wavetrainer 0.0.50__tar.gz → 0.0.52__tar.gz

{wavetrainer-0.0.50/wavetrainer.egg-info → wavetrainer-0.0.52}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wavetrainer
-Version: 0.0.50
+Version: 0.0.52
 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-0.0.50 → wavetrainer-0.0.52}/setup.py

@@ -23,7 +23,7 @@ def install_requires() -> typing.List[str]:
 
 setup(
     name='wavetrainer',
-    version='0.0.50',
+    version='0.0.52',
     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-0.0.50 → wavetrainer-0.0.52}/wavetrainer/__init__.py

@@ -2,5 +2,5 @@
 
 from .create import create
 
-__VERSION__ = "0.0.50"
+__VERSION__ = "0.0.52"
 __all__ = ("create",)

{wavetrainer-0.0.50 → wavetrainer-0.0.52}/wavetrainer/model/model_router.py

@@ -1,19 +1,23 @@
 """A model class that routes to other models."""
 
+import functools
 import json
 import os
 from typing import Self
 
 import optuna
 import pandas as pd
+from sklearn.metrics import accuracy_score  # type: ignore
 
+from ..model_type import ModelType, determine_model_type
 from .catboost.catboost_model import CatboostModel
-from .model import Model
+from .model import PREDICTION_COLUMN, PROBABILITY_COLUMN_PREFIX, Model
 from .tabpfn.tabpfn_model import TabPFNModel
 from .xgboost.xgboost_model import XGBoostModel
 
 _MODEL_ROUTER_FILE = "model_router.json"
 _MODEL_KEY = "model"
+_FALSE_POSITIVE_REDUCTION_STEPS_KEY = "false_positive_reduction_steps"
 _MODELS = {
     CatboostModel.name(): CatboostModel,
     TabPFNModel.name(): TabPFNModel,
@@ -27,10 +31,12 @@ class ModelRouter(Model):
     # pylint: disable=too-many-positional-arguments,too-many-arguments
 
     _model: Model | None
+    _false_positive_reduction_steps: int | None
 
     def __init__(self) -> None:
         super().__init__()
         self._model = None
+        self._false_positive_reduction_steps = None
 
     @classmethod
     def name(cls) -> str:
@@ -81,6 +87,9 @@ class ModelRouter(Model):
     def set_options(
         self, trial: optuna.Trial | optuna.trial.FrozenTrial, df: pd.DataFrame
     ) -> None:
+        self._false_positive_reduction_steps = trial.suggest_int(
+            _FALSE_POSITIVE_REDUCTION_STEPS_KEY, 0, 5
+        )
         model_name = trial.suggest_categorical(
             "model", [k for k, v in _MODELS.items() if v.supports_x(df)]
         )
@@ -122,7 +131,48 @@ class ModelRouter(Model):
         model = self._model
         if model is None:
             raise ValueError("model is null")
-        model.fit(df, y=y, w=w, eval_x=eval_x, eval_y=eval_y)
+        false_positive_reduction_steps = self._false_positive_reduction_steps
+        if false_positive_reduction_steps is None:
+            false_positive_reduction_steps = 0
+        for i in range(max(false_positive_reduction_steps, 1)):
+            print(f"False Positive Reduction Step: {i + 1}")
+            pred = model.fit_transform(df, y=y, w=w, eval_x=eval_x, eval_y=eval_y)
+            if (
+                w is None
+                or y is None
+                or determine_model_type(y) == ModelType.REGRESSION
+            ):
+                break
+            print(f"Accuracy: {accuracy_score(y, pred[PREDICTION_COLUMN])}")
+            pred["__wavetrain_correct"] = pred[PREDICTION_COLUMN] != y
+            pred["__wavetrain_error_weight"] = pred["__wavetrain_correct"].astype(float)
+            prob_columns = sorted(
+                [
+                    x
+                    for x in pred.columns.values.tolist()
+                    if x.startswith(PROBABILITY_COLUMN_PREFIX)
+                ]
+            )
+            if prob_columns:
+
+                def determine_error_weight(
+                    row: pd.Series, prob_columns: list[str]
+                ) -> float:
+                    nonlocal y
+                    if not row["__wavetrain_correct"]:
+                        return abs(row[prob_columns[1 - int(y.loc[row.name])]])  # type: ignore
+                    return 0.0
+
+                pred["__wavetrain_error_weight"] = pred.apply(
+                    functools.partial(
+                        determine_error_weight,
+                        prob_columns=prob_columns,
+                    ),
+                    axis=1,
+                )
+            w += pred["__wavetrain_error_weight"]
+            w = w.clip(lower=0.0)
+
         return self
 
     def transform(self, df: pd.DataFrame) -> pd.DataFrame:

{wavetrainer-0.0.50 → wavetrainer-0.0.52}/wavetrainer/model/xgboost/xgboost_model.py

@@ -125,7 +125,10 @@ class XGBoostModel(Model):
         try:
             score_dict = bst.get_booster().get_score(importance_type="weight")  # type: ignore
             total = sum(score_dict.values())  # type: ignore
-            return {k: v / total for k, v in score_dict.items()}  # type: ignore
+            return {
+                k: 0.0 if total == 0.0 else v / total  # type: ignore
+                for k, v in score_dict.items()  # type: ignore
+            }  # type: ignore
         except XGBoostError as exc:
             print(str(exc))
             return {}

{wavetrainer-0.0.50 → wavetrainer-0.0.52}/wavetrainer/trainer.py

@@ -154,7 +154,7 @@ class Trainer(Fit):
         self._dt_column = dt_column
         self._max_train_timeout = max_train_timeout
         self._cutoff_dt = cutoff_dt
-        self._embedding_cols = embedding_cols
+        self.embedding_cols = embedding_cols
 
     def _provide_study(self, column: str) -> optuna.Study:
         storage_name = f"sqlite:///{self._folder}/{column}/{_STUDYDB_FILENAME}"
@@ -249,7 +249,7 @@ class Trainer(Fit):
 
                     # Perform common reductions
                     start_reducer = time.time()
-                    reducer = CombinedReducer(self._embedding_cols)
+                    reducer = CombinedReducer(self.embedding_cols)
                     reducer.set_options(trial, x)
                     x_train = reducer.fit_transform(x_train, y=y_train)
                     x_test = reducer.transform(x_test)
@@ -514,7 +514,7 @@ class Trainer(Fit):
                 date_str = dates[-1].isoformat()
                 folder = os.path.join(column_path, date_str)
 
-                reducer = CombinedReducer(self._embedding_cols)
+                reducer = CombinedReducer(self.embedding_cols)
                 reducer.load(folder)
 
                 model = ModelRouter()

{wavetrainer-0.0.50 → wavetrainer-0.0.52/wavetrainer.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wavetrainer
-Version: 0.0.50
+Version: 0.0.52
 Summary: A library for automatically finding the optimal model within feature and hyperparameter space.
 Home-page: https://github.com/8W9aG/wavetrainer
 Author: Will Sackfield