WIP: core: support autotune

README.rst

@@ -136,6 +136,23 @@ These wrappers assume the ``X`` parameter given to ``fit``, ``predict``, and ``p
   >>> sk_clf.predict(['woof'])
   >>> sk_clf.predict(df['txt'])
 
+Hyper-parameter auto-tuning
+----------------------------
+
+It's possible to pass a validation set to ``fit`` in order to optimize the hyper-parameters.
+The `auto-tune settings <https://fasttext.cc/docs/en/autotune.html>`_ can be passed to the constructor. E.g.,
+
+.. code-block:: python
+
+  >>> from skift import SeriesFtClassifier
+  >>> df = pandas.DataFrame([['woof', 0], ['meow', 1]], columns=['txt', 'lbl'])
+  >>> df_val = pandas.DataFrame([['woof woof', 0], ['meow meow', 1]], columns=['txt', 'lbl'])
+  >>> sk_clf = SeriesFtClassifier(input_col_lbl='txt', epoch=8, autotuneDuration=5)
+  >>> sk_clf.fit(df['txt'], df['lbl'], df2['txt'], df2['lbl'])
+  >>> sk_clf.predict(['woof'])
+  >>> sk_clf.predict(df['txt'])
+
+
 
 Contributing
 ============

skift/core.py

@@ -95,7 +95,7 @@ def _validate_y(y):
     def _input_col(self, X):
         pass  # pragma: no cover
 
-    def fit(self, X, y):
+    def fit(self, X, y, X_validation=None, y_validation=None):
         """Fits the classifier
 
         Parameters
@@ -104,6 +104,10 @@ def fit(self, X, y):
             The training input samples.
         y : array-like, shape = [n_samples]
             The target values. An array of int.
+        X_validation : array-like, shape = [n_samples, n_features]
+            The validation input samples.
+        y_validation : array-like, shape = [n_samples]
+            The validation target values. An array of int.
 
         Returns
         -------
@@ -114,10 +118,16 @@ def fit(self, X, y):
         self._validate_x(X)
         y = self._validate_y(y)
         input_col = self._input_col(X)
+        if X_validation is not None:
+            self._validate_x(X_validation)
+            y_validation = self._validate_y(y_validation)
+            input_col_validation = self._input_col(X_validation)
+        else:
+            input_col_validation = None
 
-        return self._fit_input_col(input_col, y)
+        return self._fit_input_col(input_col, y, input_col_validation, y_validation)
 
-    def _fit_input_col(self, input_col, y):
+    def _fit_input_col(self, input_col, y, input_col_validation=None, y_validation=None):
         # Store the classes seen during fit
         self.classes_ = unique_labels(y)
         self.num_classes_ = len(self.classes_)
@@ -126,9 +136,23 @@ def _fit_input_col(self, input_col, y):
         # Dump training set to a fasttext-compatible file
         temp_trainset_fpath = temp_dataset_fpath()
         dump_xy_to_fasttext_format(input_col, y, temp_trainset_fpath)
-        # train
-        self.model = train_supervised(
-            input=temp_trainset_fpath, **self.kwargs)
+        if input_col_validation is not None:
+            n_classes_validation = len(unique_labels(y_validation))
+            assert n_classes_validation == self.num_classes_,\
+                "Number of validation classes doesn't match number of training classes"
+            temp_trainset_fpath_validation = temp_dataset_fpath()
+            dump_xy_to_fasttext_format(input_col_validation, y_validation, temp_trainset_fpath_validation)
+            # train
+            self.model = train_supervised(
+                input=temp_trainset_fpath, **{'autotuneValidationFile': temp_trainset_fpath_validation, **self.kwargs})
+            try:
+                os.remove(temp_trainset_fpath_validation)
+            except FileNotFoundError:  # pragma: no cover
+                pass
+        else:
+            self.model = train_supervised(
+                input=temp_trainset_fpath, **self.kwargs)
+
         # Return the classifier
         try:
             os.remove(temp_trainset_fpath)
@@ -372,7 +396,7 @@ def __init__(self,  **kwargs):
     def _input_col(self, X):
         pass
 
-    def fit(self, X, y):
+    def fit(self, X, y, X_validation=None, y_validation=None):
         """Fits the classifier
 
         Parameters
@@ -381,6 +405,10 @@ def fit(self, X, y):
             The training input samples.
         y : array-like, shape = [n_samples]
             The target values. An array of int.
+        X_validation : pd.Series
+            The validation input samples.
+        y_validation : array-like, shape = [n_samples]
+            The validation target values. An array of int.
 
         Returns
         -------
@@ -393,7 +421,15 @@ def fit(self, X, y):
         except AttributeError:
             input_col = X
         y = self._validate_y(y)
-        return self._fit_input_col(input_col, y)
+        if X_validation is not None:
+            try:
+                input_col_validation = X_validation.values
+            except AttributeError:
+                input_col_validation = X_validation
+            y_validation = self._validate_y(y_validation)
+        else:
+            input_col_validation = None
+        return self._fit_input_col(input_col, y, input_col_validation, y_validation)
 
     def _predict(self, X, k=1):
         # Ensure that fit had been called

tests/test_common.py

@@ -20,6 +20,13 @@ def _ftdf():
     )
 
 
+def _ftdf2():
+    return pd.DataFrame(
+        data=[['woof', 0], ['meow', 1]],
+        columns=['txt', 'lbl']
+    )
+
+
 def test_bad_shape():
     ft_clf = FirstColFtClassifier()
     with pytest.raises(ValueError):
@@ -28,6 +35,28 @@ def test_bad_shape():
         ft_clf.fit([[7]], [[0]])
 
 
+def test_series_predict_tune():
+    ftdf = _ftdf()
+    ftdf2 = _ftdf2()
+    ft_clf = SeriesFtClassifier(autotuneDuration=5)
+    ft_clf.fit(ftdf['txt'], ftdf['lbl'], ftdf2['txt'], ftdf2['lbl'])
+
+    preds = ft_clf.predict(ftdf['txt'])
+    assert preds[0] == 0
+    assert preds[1] == 1
+
+
+def test_series_np_predict_tune():
+    ftdf = _ftdf()
+    ftdf2 = _ftdf2()
+    ft_clf = SeriesFtClassifier(autotuneDuration=5)
+    ft_clf.fit(ftdf['txt'].values, ftdf['lbl'].values, ftdf2['txt'].values, ftdf2['lbl'].values)
+
+    preds = ft_clf.predict(ftdf['txt'])
+    assert preds[0] == 0
+    assert preds[1] == 1
+
+
 def test_series_predict():
     ftdf = _ftdf()
     ft_clf = SeriesFtClassifier()
@@ -38,6 +67,29 @@ def test_series_predict():
     assert preds[1] == 1
 
 
+def test_series_np_predict():
+    ftdf = _ftdf()
+    ft_clf = SeriesFtClassifier()
+    ft_clf.fit(ftdf['txt'].values, ftdf['lbl'].values)
+
+    preds = ft_clf.predict(ftdf['txt'].values)
+    assert preds[0] == 0
+    assert preds[1] == 1
+
+
+def test_predict_tune():
+    ftdf = _ftdf()
+    ftdf2 = _ftdf2()
+    ft_clf = FirstColFtClassifier(autotuneDuration=5)
+    ft_clf.fit(ftdf[['txt']], ftdf['lbl'], X_validation=ftdf2[['txt']], y_validation=ftdf2['lbl'])
+
+    assert ft_clf.predict([['woof woof']])[0] == 0
+    assert ft_clf.predict([['meow meow']])[0] == 1
+    assert ft_clf.predict([['meow']])[0] == 1
+    assert ft_clf.predict([['woof lol']])[0] == 0
+    assert ft_clf.predict([['meow lolz']])[0] == 1
+
+
 def test_predict():
     ftdf = _ftdf()
     ft_clf = FirstColFtClassifier()
@@ -49,7 +101,6 @@ def test_predict():
     assert ft_clf.predict([['woof lol']])[0] == 0
     assert ft_clf.predict([['meow lolz']])[0] == 1
 
-
 def test_predict_proba():
     ftdf = _ftdf()
     ft_clf = FirstColFtClassifier()