wavetrainer 0.1.8__tar.gz → 0.1.9__tar.gz

{wavetrainer-0.1.8/wavetrainer.egg-info → wavetrainer-0.1.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wavetrainer
-Version: 0.1.8
+Version: 0.1.9
 Summary: A library for automatically finding the optimal model within feature and hyperparameter space.
 Home-page: https://github.com/8W9aG/wavetrainer
 Author: Will Sackfield
@@ -28,6 +28,7 @@ Requires-Dist: xgboost>=3.0.0
 Requires-Dist: jax>=0.6.1
 Requires-Dist: tabpfn_extensions>=0.0.4
 Requires-Dist: hyperopt>=0.2.7
+Requires-Dist: pycaleva>=0.8.2
 
 # wavetrainer
 
@@ -62,6 +63,7 @@ Python 3.11.6:
 - [jax](https://github.com/jax-ml/jax)
 - [tabpfn_extensions](https://github.com/PriorLabs/tabpfn-extensions)
 - [hyperopt](https://github.com/hyperopt/hyperopt)
+- [pycaleva](https://github.com/MartinWeigl/pycaleva)
 
 ## Raison D'être :thought_balloon:
 

{wavetrainer-0.1.8 → wavetrainer-0.1.9}/README.md

@@ -31,6 +31,7 @@ Python 3.11.6:
 - [jax](https://github.com/jax-ml/jax)
 - [tabpfn_extensions](https://github.com/PriorLabs/tabpfn-extensions)
 - [hyperopt](https://github.com/hyperopt/hyperopt)
+- [pycaleva](https://github.com/MartinWeigl/pycaleva)
 
 ## Raison D'être :thought_balloon:
 

wavetrainer-0.1.8/wavetrainer.egg-info/requires.txt → wavetrainer-0.1.9/requirements.txt

@@ -15,3 +15,4 @@ xgboost>=3.0.0
 jax>=0.6.1
 tabpfn_extensions>=0.0.4
 hyperopt>=0.2.7
+pycaleva>=0.8.2

{wavetrainer-0.1.8 → wavetrainer-0.1.9}/setup.py

@@ -23,7 +23,7 @@ def install_requires() -> typing.List[str]:
 
 setup(
     name='wavetrainer',
-    version='0.1.8',
+    version='0.1.9',
     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-0.1.8 → wavetrainer-0.1.9}/wavetrainer/__init__.py

@@ -2,5 +2,5 @@
 
 from .create import create
 
-__VERSION__ = "0.1.8"
+__VERSION__ = "0.1.9"
 __all__ = ("create",)

{wavetrainer-0.1.8 → wavetrainer-0.1.9}/wavetrainer/calibrator/calibrator_router.py

@@ -5,10 +5,12 @@ import logging
 import os
 from typing import Self
 
+import numpy as np
 import optuna
 import pandas as pd
+from pycaleva import CalibrationEvaluator  # type: ignore
 
-from ..model.model import Model
+from ..model.model import PROBABILITY_COLUMN_PREFIX, Model
 from ..model_type import ModelType, determine_model_type
 from .calibrator import Calibrator
 from .vennabers_calibrator import VennabersCalibrator
@@ -26,10 +28,12 @@ class CalibratorRouter(Calibrator):
     # pylint: disable=too-many-positional-arguments,too-many-arguments
 
     _calibrator: Calibrator | None
+    _ce: CalibrationEvaluator | None
 
     def __init__(self, model: Model):
         super().__init__(model)
         self._calibrator = None
+        self._ce = None
 
     @classmethod
     def name(cls) -> str:
@@ -75,6 +79,14 @@ class CalibratorRouter(Calibrator):
                 },
                 handle,
             )
+        ce = self._ce
+        if ce is not None:
+            try:
+                ce.calibration_report(
+                    os.path.join(folder, "calibration.pdf"), "binary-classifier"
+                )
+            except ValueError as exc:
+                logging.warning(str(exc))
 
     def fit(
         self,
@@ -94,6 +106,25 @@ class CalibratorRouter(Calibrator):
             calibrator = VennabersCalibrator(self._model)
         calibrator.fit(df, y=y, w=w)
         self._calibrator = calibrator
+
+        pred_prob = calibrator.transform(df)
+        pred_prob = pred_prob.drop(
+            columns=[
+                x
+                for x in pred_prob.columns.values.tolist()
+                if not x.startswith(PROBABILITY_COLUMN_PREFIX)
+            ],
+            errors="ignore",
+        )
+        ce = CalibrationEvaluator(
+            y.to_numpy(),
+            np.max(pred_prob.to_numpy(), axis=1),
+            outsample=True,
+            n_groups="auto",
+        )
+        print(f"Hosmer Lemeshow: {ce.hosmerlemeshow()}")
+        self._ce = ce
+
         return self
 
     def transform(self, df: pd.DataFrame) -> pd.DataFrame:

{wavetrainer-0.1.8 → wavetrainer-0.1.9/wavetrainer.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: wavetrainer
-Version: 0.1.8
+Version: 0.1.9
 Summary: A library for automatically finding the optimal model within feature and hyperparameter space.
 Home-page: https://github.com/8W9aG/wavetrainer
 Author: Will Sackfield
@@ -28,6 +28,7 @@ Requires-Dist: xgboost>=3.0.0
 Requires-Dist: jax>=0.6.1
 Requires-Dist: tabpfn_extensions>=0.0.4
 Requires-Dist: hyperopt>=0.2.7
+Requires-Dist: pycaleva>=0.8.2
 
 # wavetrainer
 
@@ -62,6 +63,7 @@ Python 3.11.6:
 - [jax](https://github.com/jax-ml/jax)
 - [tabpfn_extensions](https://github.com/PriorLabs/tabpfn-extensions)
 - [hyperopt](https://github.com/hyperopt/hyperopt)
+- [pycaleva](https://github.com/MartinWeigl/pycaleva)
 
 ## Raison D'être :thought_balloon:
 

wavetrainer-0.1.8/requirements.txt → wavetrainer-0.1.9/wavetrainer.egg-info/requires.txt

@@ -14,4 +14,5 @@ pytest-is-running>=1.5.1
 xgboost>=3.0.0
 jax>=0.6.1
 tabpfn_extensions>=0.0.4
-hyperopt>=0.2.7
+hyperopt>=0.2.7
+pycaleva>=0.8.2