transform: Fix representation of a patched sampler

Show the parameters of the imbalanced-learn base class and the
additional parameters introduced in the PickableSampler class (shuffle
and random_state). Ensure that these are forwarded to the base
imbalanced-learn Sampler class if the constructor requries them.

Adapt the tests and example accordingly.

examples/plot_imblearn.py

@@ -32,6 +32,7 @@
     ('segment', SegmentXY(width=1, overlap=0)),
     ('resample', patch_sampler(RandomUnderSampler)()),
 ])
+print("Pipeline:", pipe)
 
 print("Calling a transform on the data does not change it ...")
 Xf, yf = pipe.transform(X, y)
@@ -64,6 +65,7 @@ def score(self, X, y, sample_weight=None):
     ('feature', FeatureRep(features={"min":minimum})),
     ('estimator', VerboseDummyClassifier(strategy="constant", constant=True)),
 ])
+print("Pipeline:", pipe)
 
 print("Split the data into half training and half test data:")
 X_train, X_test, y_train, y_test = temporal_split(X, y, 0.5)

seglearn/tests/test_transform.py

@@ -11,6 +11,7 @@
 from seglearn.feature_functions import all_features, mean
 from seglearn.util import get_ts_data_parts
 from sklearn.utils import shuffle
+from sklearn.base import BaseEstimator
 
 
 def test_sliding_window():
@@ -677,7 +678,9 @@ def resample(Xt):
 # MUST be defined in the global scope for pickling to work correctly
 def mock_resample(ndarray):
     return ndarray[:len(ndarray) // 2]
-class MockImblearnSampler(object):
+class MockImblearnSampler(BaseEstimator):
+    def __init__(self, mocked_param="mock"):
+        pass
     @staticmethod
     def _check_X_y(X, y):
         return X, y, True
@@ -700,6 +703,11 @@ class EmptyClass(object):
     unpickled_sampler = pickle.loads(pickled_sampler)
     assert str(patched_sampler.__class__) == str(unpickled_sampler.__class__)
 
+    # test representation
+    assert "mocked_param" in repr(patched_sampler)
+    assert "random_state" in repr(patched_sampler)
+    assert "shuffle" in repr(patched_sampler)
+
     # multivariate ts
     X = np.random.rand(100, 10, 4)
     y = np.ones(100)

seglearn/transform.py

@@ -1511,9 +1511,14 @@ def patch_sampler(sampler_class):
     '''
     Return a dynamically patched imbalanced-learn Sampler class compatible with Pype.
     '''
-    if not hasattr(sampler_class, 'fit_resample') or not hasattr(sampler_class, '_check_X_y'):
-        raise TypeError('The sampler class to be patched must have a "fit_resample" and a'
-                        ' "_check_X_y" method')
+    conditions = [
+        hasattr(sampler_class, 'fit_resample'),
+        hasattr(sampler_class, '_check_X_y'),
+        hasattr(sampler_class, '_get_param_names'),
+    ]
+    if not all(conditions):
+        raise TypeError('The sampler class to be patched must have a "fit_resample", a "_check_X_y"'
+                        ' method and a "_get_param_names" class method.')
 
     class PickableSampler(sampler_class, XyTransformerMixin):
         '''
@@ -1548,6 +1553,17 @@ def __init__(self, shuffle=False, random_state=None, **kwargs):
                 kwargs["random_state"] = random_state
             super(PickableSampler, self).__init__(**kwargs)
 
+        @classmethod
+        def _get_param_names(cls):
+            '''
+            Get parameters of the imbalanced-learn Sampler base class and the additional arguments
+            of the dynamically derived class.
+            '''
+            init_signature = signature(getattr(cls, '__init__'))
+            parameters = [p.name for p in init_signature.parameters.values()
+                          if p.name != 'self' and p.kind != p.VAR_KEYWORD]
+            return sorted(set(sampler_class._get_param_names() + parameters))
+
         @staticmethod
         def _check_X_y(Xt, yt):
             '''