CJ<@b!)nfJk9H)iqKkhODsMhpBT
z%{C^@u8?9AkF&Yn&sNInv2|;$yYjkMlERw9FY;g5JU4H8vi51eX0@|3gNBrq`$rF6
zydAF&2Y^C_A$QQZ|6QHj5Rz<GoQ7UT9Jop?=!@hI)iYeoe_y4=U39_4SYA(pq*AXj
z>IBHm-wv8ZQ5j7|QtIm4!bi`lvo!@;jeWk0bha^_MZT^l8W@Bpy8kI0=D%JEs@Sc%
znR=_z(sU^o&xm{PcBD&Oyt98k$u9%ouz|>PmGkUu@CGA&M4s0Z?Hrzx|1_wUtCd|<
zix#~@i?(ZcV@$LGw`cRGP@3y{hbrCEbZ6-;+>HNiVz+3zQr&YW=nHR@rByFu!$AJ?
z;Ai=aEB)#&wDGrBZI&w*-?!&#dws0nw(CV9?N&i|pL;`~Vu$hd*RiZd_Wl>(W#9E9
zrSI#}yUP{)`Tkk_2>4D<wf@Tk8t{H^zxn-H2zcK=>2tT&|9<#L==WJN^KxHcq7Lf2
z&r8*ptF+zf^A7k|@VWt6pI0$DA8$yz-gdiQH)+V0DCeqm5@Vk5XVE)fc5ay8A6F;4
z-mt@-N}p285ALH#cFV!5>z|+jUC)ake;%SB_csU2uS+0^!{S;&-a(Pc!q)<4g5yG0
zLNBJx=&4E5%&k!8Q=LxIyzEi@Sqvz_)Qix=1GZ&Fd!^@=pbrl<K+CUVjEu7xI%hNY
zAa}H;JMWxz-@`{Q^a}}xyO61vZHXVn*X!t=dY4>h=({)Hg+%V}{vV#cGAfRyX&ZOf
z;0}wsI|K*>cXxMpSTwl11$TFM_uwwUf&>k&-*Ugt`F_kfJ9FB$tLmz&YNop@?WBNT
zT&&4`PaBXQ@mJIA`W^y%Y%y)^pX9eMB;-CySS~Utv>dAbus_j~TZ0|Cza(qkBs8S|
z<)+ab@Hy;*sMymt=8ZEQG8HsovRqR>uzlmn$#%0#rClpfJ&Yp-I^R8W?iO}7YP}2C
z3J5%kTv)^kywB4l;BC#c%Y}jGZs!2yCeRN`dXoQ^`%)KgjSc@@v#*zF8Ox1eKfl3h
zysvXs<ID=3v;wB?|5*~AwNWz!_w~=csp3iSP2w;fejmH}^Jn1q;cTtB!qdOXYsy)X
zG!z=3Kc~+`Zf<29`tCm+_Qjmr&AMZM`%?|p$Evk29^{lBY()iAQO_y#Bz8JL)5N$I
z^TyXFH8;G2+1bvD{ASxkVc|>EIE!5#MD(9oj{@91!V-R1v|>fy;s+9v*|D?uR2xct
z^<J3DhHQnZZk`Sc0ct5Y8nN5`g&w=dvpplvtdy_`gyPB^qAZG7Q&menCpPL&6{zyh
zHN6+QaTHhc=&D63yb!3p=eWi|pswFl)$t5w9-E<wt?0KF7D@?;kh1ddcn)l(#>hxG
zEfsXNE;>4LtTOV8eWa^nNS<dEWx`AS=v?Pz%g91=>|ylk5*}-HCBiwYNDA*Vo(l@r
zt4AarA=D*!5);xyvTvd)Y^-8bjXgU&s8JUJcktr8?C^9M8Pl13<yu8Rd38?ee0&Yp
z+IkVH!IAWkcv<6bk*5)*CEkg1stRnGr{xNX-|nWWl9H!*Fks_m1><VT$*2ZLdBVS5
z<c!-?&_K)GEmjqHD=(@l$bd>~6?q>e=zd>*eY`mK_tjGVcBzlin-DLSxi+s}Km516
z{HLpR1r0%&_k}j3K+qLXAjp<8RJHLNnbQ!M*Vq;-e_4t7vN&jGeC(y9;1X}=kcvMw
zOMMh9yKn_1@y>}@7@{2UVVl=<i3Rb_%gf@<391S!#<O(Uni7E(7BMF}ji^P~TM6=V
zWZ00{7~%}@*I3RUmER&W=~leklHD>hE40kyN<5{>Vzh|_AYfyGq?QQf<_ThZ;V3up
zF9s2x1tjEBipt9qsW{!#k`ODLBp&b^Tb5N2RKFpy;tnN^ztv_ZW;Q^R)dM7y7Z<HK
zINh|%O6+^!D#K^#=n@;4T=}Hs(1+l>#@OvX3!HFrx-DC<Kgnx1ev{i;($p-3=9HB4
z%8^#qzucw+b|)lr7O~q{pE7{DUW>krEtFtw#J6&-y<SXC<Ajl6hIu-U&z|!$Y8Rue
zR9V&p^t3b;lvOz?DsN%?btSooddUHsIWP<f7sRO*GJ;wp%{G3dHc;v>OwgrpN^6&_
z``~-#(94`og)QA{vn3lLshpQ<DoQNaE@vlN5-<Xzv&Y9|OH0<PI1_DsJaYfS6d!YK
zv}9fFBOX8dOTs6qcyzBHCAXX(<`Z|)3U70E<OOv|FS}QyJeg7?n`SH9!0IbXZ>dyT
z*{M{&=jQD)O~fma(aRvvDh^CjMZjW3u>tH*YwS{fKQ)>Bh)##-##Ap;;T5zqk;_n+
zk;w!)D0*v#!-lE?ByO88s73G4Wl=MZt7p9I`0I+X87=fIk;R<J-oJ)c0ca_$&ny|M
z$9T^Yi~hLF01Lzfy8W7~`42Ii@#s3*gKkZo*;E2=E{U<oREHENCbHa0sljxT+Va5Q
zsV-Nu=ep4xIN&We>`DAH{XHCxgPb!BZLNSsUo_Um@y2Kw8A7&meSfK+X=^713dz_w
z;!>Rfj+B(YVYSfi>|}T4CAJt_`1y5(rV;eJ;<562S5_)cJpa6t(A>>R2WNvGa;>;&
zDB<+Niza2JH46xhzhft;qGPtfpPY=aB25yNl}}0Qx8cuctd<@@gBt%v&sHZdeue~8
zWXo(1c@98eQ^h(s{9=k3q^$&zqUyEg&Y1Fq)j-M&Un~dOamA8MA|;J8=HM{qgfvN9
zwzP%Ma$~vJd5km0Lx;9FI5b|3nSe35fc_lHi<d=CRF+;QJ)4esqo<4|7s-^3ElK0h
zY~ws+CkAQ(YseJAEYkW#Q*NJVJpvC%sv;sQy;0E2T6xo87grHRIAWj**S$0hQG6_u
zJ20>}#1TJETB4FtFmB0c!5DKuRmxF^<_(llk1sIM2@tOb1Ud&{ZB<(kp2o73sqhQx
zrxehN&~_AGZ(qE`xFFIYD7&IKqfL;N70JwHE*Du&Nc}=-hoetyB8NRI;?4uMpho@K
z<}Aa4YA%2}#WFo>p1TQg#!qgU&xoOuXiO<Ome)d{EBfW56*@ql<DwX#2rkO&R@`X!
zk8Y>1n%N*jf%ltovMQEg&aHkbMRLIdDP%s<kBN1%eaW|!k@E+0TUJ2?BPC@x`{h5u
zNFfCe4GhiqaiY9rW!Pql>lK#34MKH!g>r@CNLNdpo=5!_jbzLU_5~lUe)|BQ_%KCc
z(UaxC0w35&jbh6U3U$=AxZX1Pj$75brM_t7-fVdoi9#rZ7Yuc6ZUU?=J4M${Q8@}P
zFI<7ik}%f70~VRq=SbNO{`|6gD4&XVC?DhC0lhjRvxUL*xi^+|;<_VXz2m^?VRQSF
zX#<;zU*a-~7ANg0W+xVo*;H^kRzpD%+S0tC!qS4_hI-vi`$`k_x%~Jt3*vpDt@jsp
z85GDD{FbJc3C{|9Xp1_%{%F5VsT_umEb3_V`{L0&$Jl%*YYF(|CDT0i3e1Ay6g5c;
zN@l2XKz7-Th#$EmoA?w^6kSXL3Y){~k8H@i{de5+_<FLvnz7gp%N%x@)1eD^_LlZs
zMg3sqq8#>BoBH|E7U$LZZ$Z_OuL56{|3qH6X)lTWt~$-Pv@$#VrS-R9b$-Bl7vPyH
zhiW1@ZsaMa&P$M2#b$!~)<0AbBJC!{Q@Ku&J^UUnaYD^XYbqfPB&*06So%;gO|f}t
zSvN|as93@?%0XYE#iJX`Gn?I0GUikz!UkDA-JiPBf#{N83o<39AsGx?q)(*+z*2o7
z!6L)z{=;K?zB-?owOCG#^oJ66q^!&%O7K)yNtvd_=t>ktj29uLX-VWy1NDlw_P=Dl
z@kpmT(}V5Hq%>f8xRi8G1@0j4QuOcjp6a7$PQ{unY&^|UaJT1Y)iUA+ys+>*h|mGa
zIL7KH#cpfR3KkzWbqT+n52ZhmiG0D7=iZlNgiSx+w30<6)YiS0F6QPloBkz6n5Z8Z
z=AbM^=n>lfy4k$~1S&I<oWBWWHre>pKc;Tds_e)dgl+=oy5mWCzaEpq<{!&ZmEk0$
zju0nQ{yQ2!Fr!|g+zU!mKZh|7kOgP;l4i3Ho@e_7Y6#)Y?WghFYCuDAs9L9}l_41n
zb`T>#M>{?Zid}@-JI>OlYoZu0l9Jg+_*Za*1iXsh-f-z3{Kd3W5=DBMygJlSR5rB8
zZ5DE_YFnU{T%_bu<lZ_lb1GyZ$R<H)F`9AEwz&-@pq>Vp(o)o`=^10X={303PAijr
z$T(yWl&0w!8NVq7oseR7OEn4VpU40E-6TE&lbXdgAVyC2)wwfXbdqQQWEdNIElmy1
z$B7B*D6qw$p_+={8WbZl=WOpo8t_$d<Y$EdKrsN_Fy{TJt(9C0^<QtTWWVx5rLo9F
zG1VwpAQeGs8`Odrb;n);d2$;)W4kGJNti4#YdKfOeWJ9HNeA~Y<SDcBU8RAdOlxiJ
zwRHEBkOnuUdzu?5GWrhO#52qe=P|oD+%v(T6EaLa^|6ZTl49rxeijPT$Z&WaNA>mZ
zK!OGSBWKksfYD7N;l*31LK&&}QmNPZts&w}j1^guglvW0sM&apI4gx|!9#e}x|_Ck
zd>gL$Shy8vVwI@uN8EEjld-uy&O^Jx7$Wf?hR||mYbCme0`g-5#sls-Dej3wTdd-e
zvWvIB^5Tc|tj25DS(v9sGSrh<%H4rEsmrD|5kY!k%HEoAK0X?Wl=hd9ej@dNU<}HO
zFCbU;Rq^R+{Uo7E5Xxl@#}ia%nOK9;W_GA4ew>_oX?=hFXA|3lJR%iSX*39!5qayI
zN27&SUHy=46*)k{Mz$<gvr+htSXmu1KOlB2xHrU=mlB&eMD(B2!UE=lMq~@QJ}-88
zDDRa8ubGTRz7}xwiK0P?wjvr1mynSXc*zZ-moqBb!&IP32@+5{_rFp!F7YVttn|}Y
zpa(wxiP7XQnM$s_R2+3D_~O9OFWTX02UlMXNR!1C!6%H^y|Z!^MT6(J11%|i(EGc(
z!5L)IzxvE^S^{(UAVA13(~RpCd@cdf;lYskez9X4D7m?VI__0AR}+{qEHum(+>E*S
zw-7e+<YD!ql#zS#A6_00jR2#DL|0-Eth2#QSk15L?C3w%7DJ!N0w6)oT%pQR=k&wt
z@ATfMYXTIPF6fL$x^fn|P!m(hN}qk$T6OWsvI3DoV#TCiOFaTvAYqWQ-FFby*gDi>
zM<|ah<<1uRkfCg(F6V6{zXxb5!GxOL%jf`drY*;SY`hupI@v88`qDqsjj~ZDI0{2P
z@0-YT{LWAn^)GcWHTh;Er+6TnX7H%oM(ziI8Pd(<USZYW%+tY%L+e0v25~+E1&Dwa
zslWPPAK7S&kr-tvHrAF_5_~=m>&y+7Rtz~768?nb)h?0`e1a{E)lH@n9fU{<C8X@-
z>Xet=q+zh;mshcsOF2^TDvF17qb<cLu@ysitNN}QSfhA(ZZm*#6;Q9W+6i6ELAKqS
zn6qNhy5Byg;WMNgaOGtAh0;jD!Eu0s>rTz_|Bm%794oNweBK`$+U#MnLCBG0&3q&J
zSx|U}(=1bIAmmd-^+-=GakWzk@%g5@*moC)QlIoPnL*t2*D2f|C5TW)aUJ~HDgEVS
znxwQ0)yep{PbrRE*DICaZDp(r^*F;z_<TK#FSLfVw)(4|i@oLqg!#|gs4|KR$)x5_
zPX04ARvN%k#rsl73CTRQ?fhn_tu&B)#OzeA2=Z1a#Z@iOW$RJe_}o~sGN_2xWoWj{
zd%+AL1se7O%}2qhl8JLjI}1s)k+Se0*^zpXmtj*WCJsj+Y1{n~n8-_(gwCxSwkoc-
zJ1L2-iw{S*23B;*f=_&p1%X620hO&2+hFEhNCR84b=VF3M!vl)YzCq6z%nh@SPop|
z+ZX8<beTgod!P3SGu;J=wrGAchf@FlT(xcEUNCt`fky~~Rs&Po#($PW?Zn=qwe_E|
z9Z>Sh!tlZ!@+4D-v9~zbS#`1T(;_nr<pIT1zo*WwCN8nO#k6Zrc!7-ko-r3Qc@VLY
zTU$6CWL{=~oK_C~jABqbb5d7T>#e>c_tgXoqSRsObxSrSIzX@<4rW?jmKAZ1s0-QQ
z3i?LQoGXoQqh0>UDePMyd|V#@;nocQku}1f|L%PITk04Yt6ESdz!2eqW$|#ujd(xC
zc`fzY8D#Pnb62gb?3$Ymbyns6b1wzjW1!il`<rgW>HQC2#Mdkd4aAq`Gf3|)e(G;3
z;8zL7P}h0c@5h{&P)yv4MnOQMa|GC#-6u1cV*WSCF1kh=?f5mQ1(1KjwCskODE5+Z
zZwy=`NGAbtRbV=0=h|WTC+Y<bxU+LvHoN@;`O9i6#uy)88uJiYbXIJ);fJARj#!Y-
zd8y}G8o@zosn7V`guKQ^jlY$SpWLuV0D4m06g;1yRHJW8Oclx_ZEg^(Ecxm2R3~*<
zPG`}YIqU=SlF@N6tjLrSXWr&GC>8QGlSQbifh29k*t~?Y_QF5k=foHohUUd6CpW$<
z`b7WdFh8qT999>HrX1urbLxQXB%CwR^#Zj&pBv`#6J)EXb96cYVYc-#F)E;FrZn_h
zYAf%U7$7Pi>x8h%+60kHrgG-Z$WO5b#R@dCub`O=_WpAT`N_?!y=<(2=cCk&E%i<l
zaEST1Eh!8RE}zU(oVIeN|L1yzi{nw9X5y_{uoM|X2ku^7j2-gLpT}@`5Ly71Zg2=&
zy6mjXmUwxXtNG$0_a0aOvn3=&K8KSKHBEypPc~=}^3*ps)tw5koIkDt7NRkzK@*s5
zZDM0p{6yO1=L__O*TcT+S*z1d@k(B#eZyEdECq_;A<JkHp=?`L4&E`yR1O9^`&Tr@
z<82{cuV%K)%O@2u`dNpOtxcOtm`}lL|A|iItKXVJGWqyFt(WMw=0>!eC7CUanlHJ|
zPZCp4#!Ae?(mQ!RsMW>MZ;o$nB4vItRzA)tGU5b2-vG)nH>S<`%5WM|mnlu^5>314
zPPg)n+%|7OsH6zYpoLn(QFo9TKsddl2GfU&OBAEjStZ08$<9+`2MJh5)AHqUolHs%
z24?(}lb+2){uiG`@?yDmw>JJmrM?5h|ILDTs3QW~)xfO<gsew0OHL3HWw3e@V_X_D
z3^*Z~sFCO#Ab8%?@D|id9|JUNHNXf#k<H$U0g|>PQ1*5r%@zin0Elu^*_t8+uc~xt
zCTHTs%ZhBu2I3N9EgY$7SRh^Z#%p~CGy6$BF6Q2CKJbuQkP`ulqSZjr!@tqA>)3XD
zJ&<`XmD(TLkV)e-M4XT%Ny%1(GEk0-NwPwk7HD`4;IOgl`(FnbgW|xDk{_&P57h#B
z>dB&37~ImJV+Da$Al4O)*^*^vwOPgjVkICZ^Eq=Sh#7bef4`(xzNJcel-PzWE&r#1
zc~x{_jn>>YRt#aZ#bKepMwb}YrbqI6B0fO=3?lU0TnAs;f_)qL<}?1lGUdW+dUp3p
z7R?zb?}G7yJ48<giBYRL3nZ;Zc(V{tec*2Nxw(zv4z0gPH6V~L)PMKvoEjB?W2Ttt
z%Pg6A`N`LwWpycS)(QWM!`JETmCnuc4Lk7)b?w>&WDu?seJ=@6gfu@RC%Uqm?r*(`
z60Cxz1uIqwRaPd7J5xTN4SyPc_4q>~)zfw>7tG;l@*yY`M_2d#10+Sw7x*y|Y0Yq2
zYZ|K3kfro#HEBmylu!Q1r=23GpJ4x=hFYe!ZBzU+jbuwoI^f4M1w{!HIy-@mbhV!$
z4^lD5_YjrldQc1Xax&aURE(yQ(qt7ej7|CwzE<~gMKnn%84YwhI?REN&iv*(Bv+N)
zAgL!k1J87Rt9}ly3Ps9<GGnVX&)4AlL#0`YR4X4rGI?WRy;LG91^##@*MHD1R>v8G
z8i_`y46rrPv0BPpo)2=p@G<IPX1EOf1>yi=CsVK>mjF=Zt74GW);_t8f!SEKmw`oe
zDnlE;k%Xz{QYAK144S$Cy_*8yP$|*?lh;5jX@+Rw4qCnUS=P(dqCqlgX0GDCEfuHe
zUmj9jd;miKJ&(|s2tYial78zhv4BQ(F%JYpc03TNl#n0<4o)vJI3*8OU!aX+14!^m
zMcH{E{?mrs|I8XtMs45O7ttgy<w0f0ft-~@t(6W(#@0Md&jD;nuq+BIupFoR7*s0u
zWCkLg@SM(0F>)#vMdx<PGe7C9l7|2@3WK_|I<b`_afa|S8X4!e{nH660f&u_)$(p<
zC2Ec-3z>23c1n4+b4OfGqZ*$N`bB6;mOQNZEwR=$)$~5Ly+a0MBhp;`tYyN|!XVuj
zDeo1-A!>QgxE`qh(j+BhP-c-59gt#mALuBAvV7Ev5{r33nQ=aC&uj4t>OZ@&i)A3v
zg}CQQGOt06=DjaIeR4Ods-B1KP*6e&17J(LIsff+C*4WcGucWKUPYAzvFo|g3dH06
z#H@_SQgMQ+f#zNHlXo;tM3IspjK;LJDz*acaE8o~55qL^-nOTn7LEW;40HwI^iOfx
zX#lIITF*)o=??S6_w||{MG(k`WU8lA^6%X7)zPTLXVm5VmzMgfn0Ob55(%Pw?V`=^
z%}vxh;t2;Yc};a_Q3wfDz?JV4M|p0C4ImUc@HqqTf9OX1H_w!9CE-yLUJbNW9D|E&
z@!|qd?r5x+Cw7%U1EqyPRF^VOtnO|$dL@5a^dB9FfEBMp>ELfb(SR@uhSex*SBA#u
zFw3weR_T-X#$THJx;q@o2TkfBMc?G3xAX?)#JJ+B^4O=0c$4SFi6~mIw$oAAc>9v;
z-W<R0t2swnfq0R;jsv`qVE4;d40Aa2Z~>Bx7F8S&UI#$~)Rn^oi#;01H&BHde*X;$
z1Nzg+lV{Akz-Oxoh8RLXcBPcECjD02p03Jo@ZE&hSIu-KrDMhGGw{x-Ev_f*+5Y|j
z%2F-BA6^Ms>ic0?OL~-THYMr0s-OF^e(#mx2GpJmBWfurhnv0?1`~!j6C%`>8PuKw
zeLfk(P0-ZOV2BU|I69GxX_)FGd62<qwR5;&Q9k{byrN#BUY+-pc=IOO9n#g{1?lIw
zG7rohImow9<i-0MoIp_Vj&j6D=!fI!AY6Xx5>jzjOrgAoXre)_co)n0>Da+sPC4kW
z;k&*aepcdm(kD#5q(29u1~Wg2;*Hux;jk5gGBwIy5L>V^ukD9uRFX3&Uw#vsj!uV&
zO@qWq=_Or3S!=~G4r_LK;{8b$<8?1fzf9N%FE@`CfC6RI)+W+%*#DwiqZcv~Q;>4s
z%iF4L&IeKUviOuP1%txPq!h|0H?CUIN`*uzkW#9Fg7>Zhq5oWfbOYx+XykK$%2b*;
zTmEY*R2GL-LORhvWf%OAvrL62#ZS@UM?)q0e{}<+ORWPz5s-`!?OU7OoS1nsR?vAO
zy`Rj@TV~WJ^9RY|1^-~iDVS1`zC)C1p^M*mZv313cmy7q%2(NW=BEVF%2Htb=Qyfu
zt$x?<b_94E%*}i)3>p6(2W|Rk7h=$lnsqRD#6_@+k!*lNbaFh~LDTd~D^R5*)3q*r
zGHW04c@&94kSh5p!02!}fPewXm~P$#91r9<4iqO$MA_zR;ZdfxVQ&q!=W@o4>C5mg
z1#^OzQ$|4g=Na+^_^AtU2s*MLZoJgXBInRsUs)RmYJNfeF6k~zMSgbZ79>|e^@eo)
z%Ws_BwB}(2Ap6jkO&Z2oAXtcgVfzefiq@5q<OV9<`~C`2pq5}$$!d9r92=0j{K{Nl
zn8>ShbNerU30UgLtTNRjO7F%UthwL;+0tqUkZ)n8Pb!>%U91Eyoi6#RZLarjByUMz
zv=C}SCXT=)K2#A$)F|sL6^EFcX=+=<P~(Ll4ujg(1r&1Yb+dwIUZn(&N><c+^Vndi
z<Z}2wGD+^csd=KRP&tA%YL~?}4KNzuT1iQ*s>ED~?d1C}2&=9%?}}Ltk<ABcCX42e
z#-s=pON9rX{I574=mzAU=ln#fk5LX|u9W+E0kIBX2%MmNjE-?@6KIs2Nheu_{FpHy
zOTgfVX&2~`h5ICvwh`U1&~Vr+$yb&QsM^S~0y23Jz8vkacyMND$)?TS`ujG|HS7OL
z%rlw{H21)QmB_TJ7MJ*7MHF%PIDwVC9%ggr{7=f=a685h18NLSHG$Y`EDtW9JdE$!
z>Jnr1)~*IQ62yCNqY_*}^P@`KDn#;>&|4g;mqg^kc>_wsC^L4^@|+};PCiGpb1;6o
z6QLWZ@qzCw8VX3D*;^Xem>)1f>W_|(0Qw(*>v*<u_(!*IHwm!*i3VkYl6Os}E)9fx
z{O*6BYY4ynuAJ51dA)7<QQM3BtYOdrS`zv=IqANisRgV7wsNl*J)YL;yj;DV;U1;c
zqV9D2UFTEu+E~6g&a$@YL%H}?UtrJP8y#%_+}wWBE0)^8im4&)KU1hB+4ldqhyD<%
z%rkGzLjx~Cf#hJ}{olgX?#%CTgxn|rmv8i^Z9ZQlE=QLXDEncy&si}@S5@4T`o6$^
z^Zoq9Au<u*HtX3@Is*HUu4fxzny{y#oN=BtZA;R!gZl9UXyEreyMengxi3mzeMc=E
z$J`t2ptAZd@cOtT=kURweopZC@Ef#Q`pZf`R`IlXqe{O$Nh^?!&+dA5tA$7btE-_t
zB1ufC9u69#SPY&nEMHL~zb}ak+G;RL$Ni>M!6N(hcEKU6c3U%i?0oiiDfN2)hmbUv
zI&V4~mD-;Bi_;6N@5}7@{owCA#+bKslC(14JQ4!z^|-a><P_*hR3TGr*QCqLzfBs@
zL32cA*BXw<_J+Wmksk$B+rC8^g&JB8z*pdl|4WQ(Nk$j_MB21uinEHhGEQ3_L$pSA
z1Mw@Iv>6uTc6g`w@)A88TGuamysmsw*kn6J_`?vc;J=N|smdY?cSUMW9NsKg(2sGj
zT9x4|)e#c`*_iT>4U?8t)8Jv?yqVsxeXty`Nwt*our|O;6nQCZ3_WI^o<@lbU6o(?
zgwZmgreQbNCu)P>doMk_j(Tw~v7i%gIOb?ccu#og<O(Oni+R^oI|NjZGD4}u4}<pb
zq=o6gl?Q{h5+<-0cC7d9qr2yhbm4OdATS3?&7vI(x1dc7F$^)}O@;`Fn7<5CDMWC`
zB-p#QTOa|;UajKkbRdHdksDF$X=88hmp}`V14gmWHVQ+N-wqerIFJ5TXSh7<gX>*o
zvh^F=HNE5*B@C=zv$zc{eewU%{JUF1BH1p{d1OiU*aEwmkh!qg!$b%`>r;~IzvbUE
zD_I&d;T0W$*uiKbuM5*m%*Csz9@&JazKIqZ(*^;=HKK5hR}6_E1s%;5gLM@R-pevG
zeM!Y+vnju~?px<JeI}$?G5H&^NHtNY&yOQ**z2)-$Sh_<S~s^}GQ1Jb3c2#ks?Eu*
zSld63U~Atp%ahfX)7R}UAZosE!m|fZPG5B|99ZsHvYCO<nVB140Z=8c*ei;!3q}Fy
zSrBmP*i)lyuV}X(8+1<-8d8Jw`I!C7ZH?lvbiYXkD(ES+vo8R&R@68uRI4QA3O=Ni
z_RdMJXKq42(z6d4j1c?!Hpr9H4rWiSo#N-6c1ED5UN+h+*2MlS>Gm6Z$9HG8)Be6X
zN4Le`7yBs@!Lg)B_h@g=cDTi0;Siuc@?9R)JjO3lm~bYaK<`H%X2n^2q#+T(#}1xA
zZ&FM?uDyD^algEu&79B3cPz3Ef`L`%BHe;}BY}GY$9cnEr17;@b!GYD3}Y)CnGC9#
zgL1{=VrD<RG|5$KDk@=jQTf{a>0aebbm4&uUo-}GI7YCqEYRN>T)znJUfu(@RxGYA
z`^9>YKCFe}<l%AxPAmV2Rl)t@%ZGmjo5=a~sO=#4<9Sa1*hG+um3T~|#rG_E(&F^H
z@whbw2LyfnkOadyYw34S;U99bB+Q+*<^)6>+$ZeG{t4lc-$mO;M@IK{PQ@R5A|~-$
zliIj=m!;NgNbs~aJdv$7eCv&YuT)MK{>)BV`|cPk7IjV<ul!C;a$R3q@O${EuTeD$
zA0@74p<)D^e&<T3a{0mFB17+y<Iq6Ic~}GDd1(9Cc0(W{l);y$dEC!6v#}sr$<^Vy
zvpAtECh`@dQ~|2J9YKyR4jt@2-&wXo5l?1@Lb{`I-rXSM8VeSPCUY);vIrB%9cO#8
z9*UmVF0KzE@8{n<(&<>;I>cE|<UL%~Y!Bwl>Jy|Ih+M;*w^l2=#GHv;!`P~MWId>%
z{wCyHC{kydcCdqQ$_HO4&RG2-I%1t8Ss<FYRh&i)hVMK1&3jyPC~8uX3LAXUA&HqY
z!M^1d!-Sf0<fw9TqnKd<#7RX;$vK%3F=<gK>hy{}JGLj@=j+HCZOgs@B|Fb#Wi%yt
zCbCv)x4S+rmaefbiE9p`u3|AeJoscENxAsy0=oK8Fc}I}&s|5-tZ*m#DyTdlm6$|h
zHB{jWiwg;7vD5ovF$oGQ!yxTsHT9b@UBvDub6Aqhp@!9ys(*z6p%=dA79MWBa^;qn
zg8#AgrQFULX@B~kiDjY2b1FR>I<8?0eOb516S7?Y-Olj;&GQ+F>mtIVG`1xIcTjy$
zW0Z2c-#fDxZVNkx{%aWj`3bL)p;E8Bsw$NiFD>_2jSNP`sV7mz(3Ow4+K*I&^5Y}p
zf_b2NM#n0XfSQB&XgM89W{@I`sES;+WyQ_HAdB0_)^<X%%!tZpta5;?Yj{7P*sWSf
zvRoN0k53xgnue&zxx7<X7H0TrDN^Z52P31vahr{_!ZCwg{IEn99)c|#EDoNIy^d8R
z&>2m^HqZ9OW$x$WP-{0U;?6y|!nt>faz5}G`*^XY5rDNiUD8*em|pT12QD-xpEl#3
zt8ZzF1KEj)O-@`_N_GKF!!``>s~==yZQt#4*u&_<&b!;hcT}7+i8j@Lt=uTS$RYu7
z43_GS(BaOlkaWQ^@e*ZS40+H)fG<UhuoqlP@F|X|Y#7r%^;5@eITZyAtxeH0cM`z(
z<@wzXf#TsMs)_AtH$TO*eLNv(`+s=4Q?uHS-@a-K95c;=YYjBRBuRXc8R1$A>*G?b
z2#dEX&^@^dTIDiG)2yoJi#Y6AK9BfnbG1>`C>J#xCA?sc>*pU)C?7@h5c8ui?+~e8
z?}%NJ?<mdT$8RwXeWXG=qicQ5uM@|>+Ry|_rj@mIGuDTT4Fx>1zjVsZx0@ocQIqY)
zZ-fL2l<vwm`D|c56Z1pTx7Qhhr9KOWQnCXk1`Dj?*G~q;lmhT|kco;|@leKC@dr~-
z(lg85WIrgo=t#$(Cv=dcV+}OVM#J5n%yb=1wZkY!s{_c!t%PpMn&Bocv`8a>m!3PT
zv~P@~Pu?4b<Z%b@7zURc-}Um(sq4jdbWaf$l8tT3>Hi{0qvzJ-4t;YDIYX0<btEDi
zN3Q`Z`YQDf??MR6Tj61tD%6;JtC*rcB*p1A+%9c5`BqgY#M`8TIjykfYbn+e^^kFj
zrw`n+I!z??o}<bSsuHRuog2gjPWt-DmJbF3c61<=dZr<bvEEQXp>Rk3zDMSIr+InD
z9K-jneqZY~igwHTu1-E!WB3L9uPvMM6M)F{p!A{fN|)EYK@*d8y&TB;5G&-pp49X4
zua!P_%WEaWd&9ZwOpyKDJzo&*D9r@+SgzoRZ7<WqdckMYoM{lu>`h%-`vKt7RM6-M
zj|b;zN%G)S!1%`|jY)BjZK~J|Z|PfsZ|rxo#^XCkO?k(a3D>8i)=uikZdy}|4{~cW
zXGf+;n#@W~`Zq7!gQyCyq_K9W(0Whx3?Zo(9kctJe@^G#UZO_`H)4={oxBD@Ch!+r
zh8BCzj_OX%qGnM~_xBN$4ZQ@zCd$mVs=O0|pY4`VfJ+5@#Ar+WoDTh~J+({H8rf<U
z01ffti4J;=$$g1CN#r|Y{c<xUce}#a4g{8T<zy<Tsw2*p<1>9}&BZQ0b)tK9vtb%X
zC!t5@rEIJ%>vn*8u(iT^$usuijrW_ClKN%a2UZKeJJTSUb`oJQUtH4K$ZRwDe0Jx7
z3&GP9u(VLa_EO+v4&y|2b8?B`Ex&ShqupDxKX)pCM<ol;QS>L9?k)88b(U1BAZ|Qi
z%3#{yka6%}((x-EIyNp}*VLNdx`E$u3R0ThGQE(tbHvEyVZm7MWkH*~(*hHnZsGdx
zNhQ1edx8CF<R4&$GX_&dZE&qI(%q0!SVc!2Ky|U7b!vof*Q-zB-(6U0+R~ibQD=ZP
zE`G62dI#RR45qqm>O3vf^1WNP{Bm|bM^<73D?`pwOO|ENn`VX(nb2xBtbS%Kw`Tb*
zaJC<+aUyo8-Z-;G_E}{~8Silr8lsVDX1IDg**BTUl%QO06Rl{3QN3W@slJ_!Q@np<
z0n-K@*MXIbXDhXbiX8v;TeKGCs3zUvm#UUn4{E=CpO>bW1Mk>t<j9Od5&eU9shfc5
zp9ulbL=jmtKWQ@tv_}#XIv=3*P(%1S&l@BL&Wb+knL4gF+cGUx=q|dfxzn6tFLu%5
zbo0?|4{%~liw5R8tVw4#M=dq*T~%DX0M*<?PL<f(%Ys)xXBy{Ob(+=l@X)4C?(H^J
zZd@4ZKZt9PRC~U>=h>rloEDjQEO7-`#^EeAkZ3^@=L@!Zj4gqi@4gEl@RU7KlYKpL
zVtD*IA6MM`@$Z;7SxMVdwKk?H^Gs}tkbOB!wytwc%wSpexa*>B)Dn3ioQ^Mh0T?Bd
zv)t@mQSB}-l^(iGE!oMBjbk3)G^!PlI&H`XiwmiX<tKzKp*2-$rZ?g4#)sjk+B#{?
zGb$`)mW-&6H^Fu3FgjzF@w7)_cArlgV!C`iDYZ#W;e-ZvhW3k$q;;)-)m_B*oDcU-
zNyF9Q@XTAOTkxT9cuxB=pZfH04ye56D_+*leTpNhe=m|jx#Oa3QFPr}ab46X&a>OR
z@;H0AWks5;mRdV)oQzgfxBfkheX0)PkSJSc$74F2Ctc80g7<aWJZG#^uqv^Zi*V=J
zif0?Kl4JW@nkBb<?<u<3!}&D}-!8Jp?CEJU--7%K-+A>HUG!yN2YKSI10a7P7)eI2
zfG@mQ>k#+TSn(r7X8P}+^qq<BW>+mUSz=R6+6{%&SzF~-re+6-`iSle5)6ZM#Ns__
zbZA->Whj4<M1O4Ul*j+}kwD?<I{5C>5YFhZR=G)tPM}0dPb4gmL7DJdXY7@!J|I?G
z<<|Y<myoog3lnT~NBlbgg9kX*ys?zN9V$)-5A964tIeYsIiqf9GVXR^X9!7eUQ})P
z<|)-pR5dA`X=l{#DrrxzNsxrJF_206R6oj92h+oTc7AnODSltO(TCB1;*Zvh%|IsG
zU9#)sukdDnO<|ERuP+f}C<L&E>b6lDw2%CvGDj{e0Vj;jSb`|ajt#Vm&W*$|qEnXj
zxP7g$c=N$WUFcXa1N8pU`uI6A*tB!|qwj0ChLg#@Fj}CLzi?7DQpApNe$0it;I}3|
z{10}42ubCV2e@tl;i+#BUbU2P2VIho$kgZd<wK%;iU^KkV|0Z>a~7^8Z~r`su=XiL
zW9p(wrcOv{f31i@3jYEAe#qX2^*1v_-@h^e2<7}mQ4*c&E}vaapmZkR#pz3|nih(H
zdTl=9j*_?bujM312togTq!(qh0^L2fXU+CvG4(V4iqvKjMw#4nN}W8vyLwNr>DyW#
zh9z_pYR}j&;_=2{Jm>C#y*gQ*y4f@ARptyGWR;7<bhOg8fZ#YF5AqiMFn=w?E?UmE
z&<#<K#87(Ti}!Kew=Bqci?jIClPJO=Z_ThBQPN*|N)K9r%5_TJr#@aK=va&EJ0d3o
z69>FZs4Qwsx}+hG#_>&fVJHwV!G-MF9{sX{doP_(zmuy*0_%HoS;P4$dTFSwq~2m@
zfK*tky9du?`~ybH;<$IXS@F~%JGBB}iT7!c=qU5otUXPysPNG#Bbl`!a9r^IInybb
zyUc__8rJq*x*fDB#H~4Bo4SNA;q~KlH`(uF2^a^NJRtc=--OeTmV?9m)WAU(1n7bU
zT~MulYS7hN=sa9J2+Xo(_7<*|Y^1ELTs&#$iqP25tZY1>O||-;plE>2?PjV3+O3??
zot>DQTU+^-*L%dI-|9$*L1)HkZ-TbcZtKMI2s0k0)!sZT{<HZq_H?o1VEXl5x947$
z-h|Rs^;FwdSzR7S#EXU{Bql0H<x5^c#VM5g-kzySNIHK21Ni3O?CdUpt1DZLhzr~p
zJcA7iI07UTX-hzBYP<&Uy9O9HO(_}zkq9^(Oh*qwH5kq%c!fxK`@N3^DQ>{rYi)2x
z5<$rHz~;Fc64XXOpGb)YCLYb7hZ<4>oiWg6xwf`Ifw%fpb67xz&N{zW$67Ghks*Bm
zMkgJ=$n^L>Y$)(AIelP=dDZQ~oPFc?dB}^9$Cl$(*5FtiJEvtp8`k<>zGY_*cX#Kv
zTwT+qCO0fi@CIe;b!V_VCVCN3sqZhdm0W1AnZ3yLB62Tk?=&-&dn7Juymqv9er~)>
zvBI2SZD7PA8r;8f>psR}-Z|O(dR2Tlo88@<U+KXD=5QdKaE^AbZf{Lnojii;_}ka(
z`1xN`a$6Y3katgj2iB;bejd<#U?0}e{Hv$UIcPe&J1Fm)@FLKr@KBxLA794#-xyte
z?G7vF=uw_DjB8^8UawECTKW2r;a+Q5T8`|2YZYacm0*O^qK0s2wkR#WY9Z<3n_FP!
zK%-iPJnX?8cnEN|i<%msaoKzA*qwpVJGYxv>B>Yjq3<hjO7O$W<jBy2`|{&>8klBx
z0_6_bF1q=C4+o_gOc(%c@OW{tAiXk3uSm`;=*mvz%9^0R?z63suiXM~6z?JhT76v8
zP*vW@NPvriadUGA{7WfvHl)X+ks4j8+ERRPQoy-grwLCw^N9t{uM_`0fWLqG=vyYO
zOS#MK(P;<ryBr6IGF^yBtbHi?hbq3utbvh%$%FlZ2D@!&I()?Z%Utss*#6#ED|hwJ
z$=w-7GmF0s{7q!JZ-pC+gxdv`6m0Jl<@Vo`LO(E9u?-4Lo%H({{j`A;^da+Qq@fK>
zz^|Pv8|n}GXA%gOcg_;oT6XJ0D0ffPItcAT$1+}^sQyXc<Hv0B8%1hYXK-+4J4O68
z3DH|A=*QUaHz4i1_&(8<LPfN-d`;{T%13<Q=7Ffb2)!HH(ENQZ^*izhku-wsJ{VM;
z({+Q(&6Uds<J=Rw5=!l&pPq1@@Wz?i4D%K08a+MZdA|rZH(2j41M*!P=J$mAGw4@o
z{5~LADUBUz@yw0F_G>W}Y>R6HM4%nmrtd45i-yL(2S9eM>F&`!<fCJE!#cw89j*=-
z;a@|Vu}b*9HKIM(I(KJ@CznEB0WcJsU-W|dbG_gf8hzticLsenT7+D{dzP!6FU%wX
zgZJHyjA*Y8{NP})kIi*JaDiGrLl3DaVO?PTQ!8;TTEO6m@S^7vIN^+lMtsszZ~jVl
zHjzH=P_OSl2%b$5{SQ7Dr*3@Fcg7=c!L<&Ek3zT^jiH5;9b#y3!flb2-oTHGhUZ*~
ze=Ar{@15r#je#G0Hln&Ye*-T?)N6NyKJp8!QJkQ?@SJ;M!B*(spC5M=Mn13uE_Q$m
zFC@$J^YZ#awq0VnU|c^}jG??d!F98BU;RJKte+m3UTvQBPp*6a2uHzzL3x8SD)~?G
zYej~v7jyVNo-%8&Xdx)j4Ef*2i0CTN5?@F?n-b6_*j`X`Di*Mm(Xs^h{E|M5sjCMn
zqqTjDW5U|`?N!ZN@4GiIBknx8X|M#$#S)jdrj$z4qu~*EW36in&a^V1r7`GD_uwm}
z>IsF&V|ON{SX}nPofdo#^7|=kr@G#d2z&WtmRgBEy95q8ry!~*$JMEU2p_cd1@6mu
zm((fjzi6#@*dTN&s6Q9-43MeGzINRd`$(ov0!9Yhn^z98@CS@0BL4;z<{qPf`60Q4
z2m~8~n(KRfp<x9G)~5XGMtJ*3jemq@?f{gz7(4W|oY-hR)sxnjaw-gTcf`e)iIAN|
z^manyIlkd~f_EsJ%mof9cs_K~>;ERgTh1fir^Et@FY!@(m?6`eCd|dwHi4twiJj3_
z><^~LH=VM7@ASs~_{OUAOJnqaR3im1#0g@VKVv)4-gtiwkxD%DA0bC@b;*M5ix(A~
zWR=Bi@q!Dl|1`(w_)@t1;4qz5_+|Siheu!u%)XQe(j%lRru-(XkXxFW)2TRq<xi+P
z{#?`NJ2dFyCY0XqZA7ZEZ?~hloo*Ijfx7mWan_KzkPA!krTJ1`0!ninK>GvdkcF*?
zG!N~hp?Ql`wM5V>!5kS8NpThR7}W%9xP=1vwQRS278qw0g3JOKG4+LZ>|sk{)dJ{A
zFB%tIhoxm^rQAM;s=-BR=E@&kD}tWKzDSU3LQqooKKmv}hD2d1kg;>o>s&Gq5EtV0
zReA6unXH_`U#UD7xc<#N0aHpcw|zsc)P8M<RT<DPf8Dp^7kSlAQ^}H!dX$}H%@aO@
z8V`6WZqU>`S3}q(TSjtfb)Usgz?1dfu>3JyHR%>F9nA$H3R+bYH-_wU7Daq>M4x?o
z`${nwRXgl0cB4MkfA>5qB|G&sMq#yJ5IRy^4}#_V#9A<i;@%R02}tPPCe~QqTNna)
zVoYa}=qzuNyNw0Cud^qhK$_ixjK>252AA3ahG&B)vF)?leANFMvIV6*R225BCq$sO
zr4Z+`J4c@vcg7;JAh|<rI6comrtJ~iI$#{njcIM7lmm~dxG>l5Sv@iXa2t-@v9)On
zU@@;NUvl7HtScUw119-ml^(PL7Ls!jm372BuN4t`j_i(u8Ushh4n6aK=N__$VDjb2
zw(lE+L=q1#c@et)<8kfqGR%xr`5{T;BN^$MTgl0YIP*L9TDw>0hZ;Vs2x~{q?bv#Z
z_}HpY5qC<jCrPL7N*t?d<8LL2^O`C17tQeVT<uw12HZxL4Peh@FWZ`q=m_r@SB>?l
z8kR4&q{pZAODD#*Y<-xw*cO?T0_4Y6u;_M*o*F-6^1H;}?@yGvsEFihn1X8LLGOc;
zvxiTN+$Wn`#-S<*3iI`k+21TnFKmy5utV@HA~B>_ukX`o3#a&$a21Qua7!=N_!M)k
zC<S%@(Ps_=4uPG;;tH~8DV;369uW@S4|*Ghu{M*U{oFUBdK_tTnwy=~H&t{NIk($N
zr#+kg>O-vsglf5NN|5(9-+>NS(kZTT`S~S_gFn|jMR2<c-%)D%C2-BY4Vzw+hP_-}
zM=*y-&Rv_-^^3_R!@;ZxT1*b_)$Do6CQxuemr-O~LjZXzz`q+%+=z-mgYQo9EMGf?
zxW5XpygJkSg6S!pQHWIqCEGN~+`v46W4{e+d_`1rIhATeiW*5^etRlfO%!^y_8T`H
zyiTn4+GmZNg@nx&Hbv<^X?r`0h&M~geJsn~RdAdW&(VwJSGbj@n08sJ%b5(rHIKCA
zDvLk^ZE9jJzrU)kf8l&u<OFy7kJM)a!u*>-UJh%c-NiuWG4kK`yj{C;;XwB;@Gw<m
zaFrn5<Ay!jvd|}r5;5=0t5(qGbANvuGFido*Yp62mr!%~w1R$jKiqLVrkk?&QsAmn
zd}DGsIrZ6l0e@wmq5g@Lduhpo@+8y4dmFcszAp&05G3WYJYukv@5Opwr`{^2?#BA!
zZ41{OVYE<R|4*Hw7hEh;YN)x@ubZoRf<YTP0$l;hH!IbG3f{Z$_#eeC%G#~)Vry8|
zU)UPN>j8ttl@AO&Bs*q{hX%f#d^(b2h`aU4EFszM*~=*a{v3t2p<c{}Q71{g%zVxp
z7CbtTlcftScO9@hfR=NZ@zJar6Ion{-LmZFbf8BbHX-}tz(R?toBSiY^MXZt<UWa*
zm!WsPMHSX)I@Ucbhw-nAYCo+;a2`)Pt_)gXvob^AjWH{Fw91^i1-um@$#sw9o`(SO
z2{)b21c@KR+J&p$^y;-6xDaa<u7%)L7|a+jb$k%#@E<1AT|LHld+%YVgBa!ah3S<N
z?q!C|ST3rxQuTrsF+7i9EN(Uf78dj{UBSw&375UGlQRzp(!ViO;UP^I7A2I}v4h`f
z@dq=>6a3`AHCG^>bL!vowDH{y``ob=SnKYL+G}^Pl9~0*%DSi5IX5Qde{JyP>Q(v<
zyeuYE#5tN2`g_x+9SRMefvz=2NK(>UE=%b{3g>Dpy-#3IY6EsuvReTA=3KxRiYMNk
zOR*r?&K550A?v3A<%&c22POr9(vguM>c&wqgteu<E<Dg8t||fafjQ$dPLb{_=ag;T
zpv%^Z0=v!l?n0Uo=S8Z>V6FlW=0+kN(0!KAvfzuu4k{D?kBrAT?`zhTuqV9Ej1W?2
zLt_goSS0wp<1biJm5BCgruImV%?&f8fc7LaIfWz>mrDoCAR3lNLOB<lZV&iG{@<YW
zYisNd&#X<;F2OyIN;;X7e`3KSCt8}WRj<XO4u4jKVIcJ<qqg^+Nyb1_x%xFi084`!
zA>5hVuw%~`o>!vlm?qxDXz<#W!b^(}Z&=W-;h2y{Cym9<w6S|T^v&l?HhpPtcIVeR
zyTvb%RP~w7Yc`W)bRs1Y76tkDap39uo|gE={236kesDyHfF&ITTt@2(k`$giYziMZ
z;!6=uO&|s0-7#PRmw#bS`ddWZ0gGrkI3gK-jSu2#-Ag-Y(gJLSsmEOyjeeLoxEf^s
zmT0czr>0F(tN;gNzsw?UGa%Dko*;26aSAjTECLL8$T9qF&!w9+FLiBkG&auQI3jR<
z$T;xBm$g>A#qF|<c$#}epDc=3|KM0Hj3i@CVsI*$)YVw&DIz3mIqbU11h$%A;A;61
z2r~|i;}AqZr=?`tS3h;qdGB3U{>xj&l!lyzRM%{%vRzt4EX!p?f-RvVh>BxpRQi)K
zvhs%R|0BJ!-XzYKwUM;omb)HGQi+k3{MZubX9>?FIH1ZgMRdenoQNuq`7I@XW<lhK
zU_wk)h9xwavWen@uEL2zG|<#@`o!XgpxkEl4>LQwBsR_YexAheB&UQDqf9t!O%sL~
zft*%G+h6|4U)9nEO>Q3WL%i&IVA~qrHXKCS@Y_tB09&ORY2F%&Q#2b`J425MGMAnv
zj<dF7mGj;@V?Wds7>kK1#=eA^p5{a$3Y}z;@^votvDBq^zpBRiDgdp^x*I<ztZQ0^
zjO+O9e0-$8hj}wlbozZC_frI|?1#+hBWzl<Ao%k*um4hR{K3NE61HEy_wl%%P{lYe
zLwszJ#wMg4mY3*c%tj|p-;&Aa4b(=yMb`rzUE(&Tb4!glCx$vQ1)gV=)27)f73GrH
z&7M}EtAQf)le<6Fs2!*bziHO7+q3{@)fgYn9FQ-(IlEwA*JNu|#(M(eM22=x_aY=)
z0G!?#%q26Mk*3>oy;<NO;ZPU*SX%$MX3gV8>q$bP=wW~V;VH%aOpwmQ-)I=*A?|GS
zM6>RyV7M5nNpn%++V}NE+6oA^ja*J>ohzX~|E{${YT`2xxeiQBu=!xT$cB<A0dv5_
zGsKjC)GzI2$O2g9^g_4&*nt<5yrZ$wS(zS=3&h_2XPbTzt(R|b^|9Xea!;@c{Zsmw
z5?VsxgJMCzJ3S!5$bkmg90-Fua7SUG>^Oe%x<cN~bz?yy(^GPhJA$+@TEk?4veCc%
z^~jsCOI$)IatoL}KFv{`?*kb8U2$S5BNS-#-T-$?=(t9T=lE8TySQ8SOr7~#+4B^F
zzgu$uK0Y?%Qu0pCjTB*U_^@@<sczBUVfd+y{(Wn3bUgdKe6(gnH}Rh8>=w7GsTx{}
z2+F`uccl6e=b&9jw=5xCh$IrnFYQ;Qo1l6NzG-Mx0|PLYsA<e|f${+36!wQmCece$
z(D8cZ*TpH=7*cjb#;S?d5>Z$KC7;U^f-SVg*H9Kfo7&Yv&LD~Nti4oi1WBYJ!NrU<
z(uJ}7&O4w;sxMkDWMa-~P>OZDa+ucR5HUC^HMGG1O{{#K9HYP`WHCNcx+8^7$&TDA
z591)}a2Kelv|P#Yi)2=T0zW*r(RjFGT)WtyRcKQsZ~sCyJSiNHE<R-vmi!*a?Oigv
ze@nJ_QRgUyFz!T1j((nQo{X>}&d$={GI#(%n<5^)UOm(I%Ra03nMTINmQ7nelOurX
z>d{?*i}kn0G1mh&YGr%q*YT59l`L6s`MHhMrfeW0jWzV#H(mSwGZL%&0zUA+9$#<a
zT}!sh_wjP04vxA}(2lwng89><aJ*Gv<0U8HCvhBtJ<*Lr$RS{#HDb*QuZ4vq6mh*3
zwkYb?hkE&uYYepz*Y<@KFkhf>Z`uXxdUQJl8V13`i$Plr3#B^m(L9n(;=jJtQI-`W
z!Z`tHlI6^0d{<ZnMwrVp_g-*N@KJ8b%6Xh8L&;o7ML70!7{EEZ)p~yD=`-x~S#~%y
zeRu!M(>cE;QFu%FLG0tDR7LcRaKadUj-tGHGt*(8l>L`^yR|stPq(Q?Ph#bGjqNur
zDU2q5gIP}C{fOx_HT*&a-rtC#ZHTYw4`^1vmq>&A0cgM8L?D{{N3sgr?*;2;`s`Dl
zwj+D{q4Z9zpymVpjr!;B{xoSB3G||=ZyJ(H&kXm=_Z<L>k7nsK*hReZ8L--r^bLt=
zh5Fc-=5RzX?Rt7oXd(YpmVTPR8-D*ShlUnqx&sw?=$lQzdTKiy>SC4+&QDpZYQKG;
zeG?M9ux0g*fM^A}pp~0aBwbp+c%o*$O`tGTg`y^JCyz?QhNQ2?pw#~DEM0AaW4Y9p
zmp?ZZT9S5)acG)ihnzd9EowMEr~jsrfSvbzWkR0!_*2kgE>|TgQ0RLF*;ET9&FasB
zub^0C2A=UkCt`Td+hW&wHMxkK&|@BG(i~M?`~F3^o&5FSG<DY&$5|Y{h%1<A{32Ez
z_ptiUx1O*Fp%mk;X5WADHKM$W9{TBD)*_v$%RZ9jYO2NmrS8xMwKex8U3K5FgF)uq
zLRQ~Ww`y@o8-SmcEF}soxGnO*J|6Z(<c+R2IZ_RE_HSX{X6gHVd;yP|{C|a=Q*fY7
zw6$lFOeVH%+qP{x6HaWuv2EMd#I|kQw)21g#i?_7Zu+KoRo`@1?Onarv*yQV0m!t#
z4G9BMBY$ZpsqcP=?hl$EIL*n4B*o*Mk%2y46VQk|SB6S+eR4jSk>~<|fvJ^6Mg+Bv
zQw2qnNMc14wi`^vimk$Oq+}SMh{^f7{oScrDc{B4dE#(hL&x07cw|@lzhgCgTZU8;
z)XKz#ag2nBF^P-nM-w@r3+aM^b6uGqpKmX*&1F?>TjA4@Fy8Z(kb|2A8_e)}uze;e
zgJG*DwFvsGUt~Q9k={EnM1f|`BX-zxh*n-6_W1iuMbJp{G5<<%Wod5o-Ow~r4+_;?
z4L)c*a2hZTf7DlS@HU3+zYrzkvi2H%j3!1cn94kyHAfm<o~@NU(I<@qjcR^NFWY3J
z-2J<D`=l)j!IBPfN=S&1lqCd4EO<nCmvw^ZfrFG(;4<5rCzS2FUDxS;N149Yg88s#
zT3KxnGSz>BxroNDUMNAm=gHpZmFkfU3vw=lH7;l-cg-G;BzVDFHco!+w*S1Xl!|xF
zfG2cP52>d~nS!Dhd3rzv4hkmpV<q;ZrXrmGo}}<@ASq}R^>!CTeE4BREsiTMc;1<}
zdzl$_6Z^O!>HqV=M|hNJ2T=eIWu5j4ic@Iy%(Fw8i+Qn8VQvtwKQHBH3RFy+^)O+Q
zspXakFyl<+dc=G2{V%RDv0l}$ErLjOjfpR9;<2MB2xPf*8Vwv4$YWQy5rrVkUx2;A
z`@GB8qJ~^y(2lFFh$H8ug`6k456BR4I<sD&YmbK-;dxqyq(Lb|Qjt*O%~$17471eR
zcYmL#DX{Noi~!rW6=xc0x+dv{5g@GE0g2Kr3QalzGmeoa>Zu^K!+7-T#yU3&uIpmw
zpDm`N0+jzrC4<HQE%#zRIEN)Xze>uzN(U(ZXuTo5Fk*P=*9TzgN3#CaARRtRdf-{d
z0v^bsMFS9>yI+FkhzOMIor>G9uA(UfNxDDSxY<BaN3XcHA;4%Ot&AcCa&M8j8BP4R
zq0@5Lzx$mU<j?Q9krbUkqvC2)DK_fscx|xK*FJPi7b0SSZYGtShjR|3y+qa7h(#bf
z1;SA+%!&T~o2}wpq0M53(<g@hk3fEFo7-ki7r{<Yy$Po1@CeMvRQMSMM;m&Mv%z0R
z%T#y|MyRoBgX69fV|944$zzKTKf1X=7L5Ng(>#%*0`J4}IDdV?4EJNng78?%YYy#G
z>sfg8eW)4%=}T!AQyj<8jz6c#z#8lVRC4%c)P&+Gw;nTU<3;eno_SS&$xHkl`1K?z
zlRLrNy~1FG$78?9K5P&DGzHzMTZhB^UtTa?m#=?49_FoA>knNsL~tqiTCGTq@6CLN
z#IGf+Q*QYCdaCq)hwbzM%75(*yXlziE{-q}0hmNEK*Wv?F}&1^a<BGrav)a}5!zuz
zFbKH1ElttL(-L0bT<sLHAPfBLet93AeS`z>`Fg`0_=O+)U+rW;Nn@RLv^SZ4=_%Ah
zQzxv8z%wvEaIg6NyI4`?auKw27lwtU?)z#<h|Uh8Ipr}AnGwFcw^dD!0WEIJIqa?r
z4EZe&*bk7xvx6&%yK3Obivr4aAI_eQnH0yE*V9+?^r0o&RP`?vpY!tf$Bw)S>=)YF
zD%xDb8LX2p)cQvXAFp<nrza$GE^^D4yjZ_toR4VL=Fp7-REG|_byDQc^O$DQ-^qaP
zGlU)a8)jdIab04EuFPeQvV@(!8)E`bwGAquYNAx`Z7MAY^`T}}<FD@SXK@|hdu%Z0
zW`!l<Y5Cg8tJ4CW4eJOQnbPd@jy67Wq8(wt{M6Q5*{d-@VF-ee3+g!;3du%NUvCa!
z^UP-_vzMAkUebW{f2trI`fB0Dy#0?I#%`9Czt@hgy|b0BQL^#XHKXrs9X3H~uLg4f
zQ89HKUQ?<~<kl)UCt8HzTL8pyQXe$&#zgW{bBtLAPJey%i^Xb`;M`{9Cozh_bAQ~K
z^?martEcYa^y=)6+~3gtc$}XZ*hj|PvNCHBg*uX$dONz6H%}vHXv4g0=-NkE&-Rml
zqN<9{c%<P8!keT0L!aZ_Gyc`6bLRE}bm?LG(KWi{wxsBFUjq`iTx#Oev1CWtTHf89
zB(a8Q#%!9yi72B7+V+Z^yb?^SFoO}_EVjS=y+4bKuxJvvSg&WRU48tG{WDs5&y@`t
zsRzfV^N>$7_9R=3w}aVehWsc5gK=Elr6jIQFmI_F62+-(D9%CacdL=Uy*c0ke#b_n
zf7)(BeY4B#Z0-5Fd(8T6-dWlqERot()=;O|BC@YBdFTl;weKPkb2&G5g^%?J79}~`
z-tuzT_T*hH4o5`7LMmw9Dg{);{Rz#^yN+Lfy04VvTgj|_?*>&FBM;<sKHAZqrzfcz
zK`)8*Sil#Q!&#B4x5Z`Cp1u}<md~o8UkC6vyDn<}JMewoIb*Cdx?#??KQK`1a$*vw
z3LCt{StR~tlVL|zH8D;Moe})DJHUm3nEa_wm{Abjd+h@@Dy)w8mU520LOXz#%a5tS
z(jVas*C(^;7{m$sSuF)Kr4zTCLXCZlNaN?k3WQ=qI0n9NtjyeC(~Rr@mN(4TVc|oQ
zv14(XZB9R@8w~@x?h}?s<zWtu6gF{J0xIjrZFGG&A3i%b8gYM1h7K%%`2uktMj)P&
z6iu{~E?$x7T*ZS**;ljeFeh!wqS8xk^H%91t-e(*Z?wL-tP-<6PLEP8HOW4TpW|yW
z#C@=1VW`$7L$xftbfg2olt33@bJlLFl}V`%acgHyRhpzcIL=v6q&Xp(PEU-eRH++(
zI2~$hJErP1I7H2C9>ZPO{Mt^2+0?%_EIaL<Dz#!w2HpYFQ*~p;=qHdJTkpwhV4uWB
zYkxbb(3i}zDoSKWO|TWJNwJO1xh&vB^E_w3rc}g7Q5}}4&nN)!b+Wwd-F&}PCMk7V
z&^cQ>TaHQw--m7K^ZH%RPgc-%@4xa7Z$|K#?jE}|oZFALbRtsMhZ<@5Y;_>sgK0l>
z&wMmUW9eO~gQ`E*7PtEX`r5Z>0=|tPmKVwQ*He^f_xkKJyiX3JLaC&DzJ{4I6|+nl
ztw~c9<~SCg&<zCO+7^y%!mealrWF98^^QN9d1n)9uU12@mYy)7#CPOQB7$4*JBiSQ
zrF6nf_c+K&ZtyY9zuNzo^#&|i@&sq7;5;$>!7zjdA-o*<eHSd4-l#j~S?N!*bq7ic
z8Jzt7*@!Kh!!Xy*n4BE?E{mZ$ws<s;QO|%WExnq%qFyMlX$#&G)wgzBCE{t~l?x87
zMhnf4NQyv)TKz|T`=EBcuugXG%HkA~k$>?FEj=A}+JqwFW@y|Mf#qH}YqryRb#)4A
z-&7_1&16ttYo-R$#R|he<atVoQX_v2oL*4G6>TjXhK+e!D2xBcR0zaauE8Z2$tmB%
zog)L0#n}n4zb6=xBK4-KKfp&&6;E~C3<{-f+udXQ`R8m$-!k@8xK=6eWt3fa-R+wO
zacNY7UOnvEh8v#$v`G%N8cA1x_{ueXyWb()4=d}1b=N?5c@<gXc21#~)aoHXm9#g^
zR;cv*-Gkx=G{1XiNPcd-7cX6Do=p7bZ+A>G<ro>jM0V2YrTK}?$gXP0Vgr1I{}zF2
zRMADzlh3SA%mF+IozU?J({K>EQ0^zb*Zwa1F>wh$-l-6A=K<O=&0i+^)W*gZYTRPG
zKOs;CjY_>OO3sA+6JMohlih;a-WeM!Jardlzur42Yix3)iiMu{dm${4+(vqQlzP+e
z93F}PBQ?RKCqnxh7+ChHzUK|fCv1w`G^8;1|8NHN?pA*UgZX^^QYQP40Hxz>S%f_2
zO^yAvODW5fOY_)gkP54z0wXMo1DN^mCtCJhDQ(9y3*`@vneg{9>oK|AHV9Bl%kbe*
zzHZPhzm!YHO~pQiSUb3N`e+;wol%~&t?cyzTOdUE7%R`j7UenR56I3obZV%)6VV-#
zzv3#W!eaKlQGXvOYEtG)^ZByUS|PmOkvY|`ps#d-1jB4%{P@qG`YiR{y9#@A4(8K%
zGa<dv(obgH+C0wG#UmWe4SEP(mwKaO&L1n~o%z5qaYc1Z9Y^`$_{xWT@0O?;LUX17
zq^hr@3~x3GRkyDvUo-?rdF)txd@!)0ezu!_v~4k_bd6Xp!>+6Mu&%|=g~lZ1#Gu#8
zuXkuh<!wcYqw4rSDD8|{Ffs8-BstG-pBsn#1LXNG6f?g>7`)02{<gWfMLaQzE)J^8
zjYL)9gnVi0;@vozYkU}CIokva<oZ?u*S@=hvdNk;zUle#+5P0arGp|3VISvAOf1t5
zd!f$4NG`^MJF!j#g7|g2V>Z5r;9GW#v8w6o^*)GyX`ix!@NV-9p6R!K=^#%r<E4RO
zsIMyXQt$i!D)39QshD&1UdIwhK~-sD0oPbpaWAQaIItDOt#Xu?ZP4$wz1uecY*jQ5
zm(2!n3v1b6BhCRTg7W#gD?J3bc+3fPMJ3;T8+q9o{l{N?xgMuel59JzlZU%9n+Pfz
z`@W(+FxQp2Ja&NfIYy_M(UwJT<*;n>!n6|q$~W$%=}WHfb_-=mco}|#hip1`Vk%g}
zfb}~o7%zFj!QcK$?~X%9@6j_rMIKj}&<hP)<_cF-Ts+8&@ZwB5vW!8TzvJ-F4Ca!A
zvjV!lf~fB-XzJzaAkRp|zi4wV{=_IwE(u8xzqT*X&9btUDQ9?GQ*OTO1C#C4jddS=
zaY-W<d-rlh9VxIJeVQ!=u<Jwfe+Jw3kt)9Q2e$GS7#bx$M}a4sXj3M@Twg{chK3|I
zR%JH0F`@qI)m{wwp14^7f<i^0Zq0sDJ5B2bBlbM2he@;Fqq?i&Avr)%)n!r9xXZ|7
z%Tr@9gV(d*utkB~QTWQZIxd4xztq<_jV2?bm@#fZ#v~b4GQ-W4G+~72k)TA%Pu<#u
z)GG=0_;Fv@y%}+aurv(RDQgUc`~NTrzkli{r(OIb-ik3Oe(Rbx%B-TE{P3TL_Z>mg
z6|($ly|k_GheRM4|Ix$8a?j=^cFwf^X>LEUB@POvco+PX)sD5+zo1DR<zBkvt+orG
zZJ6rr8tu=3uGup(7ftV}u(%N0U&tcN-l63K5;6{=#9|-gcp@dhPr7w~?VCgY{XP9>
z&$Q-feQG_;*|V|f3b!O-XwWP~9tVbC`Vv7GrA3y6M1*NB1frBr0`7GXVP+EdyyfT_
zOxq1Xe5BQ-{jFsRVHu~*D#{u*`3l5lcvp5&5#TdL;VwGpGI!~;=PIpJpfqkk&&{D|
z_YS3)GtLDChe8UhAxgu?p3+ET*<=@dP>Dvg_!U1i#RHLsl8u<63T>;yQo1<A`4Q(U
zOV{SlAwMIUt;|N9Qm0<Zc~7TXKkpKLlI52i3PM|vh3pU->FZg&P_1^(5rE8hnA+0t
zRmnm%d)wD>PvRU~;3O^i-DHatlgDFDBB5{X3-BX%A|L@7f1e7jf*))b!dgwzXB$4p
zbS>a#FqU-J?3_1HoS9=-w=g*od;!241rE2WL9_~{HCct;DF=0``LT#p=%<5YNcGKO
zJOeAKj=E!{XXp9+&-_6DCzSX+2R%^MQa&S+l!{>PJibzCwDwPsG@Rzk=f>A?5mNDU
z{v0(khF`$>fHT5Trinvn*dR>e`{~C~YfCq@0iApr9eN6;Cei&gCTS;e@%9AY2D0)&
zCtE%UyB9ski{77<nagS;!>Y`0H8^Tcumrv=Zm>cmup*~ADpUZO?^i@eFK9j))PEf>
zS*+%THka+vXck>E*YSBZ?iIE+#%*Bbj;%^LM<Gxac>{zEnRr*Y9FP(s9*x#{JSU63
z@gzFlNT0zuf&^z!fM$w)E`!-W5|;Ji4dEl@AJ|JkHnk%&-NcHFBE%3;ON=R!dGum~
zWw*T5gtnkzLXeNRVH`APHM`=naSpuD{boHAjN{&ZhA`}Vq%+)`-3xUu7!2kk#CEsu
zr~zcB%pN{H<fP@~PKxfLz7MtY#L!Xy;|9}5B4>ml%S|w<CRy@|)RhIqqt3>Q5(jM>
zVY}lyb~CWRQ-^d;9+1(kdHoWKG2n_OW5zdBoE|52y6^3_`OB<aFgg%Ts=Rci1{qkE
zk%8x-!OK@-H~?IYrxV*GCln;VFu*?*%>Y%=3nzx@f@CVSKJrJ~rz9oCjWNdDU)tC2
zsYLHj>G0t1QjaT{n%Qe+=Esc|?$XkTflII<Eijoz4|%odF={s3_xEOTO&~wQ{FqK_
z90-4$Ay;ng@qi$gCE`<Mu`%b#{D+U{Lsu5%kBnr^@RodN*7*gR+m-YfqL{IA-U@Uy
z-u!6EJ)1%b1gWA02NO;3ZPlWCXd#+157O`b-5gpS;NxZ9-gdi`;W|*Em`#}vJKOY*
z7MVf()vXn{i{t44Hwk~`IjT-5P^!n=AY*wQ#fR#am1AYqH%_&MD@eW;ETk(};p4)T
z+IbPp<WKG)Y0gNV&^^)Uwe;B0bps@0?L#DX!&e_RE1%x0f4sD%-|NZoS2(gR3$wz$
zTKhXRIM@_&22)t5%_a1Y(fQJy`)cKn&kSLUQmm}wb_&WYB>WiWBbNE~(l49x6xTmB
z2ydNIKk+^M-5W*$`8^!7y!fX+8oAj?26X?$PG#?sQ<}T0q4!lvtBGjjuo*CtLz10X
zg!c3*dvIVW%F@)8>g3nPa(^bWGevOoo*jnczNL*jlf!!0vS#0$43ME~Sr?URaAkq2
zC{nlkz@nO*QHcp-Ur$L&P^cXwzj98okCrfh9=~+tSvV<RGz1s6uvyi~4?g0!Ws==d
z+q+0+`M0qtM3?s-dP8wR&kWGu@i(hAQ`rl%IvVilr$-j#eH|d>n(L0P#DuF1cS`FX
zC|ZrE>C#g$<~VyFUEoGxmA+r0^;qx+)e#%}IaOGj^y13k{&PL!%l-($ukjj<0!=e4
ztMs%Ae@%#j?v~=utjsbC7gV&{O15nQNNC38e|smz1s$k;$c`_lZUOIDi-yRi9$>cE
zU|^ol@r=3WKfzJH3r=DtOJHXQ#VekZsx<6qU$nUs$e~KhR$7mH71lyjBVNy6;7{$R
z^er7O1jAC5f?ZU3h6M-EAqt8gShA$mO=p_2htSiUtNE6~S1l&Hu$1P{gGu4uw8T1S
zC7YEFJaC{`j@|MTz5omYUM`d2gS@O}A{b&%>e1Jgz~(^MP}84Z-rMCDjv~l!TU{|m
z1YffbxF=*Y3vJFn2O(y2O9!iYMy*XxbD#<$j|E9QZ#!(PR9gfHcKt*0L_}YA>$S$@
zQ71_Dg43~9oC=z$Dg(Zw%F#6=a?}og@Ln5;WwrYwN_uOj{s)!!Fgx{%3H7V{&+^i>
z%SqhjNI#3TDOM&A8W1Q3Lx$d|R1o$f$HXT@qW$go-N4ZK;a!kAWC+&i`!md;-DHmO
zD~t+5<)Unskm$Ep3MUw)VLbkG&N}LIHn^duQz9QPX#I^+szR1Bsd60`^7ZGe;h?+G
zO?>DX>fpz#0T{q!`O=l+jd-BR9eX-Mm^UohJ<Sy{q-iNNzv5RiRq}8V5mvz{KORhs
z>C@M7jy9EeD{w%DuHNfEJrE9$=BHaX>zsD-H6hbsuuCDC*x)o|PFMBX1kQELY!p@j
z%e#@LV7o>Et7`R^Z&#Z<t&wIU3*8dqH@OF9W~o~i_cUNv?}J^Ihw8tC^<j-Zz!U0X
zh#U1^L!B)vpk~&PF2{aF$m;gGRjFiZFa-OK_s6c{k9Oh?W1R1q<G>VtnAJw7s)X~}
zFmWHX`}Y2%LXOp=^s`$OBo>=1+ZCkIvG=jc8z)!mNgQG2AB6cyo8TWIIprQ3Y!S`1
zchQ7GS+PLR&Qn|5{j*yaUqWB_q@HS(JIvB-+IMnuknAgj{|Ag{7}I`)^_=IaNk2Z)
zzvy&*%gh2*`F(h(*w5P*9C)8=V0M~Pa1VRK-(X2~)>}g&^W7PuQ$F^SQzgLe&wV2w
z>OLH1E=S7wK)YDKnPIY*0RM5qOOky}tS7h;8yp~E97~XU6;6R%MffteZ+3{+8E(9E
zeVlcRF?7Fhuh&2|X8>svYm<2Qw7-02?Nt2}Qka#Cuea+_5_<^A3aP!+I_+0PMyiLi
z(?xmO&ceUXDDf%FVi?eK<6*q?y7O4NqU(9I?rEgSYU0RO8#qH6z|oO&;wy2zRa&k&
zJs2nelBe{%b5>-VYEs4G3#C4z_4aQ5Fy54(9(P*Xv2BzOrbL+>`Ns@s7va>C4X8i}
zIUv=&8)wru*+IUQ2~J@>*7ldL)SHzA1W?7bq}kYwt`VV8HA`EjZ^H0sT;C2aq00?Y
zqjJX|#@6CZU}jVQ_e1R7CH$`_LzqPlX#zmLl;i-T3rzZ@yZ?{vnt5e))cvhtY==jw
zYQO2x(|M5S@aY83ESuQ^s4IsYW+0*u9U12&m?CM~Q-xw)7eiocVpki&rwM}KOA1=7
zMThGUka7YqIWpoWt@W1*Nh%uS_FGIP9ZgkeY;D5IPhYe{pT2ahR#9zFS$`dqtp}=K
z+gL)YpG~M2bC{NqjBDE&Zx$}K_-n^|Zrk_ckc!xi4po3MR*R+<2>rHK3VaVF`Gh(h
z3J%vQXkQ-k1d^CIKnHD-+u+Waf?_GiO9V0K<s-SACBDP^ZyNuQ393DK#Pd&V*yW%E
zr^gi-bJSgvm&eg@V<eWYaK5R)0U!-U0M!fG#7ypQjoS6m3$ZK$d*sg`ZTRLA;_l&5
zV$*fP{1}YI(^WigXvq(ej_x!W(JO)S<UcTl2r7kEmFK?9%>2CT=9^lX@Y5K|nhJqe
zA1|nUQ=PRk0@jJydAfEznMmSbz!@ygC&WEz#+`MfmkYa2J?ZE*wQLSv9>AgP09Ag;
zRqc8=Z??aQeWtDjSJ;$DFG9mnuu<rW9M(4OkAs7GzV}<QHH}X#<&noRD>GrE;vsJU
zp{~azA)QdDsx5qNBu4F0k!p6P0BosoX!G-(^q|XTrp2AwQPh1=qLhEX*RaDr=4;~V
z!lcie7kzv7431$xzw)S@1Sq0cQ$@v+f3E<e_|NC!$*^7gv>rD3=hP-(`U*YruUO0=
ziW9o4iCs;YcWG;+?j)qRw1XkdFqN*FQChfcv9+lbs}K}VBOR1`@?Tj3dj<Cx?Hk=`
z*{+d_IR1`NkAx!PJ*<7X1))&B%{=}^Oi$-#&&&(t5(kn1B1BoeAs|%`MpiRvDX!bT
zUQ65@A2WN4ZFFIuzjg<rEFLbCR)lzc%@}o1IbzYg%#i(W?q5mU-9mjRE^Oj|n^-oD
z|Bx&9HXxMj59-4>s)Zhr)*SXRMJ4Y|w0lO{j~pgMT1vPlZFQj)AViP83AA$wi)|YU
z`6{qsc0yb}GHFVt1%L+fC&*Af-@}B0n-kOAt>yX9heT#(zQ<g3IG7OX4Zm~>2F#|G
zYBy%Ym!1#Zs8cKJ$UpC=>8m5gGKgzU4wxnQgWi`?V3u3nspu$BcRkczLyhM}E09qR
zxv6_R(j1u)M>EmbEUcjR%yvW{9X13+?-&C^*mF!0a3{BmSik}6g+Jp?-xxEAIAVi1
zV<hZ9P(fV6elBe`x4C)>r2F4|*?3@jjVn+Z5V06&&Ez9APB?k`({XGMbh9-N(m1;M
zUdZ&xM1O~u!EEpZ47lLa<Z{OuE8=N3EA-qN@Xa8Db*{(ch(0AvPFVA%trCs+^2R6|
zjY<rzU_3AAjDVRjd4<bHv+yG8)C3m(Hr!N#rs>e?+)adu?lP<sfwYE$ly-=aX10rg
zFo=6;?Y{^47x_3p;WN2cXP<^@D{rwspae4;6sqR9aO&t5;71%kJMB(rH-^lJ<L;{6
zu^0c+M?mUwS!@}uYgv5eHheDKfVrZ~(ojD83I_o`D*$<<=?PzZ;RF+p+FJD|%kHLt
zveE+?(M+&|QL*N^#63xcN28;ppQf>5(~eyHKokz~f0w~D!=#lo8ZOLAxMluN0sqpd
zh-^~njyDx<BwnOsQbbdc31qNZvcsj?Gq(o3^D9)+tQ*6xynn%TNj~HG;!z(<4<g9R
z2^e4A&491*F7G;-^tXWlMfCVBLF;~Lw295x4QjlA?WH|l*&EUzgHH{5f0Nku_&gLh
zqTB0js>?#{tn-F9pb`3(_5roI=~w5)MNB3lUq6P~CW{t*)FtU%r`$9%9ek63mp?m%
zDdtwmAQ8cJpzNC&R&h8=-%wZ*4K?nJo|gVkf+&>&y<xVl00VdTMXJ-k=F|qls*Yn$
zBG2@cYxEnw0}fOt<tH;ZN^`Oqc%Lu`2S<Ya2t|s21vnDW`y)POr7nvi9+#W(LqL)?
z3{rHA<Z$Kw{MuCxb(Dxwsqmr~4E^VTk+U;xNf(zVdy%fypMn6ZMEIzuV#XmBJ~pyc
zv9uWt)SD`-bJVt$A<1P5Tx>IAWT$;3*7qUo<`NkzFNu?EyK$CYwx7*2#8eY8toaQZ
zpxcy2lQSeh%-;~iv2-zUNnzA0^9hT&Y%(k&v}MGKLNE}<5;eVoU-$u{PIggEh@t~Q
zF)Td0APv+}QXro{qaYswK!`~Z&lKufuz9G$mc$|)SP*F0?qH&lf~o52g2}`yqA>h8
z6~l7Dh_k@;xHMJrH|HTCen~b9?bku7K%iV8QwRJ23KSK=;A4~9AmgZu!_>rx%ENx5
zVCQC@j7j`3PuyZgvKthkty4tOCL8n1R4zsAqT+y25s;D=FhJ><M_XROuoN`37%crE
zQgD*lIB3HvhH?Cf*BeIsyI0IOFe&=VD3H_vW=3U&&5{F=rUK_a+yVRrOQ1Znpw0m6
zkZ5}lfS|`^GY^xF4D$s44eLX=K`ZtGX2XC+><>~7ZkFG}T*?^<!8j|a&ES_0D$#<?
zoAcu-g1GqpaR;XCB(Rqia^b?SNVU%>stE136-}`O9*uBC$w_DxHti|PNhryWzXE~D
z;5-my%X5B~I6O8#en``5{DI%7OadFG+(Qrz0epS5G-W}AGQ$$U9<d4oSWDx3DhDi1
z3b8`M8hiaXM>3lPpb`E?(O!B#Z)UrF9u`{<^cpZ^UWE1Uo^;@LrXa|_3i2RN4-)JA
zNP_p=YvIjcx$aHE9o#GBLdtMdb3xiQlo3z&^&xLAL7t%0djbbS1FpOFQ3rZg3)yi(
z0p#1G7f$ald)Bm#$Zz#f{_l^5`*$}c9OvKyX~t|P`S&V<IF{M3d;RaV{%z2a$Xuu%
z%2Vz_!X9DVBYISBgwAh_q?n9nEOEk&eX)-B|1t(a5Mt#9++l3qKjF{cRt5eD3R#m2
zW?0QH8|y)5waV(FE0g=dH=Uh3{YGk%1=KT?e*4#{s(U5WTF;}e<?i>pa0itC;4ZFu
z6N+&)fNnaYx~*znPj_fCCfwA1Oe#m-pIgd?`<-yqevNgTD3dc8kS%mQFyiclEMBm?
zDhWMgGcdbf66Ln(lg)yU7PMdLxq1~Qo?3+~GqgenPacK=Z!2zgnJNp<-f21Q32gT7
z9tWF@2up_)5~7N4A`ZnnjD2cs4QV97kj)A`#dinLZ5KCII9QKAK6Ftpw{D009%HTq
zep@0AO}O@qC|4m9Ov}A(9=b0?uV0;9lrv{Pgfw^Fih54PdpmlrU1vF4EMR2uJ+ATp
zno$s<%ByU__;>t{KmFYPj&Ivu07!D50jvDiLfB5jmtTW+jld=^Htc%Cu7TSo_j*qC
z-J-_?`6eN24XFl3EO;5=%+dFCm!T2Y0i7Er_kk>CDb`&`M{*X%!xi>*w)3?1%o66-
zmeW-5@}3jNhj33exxLF`XK(UTd&&N2sV7@+@UBx|e<*y(w#N))n%%4`P?(t^nF0bw
zcO~|%^`5ZfTz0d#aR_3?={|rr#)2?gdpPzb8cIS@1Ri+9Kt)h>PVxL>^g5dFnUB$r
z!!|Ez4%cEA&6_uTMszLi7|IwrW)XAMul-3qcV@?c_e1O)S7uSid+i6l5bp$by3_At
z^nSr^c5}__sSsZKpABS*K%i(eQhI;?w^NyLwT-yw<psOID%Yz0fxGe44_WihVC;i$
zX-@*?m|R<GpyrDGahC{8F5~h|=ux{tS07Bximd6Z0>+nRdydroddKqn^Yv^xr1KfC
zMEJN&^3hM5NKU|6C?K~Z;6}eJI^V8S*d0{zGW?Ar@Uu0zr3u*-;5CqzP7_~pg}t>d
zJ8sukPUdWdF_e?6Uat84w>E(QwiIg2d1paTxc6JULY&4wu2+HjnR(WkiLZB_kLLq`
zWhqZ{TjJ@e%tekegc*lwB~PK8<#CoC>L4IIY&<$5{+no!{leT$Wj?)TNmHuIy(Y|A
z=2egCz>}laaL*JBc&473U+;qCi@9m${#<<C!W7lHbA6?i-mye&#c7>z&40p`YUJcn
z2lHuL==Pc9k*B+-az3T7qj~h@+tC;SA63ZZ)68b^2MdG>HuY^dkdj9(%fcbE%W@gm
z&AuwmwZOYlHa)_W!u1ILP=#E(qcYK@Lv*YN0*9{|rtwt-&NQuBra@GA>R=}}R5#6~
z^Hm=y4US#prK){VET~)MH~F9I#^M>u28z<E;;)Zj54ln*Eq@Z>wngBEFA!V!{b*H=
zK=-=~s&DM)ad?7uW1vE!N<CJaD-huBt!g+UZb@PySqw8&%rr~v-EE3z7~#m>adRSJ
zVBkNj;bp`F7JMz@S-mt0xf0>R-eLg>>!uBLoIfW)N|^FEN?i`x5PyU^^0H=6cYS4z
z7Q^I8l4#N}LF)VnQ7?=N?R!ETW&?}o(rMyJXlXUok127^j|4*Y0t!ic?ZN8fW!ev?
zUI!dE#kH)91`zsJh%rb=@tVF55L+jX>gjNaRkgeTBf@YRj*h<!{9zdP)y19Ul6~0*
za5VUx(uKju14tNTgTF`g7f9<};NrNBxOjw|p1q#pJVcC+J8G5UGMG_yOc{<n=~#=&
zpHrI(nJn1wisU32$4uw>`khw&e>ZzzjyW6I(C?di^BxylY%kecgc%<_GfAoR|29J*
zB_2Eil7Eg+m^!JS4ItTbX(Rs5CrV+>a?Sd?1u&*fj6OmwfaVYU&S%Qs;LB2-D?qBu
zUe4cYReK(5pCql(n7)B+i7PE_y{YMs;268eJH2QKDr0ZeSuYM`OXT^}Y^W_WHn>?e
z|9C{Jdi?rtdav&=OPBQ?@p0mdB!v;0Dg~t+F!h@-x$4zkys$W`Ct|8W+pz`lrf}YY
z?EgWHo4TmUs3gRE<7b2`J$%;kFc=byAU0PkCCr_$lwx&sYUgcqnVYK6M4DX`Pxix|
z;X?%EGzsNisq&TA9xyo9vVDxM_n*+(TU>Mq?=lQ^7_AhMvrJycl`fc6l9boXGE8j&
z3eU|~`9QUVd~tE<BRs*5*!@0#nfhE=_zCaupp)d_VM<S{rA!mJ=2_g^=?RdtC0lZN
zU-cI$$xVgWOTMmAc8VNU>IGJs7EYQNcI=?>J!*<1gpK@8z%9SfJs?X`JshGQ1x0Gm
z$|jm_t5$K^K*^!AYn@W060c19vjExv=rN`-Qrl;hL6WB>>Cj3tZzinJP?c#Ess1Nn
z%<Z7*|H-g?rpKz@CN=k}C%UB@SoMB_O4XN$R!^AcKRny;LSwx#HrDdO#W`alD)Skh
zqjrgWw@#95olV|+yTatz1U4vj^Fi0I+2Y}mmSzgK;x+zA=H|(x$r)9ymxb31&}Eio
zPb)_E(LPP_!mN|`2Up<y=eXzGC$ysxE+79j$r&J`{ZvXiDJ9wPMItW59GlGy5&SF_
z6O2>=@g#owN)FRWV_YQ(6KixG0lS-}Kct|LARS`2ubudq$%uH^+F9*6#@}}<KWG|I
zg?gIvhX&f~pc32=-PmRt!_(yu*uiQUVB4;xj%}A<qs(N1{E>zxWA&64vcnx4S2xjp
zaHyP>o#PlA8%h7wvA>=?Uu~MFBQkX)d50-635T6VRc~cpS6BeTY(*g!>JOqV61*uT
zG)fIONFh4PrR7Kv$ALu1pHhHCtT{dxr#0Y*!u*GUJs|UZJ<&w|k<x}AaH}XAUw-3$
zTyw2I(+vo3+I8qBC}>c;BPqzfgJFo_NMLYW41mGGtDWwfe=X%yF4lnM<cqvmVCJU0
zV*ur@^V>X+ZO~1hO4Cun$S@i@iAu}d-WO1f8`7eQOb6$2Ay4r<bCDmO0u6MZJYTVI
zn2ETj4R_<kbm`HA^JW|da5hBt9xZA`H_{0H>z<FLG?Dm5kLXJ9RvVr#Qow>~sOdKz
z54;?GGj)}x)(dgW>uwbLdg#H3T_6i^@=ZQdZpjk_cX7vYLu{tR;TUPSk)4W3%nILM
zJ4nKk>ln?LLxoBe7_i<0E{ugVEI@G0+%YROBxm<0FS0O}@Ipla&S9p<M~akqZ(&Io
zBiPqMHg`GwyE+c#70WX=w^41mrU$XgVA@|kLTT%F$O2=JEmo+Un>}$cOFRw7?cvI1
z9nKzJ7o5XAknN<dIT5Jy&jRo})xKd5%2Y?}^%{09MGa<Smht=iP_r<`7ApdA?*mlX
z%zsyx$mNPfMn^k=4sz*ho7x%9s}x7wI7jX{SaRZw7ZRd|cIsEuzP6_@2Tt+9`ffq;
z80J~c?z>1F=b3(yM)CZGc=@U0|8R<<{<5DZovkB!o_~D$a3-|nu%d}24l0Qe>#e}T
zkC9(!%8o<2@yS1r@tNa@IX)Mv-)Z8L2aUCF9@#wJ9A_DT-iglM?{MR;bg`ER?xTCy
zL_fJNRGfQ@(2@p5GXk^TH74(|X#sm_$vCVX7}X~BwMo7Pdu9dcr%wL{I`pHOp{2Mx
z7Clqb@!x5)5%&|KThmlc<L{YF$55;nfsSy61e&{=DXinm`_wAx7PGAPOAxf)4!N1L
z8VQ>85S7Wms}t46vgsZbWlIFaH2uaA&m|V&Bue#F6P-Hog^66d<Iy4CI4LPT5C1g7
zD)Q`~gj&x*KWKg%X}_E|4Vy_yKHW1vYwQgwm*{(%2fwC=2I%49n60FVNM_Y)r<m&c
z+d@eiA`aa}CnHDhcKJ@n4CQRX+|=60tjDou82%W5F>aD^**?afD1$>&!rti*Jg>t6
z8i!y~vtX;in0K9`I*M!Qo!4;4$HufH`xU}!83~GRnRs-ch7YGe=#~S&md30+g#&BK
z!xOaoiN{l{@HD=|3QM;c=i7qj#|t-qUUM-1wmqA+yl2YNpO>Fd|43X&-8U+2yX0-3
z5Hc|TQ&R^9d1B`29LE!pm~m>s;xKv2Rn&u$ziL(<y>K&Hf2>R#g|>>&t)NQ&+Ae8Q
z5lit%B$dj3J8d<iLvHtXhW^#ZfX<dw=~#BC!}2^r`@k`u;!az_dEN#WZJ+u)S|G(;
zakwEg8=4+nb$r10K$D5bY0|tk<!L&Yy$@FnpqJQufn#+3eQ3eWiSl4i*rn+9d@z2-
za^329dp*9^|C%0GAMQWD5*YpLH~hxIecK$jH+GyB_K9%c7Tdm<hc`swO%!sIz8=Dh
zoIbHO&6&$~Tj+FOdZMctBc+viep2ldnI_-#xxV-hPG*bUN-jrs`ACm4aT(@nf}ny3
z9JQQ9pG1`ap`f=CJjnFz%O%IUw^BG^1r^yGXyki^uI`(t-J#VSsAhXQje9~m;nenz
zhezeOq1i9O4}<?OLi1V>i5bz5ZzVq{`6B4|wlB?JraQ_?co9I34pVB;^-{Fqm!4y?
zdHCJgL8Wa$s1Y;}=Pp13P^oJ3b$>(vbj?4di=zy9(Wd?8ZgEE*Onv`w*bJtKqFl&P
zeB2y8!VfUjv?WxK=a63-Y+xgu<20u<e_ox~2sDx&-P(;-HVm9P@>Evt#-d_`nlAb7
z{-ud5uj9aVo?i3uWaP}6IA89(e;qs!i?pq&c~!l*-+C$D*6r)R_8;EI-^RrUJ``?7
z2DW|&&;AJ5R33NfL@OlSV#sha&N(RgL9`5V2eNnH#xS^w@DBbrv-wl9nPKFx^ipmO
zk$LQzhWC@kdAZpj9ov3B+N0mm9`yBH0%zck9WkN8WllA=3fT8V>~4Yfv4zj!_qliZ
z$={5~r`c1WpX7~wldlWCRRMwt2r?rj1bJJwj`Qi`em-5K3`4)eWUPuLMN#=2KWbaC
zoXBF+YIjm(GVBKksujL&pd+qG0@)u9FY}`bCX$*pB1zOg%nz@B_Si!D#v2A8k)pme
zxPhXsZAE&1?h5$+^ZQ-*l^;ICT{eNA&HH}3Q!+MY!{hRw-1D>~_Jdd#C_=Xy&d-Zl
z$6SLwyA!~vzjY(Z4!?f=cY~kHsOOLGgTjy6@Xn0AwH!I2eJl4t<zwtEg!LHXBE;pa
zg+GfmAZ2#xUbB52-mrCY@${Cz?E~(KSg4G;097pBxO}{Mn`Hj`^?cvDnSO2BY=@Ts
z5CnMtV?5+mzwH^haceaK&)w5BU%Bk&hKIovc}*uFYBE^MOolTifr~OFkG8`KmiV^o
zjNScTw=NtO)AH{;RvvNUcdFI#sJP8JwJ$Btnvw4!?5F*T>*1q0!Yj+`_j=rx!=1Mq
zj~dQ=4>qPptMK$OF;vIK%<vXWjISxnPuRDy+T+680s3<GxxB%^k&AswJ7&PJtq-pk
z-})gh)SOg&;s}3KRHL-|^_<mQoK02=>bdTiM@mR+Fmm!wJLf|k#OEs4>Ipfqy`9b3
z_GM<-x>dL<Zg}zppFh<gpN)DJ--{uUJC2u>4H|JhzQgFvEeWV@4||9E?8Ec4Z=Ovf
zYy`V@ov_s8AGx#u5^?4B?%t5g++7Tu51+^Txs&m8oefJPBMvr^>00XNSya%YRF_hp
zX78M4HO{`=+sQCG`LH<o{N|9S%emKTmF)HjM!$!ccN-yymM!`So{VU!^AiyVcUhe{
z{WcF5PbG}cRv(w0QQdQ}%Bs4rGAGy6^`F~@PJScqZ|Doa^S_8lztI+M4tjggv?~~j
z)q|}~LRZF~hSc4C7{&U4h|skxl>PO9?dYuzimv_9l=FE*z{_?bE!`WA6ZJn;tv~A2
zK|hW;HG9t+Z|;X;ri=LA0~#HePKJ-C?AFkK64-z4-s?FJHgEa^xt$aIpr+G?8|_zy
z&=D&ag^Dr2Hs0sk(I^qya}b(BMc5a5h|9yY)I+ncY3JMMdpG7|HTq-4rfhc?ug~oR
zthFNt#@4xrvEw_+=<l~0G9}gGSz5A+4WW9ezn=*B9*Wr`0GUPk=aH+l(jsmGH|LkI
z6Zx~-9JJp3D}vl(Xr|fHKKdZNo_#oHr0wwj<>wwiQ=?qM&fRKfn}W06%Xz|&9liD_
z$=aTP>+W;69?{&LsJuB8sjz}e)7_x5E>niKl1$gku}~{1E9Yo&u+q1zzogHt8AyH&
z2XJ$wGhd4Z#6n$*`K;aAZB^Y1pSL#e<yw$YI3hMo?mWGrh;Oyy{P`*aDKUA{*adS9
z2?qcs4(#k_@RZzkS9?DS3xRL8{mwgV{~EChO|OwzqXJ<@VH8BY2ZNjv#{Eb?>b{5>
znSkzA1U13UEF(r_r7hd!UJ$vTMCW#EuL~tqz6`lAw7<h}=#G3-3{A?wGfW95`*?x?
zv}hsL4fO17)O7W+LP=A>9F|d4rj0_okP6`8T?pKJHn5?GeXUf}ce}@bVt*4}8@K7+
znW@RXPHqi}!g;GH%$`)YeUoYIh=<(&x15p^TZo25@)il(UZa{Z38_25Lc@f-oGp^|
zVd3ioc1_7}psxl_4`29jE#n?1Oo`i1df^Vsd?bGM*(2puU&ha3S(uUiagO<WZbl#v
zpE^uF@c~KuzcC$j{Uj~IuQdOKAgW0jzMzhAeP?L2?C%^$`$_bH?l00V`u6ZW@o=co
zs?}&LWxO>OKK^nGqQlr7&P_@q1u+^O5gl4l*p_gSP`KQY@G+w~5&6<Js#3$qj6|l9
zhvcl5LiQ_B`hZtMRq5%kG{OETg-{@5(|(+PO#fZMO0e~w-)R4P=f060OI{+2ZRyhR
z{t)DM3;EX56uG+bm$<`*YdGfjJF|^DuNU)4UfjYVR*fxk)p@og?fhA?#pAho5x~{_
ztIE_>i@i`VWn|GXP3Mr>A<yCHV5|kU)y36;{m(}7sO<OwxN1R#kf_aCuny47-HNSb
zCVhbQuMftDmbuh=<gq=Woe#EA<`U{$sgSWCnXXs~t+|hl$9LslS(b@Wf78I+x88>T
zN(*=0IWzP4Ky@u`P_rv_ZOebz?Y{-u*;MOAs&sd?aF>+Mn{$`Oe=D@Jf$J8#1R+N1
zk{dHw30JLUt^KQ8WU90HdoBQc^E*Qu;Tw!DiAhKgw4w4bKo7(hr6RURMZW${t*6!V
z*J$dYhS{|&!Zsd+rv~RJQZ2_fn!R#YHB37|ZBl^dskBU2C!rfvO~*G{V(Z4MwF*;D
z?Jshkh+AMI;qPhOye|a)t7KQIU?s^o<&WBsxdVmS7!to92dlyOgDSC9&ejgw3(w&2
zd=!wPx&bu?T`afcRhPdTuV3`~v-qlu=aOmE26?mzjj!R?;O@_(tKPWSGF9!U%Wxq%
zkEFS@vUh$aM>z8oK^G9@=D~jOTzEnzR!(LXElMbQ+5Z?<hE65~6e3(q3`{HxjEw9|
z|9`VNcph8@T!BH1MTD6{kby&pRg8n3Rg8&)g`I;zj6+aFNJx~CMVOC(=l?9x{U0(H
z0VjLPPY-Y^0Cz#sN>vqS`1Zcr_6L`pzP5RuQ!WYx%9^@4CUlX>g#dKX0#tGt>rx()
zv4bia4FnyFMShq><m`ZX{&x$D7~0T4sYsL}a!3$eBzE%9Kx!?6*z|Mr>w)VP_jT9%
z)axd)5X3hrAxUpv54PU|j-#3&lY{~HvRbS|Y|IN2P)%#5;Y+lf>}ZWCoG1Utx><u%
zo{Oa>1Vvivr&SuDT8nj-_j9R7%RJBpnsAZ1CZPCb|5P(Ikdg!Rcn5^1rvq)yBsD;q
zhZ#C{DjsPLHPy_4`jo|jHwUXf0%l!ez7Vz$lVds&C(PuI5otW_Y*qsnpWFhLt=II#
zm$nuZa1p{}(<=Sa1?;Gu`q)nP=-}hqT2Cc|wdVCs0(5J!pL+CGe!u78!}SPa_zE*X
z2-{&q_b76J5ATA(^g~4UVPpo{r}{R`80S6ee+DV%CS?XP=kn%DP;#G?NH&hX!X8iY
zP5eWPcVUG`&1JI?HWR3@hnwmbW`IT)&~||Y==yPLyQc~`MP0+M*4P)d|7A5P=*^+<
zqT<R|$BMc66ERGaPYi7>btvl;tm;`9zCr6&)ZaY1-u`;O=Y_49)Y49*=vLRjsG3#K
zFW8b}&t&2deU1+*6P4BMzkIcwI^XM{``=~DU^jvL4IK8~8LbOz&W;!rSj4yTH=Q|a
z(rQ_*q&&}qWfN%EwC?81P2DnFa#HWd%Wab5&T`?pe7KzXWoforc`xZ5V8oT>bat{Z
zYmq-MQW7SIoQ1?i>hyMuGI|>Eoc_4ETvb}2EM9UVH=Je4P3zKj{vXOnILn{omp#-T
z%D`!`GIR-Ib%8*M0IazgP&@&yzr^YsP~EzeEyZc=kYt!mX*&3P{nr7v(%)ky`4odP
zZPMwl6!AizrMd6FT;hNPvc=c@f+F0NW|fd(GKEPt;onw%kSkDGxx$9nb{-{MOl^5E
zytSY%Li7*4bSu_u?-kijF{_*FQl7LH*Gg+<lCOqEpUE<`6pz|WfVu*g!Ds_OiQD++
zl(XpPIO{YOZvDm{yi`E)yNjO1$elAlf!p-)A}%FP1H2c{DUN^7=Q~Y6CD2phHurtB
z%t2D?Tq<HU^>wlpAyPki=dxH11`P+G=9I?heA3)K7#&~+FY`s|05lE-Y<C~Sap-tq
zG>am7MlWLpaq8%*L^Vl7HQQkw?u>U!;UnDgqFbMwcdu0=G2NnF9}IVs@*i$$Kc+$}
z|L0ZY+F_cEn9GSdGFiqc=0w{3aiS<7&@4dh{^tG~qBq_Je19zMr{HdZ6F{;sLy?k-
I%8Nn$KVz=0umAu6

diff --git a/articles/index.html b/articles/index.html
index 604f542..5dc0b5d 100644
--- a/articles/index.html
+++ b/articles/index.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/authors.html b/authors.html
index 0f5a8bc..5877db0 100644
--- a/authors.html
+++ b/authors.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
@@ -64,13 +64,13 @@ <h2 id="citation">Citation</h2>
 
       <p>Gruber L (2024).
 <em>bayesianVARs: MCMC Estimation of Bayesian Vectorautoregressions</em>.
-R package version 0.1.1, https://luisgruber.github.io/bayesianVARs/, <a href="https://github.com/luisgruber/bayesianVARs" class="external-link">https://github.com/luisgruber/bayesianVARs</a>. 
+R package version 0.1.0.9000, https://luisgruber.github.io/bayesianVARs/, <a href="https://github.com/luisgruber/bayesianVARs" class="external-link">https://github.com/luisgruber/bayesianVARs</a>. 
 </p>
       <pre>@Manual{,
   title = {bayesianVARs: MCMC Estimation of Bayesian Vectorautoregressions},
   author = {Luis Gruber},
   year = {2024},
-  note = {R package version 0.1.1, https://luisgruber.github.io/bayesianVARs/},
+  note = {R package version 0.1.0.9000, https://luisgruber.github.io/bayesianVARs/},
   url = {https://github.com/luisgruber/bayesianVARs},
 }</pre>
     </div>
diff --git a/index.html b/index.html
index 2f94652..fd9c521 100644
--- a/index.html
+++ b/index.html
@@ -45,7 +45,7 @@
     
     <a class="navbar-brand me-2" href="index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/news/index.html b/news/index.html
index 282db6d..25f5a41 100644
--- a/news/index.html
+++ b/news/index.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
@@ -51,10 +51,8 @@
     </div>
 
     <div class="section level2">
-<h2 class="pkg-version" data-toc-text="0.1.1" id="bayesianvars-011">bayesianVARs 0.1.1<a class="anchor" aria-label="anchor" href="#bayesianvars-011"></a></h2><p class="text-muted">CRAN release: 2024-01-17</p>
-<ul><li>Fixed clang-UBSAN issue.</li>
-<li>Fixed undefined figure references in vignette.</li>
-</ul></div>
+<h2 class="pkg-version" data-toc-text="development version" id="bayesianvars-development-version">bayesianVARs (development version)<a class="anchor" aria-label="anchor" href="#bayesianvars-development-version"></a></h2>
+</div>
     <div class="section level2">
 <h2 class="pkg-version" data-toc-text="0.1.0" id="bayesianvars-010">bayesianVARs 0.1.0<a class="anchor" aria-label="anchor" href="#bayesianvars-010"></a></h2><p class="text-muted">CRAN release: 2024-01-13</p>
 <ul><li>Initial CRAN submission.</li>
diff --git a/pkgdown.yml b/pkgdown.yml
index 7a4b1b4..6e90210 100644
--- a/pkgdown.yml
+++ b/pkgdown.yml
@@ -3,7 +3,7 @@ pkgdown: 2.0.7
 pkgdown_sha: ~
 articles:
   bayesianVARs-vignette: bayesianVARs-vignette.pdf
-last_built: 2024-01-17T22:58Z
+last_built: 2024-01-17T23:02Z
 urls:
   reference: https://luisgruber.github.io/bayesianVARs/reference
   article: https://luisgruber.github.io/bayesianVARs/articles
diff --git a/reference/bvar.html b/reference/bvar.html
index 07fb919..ab3acd7 100644
--- a/reference/bvar.html
+++ b/reference/bvar.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/coef.html b/reference/coef.html
index d399665..7eb64ee 100644
--- a/reference/coef.html
+++ b/reference/coef.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/fitted.bayesianVARs_bvar.html b/reference/fitted.bayesianVARs_bvar.html
index fc285f9..1305f28 100644
--- a/reference/fitted.bayesianVARs_bvar.html
+++ b/reference/fitted.bayesianVARs_bvar.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/index.html b/reference/index.html
index ff803ac..c8c76c6 100644
--- a/reference/index.html
+++ b/reference/index.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/my_gig.html b/reference/my_gig.html
index 6c104e4..7bd4033 100644
--- a/reference/my_gig.html
+++ b/reference/my_gig.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/pairs_predict.html b/reference/pairs_predict.html
index 6f52faa..7dd0fd9 100644
--- a/reference/pairs_predict.html
+++ b/reference/pairs_predict.html
@@ -14,7 +14,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/plot.bayesianVARs_bvar.html b/reference/plot.bayesianVARs_bvar.html
index db2277a..c6c745f 100644
--- a/reference/plot.bayesianVARs_bvar.html
+++ b/reference/plot.bayesianVARs_bvar.html
@@ -12,7 +12,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/plot.bayesianVARs_fitted.html b/reference/plot.bayesianVARs_fitted.html
index be3ea29..dfce71c 100644
--- a/reference/plot.bayesianVARs_fitted.html
+++ b/reference/plot.bayesianVARs_fitted.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/plot.bayesianVARs_predict.html b/reference/plot.bayesianVARs_predict.html
index 5378f12..2205159 100644
--- a/reference/plot.bayesianVARs_predict.html
+++ b/reference/plot.bayesianVARs_predict.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/posterior_heatmap.html b/reference/posterior_heatmap.html
index 525fe1e..903e446 100644
--- a/reference/posterior_heatmap.html
+++ b/reference/posterior_heatmap.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/predict.bayesianVARs_bvar.html b/reference/predict.bayesianVARs_bvar.html
index 586d6ed..4c97508 100644
--- a/reference/predict.bayesianVARs_bvar.html
+++ b/reference/predict.bayesianVARs_bvar.html
@@ -14,7 +14,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/print.bayesianVARs_bvar.html b/reference/print.bayesianVARs_bvar.html
index 3d97a1c..bbfd454 100644
--- a/reference/print.bayesianVARs_bvar.html
+++ b/reference/print.bayesianVARs_bvar.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/print.bayesianVARs_predict.html b/reference/print.bayesianVARs_predict.html
index a1fdb88..2eecf65 100644
--- a/reference/print.bayesianVARs_predict.html
+++ b/reference/print.bayesianVARs_predict.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/print.summary.bayesianVARs_bvar.html b/reference/print.summary.bayesianVARs_bvar.html
index 09a0216..1fef41a 100644
--- a/reference/print.summary.bayesianVARs_bvar.html
+++ b/reference/print.summary.bayesianVARs_bvar.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/print.summary.bayesianVARs_predict.html b/reference/print.summary.bayesianVARs_predict.html
index 5ba4d7f..7b31991 100644
--- a/reference/print.summary.bayesianVARs_predict.html
+++ b/reference/print.summary.bayesianVARs_predict.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/specify_prior_phi.html b/reference/specify_prior_phi.html
index 6a3d7d5..522266a 100644
--- a/reference/specify_prior_phi.html
+++ b/reference/specify_prior_phi.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/specify_prior_sigma.html b/reference/specify_prior_sigma.html
index 6c909d4..32748dd 100644
--- a/reference/specify_prior_sigma.html
+++ b/reference/specify_prior_sigma.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/stable_bvar.html b/reference/stable_bvar.html
index 3028e44..9b28ee5 100644
--- a/reference/stable_bvar.html
+++ b/reference/stable_bvar.html
@@ -16,7 +16,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/sub-.bayesianVARs_coef.html b/reference/sub-.bayesianVARs_coef.html
index 9816e59..f32ba72 100644
--- a/reference/sub-.bayesianVARs_coef.html
+++ b/reference/sub-.bayesianVARs_coef.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/sub-.bayesianVARs_draws.html b/reference/sub-.bayesianVARs_draws.html
index 6d575a7..85ec538 100644
--- a/reference/sub-.bayesianVARs_draws.html
+++ b/reference/sub-.bayesianVARs_draws.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/summary.bayesianVARs_bvar.html b/reference/summary.bayesianVARs_bvar.html
index d03bc1b..6b8e3c0 100644
--- a/reference/summary.bayesianVARs_bvar.html
+++ b/reference/summary.bayesianVARs_bvar.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/summary.bayesianVARs_draws.html b/reference/summary.bayesianVARs_draws.html
index 217a0fe..dbc167b 100644
--- a/reference/summary.bayesianVARs_draws.html
+++ b/reference/summary.bayesianVARs_draws.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/summary.bayesianVARs_predict.html b/reference/summary.bayesianVARs_predict.html
index 08c0f83..28419ba 100644
--- a/reference/summary.bayesianVARs_predict.html
+++ b/reference/summary.bayesianVARs_predict.html
@@ -10,7 +10,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/usmacro_growth.html b/reference/usmacro_growth.html
index b4ddb9b..9d73ad3 100644
--- a/reference/usmacro_growth.html
+++ b/reference/usmacro_growth.html
@@ -16,7 +16,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/reference/vcov.bayesianVARs_bvar.html b/reference/vcov.bayesianVARs_bvar.html
index 3c18a35..27f2380 100644
--- a/reference/vcov.bayesianVARs_bvar.html
+++ b/reference/vcov.bayesianVARs_bvar.html
@@ -18,7 +18,7 @@
     
     <a class="navbar-brand me-2" href="../index.html">bayesianVARs</a>
 
-    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.1</small>
+    <small class="nav-text text-muted me-auto" data-bs-toggle="tooltip" data-bs-placement="bottom" title="">0.1.0.9000</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">
diff --git a/search.json b/search.json
index 7bbc1c1..e62a276 100644
--- a/search.json
+++ b/search.json
@@ -1 +1 @@
-[{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":null,"dir":"","previous_headings":"","what":"GNU General Public License","title":"GNU General Public License","text":"Version 3, 29 June 2007Copyright © 2007 Free Software Foundation, Inc. <http://fsf.org/> Everyone permitted copy distribute verbatim copies license document, changing allowed.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"preamble","dir":"","previous_headings":"","what":"Preamble","title":"GNU General Public License","text":"GNU General Public License free, copyleft license software kinds works. licenses software practical works designed take away freedom share change works. contrast, GNU General Public License intended guarantee freedom share change versions program–make sure remains free software users. , Free Software Foundation, use GNU General Public License software; applies also work released way authors. can apply programs, . speak free software, referring freedom, price. General Public Licenses designed make sure freedom distribute copies free software (charge wish), receive source code can get want , can change software use pieces new free programs, know can things. protect rights, need prevent others denying rights asking surrender rights. Therefore, certain responsibilities distribute copies software, modify : responsibilities respect freedom others. example, distribute copies program, whether gratis fee, must pass recipients freedoms received. must make sure , , receive can get source code. must show terms know rights. Developers use GNU GPL protect rights two steps: (1) assert copyright software, (2) offer License giving legal permission copy, distribute /modify . developers’ authors’ protection, GPL clearly explains warranty free software. users’ authors’ sake, GPL requires modified versions marked changed, problems attributed erroneously authors previous versions. devices designed deny users access install run modified versions software inside , although manufacturer can . fundamentally incompatible aim protecting users’ freedom change software. systematic pattern abuse occurs area products individuals use, precisely unacceptable. Therefore, designed version GPL prohibit practice products. problems arise substantially domains, stand ready extend provision domains future versions GPL, needed protect freedom users. Finally, every program threatened constantly software patents. States allow patents restrict development use software general-purpose computers, , wish avoid special danger patents applied free program make effectively proprietary. prevent , GPL assures patents used render program non-free. precise terms conditions copying, distribution modification follow.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_0-definitions","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"0. Definitions","title":"GNU General Public License","text":"“License” refers version 3 GNU General Public License. “Copyright” also means copyright-like laws apply kinds works, semiconductor masks. “Program” refers copyrightable work licensed License. licensee addressed “”. “Licensees” “recipients” may individuals organizations. “modify” work means copy adapt part work fashion requiring copyright permission, making exact copy. resulting work called “modified version” earlier work work “based ” earlier work. “covered work” means either unmodified Program work based Program. “propagate” work means anything , without permission, make directly secondarily liable infringement applicable copyright law, except executing computer modifying private copy. Propagation includes copying, distribution (without modification), making available public, countries activities well. “convey” work means kind propagation enables parties make receive copies. Mere interaction user computer network, transfer copy, conveying. interactive user interface displays “Appropriate Legal Notices” extent includes convenient prominently visible feature (1) displays appropriate copyright notice, (2) tells user warranty work (except extent warranties provided), licensees may convey work License, view copy License. interface presents list user commands options, menu, prominent item list meets criterion.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_1-source-code","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"1. Source Code","title":"GNU General Public License","text":"“source code” work means preferred form work making modifications . “Object code” means non-source form work. “Standard Interface” means interface either official standard defined recognized standards body, , case interfaces specified particular programming language, one widely used among developers working language. “System Libraries” executable work include anything, work whole, () included normal form packaging Major Component, part Major Component, (b) serves enable use work Major Component, implement Standard Interface implementation available public source code form. “Major Component”, context, means major essential component (kernel, window system, ) specific operating system () executable work runs, compiler used produce work, object code interpreter used run . “Corresponding Source” work object code form means source code needed generate, install, (executable work) run object code modify work, including scripts control activities. However, include work’s System Libraries, general-purpose tools generally available free programs used unmodified performing activities part work. example, Corresponding Source includes interface definition files associated source files work, source code shared libraries dynamically linked subprograms work specifically designed require, intimate data communication control flow subprograms parts work. Corresponding Source need include anything users can regenerate automatically parts Corresponding Source. Corresponding Source work source code form work.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_2-basic-permissions","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"2. Basic Permissions","title":"GNU General Public License","text":"rights granted License granted term copyright Program, irrevocable provided stated conditions met. License explicitly affirms unlimited permission run unmodified Program. output running covered work covered License output, given content, constitutes covered work. License acknowledges rights fair use equivalent, provided copyright law. may make, run propagate covered works convey, without conditions long license otherwise remains force. may convey covered works others sole purpose make modifications exclusively , provide facilities running works, provided comply terms License conveying material control copyright. thus making running covered works must exclusively behalf, direction control, terms prohibit making copies copyrighted material outside relationship . Conveying circumstances permitted solely conditions stated . Sublicensing allowed; section 10 makes unnecessary.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_3-protecting-users-legal-rights-from-anti-circumvention-law","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"3. Protecting Users’ Legal Rights From Anti-Circumvention Law","title":"GNU General Public License","text":"covered work shall deemed part effective technological measure applicable law fulfilling obligations article 11 WIPO copyright treaty adopted 20 December 1996, similar laws prohibiting restricting circumvention measures. convey covered work, waive legal power forbid circumvention technological measures extent circumvention effected exercising rights License respect covered work, disclaim intention limit operation modification work means enforcing, work’s users, third parties’ legal rights forbid circumvention technological measures.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_4-conveying-verbatim-copies","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"4. Conveying Verbatim Copies","title":"GNU General Public License","text":"may convey verbatim copies Program’s source code receive , medium, provided conspicuously appropriately publish copy appropriate copyright notice; keep intact notices stating License non-permissive terms added accord section 7 apply code; keep intact notices absence warranty; give recipients copy License along Program. may charge price price copy convey, may offer support warranty protection fee.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_5-conveying-modified-source-versions","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"5. Conveying Modified Source Versions","title":"GNU General Public License","text":"may convey work based Program, modifications produce Program, form source code terms section 4, provided also meet conditions: ) work must carry prominent notices stating modified , giving relevant date. b) work must carry prominent notices stating released License conditions added section 7. requirement modifies requirement section 4 “keep intact notices”. c) must license entire work, whole, License anyone comes possession copy. License therefore apply, along applicable section 7 additional terms, whole work, parts, regardless packaged. License gives permission license work way, invalidate permission separately received . d) work interactive user interfaces, must display Appropriate Legal Notices; however, Program interactive interfaces display Appropriate Legal Notices, work need make . compilation covered work separate independent works, nature extensions covered work, combined form larger program, volume storage distribution medium, called “aggregate” compilation resulting copyright used limit access legal rights compilation’s users beyond individual works permit. Inclusion covered work aggregate cause License apply parts aggregate.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_6-conveying-non-source-forms","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"6. Conveying Non-Source Forms","title":"GNU General Public License","text":"may convey covered work object code form terms sections 4 5, provided also convey machine-readable Corresponding Source terms License, one ways: ) Convey object code , embodied , physical product (including physical distribution medium), accompanied Corresponding Source fixed durable physical medium customarily used software interchange. b) Convey object code , embodied , physical product (including physical distribution medium), accompanied written offer, valid least three years valid long offer spare parts customer support product model, give anyone possesses object code either (1) copy Corresponding Source software product covered License, durable physical medium customarily used software interchange, price reasonable cost physically performing conveying source, (2) access copy Corresponding Source network server charge. c) Convey individual copies object code copy written offer provide Corresponding Source. alternative allowed occasionally noncommercially, received object code offer, accord subsection 6b. d) Convey object code offering access designated place (gratis charge), offer equivalent access Corresponding Source way place charge. need require recipients copy Corresponding Source along object code. place copy object code network server, Corresponding Source may different server (operated third party) supports equivalent copying facilities, provided maintain clear directions next object code saying find Corresponding Source. Regardless server hosts Corresponding Source, remain obligated ensure available long needed satisfy requirements. e) Convey object code using peer--peer transmission, provided inform peers object code Corresponding Source work offered general public charge subsection 6d. separable portion object code, whose source code excluded Corresponding Source System Library, need included conveying object code work. “User Product” either (1) “consumer product”, means tangible personal property normally used personal, family, household purposes, (2) anything designed sold incorporation dwelling. determining whether product consumer product, doubtful cases shall resolved favor coverage. particular product received particular user, “normally used” refers typical common use class product, regardless status particular user way particular user actually uses, expects expected use, product. product consumer product regardless whether product substantial commercial, industrial non-consumer uses, unless uses represent significant mode use product. “Installation Information” User Product means methods, procedures, authorization keys, information required install execute modified versions covered work User Product modified version Corresponding Source. information must suffice ensure continued functioning modified object code case prevented interfered solely modification made. convey object code work section , , specifically use , User Product, conveying occurs part transaction right possession use User Product transferred recipient perpetuity fixed term (regardless transaction characterized), Corresponding Source conveyed section must accompanied Installation Information. requirement apply neither third party retains ability install modified object code User Product (example, work installed ROM). requirement provide Installation Information include requirement continue provide support service, warranty, updates work modified installed recipient, User Product modified installed. Access network may denied modification materially adversely affects operation network violates rules protocols communication across network. Corresponding Source conveyed, Installation Information provided, accord section must format publicly documented (implementation available public source code form), must require special password key unpacking, reading copying.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_7-additional-terms","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"7. Additional Terms","title":"GNU General Public License","text":"“Additional permissions” terms supplement terms License making exceptions one conditions. Additional permissions applicable entire Program shall treated though included License, extent valid applicable law. additional permissions apply part Program, part may used separately permissions, entire Program remains governed License without regard additional permissions. convey copy covered work, may option remove additional permissions copy, part . (Additional permissions may written require removal certain cases modify work.) may place additional permissions material, added covered work, can give appropriate copyright permission. Notwithstanding provision License, material add covered work, may (authorized copyright holders material) supplement terms License terms: ) Disclaiming warranty limiting liability differently terms sections 15 16 License; b) Requiring preservation specified reasonable legal notices author attributions material Appropriate Legal Notices displayed works containing ; c) Prohibiting misrepresentation origin material, requiring modified versions material marked reasonable ways different original version; d) Limiting use publicity purposes names licensors authors material; e) Declining grant rights trademark law use trade names, trademarks, service marks; f) Requiring indemnification licensors authors material anyone conveys material (modified versions ) contractual assumptions liability recipient, liability contractual assumptions directly impose licensors authors. non-permissive additional terms considered “restrictions” within meaning section 10. Program received , part , contains notice stating governed License along term restriction, may remove term. license document contains restriction permits relicensing conveying License, may add covered work material governed terms license document, provided restriction survive relicensing conveying. add terms covered work accord section, must place, relevant source files, statement additional terms apply files, notice indicating find applicable terms. Additional terms, permissive non-permissive, may stated form separately written license, stated exceptions; requirements apply either way.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_8-termination","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"8. Termination","title":"GNU General Public License","text":"may propagate modify covered work except expressly provided License. attempt otherwise propagate modify void, automatically terminate rights License (including patent licenses granted third paragraph section 11). However, cease violation License, license particular copyright holder reinstated () provisionally, unless copyright holder explicitly finally terminates license, (b) permanently, copyright holder fails notify violation reasonable means prior 60 days cessation. Moreover, license particular copyright holder reinstated permanently copyright holder notifies violation reasonable means, first time received notice violation License (work) copyright holder, cure violation prior 30 days receipt notice. Termination rights section terminate licenses parties received copies rights License. rights terminated permanently reinstated, qualify receive new licenses material section 10.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_9-acceptance-not-required-for-having-copies","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"9. Acceptance Not Required for Having Copies","title":"GNU General Public License","text":"required accept License order receive run copy Program. Ancillary propagation covered work occurring solely consequence using peer--peer transmission receive copy likewise require acceptance. However, nothing License grants permission propagate modify covered work. actions infringe copyright accept License. Therefore, modifying propagating covered work, indicate acceptance License .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_10-automatic-licensing-of-downstream-recipients","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"10. Automatic Licensing of Downstream Recipients","title":"GNU General Public License","text":"time convey covered work, recipient automatically receives license original licensors, run, modify propagate work, subject License. responsible enforcing compliance third parties License. “entity transaction” transaction transferring control organization, substantially assets one, subdividing organization, merging organizations. propagation covered work results entity transaction, party transaction receives copy work also receives whatever licenses work party’s predecessor interest give previous paragraph, plus right possession Corresponding Source work predecessor interest, predecessor can get reasonable efforts. may impose restrictions exercise rights granted affirmed License. example, may impose license fee, royalty, charge exercise rights granted License, may initiate litigation (including cross-claim counterclaim lawsuit) alleging patent claim infringed making, using, selling, offering sale, importing Program portion .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_11-patents","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"11. Patents","title":"GNU General Public License","text":"“contributor” copyright holder authorizes use License Program work Program based. work thus licensed called contributor’s “contributor version”. contributor’s “essential patent claims” patent claims owned controlled contributor, whether already acquired hereafter acquired, infringed manner, permitted License, making, using, selling contributor version, include claims infringed consequence modification contributor version. purposes definition, “control” includes right grant patent sublicenses manner consistent requirements License. contributor grants non-exclusive, worldwide, royalty-free patent license contributor’s essential patent claims, make, use, sell, offer sale, import otherwise run, modify propagate contents contributor version. following three paragraphs, “patent license” express agreement commitment, however denominated, enforce patent (express permission practice patent covenant sue patent infringement). “grant” patent license party means make agreement commitment enforce patent party. convey covered work, knowingly relying patent license, Corresponding Source work available anyone copy, free charge terms License, publicly available network server readily accessible means, must either (1) cause Corresponding Source available, (2) arrange deprive benefit patent license particular work, (3) arrange, manner consistent requirements License, extend patent license downstream recipients. “Knowingly relying” means actual knowledge , patent license, conveying covered work country, recipient’s use covered work country, infringe one identifiable patents country reason believe valid. , pursuant connection single transaction arrangement, convey, propagate procuring conveyance , covered work, grant patent license parties receiving covered work authorizing use, propagate, modify convey specific copy covered work, patent license grant automatically extended recipients covered work works based . patent license “discriminatory” include within scope coverage, prohibits exercise , conditioned non-exercise one rights specifically granted License. may convey covered work party arrangement third party business distributing software, make payment third party based extent activity conveying work, third party grants, parties receive covered work , discriminatory patent license () connection copies covered work conveyed (copies made copies), (b) primarily connection specific products compilations contain covered work, unless entered arrangement, patent license granted, prior 28 March 2007. Nothing License shall construed excluding limiting implied license defenses infringement may otherwise available applicable patent law.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_12-no-surrender-of-others-freedom","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"12. No Surrender of Others’ Freedom","title":"GNU General Public License","text":"conditions imposed (whether court order, agreement otherwise) contradict conditions License, excuse conditions License. convey covered work satisfy simultaneously obligations License pertinent obligations, consequence may convey . example, agree terms obligate collect royalty conveying convey Program, way satisfy terms License refrain entirely conveying Program.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_13-use-with-the-gnu-affero-general-public-license","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"13. Use with the GNU Affero General Public License","title":"GNU General Public License","text":"Notwithstanding provision License, permission link combine covered work work licensed version 3 GNU Affero General Public License single combined work, convey resulting work. terms License continue apply part covered work, special requirements GNU Affero General Public License, section 13, concerning interaction network apply combination .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_14-revised-versions-of-this-license","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"14. Revised Versions of this License","title":"GNU General Public License","text":"Free Software Foundation may publish revised /new versions GNU 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 certain numbered version GNU General Public License “later version” applies , option following terms conditions either numbered version later version published Free Software Foundation. Program specify version number GNU General Public License, may choose version ever published Free Software Foundation. Program specifies proxy can decide future versions GNU General Public License can used, proxy’s public statement acceptance version permanently authorizes choose version Program. Later license versions may give additional different permissions. However, additional obligations imposed author copyright holder result choosing follow later version.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_15-disclaimer-of-warranty","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"15. Disclaimer of Warranty","title":"GNU General Public License","text":"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.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_16-limitation-of-liability","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"16. Limitation of Liability","title":"GNU General Public License","text":"EVENT UNLESS REQUIRED APPLICABLE LAW AGREED WRITING COPYRIGHT HOLDER, PARTY MODIFIES /CONVEYS 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.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_17-interpretation-of-sections-15-and-16","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"17. Interpretation of Sections 15 and 16","title":"GNU General Public License","text":"disclaimer warranty limitation liability provided given local legal effect according terms, reviewing courts shall apply local law closely approximates absolute waiver civil liability connection Program, unless warranty assumption liability accompanies copy Program return fee. END TERMS CONDITIONS","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/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 state exclusion warranty; file least “copyright” line pointer full notice found. Also add information contact electronic paper mail. program terminal interaction, make output short notice like starts interactive mode: hypothetical commands show w show c show appropriate parts General Public License. course, program’s commands might different; GUI interface, use “box”. also get employer (work programmer) school, , sign “copyright disclaimer” program, necessary. information , apply follow GNU GPL, see <http://www.gnu.org/licenses/>. GNU 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. first, please read <http://www.gnu.org/philosophy/--lgpl.html>.","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 3 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, see <http://www.gnu.org/licenses/>. <program>  Copyright (C) <year>  <name of author> This program 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."},{"path":"https://luisgruber.github.io/bayesianVARs/authors.html","id":null,"dir":"","previous_headings":"","what":"Authors","title":"Authors and Citation","text":"Luis Gruber. Copyright holder, author, maintainer.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/authors.html","id":"citation","dir":"","previous_headings":"","what":"Citation","title":"Authors and Citation","text":"Gruber L (2024). bayesianVARs: MCMC Estimation Bayesian Vectorautoregressions. R package version 0.1.1, https://luisgruber.github.io/bayesianVARs/, https://github.com/luisgruber/bayesianVARs.","code":"@Manual{,   title = {bayesianVARs: MCMC Estimation of Bayesian Vectorautoregressions},   author = {Luis Gruber},   year = {2024},   note = {R package version 0.1.1, https://luisgruber.github.io/bayesianVARs/},   url = {https://github.com/luisgruber/bayesianVARs}, }"},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"bayesianvars-","dir":"","previous_headings":"","what":"MCMC Estimation of Bayesian Vectorautoregressions","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"Estimation Bayesian vectorautoregressions /without stochastic volatility. Implements several modern hierarchical shrinkage priors, amongst Dirichlet-Laplace prior (DL), hierarchical Minnesota prior (HM), Horseshoe prior (HS), normal-gamma prior (NG), R2-induced-Dirichlet-decomposition prior (R2D2) stochastic search variable selection prior (SSVS). Concerning error-term, user can either specify order-invariant factor structure order-variant cholesky structure.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"installation","dir":"","previous_headings":"","what":"Installation","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"Install CRAN version: Install latest development version directly GitHub:","code":"install.packages(\"bayesianVARs\") devtools::install_github(\"luisgruber/bayesianVARs\")"},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"usage","dir":"","previous_headings":"","what":"Usage","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"main workhorse conduct Bayesian inference vectorautoregression models package function bvar(). features: Prediction, plotting, extraction model parameters extraction fitted values usual generic functions predict(), plot(), coef(), vcov() fitted(). Configure prior distributions helper functions specify_prior_phi() specify_prior_sigma().","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"demonstration","dir":"","previous_headings":"","what":"Demonstration","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"","code":"set.seed(537) # load package library(bayesianVARs)  # Load data train_data <-100 * usmacro_growth[1:237,c(\"GDPC1\", \"PCECC96\", \"GPDIC1\", \"AWHMAN\", \"GDPCTPI\", \"CES2000000008x\", \"FEDFUNDS\", \"GS10\", \"EXUSUKx\", \"S&P 500\")] test_data <-100 * usmacro_growth[238:241,c(\"GDPC1\", \"PCECC96\", \"GPDIC1\", \"AWHMAN\", \"GDPCTPI\", \"CES2000000008x\", \"FEDFUNDS\", \"GS10\", \"EXUSUKx\", \"S&P 500\")]                                     # Estimate model using default prior settings mod <- bvar(train_data, lags = 2L, draws = 2000, burnin = 1000, sv_keep = \"all\")  # Out of sample prediction and log-predictive-likelihood evaluation pred <- predict(mod, ahead = 1:4, LPL = TRUE, Y_obs = test_data)  # Visualize in-sample fit plus out-of-sample prediction intervals plot(mod, predictions = pred)"},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"documentation","dir":"","previous_headings":"","what":"Documentation","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"bayesianVARs - Shrinkage Priors Bayesian Vectorautoregressions R","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"bvar simulates joint posterior distribution parameters latent variables returns posterior draws.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"","code":"bvar(   data,   lags = 1L,   draws = 1000L,   burnin = 1000L,   thin = 1L,   prior_intercept = 100,   prior_phi = specify_prior_phi(data = data, lags = lags, prior = \"HS\"),   prior_sigma = specify_prior_sigma(data = data, type = \"factor\", quiet = TRUE),   sv_keep = \"last\",   quiet = FALSE,   startvals = list(),   expert = list() )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"data Data matrix (can time series object). \\(M\\) columns assumed contain single time-series length \\(T\\). lags Integer indicating order VAR, .e. number lags dependent variables included predictors. draws single integer indicating number draws burnin burnin single integer indicating number draws discarded burnin thin single integer. Every \\(thin\\)th draw stored. Default thin=1L. prior_intercept Either prior_intercept=FALSE constant term (intercept) included. numeric vector length \\(M\\) indicating (fixed) prior variances constant term. single number recycled accordingly. Default prior_intercept=100. prior_phi bayesianVARs_prior_phi object specifying prior reduced form VAR coefficients. Best use constructor specify_prior_phi. prior_sigma bayesianVARs_prior_sigma object specifying prior variance-covariance matrix VAR. Best use constructor specify_prior_sigma. sv_keep String equal \"\" \"last\". case sv_keep = \"last\", default, draws last log-variance \\(h_T\\) stored. quiet logical value indicating whether information progress sampling displayed sampling (default TRUE). startvals optional list starting values. expert optional list expert settings.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"object type bayesianVARs_bvar, list containing following objects: PHI: bayesianVARs_coef object, array, containing posterior draws VAR coefficients (including intercept). U: bayesianVARs_draws object, matrix, containing posterior draws contemporaneous coefficients (cholesky decomposition sigma specified). logvar: bayesianVARs_draws object containing log-variance draws. sv_para: baysesianVARs_draws object containing posterior draws stochastic volatility related parameters. phi_hyperparameter: matrix containing posterior draws hyperparameters conditional normal prior VAR coefficients. u_hyperparameter: matrix containing posterior draws hyperparameters conditional normal prior U (cholesky decomposition sigma specified). bench: Numerical indicating average time took generate one single draw joint posterior distribution parameters. V_prior: array containing posterior draws variances conditional normal prior VAR coefficients. facload: bayesianVARs_draws object, array, containing draws posterior distribution factor loadings matrix (factor decomposition sigma specified). fac: bayesianVARs_draws object, array, containing factor draws posterior distribution (factor decomposition sigma specified). Y: Matrix containing dependent variables used estimation. X matrix containing lagged values dependent variables, .e. covariates. lags: Integer indicating lag order VAR. intercept: Logical indicating whether constant term included. heteroscedastic logical indicating whether heteroscedasticity assumed. Yraw: Matrix containing dependent variables, including initial 'lags' observations. Traw: Integer indicating total number observations. sigma_type: Character specifying decomposition variance-covariance matrix. datamat: Matrix containing 'Y' 'X'. config: List containing information configuration parameters.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"VAR(p) model following form: \\( \\boldsymbol{y}^\\prime_t = \\boldsymbol{\\iota}^\\prime + \\boldsymbol{x}^\\prime_t\\boldsymbol{\\Phi} + \\boldsymbol{\\epsilon}^\\prime_t\\), \\(\\boldsymbol{y}_t\\) \\(M\\)-dimensional vector dependent variables \\(\\boldsymbol{\\epsilon}_t\\) error term dimension. \\(\\boldsymbol{x}_t\\) \\(K=pM\\)-dimensional vector containing lagged/past values dependent variables \\(\\boldsymbol{y}_{t-l}\\) \\(l=1,\\dots,p\\) \\(\\boldsymbol{\\iota}\\) constant term (intercept) dimension \\(M\\times 1\\). reduced-form coefficient matrix \\(\\boldsymbol{\\Phi}\\) dimension \\(K \\times M\\). bvar offers two different specifications errors: user can choose factor stochastic volatility structure cholesky stochastic volatility structure. cases disturbances \\(\\boldsymbol{\\epsilon}_t\\) assumed follow \\(M\\)-dimensional multivariate normal distribution zero mean variance-covariance matrix \\(\\boldsymbol{\\Sigma}_t\\). case cholesky specification \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{U}^{\\prime -1} \\boldsymbol{D}_t \\boldsymbol{U}^{-1}\\), \\(\\boldsymbol{U}^{-1}\\) upper unitriangular (ones diagonal). diagonal matrix \\(\\boldsymbol{D}_t\\) depends upon latent log-variances, .e. \\(\\boldsymbol{D}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\). case factor structure, \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{\\Lambda} \\boldsymbol{V}_t \\boldsymbol{\\Lambda}^\\prime + \\boldsymbol{G}_t\\). diagonal matrices \\(\\boldsymbol{V}_t\\) \\(\\boldsymbol{G}_t\\) depend upon latent log-variances, .e. \\(\\boldsymbol{G}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\) \\(\\boldsymbol{V}_t=diag(exp(h_{M+1,t}),\\dots, exp(h_{M+r,t})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\) \\(h_{M+j,t}\\sim N(\\phi_ih_{M+j,t-1},\\sigma_{M+j}^2)\\) \\(j=1,\\dots,r\\).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"mcmc-algorithm","dir":"Reference","previous_headings":"","what":"MCMC algorithm","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"sample efficiently reduced-form VAR coefficients assuming factor structure errors, equation per equation algorithm Kastner & Huber (2020) implemented. parameters latent variables associated factor-structure sampled using package factorstochvol-package's function update_fsv callable C-level . sample efficiently reduced-form VAR coefficients, assuming cholesky-structure errors, corrected triangular algorithm Carriero et al. (2021) implemented. SV parameters latent variables sampled using package stochvol's update_fast_sv function. precision parameters, .e. free -diagonal elements \\(\\boldsymbol{U}\\), sampled Cogley Sargent (2005).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"Gruber, L. Kastner, G. (2023). Forecasting macroeconomic data Bayesian VARs: Sparse dense? depends! arXiv:2206.04902. Kastner, G. Huber, F. Sparse (2020). Bayesian vector autoregressions huge dimensions. Journal Forecasting. 39, 1142--1165, doi:10.1002/.2680 . Kastner, G. (2019). Sparse Bayesian Time-Varying Covariance Estimation Many Dimensions Journal Econometrics, 210(1), 98--115, doi:10.1016/j.jeconom.2018.11.007 . Carriero, . Chan, J.  Clark, T. E. Marcellino, M. (2021). Corrigendum “Large Bayesian vector autoregressions stochastic volatility non-conjugate priors” [J. Econometrics 212 (1) (2019) 137–154]. Journal Econometrics, doi:10.1016/j.jeconom.2021.11.010 . Cogley, S. Sargent, T. (2005). Drifts volatilities: monetary policies outcomes post WWII US. Review Economic Dynamics, 8, 262--302, doi:10.1016/j.red.2004.10.009 . Hosszejni, D. Kastner, G. (2021). Modeling Univariate Multivariate Stochastic Volatility R stochvol factorstochvol. Journal Statistical Software, 100, 1–-34. doi:10.18637/jss.v100.i12 .","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Plot plot(mod)   # Summary summary(mod) #>  #> Posterior median of reduced-form coefficients: #>                 GDPC1 CPIAUCSL   FEDFUNDS #> GDPC1.l1     0.236588 0.006926  2.109e-02 #> CPIAUCSL.l1 -0.053390 0.616162 -3.646e-03 #> FEDFUNDS.l1  0.007874 0.037088  1.001e+00 #> intercept    0.005874 0.001291 -8.055e-05 #>  #> Posterior interquartile range of of reduced-form coefficients: #>                GDPC1  CPIAUCSL  FEDFUNDS #> GDPC1.l1    0.089598 0.0290897 0.0252782 #> CPIAUCSL.l1 0.111515 0.0904426 0.0150976 #> FEDFUNDS.l1 0.018416 0.0141909 0.0076726 #> intercept   0.001135 0.0007528 0.0001651"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract VAR coefficients — coef","title":"Extract VAR coefficients — coef","text":"Extracts posterior draws VAR coefficients VAR model estimated bvar().","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract VAR coefficients — coef","text":"","code":"# S3 method for bayesianVARs_bvar coef(object, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract VAR coefficients — coef","text":"object bayesianVARs_bvar object obtained bvar(). ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract VAR coefficients — coef","text":"Returns numeric array dimension \\(M \\times K \\times draws\\), M number time-series, K number covariates per equation (including intercept) draws number stored posterior draws.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract VAR coefficients — coef","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of VAR coefficients bvar_coefs <- coef(mod)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"Simulates fitted/predicted (-sample) values estimated VAR model.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar fitted(object, error_term = TRUE, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"object bayesianVARs_bvar object estimated via bvar(). error_term logical indicating whether include error term . ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"object class bayesianVARs_fitted.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate predicted historical values including the error term. pred <- fitted(mod, error_term = TRUE)  # Simulate fitted historical values not including the error term. fit <- fitted(mod, error_term = FALSE)  # Visualize plot(pred)  plot(fit)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":null,"dir":"Reference","previous_headings":"","what":"Draw from generalized inverse Gaussian — my_gig","title":"Draw from generalized inverse Gaussian — my_gig","text":"Vectorized version rgig","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Draw from generalized inverse Gaussian — my_gig","text":"","code":"my_gig(n, lambda, chi, psi)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Draw from generalized inverse Gaussian — my_gig","text":"n single integer indicating number draws generate. lambda vector shape parameters. chi vector shape/scale parameters. Must nonnegative positive lambdas positive else. psi vector shape/scale parameters. Must nonnegative negative lambdas positive else.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Draw from generalized inverse Gaussian — my_gig","text":"Matrix dimension c(n,m), m maximum length lambda, psi chi.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Draw from generalized inverse Gaussian — my_gig","text":"","code":"gigsamples <- my_gig(2, c(1,1), c(1,1), c(1,1))"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"Pairwise visualization --sample posterior predictive densities.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"","code":"# S3 method for bayesianVARs_predict pairs(x, vars, ahead, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"x object class bayesianVARs_predict obtained via predict.bayesianVARs_bvar(). vars Integer vector (coercible ) indicating variables plot. ahead Integer vector (coercible ) indicating step ahead plot. max(ahead) must smaller equal dim(x$predictions)[1]. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"Note bayesianVARs_predict can also used withing plot.bayesianVARs_bvar().","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate from posterior predictive predictions <- predict(mod, ahead = 1:3) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  485 stable posterior draws remaining for prediction!  # Visualize pairs(predictions, vars = 1:3, ahead = 1:3)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"Visualization -sample fit. Can also used display prediction intervals future values.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar plot(   x,   predictions = NULL,   quantiles = c(0.05, 0.5, 0.95),   dates = NULL,   n_col = 1,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"x object class bayesianVARs_bvar obtained via bvar(). predictions Optional array sample predictions, e.g. obtained via predict.bayesianVARs_bvar(). quantiles numeric vector indicating quantiles plot. dates optional vector dates labelling x-axis. default values NULL; case, axis labeled numbers. n_col integer indicating number columns use plotting. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"Returns x invisibly.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate from posterior predictive predictions <- predict(mod, ahead = 1:3) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  626 stable posterior draws remaining for prediction!  # Visualize plot(mod, predictions = predictions)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":null,"dir":"Reference","previous_headings":"","what":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"Visualization -sample fit estimated VAR.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"","code":"# S3 method for bayesianVARs_fitted plot(   x,   dates = NULL,   vars = \"all\",   quantiles = c(0.05, 0.5, 0.95),   n_col = 1L,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"x bayesianVARs_fitted object. dates optional vector dates labelling x-axis. default values NULL; case, axis labeled numbers. vars character vector containing names variables visualized. default \"\" indicating fit variables visualized. quantiles numeric vector indicating quantiles plot. n_col integer indicating number columns use plotting. ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"returns x invisibly","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate predicted historical values including the error term. pred <- fitted(mod, error_term = TRUE)  # Visualize plot(pred)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Fan chart — plot.bayesianVARs_predict","title":"Fan chart — plot.bayesianVARs_predict","text":"Visualization (--sample) predictive distribution.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Fan chart — plot.bayesianVARs_predict","text":"","code":"# S3 method for bayesianVARs_predict plot(   x,   dates = NULL,   vars = \"all\",   ahead = NULL,   quantiles = c(0.05, 0.25, 0.5, 0.75, 0.95),   n_col = 1L,   first_obs = 1L,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Fan chart — plot.bayesianVARs_predict","text":"x object type bayesianVARs_predict obtained via predict.bayesianVARs_bvar(). dates optional vector dates labeling x-axis. default values NULL; case, axis labeled numbers. vars character vector containing names variables visualized. default \"\" indicating variables visualized. ahead Integer vector (coercible ) indicating step ahead plot. max(ahead) must smaller equal dim(x$predictions)[1]. quantiles numeric vector indicating quantiles plot. n_col integer indicating number columns use plotting. first_obs integer indicating first observation used plotting. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Fan chart — plot.bayesianVARs_predict","text":"Returns x invisibly!","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Fan chart — plot.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate from posterior predictive predictions <- predict(mod, ahead = 1:3) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  385 stable posterior draws remaining for prediction!  # Visualize plot(predictions, vars = 1:3, ahead = 1:3)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":null,"dir":"Reference","previous_headings":"","what":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"Posterior heatmaps VAR coefficients variance-covariance matrices","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"","code":"posterior_heatmap(   x,   FUN,   ...,   colorbar = TRUE,   xlabels = NULL,   ylabels = NULL,   add_numbers = FALSE,   zlim = NULL,   colspace = NULL,   main = \"\",   cex.axis = 0.75,   cex.colbar = 1,   cex.numbers = 1,   asp = NULL )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"x array dimension \\(\\times b \\times draws\\), \\(  \\times b\\) dimension parameter visualize draws number posterior draws. FUN summary function applied margins c(1,2) x. E.g. \"median\", \"mean\", \"IQR\", \"sd\" \"var\". apply(x, 1:2, FUN, ...) must return matrix! ... optional arguments FUN. colorbar logical indicating whether display colorbar . Default TRUE. xlabels ylabels=NULL, default, indicates names dependent variables displayed. ylabels=\"\" indicates ylabels displayed. ylabels xlabels=NULL, default, indicates labels covariables (lagged values dependent variables) displayed. xlabels=\"lags\" indicates lags marked. xlabels=\"\" indicates ylabels displayed. add_numbers logical. add_numbers=TRUE, default indicates actual values summary displayed. zlim numeric vector length two indicating minimum maximum values colors plotted. default range determined maximum absolute values selected summary. colspace Optional argument. main main title plot. cex.axis magnification used y-axis annotation relative current setting cex. cex.colbar magnification used colorbar annotation relative current setting cex. cex.numbers magnification used actual values (add_numbers=TRUE) relative current setting cex. asp aspect ratio.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"Returns x invisibly.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(100*data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of VAR coefficients phi_post <- coef(mod)  # Visualize posterior median of VAR coefficients posterior_heatmap(phi_post, median)   # Extract posterior draws of variance-covariance matrices (for each point in time) sigma_post <- vcov(mod) # Visualize posterior interquartile-range of variance-covariance matrix of the first observation posterior_heatmap(sigma_post[1,,,], IQR)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"Simulates (--sample) predictive density Bayesian VARs estimated via bvar() computes log predictive likelhoods ex-post observed data supplied.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar predict(   object,   ahead = 1L,   each = 1L,   stable = TRUE,   simulate_predictive = TRUE,   LPL = FALSE,   Y_obs = NA,   LPL_VoI = NA,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"object bayesianVARs_bvar object, obtained bvar(). ahead Integer vector (coercible ), indicating number steps ahead predict. Single integer (coercible ) indicating often drawn posterior predictive distribution draw stored MCMC sampling. stable logical indicating whether consider draws posterior fulfill 'stable' criterion. Default TRUE. simulate_predictive logical, indicating whether posterior predictive distribution simulated. LPL logical indicating whether ahead-step-ahead log predictive likelihoods computed. LPL=TRUE, Y_obs specified. Y_obs Data matrix observed values computation log predictive likelihood. ncol(object$Yraw) columns assumed contain single time-series length length(ahead). LPL_VoI either integer vector character vector column-names indicating subgroup time-series object$Yraw joint log predictive likelihood shall computed. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"Object class bayesianVARs_predict, list may contain following elements: predictions array dimensions c(length(ahead), ncol(object$Yraw), * dim(object$PHI)[3]) containing simulations predictive density (simulate_predictive=TRUE). LPL vector length length(ahead) containing log-predictive-likelihoods (taking account joint distribution variables) (LPL=TRUE). LPL_univariate matrix dimension c(length(ahead), ncol(object$Yraw) containing marginalized univariate log-predictive-likelihoods series (LPL=TRUE). LPL_VoI vector length length(ahead) containing log-predictive-likelihoods subset variables (LPL=TRUE LPL_VoI != NA). Yraw matrix containing data used estimation VAR. LPL_draws matrix containing simulations log-predictive-likelihood (LPL=TRUE). PL_univariate_draws array containing simulations univariate predictive-likelihoods (LPL=TRUE). LPL_sub_draws matrix containing simulations log-predictive-likelihood subset variables (LPL=TRUE LPL_VoI != NA).","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step to 4-steps ahead posterior predictive and compute # log-predictive-likelihoods predictions <- predict(mod, ahead = 1:4, LPL = TRUE, Y_obs = test) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  617 stable posterior draws remaining for prediction!  # Summary summary(predictions) #>  #> LPL: #>   t+1   t+2   t+3   t+4  #> 5.114 9.665 9.092 6.526  #>  #> Marginal univariate LPLs: #>       GDPC1 CPIAUCSL FEDFUNDS #> t+1 -0.8729   0.4109    3.541 #> t+2  2.8555   2.6739    3.391 #> t+3  2.8363   2.4370    3.349 #> t+4  2.8931   0.2959    3.326 #>  #> Prediction quantiles: #> , , GDPC1 #>  #>          t+1        t+2       t+3       t+4 #> 5%  -0.07875 -0.0329135 -0.021393 -0.011870 #> 50% -0.01851  0.0009099  0.005688  0.006618 #> 95%  0.03798  0.0429166  0.034118  0.024602 #>  #> , , CPIAUCSL #>  #>           t+1       t+2       t+3        t+4 #> 5%  -0.018515 -0.017824 -0.015592 -0.0140807 #> 50% -0.007656 -0.005287 -0.002413 -0.0005886 #> 95%  0.002679  0.006747  0.009743  0.0106464 #>  #> , , FEDFUNDS #>  #>           t+1       t+2       t+3       t+4 #> 5%  -0.016214 -0.022544 -0.026463 -0.030055 #> 50% -0.004527 -0.005249 -0.006258 -0.006404 #> 95%  0.005912  0.011279  0.016238  0.018100 #>   # Visualize via fan-charts plot(predictions)   # \\donttest{ # In order to evaluate the joint predictive density of a subset of the # variables (variables of interest), consider specifying 'LPL_VoI': predictions <- predict(mod, ahead = 1:4, LPL = TRUE, Y_obs = test, LPL_VoI = c(\"GDPC1\",\"FEDFUNDS\")) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  617 stable posterior draws remaining for prediction! predictions$LPL_VoI #>      t+1      t+2      t+3      t+4  #> 2.542386 6.712653 6.548112 6.391536  # }"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Pretty printing of a bvar object — print.bayesianVARs_bvar","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"Pretty printing bvar object","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"x Object class bayesianVARs_bvar, usually resulting call bvar(). ... Ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Print model mod #>  #> Fitted bayesianVARs_bvar object with #>   -       3 series #>   -       1 lag(s) #>   -     246 used observations #>   -     247 total observations #>   -    1000 MCMC draws #>   -       1 thinning #>   -    1000 burn-in #>"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"Print method bayesianVARs_predict objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"","code":"# S3 method for bayesianVARs_predict print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"x bayesianVARs_predict object obtained via predict.bayesianVARs_bvar(). ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step ahead posterior predictive predictions <- predict(mod, ahead = 1L) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  595 stable posterior draws remaining for prediction! print(predictions) #>  #> Generic functions for bayesianVARs_predict objects: #>   - summary.bayesianVARs_predict(), #>   - pairs.bayesianVARs_predict(), #>   - plot.bayesianVARs_predict() (alias for pairs.bayesianVARs_predict())."},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"Print method summary.bayesianVARs_bvar objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"","code":"# S3 method for summary.bayesianVARs_bvar print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"x summary.bayesianVARs_bvar object obtained via summary.bayesianVARs_bvar(). ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"Returns x invisibly!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate model mod <- bvar(data, quiet = TRUE)  # Print summary summary(mod) #>  #> Posterior median of reduced-form coefficients: #>                 GDPC1 CPIAUCSL   FEDFUNDS #> GDPC1.l1     0.226152  0.00604  1.362e-02 #> CPIAUCSL.l1 -0.063468  0.62045 -4.773e-03 #> FEDFUNDS.l1  0.009492  0.03717  1.001e+00 #> intercept    0.005959  0.00120 -4.555e-05 #>  #> Posterior interquartile range of of reduced-form coefficients: #>                GDPC1  CPIAUCSL  FEDFUNDS #> GDPC1.l1    0.105445 0.0271618 0.0208810 #> CPIAUCSL.l1 0.118059 0.0930382 0.0131568 #> FEDFUNDS.l1 0.019176 0.0145773 0.0065391 #> intercept   0.001173 0.0007342 0.0001668"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"Print method summary.bayesianVARs_predict objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"","code":"# S3 method for summary.bayesianVARs_predict print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"x summary.bayesianVARs_predict object obtained via summary.bayesianVARs_predict(). ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step ahead posterior predictive predictions <- predict(mod, ahead = 1L) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  586 stable posterior draws remaining for prediction! sum <- summary(predictions) print(sum) #>  #> Prediction quantiles: #> , , GDPC1 #>  #>          t+1 #> 5%  -0.06812 #> 50% -0.01798 #> 95%  0.02909 #>  #> , , CPIAUCSL #>  #>           t+1 #> 5%  -0.018438 #> 50% -0.008339 #> 95%  0.003312 #>  #> , , FEDFUNDS #>  #>           t+1 #> 5%  -0.021399 #> 50% -0.004006 #> 95%  0.009963 #>"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":null,"dir":"Reference","previous_headings":"","what":"Specify prior on PHI — specify_prior_phi","title":"Specify prior on PHI — specify_prior_phi","text":"Configures prior PHI, matrix reduced-form VAR coefficients.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Specify prior on PHI — specify_prior_phi","text":"","code":"specify_prior_phi(   data = NULL,   M = ncol(data),   lags = 1L,   prior = \"HS\",   priormean = 0,   PHI_tol = 1e-18,   DL_a = \"1/K\",   DL_tol = 0,   R2D2_a = 0.1,   R2D2_b = 0.5,   R2D2_tol = 0,   NG_a = 0.1,   NG_b = 1,   NG_c = 1,   NG_tol = 0,   SSVS_c0 = 0.01,   SSVS_c1 = 100,   SSVS_semiautomatic = TRUE,   SSVS_p = 0.5,   HMP_lambda1 = c(0.01, 0.01),   HMP_lambda2 = c(0.01, 0.01),   normal_sds = 10,   global_grouping = \"global\",   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Specify prior on PHI — specify_prior_phi","text":"data Optional. Data matrix (can time series object). \\(M\\) columns assumed contain single time-series length \\(T\\). M positive integer indicating number time-series VAR. lags positive integer indicating order VAR, .e. number lags dependent variables included predictors. prior character, one \"HS\", \"R2D2\", \"NG\", \"DL\", \"SSVS\", \"HMP\" \"normal\". priormean real numbers indicating prior means VAR coefficients. One single number means prior mean -lag coefficients w.r.t. first lag equals priormean 0 else. vector length M means prior mean -lag coefficients w.r.t. first lag equals priormean 0 else. priormean matrix dimension c(lags*M,M), \\(M\\) columns assumed contain lags*M prior means VAR coefficients respective VAR equations. PHI_tol Minimum number absolute value VAR coefficient draw can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. DL_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. argument global_grouping specifies e.g. k groups, DL_a can numeric vector length k elements indicate shrinkage group.  matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. DL_a specified prior=\"DL\". DL_tol Minimum number parameter draw one shrinking parameters Dirichlet Laplace prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. DL_tol specified prior=\"DL\". R2D2_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. argument global_grouping specifies e.g. k groups, R2D2_a can numeric vector length k elements indicate shrinkage group.  matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. R2D2_a specified prior=\"R2D2\". R2D2_b (Single) positive real number. value indicates shape parameter inverse gamma prior (semi-)global scales. argument global_grouping specifies e.g. k groups, NG_b can numeric vector length k elements determine shape parameter group. R2D2_b specified prior=\"R2D2\". R2D2_tol Minimum number parameter draw one shrinking parameters R2D2 prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. R2D2_tol specified prior=\"R2D2\". NG_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. argument global_grouping specifies e.g. k groups, NG_a can numeric vector length k elements indicate shrinkage group.  matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. NG_a specified prior=\"NG\". NG_b (Single) positive real number. value indicates shape parameter inverse gamma prior (semi-)global scales. argument global_grouping specifies e.g. k groups, NG_b can numeric vector length k elements determine shape parameter group. NG_b specified prior=\"NG\". NG_c (Single) positive real number. value indicates scale parameter inverse gamma prior (semi-)global scales. argument global_grouping specifies e.g. k groups, NG_c can numeric vector length k elements determine scale parameter group. Expert option set scale parameter proportional NG_a. E.g. case discrete hyperprior NG_a chosen, desired proportion say 0.2 achieved setting NG_c=\"0.2a\" (character input!). NG_c specified prior=\"NG\". NG_tol Minimum number parameter draw one shrinking parameters normal-gamma prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. NG_tol specified prior=\"NG\". SSVS_c0 single positive number indicating (unscaled) standard deviation spike component. SSVS_c0 specified prior=\"SSVS\". \\(SSVS_{c0} \\ll SSVS_{c1}\\)! SSVS_c0 specified prior=\"SSVS\". SSVS_c1 single positive number indicating (unscaled) standard deviation slab component. SSVS_c0 specified prior=\"SSVS\". \\(SSVS_{c0} \\ll SSVS_{c1}\\)! SSVS_semiautomatic logical. SSVS_semiautomatic=TRUE SSVS_c0 SSVS_c1 scaled variances posterior PHI FLAT conjugate (dependent Normal-Wishart prior). SSVS_semiautomatic specified prior=\"SSVS\". SSVS_p Either single positive number range (0,1) indicating (fixed) prior inclusion probability coefficient. numeric vector length 2 positive entries indicating shape parameters Beta distribution. case Beta hyperprior placed prior inclusion probability. SSVS_p specified prior=\"SSVS\". HMP_lambda1 numeric vector length 2. entries must positive. first indicates shape second rate Gamma hyperprior -lag coefficients. HMP_lambda1 specified prior=\"HMP\". HMP_lambda2 numeric vector length 2. entries must positive. first indicates shape second rate Gamma hyperprior cross-lag coefficients. HMP_lambda2 specified prior=\"HMP\". normal_sds numeric vector length \\(n\\), \\(n = lags M^2\\) number VAR coefficients (excluding intercept), indicating prior variances. single number recycled accordingly! Must positive. normal_sds specified prior=\"normal\". global_grouping One \"global\", \"equation-wise\", \"covariate-wise\", \"olcl-lagwise\" \"fol\" indicating sub-groups semi-global(-local) modifications HS, R2D2, NG, DL SSVS prior. Works also user-specified indicator matrix dimension c(lags*M,M).  relevant prior=\"HS\", prior=\"DL\", prior=\"R2D2\", prior=\"NG\" prior=\"SSVS\". ... use!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Specify prior on PHI — specify_prior_phi","text":"baysianVARs_prior_phi-object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Specify prior on PHI — specify_prior_phi","text":"details concerning prior-elicitation VARs please see Gruber & Kastner (2023). Currently one can choose six hierarchical shrinkage priors normal prior: prior=\"HS\" stands Horseshoe-prior, prior=\"R2D2 R\\(^2\\)-induced-Dirichlet-decompostion-prior, prior=\"NG\" normal-gamma-prior, prior=\"DL\" Dirichlet-Laplace-prior, prior=\"SSVS\" stochastic-search-variable-selection-prior, prior=\"HMP\" semi-hierarchical Minnesota prior prior=normal normal-prior. Semi-global shrinkage, .e. group-specific shrinkage pre-specified subgroups coefficients, can achieved argument global_grouping.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Specify prior on PHI — specify_prior_phi","text":"Gruber, L. Kastner, G. (2023). Forecasting macroeconomic data Bayesian VARs: Sparse dense? depends! arXiv:2206.04902.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Specify prior on PHI — specify_prior_phi","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Horseshoe prior for a VAR(2) phi_hs <- specify_prior_phi(data = data, lags = 2L ,prior = \"HS\")  # Semi-global-local Horseshoe prior for a VAR(2) with semi-global shrinkage parameters for # cross-lag and own-lag coefficients in each lag phi_hs_sg <- specify_prior_phi(data = data, lags = 2L, prior = \"HS\", global_grouping = \"olcl-lagwise\")  # Semi-global-local Horseshoe prior for a VAR(2) with equation-wise shrinkage # construct indicator matrix for equation-wise shrinkage semi_global_mat <- matrix(1:ncol(data), 2*ncol(data), ncol(data), byrow = TRUE) phi_hs_ew <- specify_prior_phi(data = data, lags = 2L, prior = \"HS\", global_grouping = semi_global_mat) # (for equation-wise shrinkage one can also use 'global_grouping = \"equation-wise\"')  # \\donttest{ # Estimate model with your prior configuration of choice mod <- bvar(data, lags = 2L, prior_phi = phi_hs_sg, quiet = TRUE) # }"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":null,"dir":"Reference","previous_headings":"","what":"Specify prior on Sigma — specify_prior_sigma","title":"Specify prior on Sigma — specify_prior_sigma","text":"Configures prior variance-covariance VAR.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Specify prior on Sigma — specify_prior_sigma","text":"","code":"specify_prior_sigma(   data = NULL,   M = ncol(data),   type = c(\"factor\", \"cholesky\"),   factor_factors = 1L,   factor_restrict = c(\"none\", \"upper\"),   factor_priorfacloadtype = c(\"rowwiseng\", \"colwiseng\", \"normal\"),   factor_priorfacload = 0.1,   factor_facloadtol = 1e-18,   factor_priorng = c(1, 1),   factor_priormu = c(0, 10),   factor_priorphiidi = c(10, 3),   factor_priorphifac = c(10, 3),   factor_priorsigmaidi = 1,   factor_priorsigmafac = 1,   factor_priorh0idi = \"stationary\",   factor_priorh0fac = \"stationary\",   factor_heteroskedastic = TRUE,   factor_priorhomoskedastic = NA,   factor_interweaving = 4,   cholesky_U_prior = c(\"HS\", \"DL\", \"R2D2\", \"NG\", \"SSVS\", \"normal\", \"HMP\"),   cholesky_U_tol = 1e-18,   cholesky_heteroscedastic = TRUE,   cholesky_priormu = c(0, 100),   cholesky_priorphi = c(20, 1.5),   cholesky_priorsigma2 = c(0.5, 0.5),   cholesky_priorh0 = \"stationary\",   cholesky_priorhomoscedastic = as.numeric(NA),   cholesky_DL_a = \"1/n\",   cholesky_DL_tol = 0,   cholesky_R2D2_a = 0.4,   cholesky_R2D2_b = 0.5,   cholesky_R2D2_tol = 0,   cholesky_NG_a = 0.5,   cholesky_NG_b = 0.5,   cholesky_NG_c = 0.5,   cholesky_NG_tol = 0,   cholesky_SSVS_c0 = 0.001,   cholesky_SSVS_c1 = 1,   cholesky_SSVS_p = 0.5,   cholesky_HMP_lambda3 = c(0.01, 0.01),   cholesky_normal_sds = 10,   expert_sv_offset = 0,   quiet = FALSE,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Specify prior on Sigma — specify_prior_sigma","text":"data Optional. Data matrix (can time series object). \\(M\\) columns assumed contain single time-series length \\(T\\). M positive integer indicating number time-series VAR. type character, one \"factor\" (default) \"cholesky\", indicating decomposition applied covariance-matrix. factor_factors Number latent factors estimated. required type=\"factor\". factor_restrict Either \"upper\" \"none\", indicating whether factor loadings matrix restricted zeros diagonal (\"upper\") whether elements estimated data (\"none\"). Setting restrict \"upper\" often stabilizes MCMC estimation can important identifying factor loadings matrix, however, generally strong prior assumption. Setting restrict \"none\" usually preferred option identification factor loadings matrix less concern covariance estimation prediction goal. required type=\"factor\". factor_priorfacloadtype Can \"normal\", \"rowwiseng\", \"colwiseng\". required type=\"factor\". \"normal\": Normal prior. value priorfacload interpreted standard deviations Gaussian prior distributions factor loadings. \"rowwiseng\": Row-wise Normal-Gamma prior. value priorfacload interpreted shrinkage parameter . \"colwiseng\": Column-wise Normal-Gamma prior. value priorfacload interpreted shrinkage parameter . details please see Kastner (2019). factor_priorfacload Either matrix dimensions M times factor_factors positive elements single number (recycled accordingly). required type=\"factor\". meaning factor_priorfacload depends setting factor_priorfacloadtype explained . factor_facloadtol Minimum number absolute value factor loadings draw can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. required type=\"factor\". factor_priorng Two-element vector positive entries indicating Normal-Gamma prior's hyperhyperparameters c d (cf. Kastner (2019)). required type=\"factor\". factor_priormu Vector length 2 denoting prior mean standard deviation unconditional levels idiosyncratic log variance processes. required type=\"factor\". factor_priorphiidi Vector length 2, indicating shape parameters Beta prior distributions transformed parameters (phi+1)/2, phi denotes persistence idiosyncratic log variances. required type=\"factor\". factor_priorphifac Vector length 2, indicating shape parameters Beta prior distributions transformed parameters (phi+1)/2, phi denotes persistence factor log variances. required type=\"factor\". factor_priorsigmaidi Vector length M containing prior volatilities log variances. factor_priorsigmaidi exactly one element, recycled idiosyncratic log variances. required type=\"factor\". factor_priorsigmafac Vector length factor_factors containing prior volatilities log variances. factor_priorsigmafac exactly one element, recycled factor log variances. required type=\"factor\". factor_priorh0idi Vector length 1 M, containing information Gaussian prior initial idiosyncratic log variances. required type=\"factor\". element factor_priorh0idi nonnegative number, conditional prior corresponding initial log variance h0 assumed Gaussian mean 0 standard deviation factor_priorh0idi times \\(sigma\\). element factor_priorh0idi string 'stationary', prior corresponding initial log volatility taken stationary distribution, .e. h0 assumed Gaussian mean 0 variance \\(sigma^2/(1-phi^2)\\). factor_priorh0fac Vector length 1 factor_factors, containing information Gaussian prior initial factor log variances. required type=\"factor\". element factor_priorh0fac nonnegative number, conditional prior corresponding initial log variance h0 assumed Gaussian mean 0 standard deviation factor_priorh0fac times \\(sigma\\). element factor_priorh0fac string 'stationary', prior corresponding initial log volatility taken stationary distribution, .e. h0 assumed Gaussian mean 0 variance \\(sigma^2/(1-phi^2)\\). factor_heteroskedastic Vector length 1, 2, M +  factor_factors, containing logical values indicating whether time-varying (factor_heteroskedastic = TRUE) constant (factor_heteroskedastic =   FALSE) variance estimated. factor_heteroskedastic length 2 recycled accordingly, whereby first element used idiosyncratic variances second element used factor variances. required type=\"factor\". factor_priorhomoskedastic used least one element factor_heteroskedastic set FALSE. case, factor_priorhomoskedastic must matrix positive entries dimension c(M, 2). Values column 1 interpreted shape values column 2 interpreted rate parameter corresponding inverse gamma prior distribution idiosyncratic variances. required type=\"factor\". factor_interweaving following values interweaving factor loadings accepted (required type=\"factor\"): 0: interweaving. 1: Shallow interweaving diagonal entries. 2: Deep interweaving diagonal entries. 3: Shallow interweaving largest absolute entries column. 4: Deep interweaving largest absolute entries column. details please see Kastner et al. (2017). value 4 highly recommended default. cholesky_U_prior character, one \"HS\", \"R2D2\", \"NG\", \"DL\", \"SSVS\", \"HMP\" \"normal\". required type=\"cholesky\". cholesky_U_tol Minimum number absolute value free -diagonal element \\(U\\)-draw can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. required type=\"cholesky\". cholesky_heteroscedastic single logical indicating whether time-varying (cholesky_heteroscedastic = TRUE) constant (cholesky_heteroscedastic = FALSE) variance estimated. required type=\"cholesky\". cholesky_priormu Vector length 2 denoting prior mean standard deviation unconditional levels log variance processes. required type=\"cholesky\". cholesky_priorphi Vector length 2, indicating shape parameters Beta prior distributions transformed parameters (phi+1)/2, phi denotes persistence log variances. required type=\"cholesky\". cholesky_priorsigma2 Vector length 2, indicating shape rate Gamma prior distributions variance log variance processes. (Currently one global setting \\(M\\) processes supported). required type=\"cholesky\". cholesky_priorh0 Vector length 1 M, containing information Gaussian prior initial idiosyncratic log variances. required type=\"cholesky\". element cholesky_priorh0 nonnegative number, conditional prior corresponding initial log variance h0 assumed Gaussian mean 0 standard deviation cholesky_priorh0 times \\(sigma\\). element cholesky_priorh0 string 'stationary', prior corresponding initial log volatility taken stationary distribution, .e. h0 assumed Gaussian mean 0 variance \\(sigma^2/(1-phi^2)\\). cholesky_priorhomoscedastic used cholesky_heteroscedastic=FALSE. case, cholesky_priorhomoscedastic must matrix positive entries dimension c(M, 2). Values column 1 interpreted shape values column 2 interpreted scale parameter corresponding inverse gamma prior distribution variances. required type=\"cholesky\". cholesky_DL_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. cholesky_DL_a specified cholesky_U_prior=\"DL\". cholesky_DL_tol Minimum number parameter draw one shrinking parameters Dirichlet Laplace prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. DL_tol specified cholesky_U_prior=\"DL\". cholesky_R2D2_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. cholesky_R2D2_a specified cholesky_U_prior=\"R2D2\". cholesky_R2D2_b single positive number, greater values indicate heavier regularization. cholesky_R2D2_b specified cholesky_U_prior=\"R2D2\". cholesky_R2D2_tol Minimum number parameter draw one shrinking parameters R2D2 prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. cholesky_R2D2_tol specified cholesky_U_prior=\"R2D2\". cholesky_NG_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. cholesky_NG_a specified cholesky_U_prior=\"NG\". cholesky_NG_b (Single) positive real number. value indicates shape parameter inverse gamma prior global scales. cholesky_NG_b specified cholesky_U_prior=\"NG\". cholesky_NG_c (Single) positive real number. value indicates scale parameter inverse gamma prior global scales. Expert option set scale parameter proportional NG_a. E.g. case discrete hyperprior NG_a chosen, desired proportion say 0.2 achieved setting NG_c=\"0.2a\" (character input!). cholesky_NG_c specified cholesky_U_prior=\"NG\". cholesky_NG_tol Minimum number parameter draw one shrinking parameters normal-gamma prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. cholesky_NG_tol specified cholesky_U_prior=\"NG\". cholesky_SSVS_c0 single positive number indicating (unscaled) standard deviation spike component. cholesky_SSVS_c0 specified choleksy_U_prior=\"SSVS\". \\(SSVS_{c0}   \\ll SSVS_{c1}\\)! cholesky_SSVS_c1 single positive number indicating (unscaled) standard deviation slab component. cholesky_SSVS_c1 specified choleksy_U_prior=\"SSVS\". \\(SSVS_{c0} \\ll SSVS_{c1}\\)! cholesky_SSVS_p Either single positive number range (0,1) indicating (fixed) prior inclusion probability coefficient. numeric vector length 2 positive entries indicating shape parameters Beta distribution. case Beta hyperprior placed prior inclusion probability. cholesky_SSVS_p specified choleksy_U_prior=\"SSVS\". cholesky_HMP_lambda3 numeric vector length 2. entries must positive. first indicates shape second rate Gamma hyperprior contemporaneous coefficients. cholesky_HMP_lambda3 specified choleksy_U_prior=\"HMP\". cholesky_normal_sds numeric vector length \\(\\frac{M^2-M}{2}\\), indicating prior variances free -diagonal elements \\(U\\). single number recycled accordingly! Must positive. cholesky_normal_sds specified choleksy_U_prior=\"normal\". expert_sv_offset ... use! quiet logical indicating whether informative output omitted. ... use!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Specify prior on Sigma — specify_prior_sigma","text":"Object class bayesianVARs_prior_sigma.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Specify prior on Sigma — specify_prior_sigma","text":"bvar offers two different specifications errors: user can choose factor stochastic volatility structure cholesky stochastic volatility structure. cases disturbances \\(\\boldsymbol{\\epsilon}_t\\) assumed follow \\(M\\)-dimensional multivariate normal distribution zero mean variance-covariance matrix \\(\\boldsymbol{\\Sigma}_t\\). case cholesky specification \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{U}^{\\prime -1} \\boldsymbol{D}_t \\boldsymbol{U}^{-1}\\), \\(\\boldsymbol{U}^{-1}\\) upper unitriangular (ones diagonal). diagonal matrix \\(\\boldsymbol{D}_t\\) depends upon latent log-variances, .e. \\(\\boldsymbol{D}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\). case factor structure, \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{\\Lambda} \\boldsymbol{V}_t \\boldsymbol{\\Lambda}^\\prime + \\boldsymbol{G}_t\\). diagonal matrices \\(\\boldsymbol{V}_t\\) \\(\\boldsymbol{G}_t\\) depend upon latent log-variances, .e. \\(\\boldsymbol{G}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\) \\(\\boldsymbol{V}_t=diag(exp(h_{M+1,t}),\\dots, exp(h_{M+r,t})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\) \\(h_{M+j,t}\\sim N(\\phi_ih_{M+j,t-1},\\sigma_{M+j}^2)\\) \\(j=1,\\dots,r\\).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Specify prior on Sigma — specify_prior_sigma","text":"Kastner, G. (2019). Sparse Bayesian Time-Varying Covariance Estimation Many Dimensions Journal Econometrics, 210(1), 98--115, doi:10.1016/j.jeconom.2018.11.007 Kastner, G., Frühwirth-Schnatter, S., Lopes, H.F. (2017). Efficient Bayesian Inference Multivariate Factor Stochastic Volatility Models. Journal Computational Graphical Statistics, 26(4), 905--917, doi:10.1080/10618600.2017.1322091 .","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Specify prior on Sigma — specify_prior_sigma","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # examples with stochastic volatility (heteroscedasticity) ----------------- # factor-decomposition with 2 factors and colwise normal-gamma prior on the loadings sigma_factor_cng_sv <- specify_prior_sigma(data = data, type = \"factor\", factor_factors = 2L, factor_priorfacloadtype = \"colwiseng\", factor_heteroskedastic = TRUE) #>  #> Since argument 'type' is specified with 'factor', all arguments starting with 'cholesky_' are being ignored.  # cholesky-decomposition with Dirichlet-Laplace prior on U sigma_cholesky_dl_sv <- specify_prior_sigma(data = data, type = \"cholesky\", cholesky_U_prior = \"DL\", cholesky_DL_a = 0.5, cholesky_heteroscedastic = TRUE) #>  #> Since argument 'type' is specified with 'cholesky', all arguments starting with 'factor_' are being ignored.  # examples without stochastic volatility (homoscedasticity) ---------------- # factor-decomposition with 2 factors and colwise normal-gamma prior on the loadings sigma_factor_cng <- specify_prior_sigma(data = data, type = \"factor\", factor_factors = 2L, factor_priorfacloadtype = \"colwiseng\", factor_heteroskedastic = FALSE, factor_priorhomoskedastic = matrix(c(0.5,0.5), ncol(data), 2)) #>  #> Since argument 'type' is specified with 'factor', all arguments starting with 'cholesky_' are being ignored. #>  #> Cannot do deep factor_interweaving if (some) factor_factors are homoskedastic. Setting 'factor_interweaving' to 3.  # cholesky-decomposition with Horseshoe prior on U sigma_cholesky_dl <- specify_prior_sigma(data = data, type = \"cholesky\", cholesky_U_prior = \"HS\", cholesky_heteroscedastic = FALSE) #>  #> Since argument 'type' is specified with 'cholesky', all arguments starting with 'factor_' are being ignored. #>  #> Argument 'cholesky_priorhomoscedastic' not specified. Setting both shape and rate of inverse gamma prior equal to 0.01.  # \\donttest{ # Estimate model with your prior configuration of choice mod <- bvar(data, prior_sigma = sigma_factor_cng_sv, quiet = TRUE) # }"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Stable posterior draws — stable_bvar","title":"Stable posterior draws — stable_bvar","text":"stable_bvar() detects discards posterior draws bayesianVARs_bvar object fulfill stability condition: VAR(p) model considered stable eigenvalues companion form matrix lie inside unit circle.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Stable posterior draws — stable_bvar","text":"","code":"stable_bvar(object, quiet = FALSE)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Stable posterior draws — stable_bvar","text":"object bayesianVARs_bvar object obtained via bvar(). quiet logical indicating whether informative output omitted.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Stable posterior draws — stable_bvar","text":"object type bayesianVARs_bvar.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Stable posterior draws — stable_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Discard \"unstable\" draws stable_mod <- stable_bvar(mod) #>  #> Original 'bayesianVARs_bvar' object consists of 1000 posterior draws. #>  #> Detected 475 unstable draws. #>  #> Remaining draws: 525 !"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"Extract replace parts bayesianVARs_coef object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"","code":"# S3 method for bayesianVARs_coef [(x, i, j, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"x object type bayesianVARs_coef. indices j indices ... indices","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"object type bayesianVARs_coef.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract coefficients, which are of class bayesianVARs_coef phi <- coef(mod) phi[1,1,1] #> [1] 0.2644423 #> attr(,\"class\") #> [1] \"bayesianVARs_coef\"  \"bayesianVARs_draws\""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"Extract replace parts bayesianVARs_draws object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"","code":"# S3 method for bayesianVARs_draws [(x, i, j, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"x object type bayesianVARs_draws. indices j indices ... indices","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"object type bayesianVARs_draws.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract coefficients, which are of class bayesianVARs_draws phi <- coef(mod) phi[1,1,1] #> [1] 0.2662653 #> attr(,\"class\") #> [1] \"bayesianVARs_coef\"  \"bayesianVARs_draws\""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"Summary method bayesianVARs_bvar objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar summary(object, quantiles = c(0.025, 0.25, 0.5, 0.75, 0.975), ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"object bayesianVARs_bvar object obtained via bvar(). quantiles numeric vector quantiles compute. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"object type summary.bayesianVARs_bvar.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate model mod <- bvar(data, quiet = TRUE)  # Summary sum <- summary(mod)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":null,"dir":"Reference","previous_headings":"","what":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"Summary statistics bayesianVARs posterior draws.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"","code":"# S3 method for bayesianVARs_draws summary(object, quantiles = c(0.25, 0.5, 0.75), ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"object object class bayesianVARs_draws usually obtained extractors like coef.bayesianVARs_bvar() vcov.bayesianVARs_bvar(). quantiles vector quantiles evaluate. ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"list object class bayesianVARs_draws_summary holding mean: Vector matrix containing posterior mean. sd: Vector matrix containing posterior standard deviation . quantiles: Array containing posterior quantiles.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")] # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of VAR coefficients bvar_coefs <- coef(mod)  # Compute summary statistics summary_stats <- summary(bvar_coefs)  # Compute summary statistics of VAR coefficients without using coef() summary_stats <- summary(mod$PHI)  # Test which list elements of 'mod' are of class 'bayesianVARs_draws'. names(mod)[sapply(names(mod), function(x) inherits(mod[[x]], \"bayesianVARs_draws\"))] #> [1] \"PHI\"     \"U\"       \"logvar\"  \"sv_para\" \"facload\" \"fac\""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"Summary method bayesianVARs_predict objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"","code":"# S3 method for bayesianVARs_predict summary(object, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"object bayesianVARs_predict object obtained via predict.bayesianVARs_bvar(). ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"summary.bayesianVARs_predict object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step ahead posterior predictive predictions <- predict(mod, ahead = 1L) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  597 stable posterior draws remaining for prediction! summary(predictions) #>  #> Prediction quantiles: #> , , GDPC1 #>  #>          t+1 #> 5%  -0.06623 #> 50% -0.02089 #> 95%  0.02220 #>  #> , , CPIAUCSL #>  #>           t+1 #> 5%  -0.019613 #> 50% -0.008145 #> 95%  0.003785 #>  #> , , FEDFUNDS #>  #>           t+1 #> 5%  -0.022382 #> 50% -0.003965 #> 95%  0.013119 #>"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":null,"dir":"Reference","previous_headings":"","what":"Data from the US-economy — usmacro_growth","title":"Data from the US-economy — usmacro_growth","text":"21 selected quarterly time-series 1953:Q1 2021:Q2. FRED-QD data base (McCracken Ng, 2021). Release date 2021-07. Data transformed interpreted growth-rates (first log-differences exception interest rates, already growth rates).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Data from the US-economy — usmacro_growth","text":"","code":"usmacro_growth"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Data from the US-economy — usmacro_growth","text":"matrix 247 rows 21 columns.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Data from the US-economy — usmacro_growth","text":"Raw (untransformed) data available https://research.stlouisfed.org/econ/mccracken/fred-databases/, https://files.stlouisfed.org/files/htdocs/fred-md/quarterly/2021-07.csv.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Data from the US-economy — usmacro_growth","text":"McCracken, M. W. Ng, S. (2021). FRED-QD: Quarterly Database Macroeconomic Research, Review, Federal Reserve Bank St. Louis, 103(1), 1--44, doi:10.20955/r.103.1-44 .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"Returns posterior draws possibly time-varying variance-covariance matrix VAR estimated via bvar(). Returns full paths sv_keep=\"\" calling bvar(). Otherwise, draws variance-covariance matrix last observation returned, .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar vcov(object, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"object object class bayesianVARs_bvar obtained via bvar(). ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"array class bayesianVARs_draws dimension \\(T \\times M   \\times M \\times draws\\), \\(T\\) number observations, \\(M\\) number time-series \\(draws\\) number stored posterior draws.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")] # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of the variance-covariance matrix bvar_vcov <- vcov(mod)"},{"path":"https://luisgruber.github.io/bayesianVARs/news/index.html","id":"bayesianvars-011","dir":"Changelog","previous_headings":"","what":"bayesianVARs 0.1.1","title":"bayesianVARs 0.1.1","text":"CRAN release: 2024-01-17 Fixed clang-UBSAN issue. Fixed undefined figure references vignette.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/news/index.html","id":"bayesianvars-010","dir":"Changelog","previous_headings":"","what":"bayesianVARs 0.1.0","title":"bayesianVARs 0.1.0","text":"CRAN release: 2024-01-13 Initial CRAN submission.","code":""}]
+[{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":null,"dir":"","previous_headings":"","what":"GNU General Public License","title":"GNU General Public License","text":"Version 3, 29 June 2007Copyright © 2007 Free Software Foundation, Inc. <http://fsf.org/> Everyone permitted copy distribute verbatim copies license document, changing allowed.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"preamble","dir":"","previous_headings":"","what":"Preamble","title":"GNU General Public License","text":"GNU General Public License free, copyleft license software kinds works. licenses software practical works designed take away freedom share change works. contrast, GNU General Public License intended guarantee freedom share change versions program–make sure remains free software users. , Free Software Foundation, use GNU General Public License software; applies also work released way authors. can apply programs, . speak free software, referring freedom, price. General Public Licenses designed make sure freedom distribute copies free software (charge wish), receive source code can get want , can change software use pieces new free programs, know can things. protect rights, need prevent others denying rights asking surrender rights. Therefore, certain responsibilities distribute copies software, modify : responsibilities respect freedom others. example, distribute copies program, whether gratis fee, must pass recipients freedoms received. must make sure , , receive can get source code. must show terms know rights. Developers use GNU GPL protect rights two steps: (1) assert copyright software, (2) offer License giving legal permission copy, distribute /modify . developers’ authors’ protection, GPL clearly explains warranty free software. users’ authors’ sake, GPL requires modified versions marked changed, problems attributed erroneously authors previous versions. devices designed deny users access install run modified versions software inside , although manufacturer can . fundamentally incompatible aim protecting users’ freedom change software. systematic pattern abuse occurs area products individuals use, precisely unacceptable. Therefore, designed version GPL prohibit practice products. problems arise substantially domains, stand ready extend provision domains future versions GPL, needed protect freedom users. Finally, every program threatened constantly software patents. States allow patents restrict development use software general-purpose computers, , wish avoid special danger patents applied free program make effectively proprietary. prevent , GPL assures patents used render program non-free. precise terms conditions copying, distribution modification follow.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_0-definitions","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"0. Definitions","title":"GNU General Public License","text":"“License” refers version 3 GNU General Public License. “Copyright” also means copyright-like laws apply kinds works, semiconductor masks. “Program” refers copyrightable work licensed License. licensee addressed “”. “Licensees” “recipients” may individuals organizations. “modify” work means copy adapt part work fashion requiring copyright permission, making exact copy. resulting work called “modified version” earlier work work “based ” earlier work. “covered work” means either unmodified Program work based Program. “propagate” work means anything , without permission, make directly secondarily liable infringement applicable copyright law, except executing computer modifying private copy. Propagation includes copying, distribution (without modification), making available public, countries activities well. “convey” work means kind propagation enables parties make receive copies. Mere interaction user computer network, transfer copy, conveying. interactive user interface displays “Appropriate Legal Notices” extent includes convenient prominently visible feature (1) displays appropriate copyright notice, (2) tells user warranty work (except extent warranties provided), licensees may convey work License, view copy License. interface presents list user commands options, menu, prominent item list meets criterion.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_1-source-code","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"1. Source Code","title":"GNU General Public License","text":"“source code” work means preferred form work making modifications . “Object code” means non-source form work. “Standard Interface” means interface either official standard defined recognized standards body, , case interfaces specified particular programming language, one widely used among developers working language. “System Libraries” executable work include anything, work whole, () included normal form packaging Major Component, part Major Component, (b) serves enable use work Major Component, implement Standard Interface implementation available public source code form. “Major Component”, context, means major essential component (kernel, window system, ) specific operating system () executable work runs, compiler used produce work, object code interpreter used run . “Corresponding Source” work object code form means source code needed generate, install, (executable work) run object code modify work, including scripts control activities. However, include work’s System Libraries, general-purpose tools generally available free programs used unmodified performing activities part work. example, Corresponding Source includes interface definition files associated source files work, source code shared libraries dynamically linked subprograms work specifically designed require, intimate data communication control flow subprograms parts work. Corresponding Source need include anything users can regenerate automatically parts Corresponding Source. Corresponding Source work source code form work.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_2-basic-permissions","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"2. Basic Permissions","title":"GNU General Public License","text":"rights granted License granted term copyright Program, irrevocable provided stated conditions met. License explicitly affirms unlimited permission run unmodified Program. output running covered work covered License output, given content, constitutes covered work. License acknowledges rights fair use equivalent, provided copyright law. may make, run propagate covered works convey, without conditions long license otherwise remains force. may convey covered works others sole purpose make modifications exclusively , provide facilities running works, provided comply terms License conveying material control copyright. thus making running covered works must exclusively behalf, direction control, terms prohibit making copies copyrighted material outside relationship . Conveying circumstances permitted solely conditions stated . Sublicensing allowed; section 10 makes unnecessary.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_3-protecting-users-legal-rights-from-anti-circumvention-law","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"3. Protecting Users’ Legal Rights From Anti-Circumvention Law","title":"GNU General Public License","text":"covered work shall deemed part effective technological measure applicable law fulfilling obligations article 11 WIPO copyright treaty adopted 20 December 1996, similar laws prohibiting restricting circumvention measures. convey covered work, waive legal power forbid circumvention technological measures extent circumvention effected exercising rights License respect covered work, disclaim intention limit operation modification work means enforcing, work’s users, third parties’ legal rights forbid circumvention technological measures.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_4-conveying-verbatim-copies","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"4. Conveying Verbatim Copies","title":"GNU General Public License","text":"may convey verbatim copies Program’s source code receive , medium, provided conspicuously appropriately publish copy appropriate copyright notice; keep intact notices stating License non-permissive terms added accord section 7 apply code; keep intact notices absence warranty; give recipients copy License along Program. may charge price price copy convey, may offer support warranty protection fee.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_5-conveying-modified-source-versions","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"5. Conveying Modified Source Versions","title":"GNU General Public License","text":"may convey work based Program, modifications produce Program, form source code terms section 4, provided also meet conditions: ) work must carry prominent notices stating modified , giving relevant date. b) work must carry prominent notices stating released License conditions added section 7. requirement modifies requirement section 4 “keep intact notices”. c) must license entire work, whole, License anyone comes possession copy. License therefore apply, along applicable section 7 additional terms, whole work, parts, regardless packaged. License gives permission license work way, invalidate permission separately received . d) work interactive user interfaces, must display Appropriate Legal Notices; however, Program interactive interfaces display Appropriate Legal Notices, work need make . compilation covered work separate independent works, nature extensions covered work, combined form larger program, volume storage distribution medium, called “aggregate” compilation resulting copyright used limit access legal rights compilation’s users beyond individual works permit. Inclusion covered work aggregate cause License apply parts aggregate.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_6-conveying-non-source-forms","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"6. Conveying Non-Source Forms","title":"GNU General Public License","text":"may convey covered work object code form terms sections 4 5, provided also convey machine-readable Corresponding Source terms License, one ways: ) Convey object code , embodied , physical product (including physical distribution medium), accompanied Corresponding Source fixed durable physical medium customarily used software interchange. b) Convey object code , embodied , physical product (including physical distribution medium), accompanied written offer, valid least three years valid long offer spare parts customer support product model, give anyone possesses object code either (1) copy Corresponding Source software product covered License, durable physical medium customarily used software interchange, price reasonable cost physically performing conveying source, (2) access copy Corresponding Source network server charge. c) Convey individual copies object code copy written offer provide Corresponding Source. alternative allowed occasionally noncommercially, received object code offer, accord subsection 6b. d) Convey object code offering access designated place (gratis charge), offer equivalent access Corresponding Source way place charge. need require recipients copy Corresponding Source along object code. place copy object code network server, Corresponding Source may different server (operated third party) supports equivalent copying facilities, provided maintain clear directions next object code saying find Corresponding Source. Regardless server hosts Corresponding Source, remain obligated ensure available long needed satisfy requirements. e) Convey object code using peer--peer transmission, provided inform peers object code Corresponding Source work offered general public charge subsection 6d. separable portion object code, whose source code excluded Corresponding Source System Library, need included conveying object code work. “User Product” either (1) “consumer product”, means tangible personal property normally used personal, family, household purposes, (2) anything designed sold incorporation dwelling. determining whether product consumer product, doubtful cases shall resolved favor coverage. particular product received particular user, “normally used” refers typical common use class product, regardless status particular user way particular user actually uses, expects expected use, product. product consumer product regardless whether product substantial commercial, industrial non-consumer uses, unless uses represent significant mode use product. “Installation Information” User Product means methods, procedures, authorization keys, information required install execute modified versions covered work User Product modified version Corresponding Source. information must suffice ensure continued functioning modified object code case prevented interfered solely modification made. convey object code work section , , specifically use , User Product, conveying occurs part transaction right possession use User Product transferred recipient perpetuity fixed term (regardless transaction characterized), Corresponding Source conveyed section must accompanied Installation Information. requirement apply neither third party retains ability install modified object code User Product (example, work installed ROM). requirement provide Installation Information include requirement continue provide support service, warranty, updates work modified installed recipient, User Product modified installed. Access network may denied modification materially adversely affects operation network violates rules protocols communication across network. Corresponding Source conveyed, Installation Information provided, accord section must format publicly documented (implementation available public source code form), must require special password key unpacking, reading copying.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_7-additional-terms","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"7. Additional Terms","title":"GNU General Public License","text":"“Additional permissions” terms supplement terms License making exceptions one conditions. Additional permissions applicable entire Program shall treated though included License, extent valid applicable law. additional permissions apply part Program, part may used separately permissions, entire Program remains governed License without regard additional permissions. convey copy covered work, may option remove additional permissions copy, part . (Additional permissions may written require removal certain cases modify work.) may place additional permissions material, added covered work, can give appropriate copyright permission. Notwithstanding provision License, material add covered work, may (authorized copyright holders material) supplement terms License terms: ) Disclaiming warranty limiting liability differently terms sections 15 16 License; b) Requiring preservation specified reasonable legal notices author attributions material Appropriate Legal Notices displayed works containing ; c) Prohibiting misrepresentation origin material, requiring modified versions material marked reasonable ways different original version; d) Limiting use publicity purposes names licensors authors material; e) Declining grant rights trademark law use trade names, trademarks, service marks; f) Requiring indemnification licensors authors material anyone conveys material (modified versions ) contractual assumptions liability recipient, liability contractual assumptions directly impose licensors authors. non-permissive additional terms considered “restrictions” within meaning section 10. Program received , part , contains notice stating governed License along term restriction, may remove term. license document contains restriction permits relicensing conveying License, may add covered work material governed terms license document, provided restriction survive relicensing conveying. add terms covered work accord section, must place, relevant source files, statement additional terms apply files, notice indicating find applicable terms. Additional terms, permissive non-permissive, may stated form separately written license, stated exceptions; requirements apply either way.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_8-termination","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"8. Termination","title":"GNU General Public License","text":"may propagate modify covered work except expressly provided License. attempt otherwise propagate modify void, automatically terminate rights License (including patent licenses granted third paragraph section 11). However, cease violation License, license particular copyright holder reinstated () provisionally, unless copyright holder explicitly finally terminates license, (b) permanently, copyright holder fails notify violation reasonable means prior 60 days cessation. Moreover, license particular copyright holder reinstated permanently copyright holder notifies violation reasonable means, first time received notice violation License (work) copyright holder, cure violation prior 30 days receipt notice. Termination rights section terminate licenses parties received copies rights License. rights terminated permanently reinstated, qualify receive new licenses material section 10.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_9-acceptance-not-required-for-having-copies","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"9. Acceptance Not Required for Having Copies","title":"GNU General Public License","text":"required accept License order receive run copy Program. Ancillary propagation covered work occurring solely consequence using peer--peer transmission receive copy likewise require acceptance. However, nothing License grants permission propagate modify covered work. actions infringe copyright accept License. Therefore, modifying propagating covered work, indicate acceptance License .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_10-automatic-licensing-of-downstream-recipients","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"10. Automatic Licensing of Downstream Recipients","title":"GNU General Public License","text":"time convey covered work, recipient automatically receives license original licensors, run, modify propagate work, subject License. responsible enforcing compliance third parties License. “entity transaction” transaction transferring control organization, substantially assets one, subdividing organization, merging organizations. propagation covered work results entity transaction, party transaction receives copy work also receives whatever licenses work party’s predecessor interest give previous paragraph, plus right possession Corresponding Source work predecessor interest, predecessor can get reasonable efforts. may impose restrictions exercise rights granted affirmed License. example, may impose license fee, royalty, charge exercise rights granted License, may initiate litigation (including cross-claim counterclaim lawsuit) alleging patent claim infringed making, using, selling, offering sale, importing Program portion .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_11-patents","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"11. Patents","title":"GNU General Public License","text":"“contributor” copyright holder authorizes use License Program work Program based. work thus licensed called contributor’s “contributor version”. contributor’s “essential patent claims” patent claims owned controlled contributor, whether already acquired hereafter acquired, infringed manner, permitted License, making, using, selling contributor version, include claims infringed consequence modification contributor version. purposes definition, “control” includes right grant patent sublicenses manner consistent requirements License. contributor grants non-exclusive, worldwide, royalty-free patent license contributor’s essential patent claims, make, use, sell, offer sale, import otherwise run, modify propagate contents contributor version. following three paragraphs, “patent license” express agreement commitment, however denominated, enforce patent (express permission practice patent covenant sue patent infringement). “grant” patent license party means make agreement commitment enforce patent party. convey covered work, knowingly relying patent license, Corresponding Source work available anyone copy, free charge terms License, publicly available network server readily accessible means, must either (1) cause Corresponding Source available, (2) arrange deprive benefit patent license particular work, (3) arrange, manner consistent requirements License, extend patent license downstream recipients. “Knowingly relying” means actual knowledge , patent license, conveying covered work country, recipient’s use covered work country, infringe one identifiable patents country reason believe valid. , pursuant connection single transaction arrangement, convey, propagate procuring conveyance , covered work, grant patent license parties receiving covered work authorizing use, propagate, modify convey specific copy covered work, patent license grant automatically extended recipients covered work works based . patent license “discriminatory” include within scope coverage, prohibits exercise , conditioned non-exercise one rights specifically granted License. may convey covered work party arrangement third party business distributing software, make payment third party based extent activity conveying work, third party grants, parties receive covered work , discriminatory patent license () connection copies covered work conveyed (copies made copies), (b) primarily connection specific products compilations contain covered work, unless entered arrangement, patent license granted, prior 28 March 2007. Nothing License shall construed excluding limiting implied license defenses infringement may otherwise available applicable patent law.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_12-no-surrender-of-others-freedom","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"12. No Surrender of Others’ Freedom","title":"GNU General Public License","text":"conditions imposed (whether court order, agreement otherwise) contradict conditions License, excuse conditions License. convey covered work satisfy simultaneously obligations License pertinent obligations, consequence may convey . example, agree terms obligate collect royalty conveying convey Program, way satisfy terms License refrain entirely conveying Program.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_13-use-with-the-gnu-affero-general-public-license","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"13. Use with the GNU Affero General Public License","title":"GNU General Public License","text":"Notwithstanding provision License, permission link combine covered work work licensed version 3 GNU Affero General Public License single combined work, convey resulting work. terms License continue apply part covered work, special requirements GNU Affero General Public License, section 13, concerning interaction network apply combination .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_14-revised-versions-of-this-license","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"14. Revised Versions of this License","title":"GNU General Public License","text":"Free Software Foundation may publish revised /new versions GNU 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 certain numbered version GNU General Public License “later version” applies , option following terms conditions either numbered version later version published Free Software Foundation. Program specify version number GNU General Public License, may choose version ever published Free Software Foundation. Program specifies proxy can decide future versions GNU General Public License can used, proxy’s public statement acceptance version permanently authorizes choose version Program. Later license versions may give additional different permissions. However, additional obligations imposed author copyright holder result choosing follow later version.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_15-disclaimer-of-warranty","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"15. Disclaimer of Warranty","title":"GNU General Public License","text":"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.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_16-limitation-of-liability","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"16. Limitation of Liability","title":"GNU General Public License","text":"EVENT UNLESS REQUIRED APPLICABLE LAW AGREED WRITING COPYRIGHT HOLDER, PARTY MODIFIES /CONVEYS 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.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/LICENSE.html","id":"id_17-interpretation-of-sections-15-and-16","dir":"","previous_headings":"TERMS AND CONDITIONS","what":"17. Interpretation of Sections 15 and 16","title":"GNU General Public License","text":"disclaimer warranty limitation liability provided given local legal effect according terms, reviewing courts shall apply local law closely approximates absolute waiver civil liability connection Program, unless warranty assumption liability accompanies copy Program return fee. END TERMS CONDITIONS","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/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 state exclusion warranty; file least “copyright” line pointer full notice found. Also add information contact electronic paper mail. program terminal interaction, make output short notice like starts interactive mode: hypothetical commands show w show c show appropriate parts General Public License. course, program’s commands might different; GUI interface, use “box”. also get employer (work programmer) school, , sign “copyright disclaimer” program, necessary. information , apply follow GNU GPL, see <http://www.gnu.org/licenses/>. GNU 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. first, please read <http://www.gnu.org/philosophy/--lgpl.html>.","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 3 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, see <http://www.gnu.org/licenses/>. <program>  Copyright (C) <year>  <name of author> This program 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."},{"path":"https://luisgruber.github.io/bayesianVARs/authors.html","id":null,"dir":"","previous_headings":"","what":"Authors","title":"Authors and Citation","text":"Luis Gruber. Copyright holder, author, maintainer.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/authors.html","id":"citation","dir":"","previous_headings":"","what":"Citation","title":"Authors and Citation","text":"Gruber L (2024). bayesianVARs: MCMC Estimation Bayesian Vectorautoregressions. R package version 0.1.0.9000, https://luisgruber.github.io/bayesianVARs/, https://github.com/luisgruber/bayesianVARs.","code":"@Manual{,   title = {bayesianVARs: MCMC Estimation of Bayesian Vectorautoregressions},   author = {Luis Gruber},   year = {2024},   note = {R package version 0.1.0.9000, https://luisgruber.github.io/bayesianVARs/},   url = {https://github.com/luisgruber/bayesianVARs}, }"},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"bayesianvars-","dir":"","previous_headings":"","what":"MCMC Estimation of Bayesian Vectorautoregressions","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"Estimation Bayesian vectorautoregressions /without stochastic volatility. Implements several modern hierarchical shrinkage priors, amongst Dirichlet-Laplace prior (DL), hierarchical Minnesota prior (HM), Horseshoe prior (HS), normal-gamma prior (NG), R2-induced-Dirichlet-decomposition prior (R2D2) stochastic search variable selection prior (SSVS). Concerning error-term, user can either specify order-invariant factor structure order-variant cholesky structure.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"installation","dir":"","previous_headings":"","what":"Installation","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"Install CRAN version: Install latest development version directly GitHub:","code":"install.packages(\"bayesianVARs\") devtools::install_github(\"luisgruber/bayesianVARs\")"},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"usage","dir":"","previous_headings":"","what":"Usage","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"main workhorse conduct Bayesian inference vectorautoregression models package function bvar(). features: Prediction, plotting, extraction model parameters extraction fitted values usual generic functions predict(), plot(), coef(), vcov() fitted(). Configure prior distributions helper functions specify_prior_phi() specify_prior_sigma().","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"demonstration","dir":"","previous_headings":"","what":"Demonstration","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"","code":"set.seed(537) # load package library(bayesianVARs)  # Load data train_data <-100 * usmacro_growth[1:237,c(\"GDPC1\", \"PCECC96\", \"GPDIC1\", \"AWHMAN\", \"GDPCTPI\", \"CES2000000008x\", \"FEDFUNDS\", \"GS10\", \"EXUSUKx\", \"S&P 500\")] test_data <-100 * usmacro_growth[238:241,c(\"GDPC1\", \"PCECC96\", \"GPDIC1\", \"AWHMAN\", \"GDPCTPI\", \"CES2000000008x\", \"FEDFUNDS\", \"GS10\", \"EXUSUKx\", \"S&P 500\")]                                     # Estimate model using default prior settings mod <- bvar(train_data, lags = 2L, draws = 2000, burnin = 1000, sv_keep = \"all\")  # Out of sample prediction and log-predictive-likelihood evaluation pred <- predict(mod, ahead = 1:4, LPL = TRUE, Y_obs = test_data)  # Visualize in-sample fit plus out-of-sample prediction intervals plot(mod, predictions = pred)"},{"path":"https://luisgruber.github.io/bayesianVARs/index.html","id":"documentation","dir":"","previous_headings":"","what":"Documentation","title":"MCMC Estimation of Bayesian Vectorautoregressions","text":"bayesianVARs - Shrinkage Priors Bayesian Vectorautoregressions R","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"bvar simulates joint posterior distribution parameters latent variables returns posterior draws.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"","code":"bvar(   data,   lags = 1L,   draws = 1000L,   burnin = 1000L,   thin = 1L,   prior_intercept = 100,   prior_phi = specify_prior_phi(data = data, lags = lags, prior = \"HS\"),   prior_sigma = specify_prior_sigma(data = data, type = \"factor\", quiet = TRUE),   sv_keep = \"last\",   quiet = FALSE,   startvals = list(),   expert = list() )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"data Data matrix (can time series object). \\(M\\) columns assumed contain single time-series length \\(T\\). lags Integer indicating order VAR, .e. number lags dependent variables included predictors. draws single integer indicating number draws burnin burnin single integer indicating number draws discarded burnin thin single integer. Every \\(thin\\)th draw stored. Default thin=1L. prior_intercept Either prior_intercept=FALSE constant term (intercept) included. numeric vector length \\(M\\) indicating (fixed) prior variances constant term. single number recycled accordingly. Default prior_intercept=100. prior_phi bayesianVARs_prior_phi object specifying prior reduced form VAR coefficients. Best use constructor specify_prior_phi. prior_sigma bayesianVARs_prior_sigma object specifying prior variance-covariance matrix VAR. Best use constructor specify_prior_sigma. sv_keep String equal \"\" \"last\". case sv_keep = \"last\", default, draws last log-variance \\(h_T\\) stored. quiet logical value indicating whether information progress sampling displayed sampling (default TRUE). startvals optional list starting values. expert optional list expert settings.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"object type bayesianVARs_bvar, list containing following objects: PHI: bayesianVARs_coef object, array, containing posterior draws VAR coefficients (including intercept). U: bayesianVARs_draws object, matrix, containing posterior draws contemporaneous coefficients (cholesky decomposition sigma specified). logvar: bayesianVARs_draws object containing log-variance draws. sv_para: baysesianVARs_draws object containing posterior draws stochastic volatility related parameters. phi_hyperparameter: matrix containing posterior draws hyperparameters conditional normal prior VAR coefficients. u_hyperparameter: matrix containing posterior draws hyperparameters conditional normal prior U (cholesky decomposition sigma specified). bench: Numerical indicating average time took generate one single draw joint posterior distribution parameters. V_prior: array containing posterior draws variances conditional normal prior VAR coefficients. facload: bayesianVARs_draws object, array, containing draws posterior distribution factor loadings matrix (factor decomposition sigma specified). fac: bayesianVARs_draws object, array, containing factor draws posterior distribution (factor decomposition sigma specified). Y: Matrix containing dependent variables used estimation. X matrix containing lagged values dependent variables, .e. covariates. lags: Integer indicating lag order VAR. intercept: Logical indicating whether constant term included. heteroscedastic logical indicating whether heteroscedasticity assumed. Yraw: Matrix containing dependent variables, including initial 'lags' observations. Traw: Integer indicating total number observations. sigma_type: Character specifying decomposition variance-covariance matrix. datamat: Matrix containing 'Y' 'X'. config: List containing information configuration parameters.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"VAR(p) model following form: \\( \\boldsymbol{y}^\\prime_t = \\boldsymbol{\\iota}^\\prime + \\boldsymbol{x}^\\prime_t\\boldsymbol{\\Phi} + \\boldsymbol{\\epsilon}^\\prime_t\\), \\(\\boldsymbol{y}_t\\) \\(M\\)-dimensional vector dependent variables \\(\\boldsymbol{\\epsilon}_t\\) error term dimension. \\(\\boldsymbol{x}_t\\) \\(K=pM\\)-dimensional vector containing lagged/past values dependent variables \\(\\boldsymbol{y}_{t-l}\\) \\(l=1,\\dots,p\\) \\(\\boldsymbol{\\iota}\\) constant term (intercept) dimension \\(M\\times 1\\). reduced-form coefficient matrix \\(\\boldsymbol{\\Phi}\\) dimension \\(K \\times M\\). bvar offers two different specifications errors: user can choose factor stochastic volatility structure cholesky stochastic volatility structure. cases disturbances \\(\\boldsymbol{\\epsilon}_t\\) assumed follow \\(M\\)-dimensional multivariate normal distribution zero mean variance-covariance matrix \\(\\boldsymbol{\\Sigma}_t\\). case cholesky specification \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{U}^{\\prime -1} \\boldsymbol{D}_t \\boldsymbol{U}^{-1}\\), \\(\\boldsymbol{U}^{-1}\\) upper unitriangular (ones diagonal). diagonal matrix \\(\\boldsymbol{D}_t\\) depends upon latent log-variances, .e. \\(\\boldsymbol{D}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\). case factor structure, \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{\\Lambda} \\boldsymbol{V}_t \\boldsymbol{\\Lambda}^\\prime + \\boldsymbol{G}_t\\). diagonal matrices \\(\\boldsymbol{V}_t\\) \\(\\boldsymbol{G}_t\\) depend upon latent log-variances, .e. \\(\\boldsymbol{G}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\) \\(\\boldsymbol{V}_t=diag(exp(h_{M+1,t}),\\dots, exp(h_{M+r,t})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\) \\(h_{M+j,t}\\sim N(\\phi_ih_{M+j,t-1},\\sigma_{M+j}^2)\\) \\(j=1,\\dots,r\\).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"mcmc-algorithm","dir":"Reference","previous_headings":"","what":"MCMC algorithm","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"sample efficiently reduced-form VAR coefficients assuming factor structure errors, equation per equation algorithm Kastner & Huber (2020) implemented. parameters latent variables associated factor-structure sampled using package factorstochvol-package's function update_fsv callable C-level . sample efficiently reduced-form VAR coefficients, assuming cholesky-structure errors, corrected triangular algorithm Carriero et al. (2021) implemented. SV parameters latent variables sampled using package stochvol's update_fast_sv function. precision parameters, .e. free -diagonal elements \\(\\boldsymbol{U}\\), sampled Cogley Sargent (2005).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"Gruber, L. Kastner, G. (2023). Forecasting macroeconomic data Bayesian VARs: Sparse dense? depends! arXiv:2206.04902. Kastner, G. Huber, F. Sparse (2020). Bayesian vector autoregressions huge dimensions. Journal Forecasting. 39, 1142--1165, doi:10.1002/.2680 . Kastner, G. (2019). Sparse Bayesian Time-Varying Covariance Estimation Many Dimensions Journal Econometrics, 210(1), 98--115, doi:10.1016/j.jeconom.2018.11.007 . Carriero, . Chan, J.  Clark, T. E. Marcellino, M. (2021). Corrigendum “Large Bayesian vector autoregressions stochastic volatility non-conjugate priors” [J. Econometrics 212 (1) (2019) 137–154]. Journal Econometrics, doi:10.1016/j.jeconom.2021.11.010 . Cogley, S. Sargent, T. (2005). Drifts volatilities: monetary policies outcomes post WWII US. Review Economic Dynamics, 8, 262--302, doi:10.1016/j.red.2004.10.009 . Hosszejni, D. Kastner, G. (2021). Modeling Univariate Multivariate Stochastic Volatility R stochvol factorstochvol. Journal Statistical Software, 100, 1–-34. doi:10.18637/jss.v100.i12 .","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Markov Chain Monte Carlo Sampling for Bayesian Vectorautoregressions — bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Plot plot(mod)   # Summary summary(mod) #>  #> Posterior median of reduced-form coefficients: #>                 GDPC1 CPIAUCSL   FEDFUNDS #> GDPC1.l1     0.236588 0.006926  2.109e-02 #> CPIAUCSL.l1 -0.053390 0.616162 -3.646e-03 #> FEDFUNDS.l1  0.007874 0.037088  1.001e+00 #> intercept    0.005874 0.001291 -8.055e-05 #>  #> Posterior interquartile range of of reduced-form coefficients: #>                GDPC1  CPIAUCSL  FEDFUNDS #> GDPC1.l1    0.089598 0.0290897 0.0252782 #> CPIAUCSL.l1 0.111515 0.0904426 0.0150976 #> FEDFUNDS.l1 0.018416 0.0141909 0.0076726 #> intercept   0.001135 0.0007528 0.0001651"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract VAR coefficients — coef","title":"Extract VAR coefficients — coef","text":"Extracts posterior draws VAR coefficients VAR model estimated bvar().","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract VAR coefficients — coef","text":"","code":"# S3 method for bayesianVARs_bvar coef(object, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract VAR coefficients — coef","text":"object bayesianVARs_bvar object obtained bvar(). ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract VAR coefficients — coef","text":"Returns numeric array dimension \\(M \\times K \\times draws\\), M number time-series, K number covariates per equation (including intercept) draws number stored posterior draws.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/coef.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract VAR coefficients — coef","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of VAR coefficients bvar_coefs <- coef(mod)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"Simulates fitted/predicted (-sample) values estimated VAR model.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar fitted(object, error_term = TRUE, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"object bayesianVARs_bvar object estimated via bvar(). error_term logical indicating whether include error term . ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"object class bayesianVARs_fitted.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/fitted.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Simulate fitted/predicted historical values for an estimated VAR model — fitted.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate predicted historical values including the error term. pred <- fitted(mod, error_term = TRUE)  # Simulate fitted historical values not including the error term. fit <- fitted(mod, error_term = FALSE)  # Visualize plot(pred)  plot(fit)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":null,"dir":"Reference","previous_headings":"","what":"Draw from generalized inverse Gaussian — my_gig","title":"Draw from generalized inverse Gaussian — my_gig","text":"Vectorized version rgig","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Draw from generalized inverse Gaussian — my_gig","text":"","code":"my_gig(n, lambda, chi, psi)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Draw from generalized inverse Gaussian — my_gig","text":"n single integer indicating number draws generate. lambda vector shape parameters. chi vector shape/scale parameters. Must nonnegative positive lambdas positive else. psi vector shape/scale parameters. Must nonnegative negative lambdas positive else.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Draw from generalized inverse Gaussian — my_gig","text":"Matrix dimension c(n,m), m maximum length lambda, psi chi.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/my_gig.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Draw from generalized inverse Gaussian — my_gig","text":"","code":"gigsamples <- my_gig(2, c(1,1), c(1,1), c(1,1))"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"Pairwise visualization --sample posterior predictive densities.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"","code":"# S3 method for bayesianVARs_predict pairs(x, vars, ahead, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"x object class bayesianVARs_predict obtained via predict.bayesianVARs_bvar(). vars Integer vector (coercible ) indicating variables plot. ahead Integer vector (coercible ) indicating step ahead plot. max(ahead) must smaller equal dim(x$predictions)[1]. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"note","dir":"Reference","previous_headings":"","what":"Note","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"Note bayesianVARs_predict can also used withing plot.bayesianVARs_bvar().","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/pairs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Pairwise visualization of out-of-sample posterior predictive\ndensities. — pairs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate from posterior predictive predictions <- predict(mod, ahead = 1:3) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  485 stable posterior draws remaining for prediction!  # Visualize pairs(predictions, vars = 1:3, ahead = 1:3)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"Visualization -sample fit. Can also used display prediction intervals future values.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar plot(   x,   predictions = NULL,   quantiles = c(0.05, 0.5, 0.95),   dates = NULL,   n_col = 1,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"x object class bayesianVARs_bvar obtained via bvar(). predictions Optional array sample predictions, e.g. obtained via predict.bayesianVARs_bvar(). quantiles numeric vector indicating quantiles plot. dates optional vector dates labelling x-axis. default values NULL; case, axis labeled numbers. n_col integer indicating number columns use plotting. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"Returns x invisibly.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Plot method for bayesianVARs_bvar — plot.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate from posterior predictive predictions <- predict(mod, ahead = 1:3) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  626 stable posterior draws remaining for prediction!  # Visualize plot(mod, predictions = predictions)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":null,"dir":"Reference","previous_headings":"","what":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"Visualization -sample fit estimated VAR.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"","code":"# S3 method for bayesianVARs_fitted plot(   x,   dates = NULL,   vars = \"all\",   quantiles = c(0.05, 0.5, 0.95),   n_col = 1L,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"x bayesianVARs_fitted object. dates optional vector dates labelling x-axis. default values NULL; case, axis labeled numbers. vars character vector containing names variables visualized. default \"\" indicating fit variables visualized. quantiles numeric vector indicating quantiles plot. n_col integer indicating number columns use plotting. ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"returns x invisibly","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_fitted.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Visualization of in-sample fit of an estimated VAR. — plot.bayesianVARs_fitted","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate predicted historical values including the error term. pred <- fitted(mod, error_term = TRUE)  # Visualize plot(pred)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Fan chart — plot.bayesianVARs_predict","title":"Fan chart — plot.bayesianVARs_predict","text":"Visualization (--sample) predictive distribution.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Fan chart — plot.bayesianVARs_predict","text":"","code":"# S3 method for bayesianVARs_predict plot(   x,   dates = NULL,   vars = \"all\",   ahead = NULL,   quantiles = c(0.05, 0.25, 0.5, 0.75, 0.95),   n_col = 1L,   first_obs = 1L,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Fan chart — plot.bayesianVARs_predict","text":"x object type bayesianVARs_predict obtained via predict.bayesianVARs_bvar(). dates optional vector dates labeling x-axis. default values NULL; case, axis labeled numbers. vars character vector containing names variables visualized. default \"\" indicating variables visualized. ahead Integer vector (coercible ) indicating step ahead plot. max(ahead) must smaller equal dim(x$predictions)[1]. quantiles numeric vector indicating quantiles plot. n_col integer indicating number columns use plotting. first_obs integer indicating first observation used plotting. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Fan chart — plot.bayesianVARs_predict","text":"Returns x invisibly!","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/plot.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Fan chart — plot.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Simulate from posterior predictive predictions <- predict(mod, ahead = 1:3) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  385 stable posterior draws remaining for prediction!  # Visualize plot(predictions, vars = 1:3, ahead = 1:3)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":null,"dir":"Reference","previous_headings":"","what":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"Posterior heatmaps VAR coefficients variance-covariance matrices","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"","code":"posterior_heatmap(   x,   FUN,   ...,   colorbar = TRUE,   xlabels = NULL,   ylabels = NULL,   add_numbers = FALSE,   zlim = NULL,   colspace = NULL,   main = \"\",   cex.axis = 0.75,   cex.colbar = 1,   cex.numbers = 1,   asp = NULL )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"x array dimension \\(\\times b \\times draws\\), \\(  \\times b\\) dimension parameter visualize draws number posterior draws. FUN summary function applied margins c(1,2) x. E.g. \"median\", \"mean\", \"IQR\", \"sd\" \"var\". apply(x, 1:2, FUN, ...) must return matrix! ... optional arguments FUN. colorbar logical indicating whether display colorbar . Default TRUE. xlabels ylabels=NULL, default, indicates names dependent variables displayed. ylabels=\"\" indicates ylabels displayed. ylabels xlabels=NULL, default, indicates labels covariables (lagged values dependent variables) displayed. xlabels=\"lags\" indicates lags marked. xlabels=\"\" indicates ylabels displayed. add_numbers logical. add_numbers=TRUE, default indicates actual values summary displayed. zlim numeric vector length two indicating minimum maximum values colors plotted. default range determined maximum absolute values selected summary. colspace Optional argument. main main title plot. cex.axis magnification used y-axis annotation relative current setting cex. cex.colbar magnification used colorbar annotation relative current setting cex. cex.numbers magnification used actual values (add_numbers=TRUE) relative current setting cex. asp aspect ratio.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"Returns x invisibly.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/posterior_heatmap.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Posterior heatmaps for VAR coefficients or variance-covariance matrices — posterior_heatmap","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(100*data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of VAR coefficients phi_post <- coef(mod)  # Visualize posterior median of VAR coefficients posterior_heatmap(phi_post, median)   # Extract posterior draws of variance-covariance matrices (for each point in time) sigma_post <- vcov(mod) # Visualize posterior interquartile-range of variance-covariance matrix of the first observation posterior_heatmap(sigma_post[1,,,], IQR)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"Simulates (--sample) predictive density Bayesian VARs estimated via bvar() computes log predictive likelhoods ex-post observed data supplied.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar predict(   object,   ahead = 1L,   each = 1L,   stable = TRUE,   simulate_predictive = TRUE,   LPL = FALSE,   Y_obs = NA,   LPL_VoI = NA,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"object bayesianVARs_bvar object, obtained bvar(). ahead Integer vector (coercible ), indicating number steps ahead predict. Single integer (coercible ) indicating often drawn posterior predictive distribution draw stored MCMC sampling. stable logical indicating whether consider draws posterior fulfill 'stable' criterion. Default TRUE. simulate_predictive logical, indicating whether posterior predictive distribution simulated. LPL logical indicating whether ahead-step-ahead log predictive likelihoods computed. LPL=TRUE, Y_obs specified. Y_obs Data matrix observed values computation log predictive likelihood. ncol(object$Yraw) columns assumed contain single time-series length length(ahead). LPL_VoI either integer vector character vector column-names indicating subgroup time-series object$Yraw joint log predictive likelihood shall computed. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"Object class bayesianVARs_predict, list may contain following elements: predictions array dimensions c(length(ahead), ncol(object$Yraw), * dim(object$PHI)[3]) containing simulations predictive density (simulate_predictive=TRUE). LPL vector length length(ahead) containing log-predictive-likelihoods (taking account joint distribution variables) (LPL=TRUE). LPL_univariate matrix dimension c(length(ahead), ncol(object$Yraw) containing marginalized univariate log-predictive-likelihoods series (LPL=TRUE). LPL_VoI vector length length(ahead) containing log-predictive-likelihoods subset variables (LPL=TRUE LPL_VoI != NA). Yraw matrix containing data used estimation VAR. LPL_draws matrix containing simulations log-predictive-likelihood (LPL=TRUE). PL_univariate_draws array containing simulations univariate predictive-likelihoods (LPL=TRUE). LPL_sub_draws matrix containing simulations log-predictive-likelihood subset variables (LPL=TRUE LPL_VoI != NA).","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/predict.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Predict method for Bayesian VARs — predict.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step to 4-steps ahead posterior predictive and compute # log-predictive-likelihoods predictions <- predict(mod, ahead = 1:4, LPL = TRUE, Y_obs = test) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  617 stable posterior draws remaining for prediction!  # Summary summary(predictions) #>  #> LPL: #>   t+1   t+2   t+3   t+4  #> 5.114 9.665 9.092 6.526  #>  #> Marginal univariate LPLs: #>       GDPC1 CPIAUCSL FEDFUNDS #> t+1 -0.8729   0.4109    3.541 #> t+2  2.8555   2.6739    3.391 #> t+3  2.8363   2.4370    3.349 #> t+4  2.8931   0.2959    3.326 #>  #> Prediction quantiles: #> , , GDPC1 #>  #>          t+1        t+2       t+3       t+4 #> 5%  -0.07875 -0.0329135 -0.021393 -0.011870 #> 50% -0.01851  0.0009099  0.005688  0.006618 #> 95%  0.03798  0.0429166  0.034118  0.024602 #>  #> , , CPIAUCSL #>  #>           t+1       t+2       t+3        t+4 #> 5%  -0.018515 -0.017824 -0.015592 -0.0140807 #> 50% -0.007656 -0.005287 -0.002413 -0.0005886 #> 95%  0.002679  0.006747  0.009743  0.0106464 #>  #> , , FEDFUNDS #>  #>           t+1       t+2       t+3       t+4 #> 5%  -0.016214 -0.022544 -0.026463 -0.030055 #> 50% -0.004527 -0.005249 -0.006258 -0.006404 #> 95%  0.005912  0.011279  0.016238  0.018100 #>   # Visualize via fan-charts plot(predictions)   # \\donttest{ # In order to evaluate the joint predictive density of a subset of the # variables (variables of interest), consider specifying 'LPL_VoI': predictions <- predict(mod, ahead = 1:4, LPL = TRUE, Y_obs = test, LPL_VoI = c(\"GDPC1\",\"FEDFUNDS\")) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  617 stable posterior draws remaining for prediction! predictions$LPL_VoI #>      t+1      t+2      t+3      t+4  #> 2.542386 6.712653 6.548112 6.391536  # }"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Pretty printing of a bvar object — print.bayesianVARs_bvar","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"Pretty printing bvar object","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"x Object class bayesianVARs_bvar, usually resulting call bvar(). ... Ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Pretty printing of a bvar object — print.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Print model mod #>  #> Fitted bayesianVARs_bvar object with #>   -       3 series #>   -       1 lag(s) #>   -     246 used observations #>   -     247 total observations #>   -    1000 MCMC draws #>   -       1 thinning #>   -    1000 burn-in #>"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"Print method bayesianVARs_predict objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"","code":"# S3 method for bayesianVARs_predict print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"x bayesianVARs_predict object obtained via predict.bayesianVARs_bvar(). ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print method for bayesianVARs_predict objects — print.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step ahead posterior predictive predictions <- predict(mod, ahead = 1L) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  595 stable posterior draws remaining for prediction! print(predictions) #>  #> Generic functions for bayesianVARs_predict objects: #>   - summary.bayesianVARs_predict(), #>   - pairs.bayesianVARs_predict(), #>   - plot.bayesianVARs_predict() (alias for pairs.bayesianVARs_predict())."},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"Print method summary.bayesianVARs_bvar objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"","code":"# S3 method for summary.bayesianVARs_bvar print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"x summary.bayesianVARs_bvar object obtained via summary.bayesianVARs_bvar(). ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"Returns x invisibly!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print method for summary.bayesianVARs_bvar objects — print.summary.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate model mod <- bvar(data, quiet = TRUE)  # Print summary summary(mod) #>  #> Posterior median of reduced-form coefficients: #>                 GDPC1 CPIAUCSL   FEDFUNDS #> GDPC1.l1     0.226152  0.00604  1.362e-02 #> CPIAUCSL.l1 -0.063468  0.62045 -4.773e-03 #> FEDFUNDS.l1  0.009492  0.03717  1.001e+00 #> intercept    0.005959  0.00120 -4.555e-05 #>  #> Posterior interquartile range of of reduced-form coefficients: #>                GDPC1  CPIAUCSL  FEDFUNDS #> GDPC1.l1    0.105445 0.0271618 0.0208810 #> CPIAUCSL.l1 0.118059 0.0930382 0.0131568 #> FEDFUNDS.l1 0.019176 0.0145773 0.0065391 #> intercept   0.001173 0.0007342 0.0001668"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"Print method summary.bayesianVARs_predict objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"","code":"# S3 method for summary.bayesianVARs_predict print(x, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"x summary.bayesianVARs_predict object obtained via summary.bayesianVARs_predict(). ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"Returns x invisibly.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/print.summary.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Print method for summary.bayesianVARs_predict objects — print.summary.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step ahead posterior predictive predictions <- predict(mod, ahead = 1L) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  586 stable posterior draws remaining for prediction! sum <- summary(predictions) print(sum) #>  #> Prediction quantiles: #> , , GDPC1 #>  #>          t+1 #> 5%  -0.06812 #> 50% -0.01798 #> 95%  0.02909 #>  #> , , CPIAUCSL #>  #>           t+1 #> 5%  -0.018438 #> 50% -0.008339 #> 95%  0.003312 #>  #> , , FEDFUNDS #>  #>           t+1 #> 5%  -0.021399 #> 50% -0.004006 #> 95%  0.009963 #>"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":null,"dir":"Reference","previous_headings":"","what":"Specify prior on PHI — specify_prior_phi","title":"Specify prior on PHI — specify_prior_phi","text":"Configures prior PHI, matrix reduced-form VAR coefficients.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Specify prior on PHI — specify_prior_phi","text":"","code":"specify_prior_phi(   data = NULL,   M = ncol(data),   lags = 1L,   prior = \"HS\",   priormean = 0,   PHI_tol = 1e-18,   DL_a = \"1/K\",   DL_tol = 0,   R2D2_a = 0.1,   R2D2_b = 0.5,   R2D2_tol = 0,   NG_a = 0.1,   NG_b = 1,   NG_c = 1,   NG_tol = 0,   SSVS_c0 = 0.01,   SSVS_c1 = 100,   SSVS_semiautomatic = TRUE,   SSVS_p = 0.5,   HMP_lambda1 = c(0.01, 0.01),   HMP_lambda2 = c(0.01, 0.01),   normal_sds = 10,   global_grouping = \"global\",   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Specify prior on PHI — specify_prior_phi","text":"data Optional. Data matrix (can time series object). \\(M\\) columns assumed contain single time-series length \\(T\\). M positive integer indicating number time-series VAR. lags positive integer indicating order VAR, .e. number lags dependent variables included predictors. prior character, one \"HS\", \"R2D2\", \"NG\", \"DL\", \"SSVS\", \"HMP\" \"normal\". priormean real numbers indicating prior means VAR coefficients. One single number means prior mean -lag coefficients w.r.t. first lag equals priormean 0 else. vector length M means prior mean -lag coefficients w.r.t. first lag equals priormean 0 else. priormean matrix dimension c(lags*M,M), \\(M\\) columns assumed contain lags*M prior means VAR coefficients respective VAR equations. PHI_tol Minimum number absolute value VAR coefficient draw can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. DL_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. argument global_grouping specifies e.g. k groups, DL_a can numeric vector length k elements indicate shrinkage group.  matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. DL_a specified prior=\"DL\". DL_tol Minimum number parameter draw one shrinking parameters Dirichlet Laplace prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. DL_tol specified prior=\"DL\". R2D2_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. argument global_grouping specifies e.g. k groups, R2D2_a can numeric vector length k elements indicate shrinkage group.  matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. R2D2_a specified prior=\"R2D2\". R2D2_b (Single) positive real number. value indicates shape parameter inverse gamma prior (semi-)global scales. argument global_grouping specifies e.g. k groups, NG_b can numeric vector length k elements determine shape parameter group. R2D2_b specified prior=\"R2D2\". R2D2_tol Minimum number parameter draw one shrinking parameters R2D2 prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. R2D2_tol specified prior=\"R2D2\". NG_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. argument global_grouping specifies e.g. k groups, NG_a can numeric vector length k elements indicate shrinkage group.  matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. NG_a specified prior=\"NG\". NG_b (Single) positive real number. value indicates shape parameter inverse gamma prior (semi-)global scales. argument global_grouping specifies e.g. k groups, NG_b can numeric vector length k elements determine shape parameter group. NG_b specified prior=\"NG\". NG_c (Single) positive real number. value indicates scale parameter inverse gamma prior (semi-)global scales. argument global_grouping specifies e.g. k groups, NG_c can numeric vector length k elements determine scale parameter group. Expert option set scale parameter proportional NG_a. E.g. case discrete hyperprior NG_a chosen, desired proportion say 0.2 achieved setting NG_c=\"0.2a\" (character input!). NG_c specified prior=\"NG\". NG_tol Minimum number parameter draw one shrinking parameters normal-gamma prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. NG_tol specified prior=\"NG\". SSVS_c0 single positive number indicating (unscaled) standard deviation spike component. SSVS_c0 specified prior=\"SSVS\". \\(SSVS_{c0} \\ll SSVS_{c1}\\)! SSVS_c0 specified prior=\"SSVS\". SSVS_c1 single positive number indicating (unscaled) standard deviation slab component. SSVS_c0 specified prior=\"SSVS\". \\(SSVS_{c0} \\ll SSVS_{c1}\\)! SSVS_semiautomatic logical. SSVS_semiautomatic=TRUE SSVS_c0 SSVS_c1 scaled variances posterior PHI FLAT conjugate (dependent Normal-Wishart prior). SSVS_semiautomatic specified prior=\"SSVS\". SSVS_p Either single positive number range (0,1) indicating (fixed) prior inclusion probability coefficient. numeric vector length 2 positive entries indicating shape parameters Beta distribution. case Beta hyperprior placed prior inclusion probability. SSVS_p specified prior=\"SSVS\". HMP_lambda1 numeric vector length 2. entries must positive. first indicates shape second rate Gamma hyperprior -lag coefficients. HMP_lambda1 specified prior=\"HMP\". HMP_lambda2 numeric vector length 2. entries must positive. first indicates shape second rate Gamma hyperprior cross-lag coefficients. HMP_lambda2 specified prior=\"HMP\". normal_sds numeric vector length \\(n\\), \\(n = lags M^2\\) number VAR coefficients (excluding intercept), indicating prior variances. single number recycled accordingly! Must positive. normal_sds specified prior=\"normal\". global_grouping One \"global\", \"equation-wise\", \"covariate-wise\", \"olcl-lagwise\" \"fol\" indicating sub-groups semi-global(-local) modifications HS, R2D2, NG, DL SSVS prior. Works also user-specified indicator matrix dimension c(lags*M,M).  relevant prior=\"HS\", prior=\"DL\", prior=\"R2D2\", prior=\"NG\" prior=\"SSVS\". ... use!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Specify prior on PHI — specify_prior_phi","text":"baysianVARs_prior_phi-object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Specify prior on PHI — specify_prior_phi","text":"details concerning prior-elicitation VARs please see Gruber & Kastner (2023). Currently one can choose six hierarchical shrinkage priors normal prior: prior=\"HS\" stands Horseshoe-prior, prior=\"R2D2 R\\(^2\\)-induced-Dirichlet-decompostion-prior, prior=\"NG\" normal-gamma-prior, prior=\"DL\" Dirichlet-Laplace-prior, prior=\"SSVS\" stochastic-search-variable-selection-prior, prior=\"HMP\" semi-hierarchical Minnesota prior prior=normal normal-prior. Semi-global shrinkage, .e. group-specific shrinkage pre-specified subgroups coefficients, can achieved argument global_grouping.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Specify prior on PHI — specify_prior_phi","text":"Gruber, L. Kastner, G. (2023). Forecasting macroeconomic data Bayesian VARs: Sparse dense? depends! arXiv:2206.04902.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_phi.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Specify prior on PHI — specify_prior_phi","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Horseshoe prior for a VAR(2) phi_hs <- specify_prior_phi(data = data, lags = 2L ,prior = \"HS\")  # Semi-global-local Horseshoe prior for a VAR(2) with semi-global shrinkage parameters for # cross-lag and own-lag coefficients in each lag phi_hs_sg <- specify_prior_phi(data = data, lags = 2L, prior = \"HS\", global_grouping = \"olcl-lagwise\")  # Semi-global-local Horseshoe prior for a VAR(2) with equation-wise shrinkage # construct indicator matrix for equation-wise shrinkage semi_global_mat <- matrix(1:ncol(data), 2*ncol(data), ncol(data), byrow = TRUE) phi_hs_ew <- specify_prior_phi(data = data, lags = 2L, prior = \"HS\", global_grouping = semi_global_mat) # (for equation-wise shrinkage one can also use 'global_grouping = \"equation-wise\"')  # \\donttest{ # Estimate model with your prior configuration of choice mod <- bvar(data, lags = 2L, prior_phi = phi_hs_sg, quiet = TRUE) # }"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":null,"dir":"Reference","previous_headings":"","what":"Specify prior on Sigma — specify_prior_sigma","title":"Specify prior on Sigma — specify_prior_sigma","text":"Configures prior variance-covariance VAR.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Specify prior on Sigma — specify_prior_sigma","text":"","code":"specify_prior_sigma(   data = NULL,   M = ncol(data),   type = c(\"factor\", \"cholesky\"),   factor_factors = 1L,   factor_restrict = c(\"none\", \"upper\"),   factor_priorfacloadtype = c(\"rowwiseng\", \"colwiseng\", \"normal\"),   factor_priorfacload = 0.1,   factor_facloadtol = 1e-18,   factor_priorng = c(1, 1),   factor_priormu = c(0, 10),   factor_priorphiidi = c(10, 3),   factor_priorphifac = c(10, 3),   factor_priorsigmaidi = 1,   factor_priorsigmafac = 1,   factor_priorh0idi = \"stationary\",   factor_priorh0fac = \"stationary\",   factor_heteroskedastic = TRUE,   factor_priorhomoskedastic = NA,   factor_interweaving = 4,   cholesky_U_prior = c(\"HS\", \"DL\", \"R2D2\", \"NG\", \"SSVS\", \"normal\", \"HMP\"),   cholesky_U_tol = 1e-18,   cholesky_heteroscedastic = TRUE,   cholesky_priormu = c(0, 100),   cholesky_priorphi = c(20, 1.5),   cholesky_priorsigma2 = c(0.5, 0.5),   cholesky_priorh0 = \"stationary\",   cholesky_priorhomoscedastic = as.numeric(NA),   cholesky_DL_a = \"1/n\",   cholesky_DL_tol = 0,   cholesky_R2D2_a = 0.4,   cholesky_R2D2_b = 0.5,   cholesky_R2D2_tol = 0,   cholesky_NG_a = 0.5,   cholesky_NG_b = 0.5,   cholesky_NG_c = 0.5,   cholesky_NG_tol = 0,   cholesky_SSVS_c0 = 0.001,   cholesky_SSVS_c1 = 1,   cholesky_SSVS_p = 0.5,   cholesky_HMP_lambda3 = c(0.01, 0.01),   cholesky_normal_sds = 10,   expert_sv_offset = 0,   quiet = FALSE,   ... )"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Specify prior on Sigma — specify_prior_sigma","text":"data Optional. Data matrix (can time series object). \\(M\\) columns assumed contain single time-series length \\(T\\). M positive integer indicating number time-series VAR. type character, one \"factor\" (default) \"cholesky\", indicating decomposition applied covariance-matrix. factor_factors Number latent factors estimated. required type=\"factor\". factor_restrict Either \"upper\" \"none\", indicating whether factor loadings matrix restricted zeros diagonal (\"upper\") whether elements estimated data (\"none\"). Setting restrict \"upper\" often stabilizes MCMC estimation can important identifying factor loadings matrix, however, generally strong prior assumption. Setting restrict \"none\" usually preferred option identification factor loadings matrix less concern covariance estimation prediction goal. required type=\"factor\". factor_priorfacloadtype Can \"normal\", \"rowwiseng\", \"colwiseng\". required type=\"factor\". \"normal\": Normal prior. value priorfacload interpreted standard deviations Gaussian prior distributions factor loadings. \"rowwiseng\": Row-wise Normal-Gamma prior. value priorfacload interpreted shrinkage parameter . \"colwiseng\": Column-wise Normal-Gamma prior. value priorfacload interpreted shrinkage parameter . details please see Kastner (2019). factor_priorfacload Either matrix dimensions M times factor_factors positive elements single number (recycled accordingly). required type=\"factor\". meaning factor_priorfacload depends setting factor_priorfacloadtype explained . factor_facloadtol Minimum number absolute value factor loadings draw can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. required type=\"factor\". factor_priorng Two-element vector positive entries indicating Normal-Gamma prior's hyperhyperparameters c d (cf. Kastner (2019)). required type=\"factor\". factor_priormu Vector length 2 denoting prior mean standard deviation unconditional levels idiosyncratic log variance processes. required type=\"factor\". factor_priorphiidi Vector length 2, indicating shape parameters Beta prior distributions transformed parameters (phi+1)/2, phi denotes persistence idiosyncratic log variances. required type=\"factor\". factor_priorphifac Vector length 2, indicating shape parameters Beta prior distributions transformed parameters (phi+1)/2, phi denotes persistence factor log variances. required type=\"factor\". factor_priorsigmaidi Vector length M containing prior volatilities log variances. factor_priorsigmaidi exactly one element, recycled idiosyncratic log variances. required type=\"factor\". factor_priorsigmafac Vector length factor_factors containing prior volatilities log variances. factor_priorsigmafac exactly one element, recycled factor log variances. required type=\"factor\". factor_priorh0idi Vector length 1 M, containing information Gaussian prior initial idiosyncratic log variances. required type=\"factor\". element factor_priorh0idi nonnegative number, conditional prior corresponding initial log variance h0 assumed Gaussian mean 0 standard deviation factor_priorh0idi times \\(sigma\\). element factor_priorh0idi string 'stationary', prior corresponding initial log volatility taken stationary distribution, .e. h0 assumed Gaussian mean 0 variance \\(sigma^2/(1-phi^2)\\). factor_priorh0fac Vector length 1 factor_factors, containing information Gaussian prior initial factor log variances. required type=\"factor\". element factor_priorh0fac nonnegative number, conditional prior corresponding initial log variance h0 assumed Gaussian mean 0 standard deviation factor_priorh0fac times \\(sigma\\). element factor_priorh0fac string 'stationary', prior corresponding initial log volatility taken stationary distribution, .e. h0 assumed Gaussian mean 0 variance \\(sigma^2/(1-phi^2)\\). factor_heteroskedastic Vector length 1, 2, M +  factor_factors, containing logical values indicating whether time-varying (factor_heteroskedastic = TRUE) constant (factor_heteroskedastic =   FALSE) variance estimated. factor_heteroskedastic length 2 recycled accordingly, whereby first element used idiosyncratic variances second element used factor variances. required type=\"factor\". factor_priorhomoskedastic used least one element factor_heteroskedastic set FALSE. case, factor_priorhomoskedastic must matrix positive entries dimension c(M, 2). Values column 1 interpreted shape values column 2 interpreted rate parameter corresponding inverse gamma prior distribution idiosyncratic variances. required type=\"factor\". factor_interweaving following values interweaving factor loadings accepted (required type=\"factor\"): 0: interweaving. 1: Shallow interweaving diagonal entries. 2: Deep interweaving diagonal entries. 3: Shallow interweaving largest absolute entries column. 4: Deep interweaving largest absolute entries column. details please see Kastner et al. (2017). value 4 highly recommended default. cholesky_U_prior character, one \"HS\", \"R2D2\", \"NG\", \"DL\", \"SSVS\", \"HMP\" \"normal\". required type=\"cholesky\". cholesky_U_tol Minimum number absolute value free -diagonal element \\(U\\)-draw can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. required type=\"cholesky\". cholesky_heteroscedastic single logical indicating whether time-varying (cholesky_heteroscedastic = TRUE) constant (cholesky_heteroscedastic = FALSE) variance estimated. required type=\"cholesky\". cholesky_priormu Vector length 2 denoting prior mean standard deviation unconditional levels log variance processes. required type=\"cholesky\". cholesky_priorphi Vector length 2, indicating shape parameters Beta prior distributions transformed parameters (phi+1)/2, phi denotes persistence log variances. required type=\"cholesky\". cholesky_priorsigma2 Vector length 2, indicating shape rate Gamma prior distributions variance log variance processes. (Currently one global setting \\(M\\) processes supported). required type=\"cholesky\". cholesky_priorh0 Vector length 1 M, containing information Gaussian prior initial idiosyncratic log variances. required type=\"cholesky\". element cholesky_priorh0 nonnegative number, conditional prior corresponding initial log variance h0 assumed Gaussian mean 0 standard deviation cholesky_priorh0 times \\(sigma\\). element cholesky_priorh0 string 'stationary', prior corresponding initial log volatility taken stationary distribution, .e. h0 assumed Gaussian mean 0 variance \\(sigma^2/(1-phi^2)\\). cholesky_priorhomoscedastic used cholesky_heteroscedastic=FALSE. case, cholesky_priorhomoscedastic must matrix positive entries dimension c(M, 2). Values column 1 interpreted shape values column 2 interpreted scale parameter corresponding inverse gamma prior distribution variances. required type=\"cholesky\". cholesky_DL_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. cholesky_DL_a specified cholesky_U_prior=\"DL\". cholesky_DL_tol Minimum number parameter draw one shrinking parameters Dirichlet Laplace prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. DL_tol specified cholesky_U_prior=\"DL\". cholesky_R2D2_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. cholesky_R2D2_a specified cholesky_U_prior=\"R2D2\". cholesky_R2D2_b single positive number, greater values indicate heavier regularization. cholesky_R2D2_b specified cholesky_U_prior=\"R2D2\". cholesky_R2D2_tol Minimum number parameter draw one shrinking parameters R2D2 prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. cholesky_R2D2_tol specified cholesky_U_prior=\"R2D2\". cholesky_NG_a (Single) positive real number. value interpreted concentration parameter local scales. Smaller values enforce heavier shrinkage. matrix dimension c(s,2) specifies discrete hyperprior, first column contains s support points second column contains associated prior probabilities. cholesky_NG_a specified cholesky_U_prior=\"NG\". cholesky_NG_b (Single) positive real number. value indicates shape parameter inverse gamma prior global scales. cholesky_NG_b specified cholesky_U_prior=\"NG\". cholesky_NG_c (Single) positive real number. value indicates scale parameter inverse gamma prior global scales. Expert option set scale parameter proportional NG_a. E.g. case discrete hyperprior NG_a chosen, desired proportion say 0.2 achieved setting NG_c=\"0.2a\" (character input!). cholesky_NG_c specified cholesky_U_prior=\"NG\". cholesky_NG_tol Minimum number parameter draw one shrinking parameters normal-gamma prior can take. Prevents numerical issues can appear strong shrinkage enforced chosen greater zero. cholesky_NG_tol specified cholesky_U_prior=\"NG\". cholesky_SSVS_c0 single positive number indicating (unscaled) standard deviation spike component. cholesky_SSVS_c0 specified choleksy_U_prior=\"SSVS\". \\(SSVS_{c0}   \\ll SSVS_{c1}\\)! cholesky_SSVS_c1 single positive number indicating (unscaled) standard deviation slab component. cholesky_SSVS_c1 specified choleksy_U_prior=\"SSVS\". \\(SSVS_{c0} \\ll SSVS_{c1}\\)! cholesky_SSVS_p Either single positive number range (0,1) indicating (fixed) prior inclusion probability coefficient. numeric vector length 2 positive entries indicating shape parameters Beta distribution. case Beta hyperprior placed prior inclusion probability. cholesky_SSVS_p specified choleksy_U_prior=\"SSVS\". cholesky_HMP_lambda3 numeric vector length 2. entries must positive. first indicates shape second rate Gamma hyperprior contemporaneous coefficients. cholesky_HMP_lambda3 specified choleksy_U_prior=\"HMP\". cholesky_normal_sds numeric vector length \\(\\frac{M^2-M}{2}\\), indicating prior variances free -diagonal elements \\(U\\). single number recycled accordingly! Must positive. cholesky_normal_sds specified choleksy_U_prior=\"normal\". expert_sv_offset ... use! quiet logical indicating whether informative output omitted. ... use!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Specify prior on Sigma — specify_prior_sigma","text":"Object class bayesianVARs_prior_sigma.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"details","dir":"Reference","previous_headings":"","what":"Details","title":"Specify prior on Sigma — specify_prior_sigma","text":"bvar offers two different specifications errors: user can choose factor stochastic volatility structure cholesky stochastic volatility structure. cases disturbances \\(\\boldsymbol{\\epsilon}_t\\) assumed follow \\(M\\)-dimensional multivariate normal distribution zero mean variance-covariance matrix \\(\\boldsymbol{\\Sigma}_t\\). case cholesky specification \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{U}^{\\prime -1} \\boldsymbol{D}_t \\boldsymbol{U}^{-1}\\), \\(\\boldsymbol{U}^{-1}\\) upper unitriangular (ones diagonal). diagonal matrix \\(\\boldsymbol{D}_t\\) depends upon latent log-variances, .e. \\(\\boldsymbol{D}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\). case factor structure, \\(\\boldsymbol{\\Sigma}_t = \\boldsymbol{\\Lambda} \\boldsymbol{V}_t \\boldsymbol{\\Lambda}^\\prime + \\boldsymbol{G}_t\\). diagonal matrices \\(\\boldsymbol{V}_t\\) \\(\\boldsymbol{G}_t\\) depend upon latent log-variances, .e. \\(\\boldsymbol{G}_t=diag(exp(h_{1t}),\\dots, exp(h_{Mt})\\) \\(\\boldsymbol{V}_t=diag(exp(h_{M+1,t}),\\dots, exp(h_{M+r,t})\\). log-variances follow priori independent autoregressive processes \\(h_{}\\sim N(\\mu_i + \\phi_i(h_{,t-1}-\\mu_i),\\sigma_i^2)\\) \\(=1,\\dots,M\\) \\(h_{M+j,t}\\sim N(\\phi_ih_{M+j,t-1},\\sigma_{M+j}^2)\\) \\(j=1,\\dots,r\\).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Specify prior on Sigma — specify_prior_sigma","text":"Kastner, G. (2019). Sparse Bayesian Time-Varying Covariance Estimation Many Dimensions Journal Econometrics, 210(1), 98--115, doi:10.1016/j.jeconom.2018.11.007 Kastner, G., Frühwirth-Schnatter, S., Lopes, H.F. (2017). Efficient Bayesian Inference Multivariate Factor Stochastic Volatility Models. Journal Computational Graphical Statistics, 26(4), 905--917, doi:10.1080/10618600.2017.1322091 .","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/specify_prior_sigma.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Specify prior on Sigma — specify_prior_sigma","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # examples with stochastic volatility (heteroscedasticity) ----------------- # factor-decomposition with 2 factors and colwise normal-gamma prior on the loadings sigma_factor_cng_sv <- specify_prior_sigma(data = data, type = \"factor\", factor_factors = 2L, factor_priorfacloadtype = \"colwiseng\", factor_heteroskedastic = TRUE) #>  #> Since argument 'type' is specified with 'factor', all arguments starting with 'cholesky_' are being ignored.  # cholesky-decomposition with Dirichlet-Laplace prior on U sigma_cholesky_dl_sv <- specify_prior_sigma(data = data, type = \"cholesky\", cholesky_U_prior = \"DL\", cholesky_DL_a = 0.5, cholesky_heteroscedastic = TRUE) #>  #> Since argument 'type' is specified with 'cholesky', all arguments starting with 'factor_' are being ignored.  # examples without stochastic volatility (homoscedasticity) ---------------- # factor-decomposition with 2 factors and colwise normal-gamma prior on the loadings sigma_factor_cng <- specify_prior_sigma(data = data, type = \"factor\", factor_factors = 2L, factor_priorfacloadtype = \"colwiseng\", factor_heteroskedastic = FALSE, factor_priorhomoskedastic = matrix(c(0.5,0.5), ncol(data), 2)) #>  #> Since argument 'type' is specified with 'factor', all arguments starting with 'cholesky_' are being ignored. #>  #> Cannot do deep factor_interweaving if (some) factor_factors are homoskedastic. Setting 'factor_interweaving' to 3.  # cholesky-decomposition with Horseshoe prior on U sigma_cholesky_dl <- specify_prior_sigma(data = data, type = \"cholesky\", cholesky_U_prior = \"HS\", cholesky_heteroscedastic = FALSE) #>  #> Since argument 'type' is specified with 'cholesky', all arguments starting with 'factor_' are being ignored. #>  #> Argument 'cholesky_priorhomoscedastic' not specified. Setting both shape and rate of inverse gamma prior equal to 0.01.  # \\donttest{ # Estimate model with your prior configuration of choice mod <- bvar(data, prior_sigma = sigma_factor_cng_sv, quiet = TRUE) # }"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Stable posterior draws — stable_bvar","title":"Stable posterior draws — stable_bvar","text":"stable_bvar() detects discards posterior draws bayesianVARs_bvar object fulfill stability condition: VAR(p) model considered stable eigenvalues companion form matrix lie inside unit circle.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Stable posterior draws — stable_bvar","text":"","code":"stable_bvar(object, quiet = FALSE)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Stable posterior draws — stable_bvar","text":"object bayesianVARs_bvar object obtained via bvar(). quiet logical indicating whether informative output omitted.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Stable posterior draws — stable_bvar","text":"object type bayesianVARs_bvar.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/stable_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Stable posterior draws — stable_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Discard \"unstable\" draws stable_mod <- stable_bvar(mod) #>  #> Original 'bayesianVARs_bvar' object consists of 1000 posterior draws. #>  #> Detected 475 unstable draws. #>  #> Remaining draws: 525 !"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"Extract replace parts bayesianVARs_coef object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"","code":"# S3 method for bayesianVARs_coef [(x, i, j, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"x object type bayesianVARs_coef. indices j indices ... indices","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"object type bayesianVARs_coef.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_coef.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract or Replace Parts of a bayesianVARs_coef object — [.bayesianVARs_coef","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract coefficients, which are of class bayesianVARs_coef phi <- coef(mod) phi[1,1,1] #> [1] 0.2644423 #> attr(,\"class\") #> [1] \"bayesianVARs_coef\"  \"bayesianVARs_draws\""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"Extract replace parts bayesianVARs_draws object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"","code":"# S3 method for bayesianVARs_draws [(x, i, j, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"x object type bayesianVARs_draws. indices j indices ... indices","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"object type bayesianVARs_draws.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/sub-.bayesianVARs_draws.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract or Replace Parts of a bayesianVARs_draws object — [.bayesianVARs_draws","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract coefficients, which are of class bayesianVARs_draws phi <- coef(mod) phi[1,1,1] #> [1] 0.2662653 #> attr(,\"class\") #> [1] \"bayesianVARs_coef\"  \"bayesianVARs_draws\""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"Summary method bayesianVARs_bvar objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar summary(object, quantiles = c(0.025, 0.25, 0.5, 0.75, 0.975), ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"object bayesianVARs_bvar object obtained via bvar(). quantiles numeric vector quantiles compute. ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"object type summary.bayesianVARs_bvar.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Summary method for bayesianVARs_bvar objects — summary.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Estimate model mod <- bvar(data, quiet = TRUE)  # Summary sum <- summary(mod)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":null,"dir":"Reference","previous_headings":"","what":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"Summary statistics bayesianVARs posterior draws.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"","code":"# S3 method for bayesianVARs_draws summary(object, quantiles = c(0.25, 0.5, 0.75), ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"object object class bayesianVARs_draws usually obtained extractors like coef.bayesianVARs_bvar() vcov.bayesianVARs_bvar(). quantiles vector quantiles evaluate. ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"list object class bayesianVARs_draws_summary holding mean: Vector matrix containing posterior mean. sd: Vector matrix containing posterior standard deviation . quantiles: Array containing posterior quantiles.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_draws.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Summary statistics for bayesianVARs posterior draws. — summary.bayesianVARs_draws","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")] # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of VAR coefficients bvar_coefs <- coef(mod)  # Compute summary statistics summary_stats <- summary(bvar_coefs)  # Compute summary statistics of VAR coefficients without using coef() summary_stats <- summary(mod$PHI)  # Test which list elements of 'mod' are of class 'bayesianVARs_draws'. names(mod)[sapply(names(mod), function(x) inherits(mod[[x]], \"bayesianVARs_draws\"))] #> [1] \"PHI\"     \"U\"       \"logvar\"  \"sv_para\" \"facload\" \"fac\""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":null,"dir":"Reference","previous_headings":"","what":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"Summary method bayesianVARs_predict objects.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"","code":"# S3 method for bayesianVARs_predict summary(object, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"object bayesianVARs_predict object obtained via predict.bayesianVARs_bvar(). ... Currently ignored!","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"summary.bayesianVARs_predict object.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/summary.bayesianVARs_predict.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Summary method for bayesianVARs_predict objects — summary.bayesianVARs_predict","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")]  # Split data in train and test train <- data[1:(nrow(data)-4),] test <- data[-c(1:(nrow(data)-4)),]  # Estimate model using train data only mod <- bvar(train, quiet = TRUE)  # Simulate from 1-step ahead posterior predictive predictions <- predict(mod, ahead = 1L) #> 'stable=TRUE': Calling 'stable_bvar()' to discard those posterior #>           draws, that do not fulfill the stable criterion. #>  #>  597 stable posterior draws remaining for prediction! summary(predictions) #>  #> Prediction quantiles: #> , , GDPC1 #>  #>          t+1 #> 5%  -0.06623 #> 50% -0.02089 #> 95%  0.02220 #>  #> , , CPIAUCSL #>  #>           t+1 #> 5%  -0.019613 #> 50% -0.008145 #> 95%  0.003785 #>  #> , , FEDFUNDS #>  #>           t+1 #> 5%  -0.022382 #> 50% -0.003965 #> 95%  0.013119 #>"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":null,"dir":"Reference","previous_headings":"","what":"Data from the US-economy — usmacro_growth","title":"Data from the US-economy — usmacro_growth","text":"21 selected quarterly time-series 1953:Q1 2021:Q2. FRED-QD data base (McCracken Ng, 2021). Release date 2021-07. Data transformed interpreted growth-rates (first log-differences exception interest rates, already growth rates).","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Data from the US-economy — usmacro_growth","text":"","code":"usmacro_growth"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"format","dir":"Reference","previous_headings":"","what":"Format","title":"Data from the US-economy — usmacro_growth","text":"matrix 247 rows 21 columns.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"source","dir":"Reference","previous_headings":"","what":"Source","title":"Data from the US-economy — usmacro_growth","text":"Raw (untransformed) data available https://research.stlouisfed.org/econ/mccracken/fred-databases/, https://files.stlouisfed.org/files/htdocs/fred-md/quarterly/2021-07.csv.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/usmacro_growth.html","id":"references","dir":"Reference","previous_headings":"","what":"References","title":"Data from the US-economy — usmacro_growth","text":"McCracken, M. W. Ng, S. (2021). FRED-QD: Quarterly Database Macroeconomic Research, Review, Federal Reserve Bank St. Louis, 103(1), 1--44, doi:10.20955/r.103.1-44 .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":null,"dir":"Reference","previous_headings":"","what":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"Returns posterior draws possibly time-varying variance-covariance matrix VAR estimated via bvar(). Returns full paths sv_keep=\"\" calling bvar(). Otherwise, draws variance-covariance matrix last observation returned, .","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"ref-usage","dir":"Reference","previous_headings":"","what":"Usage","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"","code":"# S3 method for bayesianVARs_bvar vcov(object, ...)"},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"arguments","dir":"Reference","previous_headings":"","what":"Arguments","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"object object class bayesianVARs_bvar obtained via bvar(). ... Currently ignored.","code":""},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"value","dir":"Reference","previous_headings":"","what":"Value","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"array class bayesianVARs_draws dimension \\(T \\times M   \\times M \\times draws\\), \\(T\\) number observations, \\(M\\) number time-series \\(draws\\) number stored posterior draws.","code":""},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/reference/vcov.bayesianVARs_bvar.html","id":"ref-examples","dir":"Reference","previous_headings":"","what":"Examples","title":"Extract posterior draws of the (time-varying) variance-covariance matrix for\na VAR model — vcov.bayesianVARs_bvar","text":"","code":"# Access a subset of the usmacro_growth dataset data <- usmacro_growth[,c(\"GDPC1\", \"CPIAUCSL\", \"FEDFUNDS\")] # Estimate a model mod <- bvar(data, sv_keep = \"all\", quiet = TRUE)  # Extract posterior draws of the variance-covariance matrix bvar_vcov <- vcov(mod)"},{"path":[]},{"path":"https://luisgruber.github.io/bayesianVARs/news/index.html","id":"bayesianvars-010","dir":"Changelog","previous_headings":"","what":"bayesianVARs 0.1.0","title":"bayesianVARs 0.1.0","text":"CRAN release: 2024-01-13 Initial CRAN submission.","code":""}]