[WIP] Sklearn wrapper for RandomProjections Model

gensim/sklearn_integration/sklearn_wrapper_gensim_ldamodel.py

@@ -76,7 +76,7 @@ def set_params(self, **parameters):
     def fit(self, X, y=None):
         """
         For fitting corpus into the class object.
-        Calls gensim.model.LdaModel:
+        Calls gensim.models.LdaModel:
         >>> gensim.models.LdaModel(corpus=corpus, num_topics=num_topics, id2word=id2word, passes=passes, update_every=update_every, alpha=alpha, iterations=iterations, eta=eta, random_state=random_state)
         """
         if sparse.issparse(X):

gensim/sklearn_integration/sklearn_wrapper_gensim_lsimodel.py

@@ -58,7 +58,7 @@ def set_params(self, **parameters):
     def fit(self, X, y=None):
         """
         For fitting corpus into the class object.
-        Calls gensim.model.LsiModel:
+        Calls gensim.models.LsiModel:
         >>>gensim.models.LsiModel(corpus=corpus, num_topics=num_topics, id2word=id2word, chunksize=chunksize, decay=decay, onepass=onepass, power_iters=power_iters, extra_samples=extra_samples)
         """
         if sparse.issparse(X):

gensim/sklearn_integration/sklearn_wrapper_gensim_rpmodel.py

@@ -0,0 +1,57 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2011 Radim Rehurek <[email protected]>
+# Licensed under the GNU LGPL v2.1 - http://www.gnu.org/licenses/lgpl.html
+#
+"""
+Scikit learn interface for gensim for easy use of gensim with scikit-learn
+Follows scikit-learn API conventions
+"""
+from gensim import models
+from gensim.sklearn_integration import base_sklearn_wrapper
+from sklearn.base import TransformerMixin, BaseEstimator
+
+
+class SklearnWrapperRpModel(models.RpModel, base_sklearn_wrapper.BaseSklearnWrapper, TransformerMixin, BaseEstimator):
+    """
+    Base RP module
+    """
+
+    def __init__(self, corpus, id2word=None, num_topics=300):
+        """
+        Sklearn wrapper for RP model. Class derived from gensim.models.RpModel.
+        """
+        self.corpus = corpus
+        self.id2word = id2word
+        self.num_topics = num_topics
+
+    def get_params(self, deep=True):
+        """
+        Returns all parameters as dictionary.
+        """
+        return {"corpus": self.corpus, "id2word": self.id2word, "num_topics": self.num_topics}
+
+    def set_params(self, **parameters):
+        """
+        Set all parameters.
+        """
+        super(SklearnWrapperRpModel, self).set_params(**parameters)
+
+    def fit(self, X, y=None):
+        """
+        For fitting corpus into class object.
+        Calls gensim.models.RpModel
+        >>>gensim.models.RpModel(corpus=self.corpus, id2word=self.id2word, num_topics=self.num_topics)
+        """
+        super(SklearnWrapperRpModel, self).__init__(corpus=self.corpus, id2word=self.id2word, num_topics=self.num_topics)
+
+    def transform(self, doc):
+        """
+        Take document/corpus as input.
+        Return RP representation of the input document/corpus.
+        """
+        return self[doc]
+
+    def partial_fit(self, X):
+        raise NotImplementedError("'partial_fit' has not been implemented for the RandomProjections model")

gensim/test/test_sklearn_integration.py

@@ -16,6 +16,7 @@
 
 from gensim.sklearn_integration.sklearn_wrapper_gensim_ldamodel import SklearnWrapperLdaModel
 from gensim.sklearn_integration.sklearn_wrapper_gensim_lsimodel import SklearnWrapperLsiModel
+from gensim.sklearn_integration.sklearn_wrapper_gensim_rpmodel import SklearnWrapperRpModel
 from gensim.corpora import Dictionary
 from gensim import matutils
 
@@ -192,5 +193,27 @@ def testSetGetParams(self):
             self.assertEqual(model_params[key], param_dict[key])
 
 
+class TestSklearnRpModelWrapper(unittest.TestCase):
+    def setUp(self):
+        numpy.random.seed(13)
+        self.model = SklearnWrapperRpModel(corpus, num_topics=2)
+        self.model.fit(corpus)
+
+    def testTransform(self):
+        # transform one document
+        doc = list(self.model.corpus)[0]
+        transformed_doc = self.model.transform(doc)
+        vec = matutils.sparse2full(transformed_doc, 2)  # convert to dense vector, for easier equality tests
+
+        expected_vec = numpy.array([-0.70710677, 0.70710677])
+        self.assertTrue(numpy.allclose(vec, expected_vec))  # transformed entries must be equal up to sign
+
+    def testSetGetParams(self):
+        # updating only one param
+        self.model.set_params(num_topics=3)
+        model_params = self.model.get_params()
+        self.assertEqual(model_params["num_topics"], 3)
+
+
 if __name__ == '__main__':
     unittest.main()