Merge branch 'main' into dev_docs

README.rst

@@ -22,7 +22,7 @@ Hawk
     :alt: Join the chat at https://gitter.im/bird-house/birdhouse
 
 Hawk (the bird)
-  *Hawk is a bird designed to perform causal analysis for climate data or, in general, for spatio-temporal data.*
+  *Hawk is a bird designed to perform causal analysis for climate data or, in general, for time-series.*
 
 
 Documentation
@@ -44,7 +44,7 @@ License
 -------
 
 * Free software: GNU General Public License v3
-* Documentation: https://hawk.readthedocs.io.
+* Documentation: https://clint-hawk.readthedocs.io.
 
 
 Credits
@@ -59,4 +59,4 @@ The two Python libraries TEFS_ and tigramite_ have been exploited to perform the
 .. _`Developer Guide`: https://hawk.readthedocs.io/en/latest/dev_guide.html
 .. _bump2version: https://hawk.readthedocs.io/en/latest/dev_guide.html#bump-a-new-version
 .. _tigramite: https://github.com/jakobrunge/tigramite
-.. _TEFS: https://github.com/teobucci/tefs
+.. _TEFS: https://github.com/teobucci/tefs

environment.yml

@@ -11,7 +11,7 @@ dependencies:
 - psutil
 - birdy
 #- tigramite
-#- tefs
+#- tefs==0.3.1
 - pandas
 - scikit-learn
 - numpy

hawk/analysis/main.py

@@ -2,11 +2,11 @@
 import os
 
 import pandas as pd
+from tefs.metrics import regression_analysis
 from tigramite.independence_tests.cmiknn import CMIknn
 from tigramite.independence_tests.parcorr import ParCorr
 
 from .file_management import save_to_pkl_file
-from .metrics import regression_analysis
 from .pcmci_tools import initialize_tigramite_df
 from .postprocessing import (
     run_postprocessing_pcmci,
@@ -63,7 +63,8 @@ def __init__(
         self.tefs_features_lags = []
         if self.tefs_use_contemporary_features:
             self.tefs_features_lags.append(0)
-        self.tefs_features_lags.extend(list(range(1, self.tefs_max_lag_features + 1)))
+        if self.tefs_max_lag_features > 0:
+            self.tefs_features_lags.extend(list(range(1, self.tefs_max_lag_features + 1)))
 
         self.tefs_target_lags = list(range(1, self.tefs_max_lag_target + 1))
 
@@ -84,15 +85,24 @@ def run_baseline_analysis(self):
 
         configs = []
 
-        # Autoregressive baselines
-        for i in range(1, self.tefs_max_lag_target):
+        # Only autoregressive baselines from 1 to the maximum target lag
+        for i in range(1, self.tefs_max_lag_target+1):
             configs.append((f"AR({i})", {self.target_column_name: list(range(1, i + 1))}))
 
-        # With all features
+        # All features without AR
         configs.append(
             (
                 "All features",
-                {feature: self.tefs_features_lags for feature in features_names},
+                {feature: self.tefs_features_lags for feature in features_names if feature != self.target_column_name},
+            )
+        )
+
+        # All features with AR
+        configs.append(
+            (
+                "All features and AR",
+                {feature: self.tefs_features_lags if feature != self.target_column_name
+                    else list(range(1, self.tefs_max_lag_target+1)) for feature in features_names},
             )
         )
 
@@ -168,7 +178,7 @@ def run_tefs_analysis(
     def run_pcmci_analysis(
         self,
     ):
-        lag_options = self.pcmci_features_lags  # max lag
+        lag_options = self.pcmci_features_lags  # list from 0 to max_lag
 
         # Define the tests
         parcorr = ParCorr(significance="analytic")
@@ -179,6 +189,7 @@ def run_pcmci_analysis(
             transform="ranks",
             sig_samples=200,
         )
+        # cmiknn = CMIknn()
 
         # Create the dictionary of tests
         independence_tests = {

hawk/analysis/postprocessing.py

@@ -6,10 +6,10 @@
 import numpy as np
 import pandas as pd
 import seaborn as sns
+from tefs.metrics import regression_analysis
 
 # from tigramite import plotting as tp
 from .file_management import save_to_pkl_file
-from .metrics import regression_analysis
 from .pcmci_tools import get_connected_variables
 
 
@@ -257,7 +257,7 @@ def run_postprocessing_pcmci(
     target_file_plots = {}
     for image_format in image_formats:
         target_file_plot = os.path.join(
-            destination_path, "algorithm_results", "pcmci", f"feature_presence.{image_format}"
+            destination_path, "algorithm_results", "pcmci", f"feature_presence_pcmci.{image_format}"
         )
         os.makedirs(os.path.dirname(target_file_plot), exist_ok=True)
         plt.savefig(target_file_plot, bbox_inches="tight")
@@ -384,7 +384,9 @@ def run_postprocessing_tefs(
     )
     target_file_plots = {}
     for image_format in image_formats:
-        target_file_plot = os.path.join(destination_path, "algorithm_results", "te", f"feature_presence.{image_format}")
+        target_file_plot = os.path.join(
+            destination_path, "algorithm_results", "te", f"feature_presence_tefs.{image_format}"
+        )
         os.makedirs(os.path.dirname(target_file_plot), exist_ok=True)
         plt.savefig(target_file_plot, bbox_inches="tight")
         target_file_plots[image_format] = target_file_plot