wavetrainer 0.0.26__tar.gz → 0.0.28__tar.gz

{wavetrainer-0.0.26/wavetrainer.egg-info → wavetrainer-0.0.28}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wavetrainer
-Version: 0.0.26
+Version: 0.0.28
 Summary: A library for automatically finding the optimal model within feature and hyperparameter space.
 Home-page: https://github.com/8W9aG/wavetrainer
 Author: Will Sackfield

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/setup.py

@@ -23,7 +23,7 @@ def install_requires() -> typing.List[str]:
 
 setup(
     name='wavetrainer',
-    version='0.0.26',
+    version='0.0.28',
     description='A library for automatically finding the optimal model within feature and hyperparameter space.',
     long_description=long_description,
     long_description_content_type='text/markdown',

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/__init__.py

@@ -2,5 +2,5 @@
 
 from .create import create
 
-__VERSION__ = "0.0.26"
+__VERSION__ = "0.0.28"
 __all__ = ("create",)

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/calibrator/calibrator_router.py

@@ -36,8 +36,13 @@ class CalibratorRouter(Calibrator):
     def name(cls) -> str:
         return "router"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
-        pass
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
+        calibrator = self._calibrator
+        if calibrator is None:
+            return
+        calibrator.set_options(trial, df)
 
     def load(self, folder: str) -> None:
         with open(
@@ -48,11 +53,11 @@ class CalibratorRouter(Calibrator):
         calibrator.load(folder)
         self._calibrator = calibrator
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         calibrator = self._calibrator
         if calibrator is None:
             raise ValueError("calibrator is null.")
-        calibrator.save(folder)
+        calibrator.save(folder, trial)
         with open(
             os.path.join(folder, _CALIBRATOR_ROUTER_FILE), "w", encoding="utf8"
         ) as handle:

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/calibrator/mapie_calibrator.py

@@ -29,13 +29,15 @@ class MAPIECalibrator(Calibrator):
     def name(cls) -> str:
         return "mapie"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         self._mapie = joblib.load(os.path.join(folder, _CALIBRATOR_FILENAME))
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         joblib.dump(self._mapie, os.path.join(folder, _CALIBRATOR_FILENAME))
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/calibrator/vennabers_calibrator.py

@@ -27,13 +27,15 @@ class VennabersCalibrator(Calibrator):
     def name(cls) -> str:
         return "vennabers"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         self._vennabers = joblib.load(os.path.join(folder, _CALIBRATOR_FILENAME))
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         joblib.dump(self._vennabers, os.path.join(folder, _CALIBRATOR_FILENAME))
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/create.py

@@ -15,7 +15,6 @@ def create(
     dt_column: str | None = None,
     max_train_timeout: datetime.timedelta | None = None,
     cutoff_dt: datetime.datetime | None = None,
-    max_features: int | None = None,
 ) -> Trainer:
     """Create a trainer."""
     return Trainer(
@@ -26,5 +25,4 @@ def create(
         dt_column=dt_column,
         max_train_timeout=max_train_timeout,
         cutoff_dt=cutoff_dt,
-        max_features=max_features,
     )

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/model/catboost_model.py

@@ -26,6 +26,7 @@ _L2_LEAF_REG_KEY = "l2_leaf_reg"
 _BOOSTING_TYPE_KEY = "boosting_type"
 _MODEL_TYPE_KEY = "model_type"
 _EARLY_STOPPING_ROUNDS = "early_stopping_rounds"
+_BEST_ITERATION_KEY = "best_iteration"
 
 
 class CatboostModel(Model):
@@ -41,11 +42,16 @@ class CatboostModel(Model):
     _boosting_type: None | str
     _model_type: None | ModelType
     _early_stopping_rounds: None | int
+    _best_iteration: None | int
 
     @classmethod
     def name(cls) -> str:
         return "catboost"
 
+    @classmethod
+    def supports_x(cls, df: pd.DataFrame) -> bool:
+        return True
+
     def __init__(self) -> None:
         super().__init__()
         self._catboost = None
@@ -56,6 +62,7 @@ class CatboostModel(Model):
         self._boosting_type = None
         self._model_type = None
         self._early_stopping_rounds = None
+        self._best_iteration = None
 
     @property
     def estimator(self) -> Any:
@@ -83,7 +90,9 @@ class CatboostModel(Model):
             "sample_weight": w,
         }
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         self._iterations = trial.suggest_int(_ITERATIONS_KEY, 100, 10000)
         self._learning_rate = trial.suggest_float(_LEARNING_RATE_KEY, 0.001, 0.3)
         self._depth = trial.suggest_int(_DEPTH_KEY, 1, 10)
@@ -92,6 +101,7 @@ class CatboostModel(Model):
             _BOOSTING_TYPE_KEY, ["Ordered", "Plain"]
         )
         self._early_stopping_rounds = trial.suggest_int(_EARLY_STOPPING_ROUNDS, 10, 500)
+        self._best_iteration = trial.user_attrs.get(_BEST_ITERATION_KEY)
 
     def load(self, folder: str) -> None:
         with open(
@@ -105,10 +115,11 @@ class CatboostModel(Model):
             self._boosting_type = params[_BOOSTING_TYPE_KEY]
             self._model_type = ModelType(params[_MODEL_TYPE_KEY])
             self._early_stopping_rounds = params[_EARLY_STOPPING_ROUNDS]
+            self._best_iteration = params.get(_BEST_ITERATION_KEY)
         catboost = self._provide_catboost()
         catboost.load_model(os.path.join(folder, _MODEL_FILENAME))
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         with open(
             os.path.join(folder, _MODEL_PARAMS_FILENAME), "w", encoding="utf8"
         ) as handle:
@@ -121,11 +132,13 @@ class CatboostModel(Model):
                     _BOOSTING_TYPE_KEY: self._boosting_type,
                     _MODEL_TYPE_KEY: str(self._model_type),
                     _EARLY_STOPPING_ROUNDS: self._early_stopping_rounds,
+                    _BEST_ITERATION_KEY: self._best_iteration,
                 },
                 handle,
             )
         catboost = self._provide_catboost()
         catboost.save_model(os.path.join(folder, _MODEL_FILENAME))
+        trial.user_attrs[_BEST_ITERATION_KEY] = self._best_iteration
 
     def fit(
         self,
@@ -137,8 +150,6 @@ class CatboostModel(Model):
     ) -> Self:
         if y is None:
             raise ValueError("y is null.")
-        if eval_x is None:
-            raise ValueError("eval_x is null.")
         self._model_type = determine_model_type(y)
         catboost = self._provide_catboost()
 
@@ -148,10 +159,14 @@ class CatboostModel(Model):
             weight=w,
             cat_features=df.select_dtypes(include="category").columns.tolist(),
         )
-        eval_pool = Pool(
-            eval_x,
-            label=eval_y,
-            cat_features=eval_x.select_dtypes(include="category").columns.tolist(),
+        eval_pool = (
+            Pool(
+                eval_x,
+                label=eval_y,
+                cat_features=eval_x.select_dtypes(include="category").columns.tolist(),
+            )
+            if eval_x is not None
+            else None
         )
         catboost.fit(
             train_pool,
@@ -162,6 +177,7 @@ class CatboostModel(Model):
         )
         importances = catboost.get_feature_importance(prettified=True)
         logging.info("Importances:\n%s", importances)
+        self._best_iteration = catboost.get_best_iteration()
         return self
 
     def transform(self, df: pd.DataFrame) -> pd.DataFrame:
@@ -186,10 +202,14 @@ class CatboostModel(Model):
     def _provide_catboost(self) -> CatBoost:
         catboost = self._catboost
         if catboost is None:
+            best_iteration = self._best_iteration
+            iterations = (
+                best_iteration if best_iteration is not None else self._iterations
+            )
             match self._model_type:
                 case ModelType.BINARY:
                     catboost = CatBoostClassifierWrapper(
-                        iterations=self._iterations,
+                        iterations=iterations,
                         learning_rate=self._learning_rate,
                         depth=self._depth,
                         l2_leaf_reg=self._l2_leaf_reg,
@@ -201,7 +221,7 @@ class CatboostModel(Model):
                     )
                 case ModelType.REGRESSION:
                     catboost = CatBoostRegressorWrapper(
-                        iterations=self._iterations,
+                        iterations=iterations,
                         learning_rate=self._learning_rate,
                         depth=self._depth,
                         l2_leaf_reg=self._l2_leaf_reg,
@@ -213,7 +233,7 @@ class CatboostModel(Model):
                     )
                 case ModelType.BINNED_BINARY:
                     catboost = CatBoostClassifierWrapper(
-                        iterations=self._iterations,
+                        iterations=iterations,
                         learning_rate=self._learning_rate,
                         depth=self._depth,
                         l2_leaf_reg=self._l2_leaf_reg,
@@ -225,7 +245,7 @@ class CatboostModel(Model):
                     )
                 case ModelType.MULTI_CLASSIFICATION:
                     catboost = CatBoostClassifierWrapper(
-                        iterations=self._iterations,
+                        iterations=iterations,
                         learning_rate=self._learning_rate,
                         depth=self._depth,
                         l2_leaf_reg=self._l2_leaf_reg,

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/model/model.py

@@ -20,6 +20,11 @@ class Model(Params, Fit):
         """The name of the model."""
         raise NotImplementedError("name not implemented in parent class.")
 
+    @classmethod
+    def supports_x(cls, df: pd.DataFrame) -> bool:
+        """Whether the model supports the X values."""
+        raise NotImplementedError("supports_x not implemented in parent class.")
+
     @property
     def estimator(self) -> Any:
         """The estimator backing the model."""

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/model/model_router.py

@@ -34,6 +34,10 @@ class ModelRouter(Model):
     def name(cls) -> str:
         return "router"
 
+    @classmethod
+    def supports_x(cls, df: pd.DataFrame) -> bool:
+        return True
+
     @property
     def estimator(self) -> Any:
         model = self._model
@@ -61,9 +65,15 @@ class ModelRouter(Model):
             raise ValueError("model is null")
         return model.pre_fit(df, y=y, eval_x=eval_x, eval_y=eval_y, w=w)
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
-        model = _MODELS[trial.suggest_categorical("model", list(_MODELS.keys()))]()
-        model.set_options(trial)
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
+        model = _MODELS[
+            trial.suggest_categorical(
+                "model", [k for k, v in _MODELS.items() if v.supports_x(df)]
+            )
+        ]()
+        model.set_options(trial, df)
         self._model = model
 
     def load(self, folder: str) -> None:
@@ -73,11 +83,11 @@ class ModelRouter(Model):
         model.load(folder)
         self._model = model
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         model = self._model
         if model is None:
             raise ValueError("model is null")
-        model.save(folder)
+        model.save(folder, trial)
         with open(
             os.path.join(folder, _MODEL_ROUTER_FILE), "w", encoding="utf8"
         ) as handle:

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/model/tabpfn_model.py

@@ -31,6 +31,10 @@ class TabPFNModel(Model):
     def name(cls) -> str:
         return "tabpfn"
 
+    @classmethod
+    def supports_x(cls, df: pd.DataFrame) -> bool:
+        return len(df.columns.values) < 500
+
     def __init__(self) -> None:
         super().__init__()
         self._tabpfn = None
@@ -57,7 +61,9 @@ class TabPFNModel(Model):
         self._model_type = determine_model_type(y)
         return {}
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
@@ -69,7 +75,7 @@ class TabPFNModel(Model):
             params = json.load(handle)
             self._model_type = ModelType(params[_MODEL_TYPE_KEY])
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         with open(os.path.join(folder, _MODEL_FILENAME), "wb") as f:
             pickle.dump(self._tabpfn, f)
         with open(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/params.py

@@ -1,12 +1,15 @@
 """A class for loading/saving parameters."""
 
 import optuna
+import pandas as pd
 
 
 class Params:
     """The params prototype class."""
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         """Set the options used in the object."""
         raise NotImplementedError("set_options not implemented in parent class.")
 
@@ -14,6 +17,6 @@ class Params:
         """Loads the objects from a folder."""
         raise NotImplementedError("load not implemented in parent class.")
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         """Saves the objects into a folder."""
         raise NotImplementedError("save not implemented in parent class.")

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/reducer/base_selector_reducer.py

@@ -32,14 +32,16 @@ class BaseSelectorReducer(Reducer):
         """Whether the class should raise its exception if it encounters it."""
         return True
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         file_path = os.path.join(folder, self._file_name)
         self._base_selector = joblib.load(file_path)
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         file_path = os.path.join(folder, self._file_name)
         joblib.dump(self._base_selector, file_path)
 

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/reducer/combined_reducer.py

@@ -12,8 +12,8 @@ from .constant_reducer import ConstantReducer
 from .correlation_reducer import CorrelationReducer
 from .duplicate_reducer import DuplicateReducer
 from .nonnumeric_reducer import NonNumericReducer
-from .pca_reducer import PCAReducer
 from .reducer import Reducer
+from .smart_correlation_reducer import SmartCorrelationReducer
 from .unseen_reducer import UnseenReducer
 
 _COMBINED_REDUCER_FILE = "combined_reducer.json"
@@ -25,25 +25,26 @@ class CombinedReducer(Reducer):
 
     # pylint: disable=too-many-positional-arguments,too-many-arguments
 
-    def __init__(self, max_features: int | None):
+    def __init__(self):
         super().__init__()
-        self._max_features = max_features
         self._reducers = [
             UnseenReducer(),
             NonNumericReducer(),
             ConstantReducer(),
             DuplicateReducer(),
             CorrelationReducer(),
-            PCAReducer(max_features),
+            SmartCorrelationReducer(),
         ]
 
     @classmethod
     def name(cls) -> str:
         return "combined"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         for reducer in self._reducers:
-            reducer.set_options(trial)
+            reducer.set_options(trial, df)
 
     def load(self, folder: str) -> None:
         self._reducers = []
@@ -62,12 +63,12 @@ class CombinedReducer(Reducer):
                     self._reducers.append(NonNumericReducer())
                 elif reducer_name == UnseenReducer.name():
                     self._reducers.append(UnseenReducer())
-                elif reducer_name == PCAReducer.name():
-                    self._reducers.append(PCAReducer(self._max_features))
+                elif reducer_name == SmartCorrelationReducer.name():
+                    self._reducers.append(SmartCorrelationReducer())
         for reducer in self._reducers:
             reducer.load(folder)
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         with open(
             os.path.join(folder, _COMBINED_REDUCER_FILE), "w", encoding="utf8"
         ) as handle:
@@ -78,7 +79,7 @@ class CombinedReducer(Reducer):
                 handle,
             )
         for reducer in self._reducers:
-            reducer.save(folder)
+            reducer.save(folder, trial)
 
     def fit(
         self,

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/reducer/correlation_reducer.py

@@ -1,18 +1,22 @@
 """A reducer that removes correlation features."""
 
+import optuna
+import pandas as pd
 from feature_engine.selection import DropCorrelatedFeatures
 
 from .base_selector_reducer import BaseSelectorReducer
 
 _CORRELATION_REDUCER_FILENAME = "correlation_reducer.joblib"
+_CORRELATION_REDUCER_THRESHOLD = "correlation_reducer_threshold"
 
 
 class CorrelationReducer(BaseSelectorReducer):
     """A class that removes correlated values from a dataset."""
 
     def __init__(self) -> None:
+        self._correlation_selector = DropCorrelatedFeatures(missing_values="ignore")
         super().__init__(
-            DropCorrelatedFeatures(missing_values="ignore"),
+            self._correlation_selector,
             _CORRELATION_REDUCER_FILENAME,
         )
 
@@ -23,3 +27,10 @@ class CorrelationReducer(BaseSelectorReducer):
     @classmethod
     def should_raise(cls) -> bool:
         return False
+
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
+        self._correlation_selector.threshold = trial.suggest_float(
+            _CORRELATION_REDUCER_THRESHOLD, 0.1, 0.9
+        )

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/reducer/nonnumeric_reducer.py

@@ -17,13 +17,15 @@ class NonNumericReducer(Reducer):
     def name(cls) -> str:
         return "nonnumeric"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         pass
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         pass
 
     def fit(

wavetrainer-0.0.28/wavetrainer/reducer/smart_correlation_reducer.py

@@ -0,0 +1,32 @@
+"""A reducer that removes correlation features via further heuristics."""
+
+import optuna
+import pandas as pd
+from feature_engine.selection import SmartCorrelatedSelection
+
+from .base_selector_reducer import BaseSelectorReducer
+
+_SMART_CORRELATION_REDUCER_FILENAME = "smart_correlation_reducer.joblib"
+_SMART_CORRELATION_REDUCER_THRESHOLD = "smart_correlation_reducer_threshold"
+
+
+class SmartCorrelationReducer(BaseSelectorReducer):
+    """A class that removes smart correlated values from a dataset."""
+
+    def __init__(self) -> None:
+        self._correlation_selector = SmartCorrelatedSelection(missing_values="ignore")
+        super().__init__(
+            self._correlation_selector,
+            _SMART_CORRELATION_REDUCER_FILENAME,
+        )
+
+    @classmethod
+    def name(cls) -> str:
+        return "smart_correlation"
+
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
+        self._correlation_selector.threshold = trial.suggest_float(
+            _SMART_CORRELATION_REDUCER_THRESHOLD, 0.1, 0.9
+        )

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/reducer/unseen_reducer.py

@@ -25,7 +25,9 @@ class UnseenReducer(Reducer):
     def name(cls) -> str:
         return "unseen"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
@@ -34,7 +36,7 @@ class UnseenReducer(Reducer):
         ) as handle:
             self._seen_features = json.load(handle)
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         with open(
             os.path.join(folder, _UNSEEN_REDUCER_FILE), "w", encoding="utf8"
         ) as handle:

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/selector/selector.py

@@ -31,14 +31,16 @@ class Selector(Params, Fit):
         self._steps = 0
         self._selector = None
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         self._feature_ratio = trial.suggest_float("feature_ratio", 0.0, 1.0)
         self._steps = trial.suggest_int("steps", 1, 10)
 
     def load(self, folder: str) -> None:
         self._selector = joblib.load(os.path.join(folder, _SELECTOR_FILE))
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         joblib.dump(self._selector, os.path.join(folder, _SELECTOR_FILE))
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/trainer.py

@@ -36,7 +36,6 @@ _TEST_SIZE_KEY = "test_size"
 _VALIDATION_SIZE_KEY = "validation_size"
 _IDX_USR_ATTR_KEY = "idx"
 _DT_COLUMN_KEY = "dt_column"
-_MAX_FEATURES_KEY = "max_features"
 
 
 class Trainer(Fit):
@@ -54,7 +53,6 @@ class Trainer(Fit):
         dt_column: str | None = None,
         max_train_timeout: datetime.timedelta | None = None,
         cutoff_dt: datetime.datetime | None = None,
-        max_features: int | None = None,
     ):
         tqdm.tqdm.pandas()
 
@@ -105,7 +103,6 @@ class Trainer(Fit):
                         )
                 if dt_column is None:
                     dt_column = params[_DT_COLUMN_KEY]
-                max_features = params.get(_MAX_FEATURES_KEY)
         else:
             with open(params_file, "w", encoding="utf8") as handle:
                 validation_size_value = None
@@ -136,7 +133,6 @@ class Trainer(Fit):
                         _TEST_SIZE_KEY: test_size_value,
                         _VALIDATION_SIZE_KEY: validation_size_value,
                         _DT_COLUMN_KEY: dt_column,
-                        _MAX_FEATURES_KEY: max_features,
                     },
                     handle,
                 )
@@ -147,7 +143,6 @@ class Trainer(Fit):
         self._dt_column = dt_column
         self._max_train_timeout = max_train_timeout
         self._cutoff_dt = cutoff_dt
-        self._max_features = max_features
 
     def _provide_study(self, column: str) -> optuna.Study:
         storage_name = f"sqlite:///{self._folder}/{column}/{_STUDYDB_FILENAME}"
@@ -213,7 +208,7 @@ class Trainer(Fit):
                 try:
                     # Window the data
                     windower = Windower(self._dt_column)
-                    windower.set_options(trial)
+                    windower.set_options(trial, x)
                     x_train = windower.fit_transform(x_train)
                     y_train = y_train[-len(x_train) :]
                     if len(y_train.unique()) <= 1:
@@ -221,25 +216,25 @@ class Trainer(Fit):
                         return -1.0
 
                     # Perform common reductions
-                    reducer = CombinedReducer(self._max_features)
-                    reducer.set_options(trial)
+                    reducer = CombinedReducer()
+                    reducer.set_options(trial, x)
                     x_train = reducer.fit_transform(x_train)
                     x_test = reducer.transform(x_test)
 
                     # Calculate the row weights
                     weights = CombinedWeights()
-                    weights.set_options(trial)
+                    weights.set_options(trial, x)
                     w = weights.fit(x_train, y=y_train).transform(y_train.to_frame())[
                         WEIGHTS_COLUMN
                     ]
 
                     # Create model
                     model = ModelRouter()
-                    model.set_options(trial)
+                    model.set_options(trial, x)
 
                     # Train
                     selector = Selector(model)
-                    selector.set_options(trial)
+                    selector.set_options(trial, x)
                     selector.fit(x_train, y=y_train, w=w, eval_x=x_test, eval_y=y_test)
                     x_train = selector.transform(x_train)
                     x_test = selector.transform(x_test)
@@ -249,7 +244,7 @@ class Trainer(Fit):
 
                     # Calibrate
                     calibrator = CalibratorRouter(model)
-                    calibrator.set_options(trial)
+                    calibrator.set_options(trial, x)
                     calibrator.fit(x_pred, y=y_train)
 
                     if save:
@@ -258,12 +253,12 @@ class Trainer(Fit):
                         )
                         if not os.path.exists(folder):
                             os.mkdir(folder)
-                        windower.save(folder)
-                        reducer.save(folder)
-                        weights.save(folder)
-                        model.save(folder)
-                        selector.save(folder)
-                        calibrator.save(folder)
+                        windower.save(folder, trial)
+                        reducer.save(folder, trial)
+                        weights.save(folder, trial)
+                        model.save(folder, trial)
+                        selector.save(folder, trial)
+                        calibrator.save(folder, trial)
 
                     y_pred = model.transform(x_test)
                     y_pred = calibrator.transform(y_pred)
@@ -380,6 +375,7 @@ class Trainer(Fit):
 
     def transform(self, df: pd.DataFrame) -> pd.DataFrame:
         """Predict the expected values of the data."""
+        tqdm.tqdm.pandas(desc="Inferring...")
         input_df = df.copy()
         df = df.reindex(sorted(df.columns), axis=1)
         feature_columns = df.columns.values
@@ -430,7 +426,7 @@ class Trainer(Fit):
                 date_str = dates[-1].isoformat()
                 folder = os.path.join(column_path, date_str)
 
-                reducer = CombinedReducer(self._max_features)
+                reducer = CombinedReducer()
                 reducer.load(folder)
 
                 model = ModelRouter()

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/weights/class_weights.py

@@ -27,13 +27,15 @@ class ClassWeights(Weights):
         """The name of the weight class."""
         return "class"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         pass
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         pass
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/weights/combined_weights.py

@@ -23,17 +23,19 @@ class CombinedWeights(Weights):
     def name(cls) -> str:
         return "combined"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         for weights in self._weights:
-            weights.set_options(trial)
+            weights.set_options(trial, df)
 
     def load(self, folder: str) -> None:
         for weights in self._weights:
             weights.load(folder)
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         for weights in self._weights:
-            weights.save(folder)
+            weights.save(folder, trial)
 
     def fit(
         self,

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/weights/exponential_weights.py

@@ -19,13 +19,15 @@ class ExponentialWeights(Weights):
         """The name of the weight class."""
         return "exponential"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         pass
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         pass
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/weights/linear_weights.py

@@ -19,13 +19,15 @@ class LinearWeights(Weights):
         """The name of the weight class."""
         return "linear"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         pass
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         pass
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/weights/noop_weights.py

@@ -19,13 +19,15 @@ class NoopWeights(Weights):
         """The name of the weight class."""
         return "noop"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         pass
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         pass
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/weights/sigmoid_weights.py

@@ -20,13 +20,15 @@ class SigmoidWeights(Weights):
         """The name of the weight class."""
         return "sigmoid"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         pass
 
     def load(self, folder: str) -> None:
         pass
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         pass
 
     def fit(

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/weights/weights_router.py

@@ -38,7 +38,9 @@ class WeightsRouter(Weights):
     def name(cls) -> str:
         return "router"
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         self._weights = _WEIGHTS[
             trial.suggest_categorical("weights", list(_WEIGHTS.keys()))
         ]()
@@ -54,10 +56,11 @@ class WeightsRouter(Weights):
             weights = _WEIGHTS[params[_WEIGHTS_KEY]]()
         self._weights = weights
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         weights = self._weights
         if weights is None:
             raise ValueError("weights is null")
+        weights.save(folder, trial)
         with open(
             os.path.join(folder, _WEIGHTS_ROUTER_FILE), "w", encoding="utf8"
         ) as handle:

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer/windower/windower.py

@@ -28,7 +28,9 @@ class Windower(Params, Fit):
         self._lookback_ratio = None
         self._dt_column = dt_column
 
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
+    def set_options(
+        self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
+    ) -> None:
         self._lookback_ratio = trial.suggest_float("lookback", 0.1, 1.0)
 
     def load(self, folder: str) -> None:
@@ -36,7 +38,7 @@ class Windower(Params, Fit):
             params = json.load(handle)
             self._lookback = params[_LOOKBACK_KEY]
 
-    def save(self, folder: str) -> None:
+    def save(self, folder: str, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
         with open(os.path.join(folder, _WINDOWER_FILE), "w", encoding="utf8") as handle:
             json.dump(
                 {

{wavetrainer-0.0.26 → wavetrainer-0.0.28/wavetrainer.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wavetrainer
-Version: 0.0.26
+Version: 0.0.28
 Summary: A library for automatically finding the optimal model within feature and hyperparameter space.
 Home-page: https://github.com/8W9aG/wavetrainer
 Author: Will Sackfield

{wavetrainer-0.0.26 → wavetrainer-0.0.28}/wavetrainer.egg-info/SOURCES.txt

@@ -40,8 +40,8 @@ wavetrainer/reducer/constant_reducer.py
 wavetrainer/reducer/correlation_reducer.py
 wavetrainer/reducer/duplicate_reducer.py
 wavetrainer/reducer/nonnumeric_reducer.py
-wavetrainer/reducer/pca_reducer.py
 wavetrainer/reducer/reducer.py
+wavetrainer/reducer/smart_correlation_reducer.py
 wavetrainer/reducer/unseen_reducer.py
 wavetrainer/selector/__init__.py
 wavetrainer/selector/selector.py

wavetrainer-0.0.26/wavetrainer/reducer/pca_reducer.py

@@ -1,77 +0,0 @@
-"""A reducer that removes low variance columns."""
-
-import os
-from typing import Self
-
-import joblib  # type: ignore
-import optuna
-import pandas as pd
-from sklearn.decomposition import PCA  # type: ignore
-from sklearn.preprocessing import StandardScaler  # type: ignore
-
-from .reducer import Reducer
-
-_PCA_FILE = "pca.joblib"
-_PCA_SCALER_FILE = "pca_scaler.joblib"
-
-
-class PCAReducer(Reducer):
-    """A class that removes low variance columns from a dataframe."""
-
-    # pylint: disable=too-many-positional-arguments,too-many-arguments
-
-    def __init__(self, max_features: int | None):
-        super().__init__()
-        self._max_features = max_features
-        if max_features is not None:
-            self._scaler = StandardScaler()
-            self._pca = PCA(n_components=max_features)
-        else:
-            self._scaler = None
-            self._pca = None
-
-    @classmethod
-    def name(cls) -> str:
-        return "pca"
-
-    def set_options(self, trial: optuna.Trial | optuna.trial.FrozenTrial) -> None:
-        pass
-
-    def load(self, folder: str) -> None:
-        pca_scaler_file = os.path.join(folder, _PCA_SCALER_FILE)
-        pca_file = os.path.join(folder, _PCA_FILE)
-        if os.path.exists(pca_scaler_file):
-            self._scaler = joblib.load(pca_scaler_file)
-        if os.path.exists(pca_file):
-            self._pca = joblib.load(pca_file)
-
-    def save(self, folder: str) -> None:
-        if self._scaler is not None:
-            joblib.dump(self._scaler, os.path.join(folder, _PCA_SCALER_FILE))
-        if self._pca is not None:
-            joblib.dump(self._pca, os.path.join(folder, _PCA_FILE))
-
-    def fit(
-        self,
-        df: pd.DataFrame,
-        y: pd.Series | pd.DataFrame | None = None,
-        w: pd.Series | None = None,
-        eval_x: pd.DataFrame | None = None,
-        eval_y: pd.Series | pd.DataFrame | None = None,
-    ) -> Self:
-        pca = self._pca
-        scaler = self._scaler
-        if pca is None or scaler is None:
-            return self
-        if len(df.columns.values) < pca.n_components:  # type: ignore
-            return self
-        x_scaled = scaler.fit_transform(df)
-        pca.fit(x_scaled)
-        return self
-
-    def transform(self, df: pd.DataFrame) -> pd.DataFrame:
-        if self._pca is None:
-            return df
-        if len(df.columns.values) < self._pca.n_components:  # type: ignore
-            return df
-        return self._pca.transform(df)