feat: add osm data

research/models/ensemble_model.py

@@ -82,7 +82,9 @@ class EnsembleModel(TraditionalModel):
         # Soft voting averages probabilities (preferred when members are calibrated);
         # hard voting uses majority class. Parallelize member predictions.
         voting_type = params.get("voting", "soft")  # 'hard' or 'soft'
-        return VotingClassifier(estimators=estimators, voting=voting_type, n_jobs=params.get("n_jobs", -1))
+        return VotingClassifier(
+            estimators=estimators, voting=voting_type, n_jobs=params.get("n_jobs", -1)
+        )
 
     def prepare_features(self, X: pd.DataFrame) -> np.ndarray:
         text_features = []

research/models/random_forest_model.py

@@ -55,7 +55,9 @@ class RandomForestModel(TraditionalModel):
                         encoder = self.label_encoders[feature_key]
                         column_clean = column.fillna("unknown").astype(str)
                         known_classes = set(encoder.classes_)
-                        default_class = "unknown" if "unknown" in known_classes else encoder.classes_[0]
+                        default_class = (
+                            "unknown" if "unknown" in known_classes else encoder.classes_[0]
+                        )
                         column_mapped = column_clean.apply(
                             lambda value: value if value in known_classes else default_class
                         )

research/models/transformer_model.py

@@ -36,9 +36,9 @@ class TransformerModel(NeuralNetworkModel):
 
         # Add positional encoding
         positions = tf.range(start=0, limit=params.get("max_len", 8), delta=1)
-        pos_embedding = Embedding(input_dim=params.get("max_len", 8), output_dim=params.get("embedding_dim", 64))(
-            positions
-        )
+        pos_embedding = Embedding(
+            input_dim=params.get("max_len", 8), output_dim=params.get("embedding_dim", 64)
+        )(positions)
         x = x + pos_embedding
 
         x = self._transformer_encoder(x, params)