wavetrainer 0.0.40__tar.gz → 0.0.41__tar.gz

{wavetrainer-0.0.40/wavetrainer.egg-info → wavetrainer-0.0.41}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wavetrainer
-Version: 0.0.40
+Version: 0.0.41
 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-0.0.40 → wavetrainer-0.0.41}/setup.py

@@ -23,7 +23,7 @@ def install_requires() -> typing.List[str]:
 
 setup(
     name='wavetrainer',
-    version='0.0.40',
+    version='0.0.41',
     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-0.0.40 → wavetrainer-0.0.41}/wavetrainer/__init__.py

@@ -2,5 +2,5 @@
 
 from .create import create
 
-__VERSION__ = "0.0.40"
+__VERSION__ = "0.0.41"
 __all__ = ("create",)

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/calibrator/calibrator.py

@@ -1,5 +1,7 @@
 """The prototype calibrator class."""
 
+import pandas as pd
+
 from ..fit import Fit
 from ..model.model import Model
 from ..params import Params
@@ -15,3 +17,7 @@ class Calibrator(Params, Fit):
     def name(cls) -> str:
         """The name of the calibrator."""
         raise NotImplementedError("name not implemented in parent class.")
+
+    def predictions_as_x(self, y: pd.Series | pd.DataFrame | None = None) -> bool:
+        """Whether the calibrator wants predictions as X rather than features."""
+        raise NotImplementedError("predictions_as_x not implemented in parent class.")

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/calibrator/calibrator_router.py

@@ -36,6 +36,13 @@ class CalibratorRouter(Calibrator):
     def name(cls) -> str:
         return "router"
 
+    def predictions_as_x(self, y: pd.Series | pd.DataFrame | None = None) -> bool:
+        if y is None:
+            raise ValueError("y is null")
+        if determine_model_type(y) == ModelType.REGRESSION:
+            return False
+        return True
+
     def set_options(
         self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
     ) -> None:

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/calibrator/mapie_calibrator.py

@@ -1,13 +1,11 @@
 """A calibrator that implements MAPIE."""
 
-import logging
 import os
 from typing import Self
 
 import joblib  # type: ignore
 import optuna
 import pandas as pd
-import sklearn  # type: ignore
 from mapie.regression import MapieRegressor  # type: ignore
 
 from ..model.model import PROBABILITY_COLUMN_PREFIX, Model
@@ -23,12 +21,15 @@ class MAPIECalibrator(Calibrator):
 
     def __init__(self, model: Model):
         super().__init__(model)
-        self._mapie = MapieRegressor(model.estimator, method="plus")
+        self._mapie = MapieRegressor(model.create_estimator(), method="plus")
 
     @classmethod
     def name(cls) -> str:
         return "mapie"
 
+    def predictions_as_x(self, y: pd.Series | pd.DataFrame | None = None) -> bool:
+        return False
+
     def set_options(
         self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
     ) -> None:
@@ -59,20 +60,18 @@ class MAPIECalibrator(Calibrator):
         return self
 
     def transform(self, df: pd.DataFrame) -> pd.DataFrame:
-        try:
-            alpha = []
-            for potential_alpha in [0.05, 0.32]:
-                if len(df) > int(1.0 / potential_alpha) + 1:
-                    alpha.append(potential_alpha)
-            if alpha:
-                _, y_pis = self._mapie.predict(df, alpha=alpha)
-                for i in range(y_pis.shape[1]):
-                    if i >= len(alpha):
-                        continue
-                    for ii in range(y_pis.shape[2]):
-                        alpha_val = alpha[i]
-                        values = y_pis[:, i, ii].flatten().tolist()
-                        df[f"{PROBABILITY_COLUMN_PREFIX}{alpha_val}_{ii == 1}"] = values
-        except sklearn.exceptions.NotFittedError as exc:  # type: ignore
-            logging.warning(str(exc))
-        return df
+        alpha = []
+        for potential_alpha in [0.05, 0.32]:
+            if len(df) > int(1.0 / potential_alpha) + 1:
+                alpha.append(potential_alpha)
+        ret_df = pd.DataFrame(index=df.index)
+        if alpha:
+            _, y_pis = self._mapie.predict(df, alpha=alpha)
+            for i in range(y_pis.shape[1]):
+                if i >= len(alpha):
+                    continue
+                for ii in range(y_pis.shape[2]):
+                    alpha_val = alpha[i]
+                    values = y_pis[:, i, ii].flatten().tolist()
+                    ret_df[f"{PROBABILITY_COLUMN_PREFIX}{alpha_val}_{ii == 1}"] = values
+        return ret_df

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/calibrator/vennabers_calibrator.py

@@ -28,6 +28,9 @@ class VennabersCalibrator(Calibrator):
     def name(cls) -> str:
         return "vennabers"
 
+    def predictions_as_x(self, y: pd.Series | pd.DataFrame | None = None) -> bool:
+        return True
+
     def set_options(
         self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
     ) -> None:

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/model/catboost/catboost_kwargs.py

@@ -38,12 +38,13 @@ def handle_fit_kwargs(*args, **kwargs) -> tuple[tuple[Any, ...], dict[str, Any]]
         args_list[0] = df[included_columns]
         args = tuple(args_list)
 
-        eval_x = eval_x[included_columns]
-        kwargs[EVAL_SET_ARG_KEY] = Pool(
-            eval_x,
-            label=eval_y,
-            cat_features=cat_features,
-        )
+        if eval_x is not None:
+            eval_x = eval_x[included_columns]
+            kwargs[EVAL_SET_ARG_KEY] = Pool(
+                eval_x,
+                label=eval_y,
+                cat_features=cat_features,
+            )
         kwargs[CAT_FEATURES_ARG_KEY] = cat_features
 
         del kwargs[ORIGINAL_X_ARG_KEY]

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/model/catboost/catboost_model.py

@@ -1,9 +1,10 @@
 """A model that wraps catboost."""
 
+# pylint: disable=line-too-long
 import json
 import logging
 import os
-from typing import Any, Self
+from typing import Self
 
 import optuna
 import pandas as pd
@@ -13,8 +14,6 @@ from catboost import CatBoost, Pool  # type: ignore
 from ...model_type import ModelType, determine_model_type
 from ..model import PREDICTION_COLUMN, PROBABILITY_COLUMN_PREFIX, Model
 from .catboost_classifier_wrap import CatBoostClassifierWrapper
-from .catboost_kwargs import (CAT_FEATURES_ARG_KEY, EVAL_SET_ARG_KEY,
-                              ORIGINAL_X_ARG_KEY)
 from .catboost_regressor_wrap import CatBoostRegressorWrapper
 
 _MODEL_FILENAME = "model.cbm"
@@ -64,10 +63,6 @@ class CatboostModel(Model):
         self._early_stopping_rounds = None
         self._best_iteration = None
 
-    @property
-    def estimator(self) -> Any:
-        return self._provide_catboost()
-
     @property
     def supports_importances(self) -> bool:
         return True
@@ -82,23 +77,11 @@ class CatboostModel(Model):
         importances = importances["Importances"].to_list()  # type: ignore
         return {feature_ids[x]: importances[x] for x in range(len(feature_ids))}
 
-    def pre_fit(
-        self,
-        df: pd.DataFrame,
-        y: pd.Series | pd.DataFrame | None,
-        eval_x: pd.DataFrame | None = None,
-        eval_y: pd.Series | pd.DataFrame | None = None,
-        w: pd.Series | None = None,
-    ):
-        if y is None:
-            raise ValueError("y is null.")
-        self._model_type = determine_model_type(y)
-        return {
-            EVAL_SET_ARG_KEY: (eval_x, eval_y),
-            CAT_FEATURES_ARG_KEY: df.select_dtypes(include="category").columns.tolist(),
-            ORIGINAL_X_ARG_KEY: df,
-            "sample_weight": w,
-        }
+    def provide_estimator(self):
+        return self._provide_catboost()
+
+    def create_estimator(self):
+        return self._create_catboost()
 
     def set_options(
         self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
@@ -214,66 +197,66 @@ 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
-            )
-            logging.info(
-                "Creating catboost model with depth %d, boosting type %s, best iteration %d",
-                self._depth,
-                self._boosting_type,
-                -1 if best_iteration is None else best_iteration,
-            )
-            match self._model_type:
-                case ModelType.BINARY:
-                    catboost = CatBoostClassifierWrapper(
-                        iterations=iterations,
-                        learning_rate=self._learning_rate,
-                        depth=self._depth,
-                        l2_leaf_reg=self._l2_leaf_reg,
-                        boosting_type=self._boosting_type,
-                        early_stopping_rounds=self._early_stopping_rounds,
-                        metric_period=100,
-                        task_type="GPU" if torch.cuda.is_available() else "CPU",
-                        devices="0" if torch.cuda.is_available() else None,
-                    )
-                case ModelType.REGRESSION:
-                    catboost = CatBoostRegressorWrapper(
-                        iterations=iterations,
-                        learning_rate=self._learning_rate,
-                        depth=self._depth,
-                        l2_leaf_reg=self._l2_leaf_reg,
-                        boosting_type=self._boosting_type,
-                        early_stopping_rounds=self._early_stopping_rounds,
-                        metric_period=100,
-                        task_type="GPU" if torch.cuda.is_available() else "CPU",
-                        devices="0" if torch.cuda.is_available() else None,
-                    )
-                case ModelType.BINNED_BINARY:
-                    catboost = CatBoostClassifierWrapper(
-                        iterations=iterations,
-                        learning_rate=self._learning_rate,
-                        depth=self._depth,
-                        l2_leaf_reg=self._l2_leaf_reg,
-                        boosting_type=self._boosting_type,
-                        early_stopping_rounds=self._early_stopping_rounds,
-                        metric_period=100,
-                        task_type="GPU" if torch.cuda.is_available() else "CPU",
-                        devices="0" if torch.cuda.is_available() else None,
-                    )
-                case ModelType.MULTI_CLASSIFICATION:
-                    catboost = CatBoostClassifierWrapper(
-                        iterations=iterations,
-                        learning_rate=self._learning_rate,
-                        depth=self._depth,
-                        l2_leaf_reg=self._l2_leaf_reg,
-                        boosting_type=self._boosting_type,
-                        early_stopping_rounds=self._early_stopping_rounds,
-                        metric_period=100,
-                        task_type="GPU" if torch.cuda.is_available() else "CPU",
-                        devices="0" if torch.cuda.is_available() else None,
-                    )
+            catboost = self._create_catboost()
             self._catboost = catboost
         if catboost is None:
             raise ValueError("catboost is null")
         return catboost
+
+    def _create_catboost(self) -> CatBoost:
+        best_iteration = self._best_iteration
+        iterations = best_iteration if best_iteration is not None else self._iterations
+        print(
+            f"Creating catboost model with depth {self._depth}, boosting type {self._boosting_type}, best iteration {best_iteration}",
+        )
+        match self._model_type:
+            case ModelType.BINARY:
+                return CatBoostClassifierWrapper(
+                    iterations=iterations,
+                    learning_rate=self._learning_rate,
+                    depth=self._depth,
+                    l2_leaf_reg=self._l2_leaf_reg,
+                    boosting_type=self._boosting_type,
+                    early_stopping_rounds=self._early_stopping_rounds,
+                    metric_period=100,
+                    task_type="GPU" if torch.cuda.is_available() else "CPU",
+                    devices="0" if torch.cuda.is_available() else None,
+                )
+            case ModelType.REGRESSION:
+                return CatBoostRegressorWrapper(
+                    iterations=iterations,
+                    learning_rate=self._learning_rate,
+                    depth=self._depth,
+                    l2_leaf_reg=self._l2_leaf_reg,
+                    boosting_type=self._boosting_type,
+                    early_stopping_rounds=self._early_stopping_rounds,
+                    metric_period=100,
+                    task_type="GPU" if torch.cuda.is_available() else "CPU",
+                    devices="0" if torch.cuda.is_available() else None,
+                )
+            case ModelType.BINNED_BINARY:
+                return CatBoostClassifierWrapper(
+                    iterations=iterations,
+                    learning_rate=self._learning_rate,
+                    depth=self._depth,
+                    l2_leaf_reg=self._l2_leaf_reg,
+                    boosting_type=self._boosting_type,
+                    early_stopping_rounds=self._early_stopping_rounds,
+                    metric_period=100,
+                    task_type="GPU" if torch.cuda.is_available() else "CPU",
+                    devices="0" if torch.cuda.is_available() else None,
+                )
+            case ModelType.MULTI_CLASSIFICATION:
+                return CatBoostClassifierWrapper(
+                    iterations=iterations,
+                    learning_rate=self._learning_rate,
+                    depth=self._depth,
+                    l2_leaf_reg=self._l2_leaf_reg,
+                    boosting_type=self._boosting_type,
+                    early_stopping_rounds=self._early_stopping_rounds,
+                    metric_period=100,
+                    task_type="GPU" if torch.cuda.is_available() else "CPU",
+                    devices="0" if torch.cuda.is_available() else None,
+                )
+            case _:
+                raise ValueError(f"Unrecognised model type: {self._model_type}")

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/model/model.py

@@ -25,11 +25,6 @@ class Model(Params, Fit):
         """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."""
-        raise NotImplementedError("estimator not implemented in parent class.")
-
     @property
     def supports_importances(self) -> bool:
         """Whether this model supports feature importances."""
@@ -44,13 +39,10 @@ class Model(Params, Fit):
             "feature_importances not implemented in parent class."
         )
 
-    def pre_fit(
-        self,
-        df: pd.DataFrame,
-        y: pd.Series | pd.DataFrame | None,
-        eval_x: pd.DataFrame | None = None,
-        eval_y: pd.Series | pd.DataFrame | None = None,
-        w: pd.Series | None = None,
-    ) -> dict[str, Any]:
-        """A call to make sure the model is prepared for the target type."""
-        raise NotImplementedError("pre_fit not implemented in parent class.")
+    def provide_estimator(self) -> Any:
+        """Provides the current estimator."""
+        raise NotImplementedError("provides_estimator not implemented in parent class.")
+
+    def create_estimator(self) -> Any:
+        """Creates a new estimator."""
+        raise NotImplementedError("creates_estimator not implemented in parent class.")

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/model/model_router.py

@@ -2,7 +2,7 @@
 
 import json
 import os
-from typing import Any, Self
+from typing import Self
 
 import optuna
 import pandas as pd
@@ -40,13 +40,6 @@ class ModelRouter(Model):
     def supports_x(cls, df: pd.DataFrame) -> bool:
         return True
 
-    @property
-    def estimator(self) -> Any:
-        model = self._model
-        if model is None:
-            raise ValueError("model is null")
-        return model.estimator
-
     @property
     def supports_importances(self) -> bool:
         model = self._model
@@ -61,18 +54,17 @@ class ModelRouter(Model):
             raise ValueError("model is null")
         return model.feature_importances
 
-    def pre_fit(
-        self,
-        df: pd.DataFrame,
-        y: pd.Series | pd.DataFrame | None,
-        eval_x: pd.DataFrame | None = None,
-        eval_y: pd.Series | pd.DataFrame | None = None,
-        w: pd.Series | None = None,
-    ) -> dict[str, Any]:
+    def provide_estimator(self):
+        model = self._model
+        if model is None:
+            raise ValueError("model is null")
+        return model.provide_estimator()
+
+    def create_estimator(self):
         model = self._model
         if model is None:
             raise ValueError("model is null")
-        return model.pre_fit(df, y=y, eval_x=eval_x, eval_y=eval_y, w=w)
+        return model.create_estimator()
 
     def set_options(
         self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/model/tabpfn/tabpfn_model.py

@@ -5,7 +5,7 @@ import json
 import logging
 import os
 import pickle
-from typing import Any, Self
+from typing import Self
 
 import optuna
 import pandas as pd
@@ -42,10 +42,6 @@ class TabPFNModel(Model):
         self._tabpfn = None
         self._model_type = None
 
-    @property
-    def estimator(self) -> Any:
-        return self._provide_tabpfn()
-
     @property
     def supports_importances(self) -> bool:
         return False
@@ -54,18 +50,11 @@ class TabPFNModel(Model):
     def feature_importances(self) -> dict[str, float]:
         return {}
 
-    def pre_fit(
-        self,
-        df: pd.DataFrame,
-        y: pd.Series | pd.DataFrame | None,
-        eval_x: pd.DataFrame | None = None,
-        eval_y: pd.Series | pd.DataFrame | None = None,
-        w: pd.Series | None = None,
-    ):
-        if y is None:
-            raise ValueError("y is null.")
-        self._model_type = determine_model_type(y)
-        return {}
+    def provide_estimator(self):
+        return self._provide_tabpfn()
+
+    def create_estimator(self):
+        return self._create_tabpfn()
 
     def set_options(
         self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
@@ -133,29 +122,34 @@ class TabPFNModel(Model):
     def _provide_tabpfn(self) -> AutoTabPFNClassifier | AutoTabPFNRegressor:
         tabpfn = self._tabpfn
         if tabpfn is None:
-            max_time = 1 if pytest_is_running.is_running() else 120
-            match self._model_type:
-                case ModelType.BINARY:
-                    tabpfn = AutoTabPFNClassifier(
-                        max_time=max_time,
-                        device="cuda" if torch.cuda.is_available() else "cpu",
-                    )
-                case ModelType.REGRESSION:
-                    tabpfn = AutoTabPFNRegressor(
-                        max_time=max_time,
-                        device="cuda" if torch.cuda.is_available() else "cpu",
-                    )
-                case ModelType.BINNED_BINARY:
-                    tabpfn = AutoTabPFNClassifier(
-                        max_time=max_time,
-                        device="cuda" if torch.cuda.is_available() else "cpu",
-                    )
-                case ModelType.MULTI_CLASSIFICATION:
-                    tabpfn = AutoTabPFNClassifier(
-                        max_time=max_time,
-                        device="cuda" if torch.cuda.is_available() else "cpu",
-                    )
+            tabpfn = self._create_tabpfn()
             self._tabpfn = tabpfn
         if tabpfn is None:
             raise ValueError("tabpfn is null")
         return tabpfn
+
+    def _create_tabpfn(self) -> AutoTabPFNClassifier | AutoTabPFNRegressor:
+        max_time = 1 if pytest_is_running.is_running() else 120
+        match self._model_type:
+            case ModelType.BINARY:
+                return AutoTabPFNClassifier(
+                    max_time=max_time,
+                    device="cuda" if torch.cuda.is_available() else "cpu",
+                )
+            case ModelType.REGRESSION:
+                return AutoTabPFNRegressor(
+                    max_time=max_time,
+                    device="cuda" if torch.cuda.is_available() else "cpu",
+                )
+            case ModelType.BINNED_BINARY:
+                return AutoTabPFNClassifier(
+                    max_time=max_time,
+                    device="cuda" if torch.cuda.is_available() else "cpu",
+                )
+            case ModelType.MULTI_CLASSIFICATION:
+                return AutoTabPFNClassifier(
+                    max_time=max_time,
+                    device="cuda" if torch.cuda.is_available() else "cpu",
+                )
+            case _:
+                raise ValueError(f"Unrecognised model type: {self._model_type}")

{wavetrainer-0.0.40 → wavetrainer-0.0.41}/wavetrainer/model/xgboost/xgboost_logger.py

@@ -15,6 +15,10 @@ class XGBoostEpochsLogger(TrainingCallback):
             return False
         log_items = []
         for dataset, metrics in evals_log.items():
+            if dataset == "validation_0":
+                dataset = "validation"
+            elif dataset == "validation_1":
+                dataset = "train"
             for metric_name, values in metrics.items():
                 current_val = values[-1]
                 log_items.append(f"{dataset}-{metric_name}: {current_val:.5f}")