geocif 0.1.95__tar.gz → 0.1.97__tar.gz

{geocif-0.1.95/geocif.egg-info → geocif-0.1.97}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: geocif
-Version: 0.1.95
+Version: 0.1.97
 Summary: Models to visualize and forecast crop conditions and yields
 Home-page: https://ritviksahajpal.github.io/yield_forecasting/
 Author: Ritvik Sahajpal

{geocif-0.1.95 → geocif-0.1.97}/geocif/geocif.py

@@ -18,7 +18,6 @@ from geocif import logger as log
 from geocif import utils
 from .cei import definitions as di
 from .ml import correlations
-from .ml import spatial_autocorrelation as sa
 from .ml import feature_engineering as fe
 from .ml import feature_selection as fs
 from .ml import output
@@ -1171,6 +1170,7 @@ class Geocif:
         dict_kwargs["correlation_threshold"] = self.correlation_threshold
 
         if self.spatial_autocorrelation:
+            from .ml import spatial_autocorrelation as sa
             sa.compute_spatial_autocorrelation(self.df_inputs, **dict_kwargs)
 
         dict_selected_features = {}

{geocif-0.1.95 → geocif-0.1.97}/geocif/ml/feature_selection.py

@@ -133,6 +133,27 @@ def select_features(X, y, method="RFE", min_features_to_select=3, threshold_nan=
         selected_features = (
             shap_importance_df["feature"].head(optimal_N).values.tolist()
         )
+    elif method == "stabl":
+        from stabl.stabl import Stabl
+        from sklearn.linear_model import LogisticRegression
+
+        lasso = LogisticRegression(
+            penalty="l1", class_weight="balanced", max_iter=int(1e6), solver="liblinear", random_state=42
+        )
+        stabl = Stabl(
+            base_estimator=lasso,
+            n_bootstraps=100,
+            artificial_type="knockoff",
+            artificial_proportion=.5,
+            replace=False,
+            fdr_threshold_range=np.arange(0.1, 1, 0.01),
+            sample_fraction=0.5,
+            random_state=42,
+            lambda_grid={"C": np.linspace(0.004, 0.4, 30)},
+            verbose=1
+        )
+        stabl.fit(X, y)
+        selected_features = stabl.get_feature_names_out()
     elif method == "feature_engine":
         from feature_engine.selection import SmartCorrelatedSelection
 
@@ -304,7 +325,8 @@ def select_features(X, y, method="RFE", min_features_to_select=3, threshold_nan=
     else:
         raise ValueError("Method not recognized. Use BorutaPy, Genetic, or RFE")
     # tentative_features = X.columns[selector.support_weak_].tolist()
-
+    print(selected_features)
+    breakpoint()
     non_eo = are_all_features_non_eo(selected_features)
     if non_eo or method == "SelectKBest":
         from sklearn.feature_selection import SelectKBest, f_regression

{geocif-0.1.95 → geocif-0.1.97/geocif.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: geocif
-Version: 0.1.95
+Version: 0.1.97
 Summary: Models to visualize and forecast crop conditions and yields
 Home-page: https://ritviksahajpal.github.io/yield_forecasting/
 Author: Ritvik Sahajpal

{geocif-0.1.95 → geocif-0.1.97}/setup.py

@@ -50,6 +50,6 @@ setup(
     test_suite="tests",
     tests_require=test_requirements,
     url="https://ritviksahajpal.github.io/yield_forecasting/",
-    version="0.1.95",
+    version="0.1.97",
     zip_safe=False,
 )