lecrapaud 0.9.2__tar.gz → 0.9.4__tar.gz

{lecrapaud-0.9.2 → lecrapaud-0.9.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: lecrapaud
-Version: 0.9.2
+Version: 0.9.4
 Summary: Framework for machine and deep learning, with regression, classification and time series analysis
 License: Apache License
 Author: Pierre H. Gallet

{lecrapaud-0.9.2 → lecrapaud-0.9.4}/lecrapaud/api.py

@@ -149,13 +149,13 @@ class ExperimentEngine:
                 y_pred.name = "PRED"
 
             # evaluate if TARGET is in columns
+            new_data.columns = new_data.columns.str.upper()
             if f"TARGET_{target_number}" in new_data.columns:
                 y_true = new_data[f"TARGET_{target_number}"]
                 prediction = pd.concat([y_true, y_pred], axis=1)
                 prediction.rename(
                     columns={f"TARGET_{target_number}": "TARGET"}, inplace=True
                 )
-                print(prediction)
                 score = evaluate(
                     prediction,
                     target_type=model.target_type,

lecrapaud-0.9.4/lecrapaud/db/alembic/versions/2025_06_25_1759-72aa496ca65b_.py

@@ -0,0 +1,30 @@
+"""
+
+Revision ID: 72aa496ca65b
+Revises: 86457e2f333f
+Create Date: 2025-06-25 17:59:28.544283
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = '72aa496ca65b'
+down_revision: Union[str, None] = '86457e2f333f'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('lecrapaud_model_selections', sa.Column('best_thresholds', sa.JSON(), nullable=True))
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_column('lecrapaud_model_selections', 'best_thresholds')
+    # ### end Alembic commands ###

{lecrapaud-0.9.2 → lecrapaud-0.9.4}/lecrapaud/db/models/model_selection.py

@@ -34,6 +34,7 @@ class ModelSelection(Base):
         nullable=False,
     )
     best_model_params = Column(JSON)
+    best_thresholds = Column(JSON)
     best_model_path = Column(String(255))
     best_model_id = Column(
         BigInteger, ForeignKey("lecrapaud_models.id", ondelete="CASCADE")

{lecrapaud-0.9.2 → lecrapaud-0.9.4}/lecrapaud/experiment.py

@@ -39,7 +39,9 @@ def create_experiment(
 
     with get_db() as db:
         all_targets = Target.get_all(db=db)
-        targets = [target for target in all_targets if target.name in data.columns]
+        targets = [
+            target for target in all_targets if target.name in data.columns.str.upper()
+        ]
         experiment_name = f"{experiment_name}_{groups["number_of_groups"] if group_column else 'ng'}_{corr_threshold}_{percentile}_{max_features}_{dates['start_date'].date() if date_column else 'nd'}_{dates['end_date'].date() if date_column else 'nd'}"
 
         experiment_dir = f"{tmp_dir}/{experiment_name}"

{lecrapaud-0.9.2 → lecrapaud-0.9.4}/lecrapaud/feature_engineering.py

@@ -493,6 +493,13 @@ class PreprocessFeature:
         for name, data in zip(["train", "val", "test"], [train, val, test]):
             logger.info(f"{data.shape} {name} data")
 
+        Experiment.upsert(
+            match_fields=["id"],
+            id=self.experiment_id,
+            train_size=len(train),
+            val_size=len(val),
+            test_size=len(test),
+        )
         return (
             train.reset_index(drop=True),
             val.reset_index(drop=True),
@@ -813,7 +820,7 @@ def traditional_descriptive_analysis(df: pd.DataFrame, group_column: str | None
 
         # Create a copy of the DataFrame to avoid modifying the original
         df_check = df.copy()
-        
+
         # Convert numpy arrays to tuples for hashing
         for col in df_check.columns:
             if df_check[col].apply(lambda x: isinstance(x, np.ndarray)).any():
@@ -830,10 +837,10 @@ def traditional_descriptive_analysis(df: pd.DataFrame, group_column: str | None
             duplicated_cols = []
             cols = df_check.columns
             for i, col1 in enumerate(cols):
-                for col2 in cols[i+1:]:
+                for col2 in cols[i + 1 :]:
                     if df_check[col1].equals(df_check[col2]):
                         duplicated_cols.append(f"{col1} = {col2}")
-            
+
             results["Duplicated columns"] = (
                 ", ".join(duplicated_cols) if duplicated_cols else "None"
             )

{lecrapaud-0.9.2 → lecrapaud-0.9.4}/lecrapaud/model_selection.py

@@ -569,14 +569,6 @@ class ModelEngine:
 
         target_dir = Path(self.path)
 
-        # Load threshold
-        scores_tracking = pd.read_csv(f"{target_dir}/scores_tracking.csv")
-        self.threshold = (
-            scores_tracking["THRESHOLD"].values[0]
-            if "THRESHOLD" in scores_tracking.columns
-            else None
-        )
-
         # Search for files that contain '.best' or '.keras' in the name
         best_files = list(target_dir.glob("*.best*")) + list(
             target_dir.glob("*.keras*")
@@ -589,17 +581,11 @@ class ModelEngine:
             try:
                 # Attempt to load the file as a scikit-learn, XGBoost, or LightGBM model (Pickle format)
                 self._model = joblib.load(file_path)
-                logger.info(
-                    f"Loaded model {self._model.model_name} and threshold {self.threshold}"
-                )
             except (pickle.UnpicklingError, EOFError):
                 # If it's not a pickle file, try loading it as a Keras model
                 try:
                     # Attempt to load the file as a Keras model
                     self._model = keras.models.load_model(file_path)
-                    logger.info(
-                        f"Loaded model {self._model.model_name} and threshold {self.threshold}"
-                    )
                 except Exception as e:
                     raise FileNotFoundError(
                         f"Model could not be loaded from path: {file_path}: {e}"
@@ -612,6 +598,17 @@ class ModelEngine:
         self.model_name = self._model.model_name
         self.target_type = self._model.target_type
 
+        # Load threshold
+        self.threshold = (
+            joblib.load(f"{target_dir}/thresholds.pkl")
+            if self.target_type == "classification"
+            else None
+        )
+
+        logger.info(
+            f"Loaded model {self._model.model_name} and threshold {self.threshold}"
+        )
+
 
 def trainable(
     params,
@@ -1072,6 +1069,11 @@ class ModelSelectionEngine:
         scores_tracking = pd.read_csv(scores_tracking_path)
         best_score_overall = scores_tracking.iloc[0, :]
         best_model_name = best_score_overall["MODEL_NAME"]
+        if self.target_type == "classification":
+            best_thresholds = best_score_overall["THRESHOLDS"]
+            joblib.dump(best_thresholds, f"{self.target_dir}/thresholds.pkl")
+        else:
+            best_thresholds = None
 
         # Remove any .best or .keras files
         for file_path in glob.glob(os.path.join(self.target_dir, "*.best")) + glob.glob(
@@ -1096,6 +1098,7 @@ class ModelSelectionEngine:
             name=best_score_overall["MODEL_NAME"], type=self.target_type
         ).id
         model_selection.best_model_params = best_model_params
+        model_selection.best_thresholds = best_thresholds
         model_selection.best_model_path = best_model_path
         model_selection.save()
 

{lecrapaud-0.9.2 → lecrapaud-0.9.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "lecrapaud"
-version = "0.9.2"
+version = "0.9.4"
 description = "Framework for machine and deep learning, with regression, classification and time series analysis"
 authors = [
     {name = "Pierre H. Gallet"}