Merge pull request #42 from lensacom/dev

kszucs · kszucs · commit 6ca0df32e272 · 2015-06-24T08:22:46.000+02:00
Fixed dict vectorizer heck_rdd; DictRDD transform dtype issue
diff --git a/README.rst b/README.rst
@@ -1,7 +1,7 @@
 Sparkit-learn
 =============
 
-|Build Status| |PyPi|
+|Build Status| |PyPi| |Gitter|
 
 **PySpark + Scikit-learn = Sparkit-learn**
 
@@ -448,3 +448,7 @@ Special thanks
    :target: https://travis-ci.org/lensacom/sparkit-learn
 .. |PyPi| image:: https://img.shields.io/pypi/v/sparkit-learn.svg
    :target: https://pypi.python.org/pypi/sparkit-learn
+.. |Gitter| image:: https://badges.gitter.im/Join%20Chat.svg
+   :alt: Join the chat at https://gitter.im/lensacom/sparkit-learn
+   :target: https://gitter.im/lensacom/sparkit-learn?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge
+
diff --git a/setup.py b/setup.py
@@ -2,8 +2,7 @@
 
 import sys
 
-from setuptools import setup
-from setuptools import find_packages
+from setuptools import find_packages, setup
 
 
 def is_numpy_installed():
@@ -17,7 +16,7 @@ def is_numpy_installed():
 def setup_package():
     metadata = dict(
         name='sparkit-learn',
-        version="0.2.4",
+        version='0.2.5',
         description='Scikit-learn on PySpark',
         author='Krisztian Szucs, Andras Fulop',
         author_email='krisztian.szucs@lensa.com, andras.fulop@lensa.com',
@@ -33,7 +32,8 @@ def setup_package():
         if is_numpy_installed() is False:
             raise ImportError("Numerical Python (NumPy) is not installed.\n"
                               "sparkit-learn requires NumPy.\n"
-                              "Installation instructions are available on scikit-learn website: "
+                              "Installation instructions are available on "
+                              "scikit-learn website: "
                               "http://scikit-learn.org/stable/install.html\n")
 
     setup(**metadata)
diff --git a/splearn/feature_extraction/dict_vectorizer.py b/splearn/feature_extraction/dict_vectorizer.py
@@ -2,13 +2,13 @@
 
 import numpy as np
 import scipy.sparse as sp
-from sklearn.feature_extraction import DictVectorizer
+from pyspark import AccumulatorParam
 from sklearn.externals import six
+from sklearn.feature_extraction import DictVectorizer
 
+from ..base import SparkBroadcasterMixin
 from ..rdd import DictRDD
-from pyspark import AccumulatorParam
 from ..utils.validation import check_rdd
-from ..base import SparkBroadcasterMixin
 
 
 class SparkDictVectorizer(DictVectorizer, SparkBroadcasterMixin):
@@ -87,7 +87,6 @@ def fit(self, Z):
         self
         """
         X = Z[:, 'X'] if isinstance(Z, DictRDD) else Z
-        check_rdd(X, (np.ndarray,))
 
         """Create vocabulary
         """
@@ -142,9 +141,6 @@ def transform(self, Z):
         Z : transformed, containing {array, sparse matrix}
             Feature vectors; always 2-d.
         """
-        X = Z[:, 'X'] if isinstance(Z, DictRDD) else Z
-        check_rdd(X, (np.ndarray, sp.spmatrix))
-
         mapper = self.broadcast(super(SparkDictVectorizer, self).transform,
                                 Z.context)
         dtype = sp.spmatrix if self.sparse else np.ndarray
diff --git a/splearn/feature_extraction/tests/test_dict_vectorizer.py b/splearn/feature_extraction/tests/test_dict_vectorizer.py
@@ -1,5 +1,5 @@
-import scipy.sparse as sp
 import numpy as np
+import scipy.sparse as sp
 from sklearn.feature_extraction import DictVectorizer
 from splearn.feature_extraction import SparkDictVectorizer
 from splearn.rdd import ArrayRDD
diff --git a/splearn/feature_extraction/tests/test_text.py b/splearn/feature_extraction/tests/test_text.py
@@ -1,13 +1,14 @@
-import scipy.sparse as sp
 import numpy as np
+import scipy.sparse as sp
 from sklearn.feature_extraction.text import (CountVectorizer,
                                              HashingVectorizer,
                                              TfidfTransformer)
 from splearn.feature_extraction.text import (SparkCountVectorizer,
                                              SparkHashingVectorizer,
                                              SparkTfidfTransformer)
 from splearn.utils.testing import (SplearnTestCase, assert_array_almost_equal,
-                                   assert_array_equal, assert_equal, assert_true)
+                                   assert_array_equal, assert_equal,
+                                   assert_true)
 from splearn.utils.validation import check_rdd_dtype
 
 
diff --git a/splearn/feature_selection/tests/test_variance_threshold.py b/splearn/feature_selection/tests/test_variance_threshold.py
@@ -80,6 +80,6 @@ def test_same_transform_with_treshold(self):
                                   result_dist.toarray())
 
         result_dist = dist.fit_transform(Z_rdd)[:, 'X']
-        assert_true(check_rdd_dtype(result_dist, (sp.spmatrix,)))        
+        assert_true(check_rdd_dtype(result_dist, (sp.spmatrix,)))
         assert_array_almost_equal(result_local.toarray(),
                                   result_dist.toarray())
diff --git a/splearn/feature_selection/variance_threshold.py b/splearn/feature_selection/variance_threshold.py
@@ -5,8 +5,8 @@
 from sklearn.utils.sparsefuncs import mean_variance_axis
 
 from ..rdd import DictRDD
-from .base import SparkSelectorMixin
 from ..utils.validation import check_rdd
+from .base import SparkSelectorMixin
 
 
 class SparkVarianceThreshold(VarianceThreshold, SparkSelectorMixin):
diff --git a/splearn/linear_model/base.py b/splearn/linear_model/base.py
@@ -1,8 +1,7 @@
 # encoding: utf-8
 
-import scipy.sparse as sp
 import numpy as np
-
+import scipy.sparse as sp
 from sklearn.base import copy
 from sklearn.linear_model.base import LinearRegression
 
diff --git a/splearn/linear_model/logistic.py b/splearn/linear_model/logistic.py
@@ -4,8 +4,8 @@
 import scipy.sparse as sp
 from sklearn.linear_model import LogisticRegression
 
-from .base import SparkLinearModelMixin
 from ..utils.validation import check_rdd
+from .base import SparkLinearModelMixin
 
 
 class SparkLogisticRegression(LogisticRegression, SparkLinearModelMixin):
diff --git a/splearn/linear_model/stochastic_gradient.py b/splearn/linear_model/stochastic_gradient.py
@@ -4,8 +4,8 @@
 import scipy.sparse as sp
 from sklearn.linear_model import SGDClassifier
 
-from .base import SparkLinearModelMixin
 from ..utils.validation import check_rdd
+from .base import SparkLinearModelMixin
 
 
 class SparkSGDClassifier(SGDClassifier, SparkLinearModelMixin):
diff --git a/splearn/linear_model/tests/test_base.py b/splearn/linear_model/tests/test_base.py
@@ -1,7 +1,8 @@
 import numpy as np
 from sklearn.linear_model import LinearRegression
 from splearn.linear_model import SparkLinearRegression
-from splearn.utils.testing import SplearnTestCase, assert_array_almost_equal, assert_true
+from splearn.utils.testing import (SplearnTestCase, assert_array_almost_equal,
+                                   assert_true)
 from splearn.utils.validation import check_rdd_dtype
 
 
diff --git a/splearn/rdd.py b/splearn/rdd.py
@@ -621,7 +621,7 @@ def zipper(a_b):
             raise ValueError("Columns and dtype lengths must be equal!")
 
         self.columns = tuple(columns)
-        super(DictRDD, self).__init__(rdd, bsize, dtype, noblock)
+        super(DictRDD, self).__init__(rdd, bsize, tuple(dtype), noblock)
 
     def _block(self, rdd, bsize, dtype):
         """Execute the blocking process on the given rdd.
@@ -739,18 +739,14 @@ def transform(self, fn, column=None, dtype=None):
         if column is None:
             indices = list(range(len(self.columns)))
         else:
-            # TODO: find a better way
-            #if not hasattr(column, '__iter__'):
             if not type(column) in (list, tuple):
                 column = [column]
             indices = [self.columns.index(c) for c in column]
 
         if dtype is not None:
-            # TODO: find a better way!
-            #if not hasattr(dtype, '__iter__'):
             if not type(dtype) in (list, tuple):
                 dtype = [dtype]
-            dtypes = [dtype[i] if i in indices else t
+            dtypes = [dtype[indices.index(i)] if i in indices else t
                       for i, t in enumerate(self.dtype)]
 
         def mapper(values):
diff --git a/splearn/svm/classes.py b/splearn/svm/classes.py
@@ -4,6 +4,7 @@
 import scipy.sparse as sp
 from sklearn.svm import LinearSVC
 from splearn.linear_model.base import SparkLinearModelMixin
+
 from ..utils.validation import check_rdd
 
 
diff --git a/splearn/tests/test_naive_bayes.py b/splearn/tests/test_naive_bayes.py
@@ -1,7 +1,8 @@
 import numpy as np
 from sklearn.naive_bayes import GaussianNB, MultinomialNB
 from splearn.naive_bayes import SparkGaussianNB, SparkMultinomialNB
-from splearn.utils.testing import SplearnTestCase, assert_array_almost_equal, assert_true
+from splearn.utils.testing import (SplearnTestCase, assert_array_almost_equal,
+                                   assert_true)
 from splearn.utils.validation import check_rdd_dtype
 
 
@@ -32,5 +33,5 @@ def test_same_prediction(self):
         y_local = local.fit(X, y).predict(X)
         y_dist = dist.fit(Z, classes=np.unique(y)).predict(Z[:, 'X'])
 
-        assert_true(check_rdd_dtype(y_dist, (np.ndarray,)))        
+        assert_true(check_rdd_dtype(y_dist, (np.ndarray,)))
         assert_array_almost_equal(y_local, y_dist.toarray())
diff --git a/splearn/tests/test_rdd.py b/splearn/tests/test_rdd.py
@@ -1,13 +1,14 @@
 import numpy as np
 import scipy.sparse as sp
 from pyspark import RDD
-from splearn.rdd import ArrayRDD, SparseRDD, BlockRDD, DictRDD, block
+from splearn.rdd import ArrayRDD, BlockRDD, DictRDD, SparseRDD, block
 from splearn.utils.testing import (SplearnTestCase, assert_almost_equal,
                                    assert_array_almost_equal,
                                    assert_array_equal, assert_equal,
                                    assert_is_instance,
                                    assert_multiple_tuples_equal, assert_raises,
                                    assert_true, assert_tuple_equal)
+from splearn.utils.validation import check_rdd_dtype
 
 
 class TestBlocking(SplearnTestCase):
@@ -633,6 +634,24 @@ def test_initialization(self):
         assert_is_instance(DictRDD(rdd, bsize=None), DictRDD)
         assert_is_instance(DictRDD(rdd), BlockRDD)
 
+    def test_creation_from_zipped_rdd(self):
+        x = np.arange(80).reshape((40, 2))
+        y = range(40)
+        x_rdd = self.sc.parallelize(x, 4)
+        y_rdd = self.sc.parallelize(y, 4)
+        zipped_rdd = x_rdd.zip(y_rdd)
+
+        expected = (np.arange(20).reshape(10, 2), tuple(range(10)))
+
+        rdd = DictRDD(zipped_rdd)
+        assert_tuple_equal(rdd.first(), expected)
+        rdd = DictRDD(zipped_rdd, columns=('x', 'y'))
+        assert_tuple_equal(rdd.first(), expected)
+        rdd = DictRDD(zipped_rdd, dtype=(np.ndarray, list))
+        first = rdd.first()
+        assert_tuple_equal(first, expected)
+        assert_is_instance(first[1], list)
+
     def test_creation_from_rdds(self):
         x = np.arange(80).reshape((40, 2))
         y = np.arange(40)
@@ -676,6 +695,28 @@ def test_creation_from_blocked_rdds(self):
         assert_tuple_equal(first, expected)
         assert_is_instance(first[2], list)
 
+    def test_auto_dtype(self):
+        x = np.arange(80).reshape((40, 2))
+        y = tuple(range(40))
+        z = list(range(40))
+        x_rdd = self.sc.parallelize(x, 4)
+        y_rdd = self.sc.parallelize(y, 4)
+        z_rdd = self.sc.parallelize(z, 4)
+
+        expected = (np.arange(20).reshape(10, 2), tuple(range(10)),
+                    list(range(10)))
+
+        rdd = DictRDD([x_rdd, y_rdd, z_rdd])
+        assert_tuple_equal(rdd.first(), expected)
+        assert_equal(rdd.dtype, (np.ndarray, tuple, tuple))
+        assert_true(check_rdd_dtype(rdd, {0: np.ndarray, 1: tuple, 2: tuple}))
+
+        rdd = DictRDD([x_rdd, y_rdd, z_rdd], columns=('x', 'y', 'z'))
+        assert_tuple_equal(rdd.first(), expected)
+        assert_equal(rdd.dtype, (np.ndarray, tuple, tuple))
+        assert_true(check_rdd_dtype(rdd, {'x': np.ndarray, 'y': tuple,
+                                          'z': tuple}))
+
     def test_get_single_tuple(self):
         x, y = np.arange(80).reshape((40, 2)), np.arange(40)
         x_rdd = self.sc.parallelize(x, 2)
@@ -772,16 +813,16 @@ def test_transform(self):
         X = DictRDD(rdd1.zip(rdd2), bsize=5)
 
         X1 = [(x[0], x[1] ** 2) for x in X.collect()]
-        X2 = X.transform(lambda a, b: (a, b ** 2)).collect()
-        assert_multiple_tuples_equal(X1, X2)
+        X2 = X.transform(lambda a, b: (a, b ** 2))
+        assert_multiple_tuples_equal(X1, X2.collect())
 
         X1 = [(x[0], x[1] ** 2) for x in X.collect()]
-        X2 = X.transform(lambda x: x ** 2, column=1).collect()
-        assert_multiple_tuples_equal(X1, X2)
+        X2 = X.transform(lambda x: x ** 2, column=1)
+        assert_multiple_tuples_equal(X1, X2.collect())
 
         X1 = [(x[0] ** 2, x[1]) for x in X.collect()]
-        X2 = X.transform(lambda x: x ** 2, column=0).collect()
-        assert_multiple_tuples_equal(X1, X2)
+        X2 = X.transform(lambda x: x ** 2, column=0)
+        assert_multiple_tuples_equal(X1, X2.collect())
 
         X1 = [(x[0] ** 2, x[1] ** 0.5) for x in X.collect()]
         X2 = X.transform(lambda a, b: (a ** 2, b ** 0.5), column=[0, 1])
@@ -790,3 +831,32 @@ def test_transform(self):
         X1 = [(x[0] ** 2, x[1] ** 0.5) for x in X.collect()]
         X2 = X.transform(lambda b, a: (b ** 0.5, a ** 2), column=[1, 0])
         assert_multiple_tuples_equal(X1, X2.collect())
+
+    def test_transform_with_dtype(self):
+        data1 = np.arange(400).reshape((100, 4))
+        data2 = np.arange(200).reshape((100, 2))
+        rdd1 = self.sc.parallelize(data1, 4)
+        rdd2 = self.sc.parallelize(data2, 4)
+
+        X = DictRDD(rdd1.zip(rdd2), bsize=5)
+
+        X2 = X.transform(lambda x: x ** 2, column=0)
+        assert_equal(X2.dtype, (np.ndarray, np.ndarray))
+
+        X2 = X.transform(lambda x: tuple((x ** 2).tolist()), column=0,
+                         dtype=tuple)
+        assert_equal(X2.dtype, (tuple, np.ndarray))
+        assert_true(check_rdd_dtype(X2, {0: tuple, 1: np.ndarray}))
+
+        X2 = X.transform(lambda x: x ** 2, column=1, dtype=list)
+        assert_equal(X2.dtype, (np.ndarray, list))
+        assert_true(check_rdd_dtype(X2, {0: np.ndarray, 1: list}))
+
+        X2 = X.transform(lambda a, b: (a ** 2, (b ** 0.5).tolist()),
+                         column=[0, 1], dtype=(np.ndarray, list))
+        assert_true(check_rdd_dtype(X2, {0: np.ndarray, 1: list}))
+
+        X2 = X.transform(lambda b, a: ((b ** 0.5).tolist(), a ** 2),
+                         column=[1, 0], dtype=(list, np.ndarray))
+        assert_equal(X2.dtype, (np.ndarray, list))
+        assert_true(check_rdd_dtype(X2, {0: np.ndarray, 1: list}))
diff --git a/splearn/utils/tests/test_utils.py b/splearn/utils/tests/test_utils.py
@@ -1,9 +1,10 @@
 import numpy as np
 import scipy.sparse as sp
-from splearn.utils.validation import check_rdd_dtype, check_rdd
+from sklearn.utils.testing import assert_false, assert_raises, assert_true
 from splearn.rdd import ArrayRDD, DictRDD, SparseRDD
-from sklearn.utils.testing import assert_false, assert_true, assert_raises
 from splearn.utils.testing import SplearnTestCase
+from splearn.utils.validation import check_rdd, check_rdd_dtype
+
 
 class TestUtilities(SplearnTestCase):
 
@@ -53,4 +54,3 @@ def test_check_rdd(self):
         assert_raises(TypeError, check_rdd, (dict_rdd, {'X': (array, spmat)}))
         assert_raises(TypeError, check_rdd, (dict_rdd, (tuple,)))
         assert_raises(TypeError, check_rdd, (np.arange(20), (array,)))
-
diff --git a/splearn/utils/validation.py b/splearn/utils/validation.py
@@ -1,4 +1,4 @@
-from splearn.rdd import DictRDD, BlockRDD
+from splearn.rdd import BlockRDD, DictRDD
 
 
 def check_rdd_dtype(rdd, expected_dtype):
@@ -38,6 +38,7 @@ def check_rdd_dtype(rdd, expected_dtype):
 
     return rdd.dtype in expected_dtype
 
+
 def check_rdd(rdd, expected_dtype):
     """Wrapper function to check_rdd_dtype. Raises TypeError in case of dtype
     mismatch.
