bump v0.4.0

Author: thierrymoudiki

README.md

@@ -3,10 +3,9 @@ tisthemachinelearner
 
 <hr>
 
-Lightweight interface to scikit-learn with 2 classes, Classifier and Regressor. 
+Lightweight interface to scikit-learn with 2 classes, `Classifier` and `Regressor. 
 
-![PyPI](https://img.shields.io/pypi/v/tisthemachinelearner) [![PyPI - License](https://img.shields.io/pypi/l/tisthemachinelearner)](./LICENSE) [![Downloads](https://pepy.tech/badge/tisthemachinelearner)](https://pepy.tech/project/tisthemachinelearner) 
-[![HitCount](https://hits.dwyl.com/Techtonique/tisthemachinelearner.svg?style=flat-square)](http://hits.dwyl.com/Techtonique/tisthemachinelearner)
+![PyPI](https://img.shields.io/pypi/v/tisthemachinelearner) [![PyPI - License](https://img.shields.io/pypi/l/tisthemachinelearner)](LICENSE) [![Downloads](https://pepy.tech/badge/tisthemachinelearner)](https://pepy.tech/project/tisthemachinelearner) 
 [![Documentation](https://img.shields.io/badge/documentation-is_here-green)](https://techtonique.github.io/tisthemachinelearner/)
 
 

examples/example.py

@@ -29,3 +29,37 @@
 print(reg.predict(X_test))
 print(np.sqrt(np.mean((reg.predict(X_test) - y_test) ** 2)))
 
+# custom 
+
+X, y = load_breast_cancer(return_X_y=True)
+X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
+
+
+clf = Classifier("LogisticRegression", custom=False, calibrate=False)
+clf.fit(X_train, y_train)
+print(clf.predict(X_test))
+print(clf.score(X_test, y_test))
+
+clf = Classifier("LogisticRegression", custom=True, calibrate=False)
+clf.fit(X_train, y_train)
+print(clf.predict(X_test))
+print(clf.score(X_test, y_test))
+
+clf = Classifier("LogisticRegression", custom=True, n_hidden_features=3, random_state=42, calibrate=False)
+clf.fit(X_train, y_train)
+print(clf.predict(X_test))
+print(clf.score(X_test, y_test))
+
+X, y = load_diabetes(return_X_y=True)
+X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
+
+reg = Regressor("RidgeCV", alphas=[0.01, 0.1, 1, 10], custom=True, n_hidden_features=3)
+reg.fit(X_train, y_train)
+print(reg.predict(X_test))
+print(np.sqrt(np.mean((reg.predict(X_test) - y_test) ** 2)))
+
+reg = Regressor("RidgeCV", alphas=[0.01, 0.1, 1, 10], custom=True, n_hidden_features=10)
+reg.fit(X_train, y_train)
+print(reg.predict(X_test))
+print(np.sqrt(np.mean((reg.predict(X_test) - y_test) ** 2)))
+

pyproject.toml

@@ -1,10 +1,10 @@
 [project]
 name = "tisthemachinelearner"
-version = "0.3.3"
-description = "git"
+version = "0.4.0"
+description = "Lightweight interface to scikit-learn, with 2 classes"
 readme = "README.md"
 requires-python = ">=3.8"
-dependencies = ["scikit-learn"]
+dependencies = ["scikit-learn", "nnetsauce"]
 
 [build-system]
 requires = ["hatchling"]

src/tisthemachinelearner/base.py

@@ -1,18 +1,32 @@
-from sklearn.base import BaseEstimator, ClassifierMixin, RegressorMixin
 import importlib
+import nnetsauce as ns 
+from sklearn.base import BaseEstimator
 
 
 class BaseModel(BaseEstimator):
     """
     Base class for dynamically loading and wrapping scikit-learn models.
     """
-    def __init__(self, base_model, **kwargs):
+    # Custom parameters that should only be passed to nnetsauce models
+    CUSTOM_PARAMS = [
+        'n_hidden_features',
+        'activation_name',
+        'bias',
+        'dropout',
+        'direct_link',
+        'n_clusters',
+        'cluster_encode',
+        'type_clust'
+    ]
+
+    def __init__(self, base_model, custom=False, **kwargs):
         """
         Initialize a scikit-learn model dynamically.
 
         Parameters:
         - base_model (str): The class name of the scikit-learn model (e.g., 'LogisticRegression').
-        - **kwargs: Additional parameters to pass to the scikit-learn model constructor.
+        - custom (bool): Whether the model is a custom nnetsauce model.
+        - **kwargs: Additional parameters to pass to the model constructor.
         """
         sklearn_modules = [
             "linear_model",
@@ -27,8 +41,16 @@ def __init__(self, base_model, **kwargs):
             "kernel_ridge",
         ]
         self.base_model = base_model
-        self.model_params = kwargs
-        self.model = self._load_model(base_model, sklearn_modules)(**kwargs)
+        self.custom = custom
+
+        # Split kwargs into base and custom parameters
+        self.base_kwargs = {k: v for k, v in kwargs.items() if k not in self.CUSTOM_PARAMS}
+        self.custom_kwargs = {k: v for k, v in kwargs.items() if k in self.CUSTOM_PARAMS}
+
+        # Initialize only the base model here
+        self.model = self._load_model(base_model, sklearn_modules)(**self.base_kwargs)
+        
+        # Custom model wrapping is handled in derived classes
 
     def _load_model(self, base_model, modules):
         """
@@ -57,7 +79,7 @@ def fit(self, X, y, **kwargs):
         Parameters:
         - X (array-like): Training data features.
         - y (array-like): Target values.
-        - **kwargs: Additional parameters to pass to the scikit-learn model fit method.
+        - **kwargs: Additional parameters to pass to the model fit method.
         """
         self.model.fit(X, y, **kwargs)
         return self
@@ -68,7 +90,7 @@ def predict(self, X, **kwargs):
 
         Parameters:
         - X (array-like): Input data.
-        - **kwargs: Additional parameters to pass to the scikit-learn model predict method.
+        - **kwargs: Additional parameters to pass to the model predict method.
         Returns:
         - array-like: Predictions.
         """
@@ -81,7 +103,7 @@ def score(self, X, y, **kwargs):
         Parameters:
         - X (array-like): Test data features.
         - y (array-like): True labels.
-        - **kwargs: Additional parameters to pass to the scikit-learn model score method.
+        - **kwargs: Additional parameters to pass to the model score method.
         
         Returns:
         - float: The score.

src/tisthemachinelearner/classifier.py

@@ -1,5 +1,7 @@
+import nnetsauce as ns
 from .base import BaseModel
 from sklearn.base import ClassifierMixin
+from .wrappedcalibratedclassifier import WrappedCalibratedClassifier
 
 class Classifier(BaseModel, ClassifierMixin):
     """
@@ -29,4 +31,30 @@ class Classifier(BaseModel, ClassifierMixin):
         print(clf.score(X_test, y_test))
         ```
     """
-    pass
+    def __init__(self, base_model, custom=False, calibrate=False, **kwargs):
+        super().__init__(base_model, custom, **kwargs)
+        if self.custom:
+            self.model = ns.CustomClassifier(self.model, **self.custom_kwargs)
+        self.calibrate = calibrate
+        if self.calibrate: 
+            raise NotImplementedError
+
+    def fit(self, X, y, **kwargs):
+        """Fit the model."""
+        super().fit(X, y, **kwargs)
+        if self.calibrate:
+            self.model = WrappedCalibratedClassifier(self.model, method='sigmoid', cv='prefit')
+            self.model.fit(X, y)
+        return self
+    
+    def predict_proba(self, X):
+        """
+        Predict class probabilities for the input data.
+
+        Parameters:
+        - X (array-like): Input data features.
+
+        Returns:
+        - array-like: Predicted class probabilities.
+        """
+        return self.model.predict_proba(X)

src/tisthemachinelearner/regressor.py

@@ -1,3 +1,4 @@
+import nnetsauce as ns
 from .base import BaseModel
 from sklearn.base import RegressorMixin
 
@@ -29,4 +30,14 @@ class Regressor(BaseModel, RegressorMixin):
         print(np.sqrt(np.mean((reg.predict(X_test) - y_test) ** 2)))
         ```
     """
-    pass
+    def __init__(self, base_model, custom=False, **kwargs):
+        super().__init__(base_model, custom, **kwargs)
+        if self.custom:
+            self.model = ns.CustomRegressor(self.model, **self.custom_kwargs)
+
+    def fit(self, X, y, **kwargs):
+        """Fit the model."""
+        super().fit(X, y, **kwargs)
+        if self.custom:
+            self.model.fit(X, y)
+        return self

src/tisthemachinelearner/wrappedcalibratedclassifier.py

@@ -0,0 +1,31 @@
+from sklearn.base import BaseEstimator, ClassifierMixin
+from sklearn.calibration import CalibratedClassifierCV
+
+class WrappedCalibratedClassifier(BaseEstimator, ClassifierMixin):
+    def __init__(self, base_estimator, method='sigmoid', cv=5):
+        self.base_estimator = base_estimator
+        self.method = method
+        self.cv = cv
+
+    def fit(self, X, y):
+        self.calibrated_clf_ = CalibratedClassifierCV(
+            estimator=self.base_estimator,
+            method=self.method,
+            cv=self.cv
+        )
+        self.calibrated_clf_.fit(X, y)
+        return self
+
+    def predict(self, X):
+        return self.calibrated_clf_.predict(X)
+
+    def predict_proba(self, X):
+        return self.calibrated_clf_.predict_proba(X)
+
+    @property
+    def classes_(self):
+        return self.calibrated_clf_.classes_
+
+    @property
+    def _estimator_type(self):
+        return "classifier"