2015-01-28 20:14:23 -05:00
|
|
|
#
|
|
|
|
# Licensed to the Apache Software Foundation (ASF) under one or more
|
|
|
|
# contributor license agreements. See the NOTICE file distributed with
|
|
|
|
# this work for additional information regarding copyright ownership.
|
|
|
|
# The ASF licenses this file to You under the Apache License, Version 2.0
|
|
|
|
# (the "License"); you may not use this file except in compliance with
|
|
|
|
# the License. You may obtain a copy of the License at
|
|
|
|
#
|
|
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
#
|
|
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
# See the License for the specific language governing permissions and
|
|
|
|
# limitations under the License.
|
|
|
|
#
|
|
|
|
|
|
|
|
"""
|
|
|
|
Unit tests for Spark ML Python APIs.
|
|
|
|
"""
|
2016-03-23 14:20:44 -04:00
|
|
|
import array
|
2015-01-28 20:14:23 -05:00
|
|
|
import sys
|
2016-03-23 14:20:44 -04:00
|
|
|
if sys.version > '3':
|
|
|
|
xrange = range
|
|
|
|
|
2015-10-22 18:27:11 -04:00
|
|
|
try:
|
|
|
|
import xmlrunner
|
|
|
|
except ImportError:
|
|
|
|
xmlrunner = None
|
2015-01-28 20:14:23 -05:00
|
|
|
|
|
|
|
if sys.version_info[:2] <= (2, 6):
|
|
|
|
try:
|
|
|
|
import unittest2 as unittest
|
|
|
|
except ImportError:
|
|
|
|
sys.stderr.write('Please install unittest2 to test with Python 2.6 or earlier')
|
|
|
|
sys.exit(1)
|
|
|
|
else:
|
|
|
|
import unittest
|
|
|
|
|
2016-01-29 12:22:24 -05:00
|
|
|
from shutil import rmtree
|
|
|
|
import tempfile
|
2016-03-23 14:20:44 -04:00
|
|
|
import numpy as np
|
2016-01-29 12:22:24 -05:00
|
|
|
|
2016-03-16 16:49:40 -04:00
|
|
|
from pyspark.ml import Estimator, Model, Pipeline, PipelineModel, Transformer
|
2016-01-06 13:43:03 -05:00
|
|
|
from pyspark.ml.classification import LogisticRegression
|
2016-02-11 18:53:45 -05:00
|
|
|
from pyspark.ml.clustering import KMeans
|
2015-08-28 02:59:30 -04:00
|
|
|
from pyspark.ml.evaluation import RegressionEvaluator
|
2016-01-29 12:22:24 -05:00
|
|
|
from pyspark.ml.feature import *
|
2016-03-23 14:20:44 -04:00
|
|
|
from pyspark.ml.param import Param, Params, TypeConverters
|
2015-05-20 18:16:12 -04:00
|
|
|
from pyspark.ml.param.shared import HasMaxIter, HasInputCol, HasSeed
|
2016-01-29 12:22:24 -05:00
|
|
|
from pyspark.ml.regression import LinearRegression
|
2016-03-03 12:50:05 -05:00
|
|
|
from pyspark.ml.tuning import *
|
2016-01-29 12:22:24 -05:00
|
|
|
from pyspark.ml.util import keyword_only
|
2016-03-22 15:11:23 -04:00
|
|
|
from pyspark.ml.wrapper import JavaWrapper
|
2016-03-23 14:20:44 -04:00
|
|
|
from pyspark.mllib.linalg import DenseVector, SparseVector
|
2016-01-29 12:22:24 -05:00
|
|
|
from pyspark.sql import DataFrame, SQLContext, Row
|
|
|
|
from pyspark.sql.functions import rand
|
|
|
|
from pyspark.tests import ReusedPySparkTestCase as PySparkTestCase
|
2015-01-28 20:14:23 -05:00
|
|
|
|
|
|
|
|
|
|
|
class MockDataset(DataFrame):
|
|
|
|
|
|
|
|
def __init__(self):
|
|
|
|
self.index = 0
|
|
|
|
|
|
|
|
|
2015-05-18 15:02:18 -04:00
|
|
|
class HasFake(Params):
|
|
|
|
|
|
|
|
def __init__(self):
|
|
|
|
super(HasFake, self).__init__()
|
|
|
|
self.fake = Param(self, "fake", "fake param")
|
|
|
|
|
|
|
|
def getFake(self):
|
|
|
|
return self.getOrDefault(self.fake)
|
|
|
|
|
|
|
|
|
|
|
|
class MockTransformer(Transformer, HasFake):
|
2015-01-28 20:14:23 -05:00
|
|
|
|
|
|
|
def __init__(self):
|
|
|
|
super(MockTransformer, self).__init__()
|
|
|
|
self.dataset_index = None
|
|
|
|
|
2015-05-18 15:02:18 -04:00
|
|
|
def _transform(self, dataset):
|
2015-01-28 20:14:23 -05:00
|
|
|
self.dataset_index = dataset.index
|
|
|
|
dataset.index += 1
|
|
|
|
return dataset
|
|
|
|
|
|
|
|
|
2015-05-18 15:02:18 -04:00
|
|
|
class MockEstimator(Estimator, HasFake):
|
2015-01-28 20:14:23 -05:00
|
|
|
|
|
|
|
def __init__(self):
|
|
|
|
super(MockEstimator, self).__init__()
|
|
|
|
self.dataset_index = None
|
|
|
|
|
2015-05-18 15:02:18 -04:00
|
|
|
def _fit(self, dataset):
|
2015-01-28 20:14:23 -05:00
|
|
|
self.dataset_index = dataset.index
|
|
|
|
model = MockModel()
|
2015-05-18 15:02:18 -04:00
|
|
|
self._copyValues(model)
|
2015-01-28 20:14:23 -05:00
|
|
|
return model
|
|
|
|
|
|
|
|
|
2015-05-18 15:02:18 -04:00
|
|
|
class MockModel(MockTransformer, Model, HasFake):
|
|
|
|
pass
|
2015-01-28 20:14:23 -05:00
|
|
|
|
|
|
|
|
2016-01-06 13:43:03 -05:00
|
|
|
class ParamTypeConversionTests(PySparkTestCase):
|
|
|
|
"""
|
|
|
|
Test that param type conversion happens.
|
|
|
|
"""
|
|
|
|
|
2016-03-23 14:20:44 -04:00
|
|
|
def test_int(self):
|
|
|
|
lr = LogisticRegression(maxIter=5.0)
|
|
|
|
self.assertEqual(lr.getMaxIter(), 5)
|
|
|
|
self.assertTrue(type(lr.getMaxIter()) == int)
|
|
|
|
self.assertRaises(TypeError, lambda: LogisticRegression(maxIter="notAnInt"))
|
|
|
|
self.assertRaises(TypeError, lambda: LogisticRegression(maxIter=5.1))
|
|
|
|
|
|
|
|
def test_float(self):
|
|
|
|
lr = LogisticRegression(tol=1)
|
|
|
|
self.assertEqual(lr.getTol(), 1.0)
|
|
|
|
self.assertTrue(type(lr.getTol()) == float)
|
|
|
|
self.assertRaises(TypeError, lambda: LogisticRegression(tol="notAFloat"))
|
|
|
|
|
|
|
|
def test_vector(self):
|
|
|
|
ewp = ElementwiseProduct(scalingVec=[1, 3])
|
|
|
|
self.assertEqual(ewp.getScalingVec(), DenseVector([1.0, 3.0]))
|
|
|
|
ewp = ElementwiseProduct(scalingVec=np.array([1.2, 3.4]))
|
|
|
|
self.assertEqual(ewp.getScalingVec(), DenseVector([1.2, 3.4]))
|
|
|
|
self.assertRaises(TypeError, lambda: ElementwiseProduct(scalingVec=["a", "b"]))
|
|
|
|
|
|
|
|
def test_list(self):
|
|
|
|
l = [0, 1]
|
|
|
|
for lst_like in [l, np.array(l), DenseVector(l), SparseVector(len(l), range(len(l)), l),
|
|
|
|
array.array('l', l), xrange(2), tuple(l)]:
|
|
|
|
converted = TypeConverters.toList(lst_like)
|
|
|
|
self.assertEqual(type(converted), list)
|
|
|
|
self.assertListEqual(converted, l)
|
|
|
|
|
|
|
|
def test_list_int(self):
|
|
|
|
for indices in [[1.0, 2.0], np.array([1.0, 2.0]), DenseVector([1.0, 2.0]),
|
|
|
|
SparseVector(2, {0: 1.0, 1: 2.0}), xrange(1, 3), (1.0, 2.0),
|
|
|
|
array.array('d', [1.0, 2.0])]:
|
|
|
|
vs = VectorSlicer(indices=indices)
|
|
|
|
self.assertListEqual(vs.getIndices(), [1, 2])
|
|
|
|
self.assertTrue(all([type(v) == int for v in vs.getIndices()]))
|
|
|
|
self.assertRaises(TypeError, lambda: VectorSlicer(indices=["a", "b"]))
|
|
|
|
|
|
|
|
def test_list_float(self):
|
|
|
|
b = Bucketizer(splits=[1, 4])
|
|
|
|
self.assertEqual(b.getSplits(), [1.0, 4.0])
|
|
|
|
self.assertTrue(all([type(v) == float for v in b.getSplits()]))
|
|
|
|
self.assertRaises(TypeError, lambda: Bucketizer(splits=["a", 1.0]))
|
|
|
|
|
|
|
|
def test_list_string(self):
|
|
|
|
for labels in [np.array(['a', u'b']), ['a', u'b'], np.array(['a', 'b'])]:
|
|
|
|
idx_to_string = IndexToString(labels=labels)
|
|
|
|
self.assertListEqual(idx_to_string.getLabels(), ['a', 'b'])
|
|
|
|
self.assertRaises(TypeError, lambda: IndexToString(labels=['a', 2]))
|
|
|
|
|
|
|
|
def test_string(self):
|
|
|
|
lr = LogisticRegression()
|
|
|
|
for col in ['features', u'features', np.str_('features')]:
|
|
|
|
lr.setFeaturesCol(col)
|
|
|
|
self.assertEqual(lr.getFeaturesCol(), 'features')
|
|
|
|
self.assertRaises(TypeError, lambda: LogisticRegression(featuresCol=2.3))
|
|
|
|
|
|
|
|
def test_bool(self):
|
|
|
|
self.assertRaises(TypeError, lambda: LogisticRegression(fitIntercept=1))
|
|
|
|
self.assertRaises(TypeError, lambda: LogisticRegression(fitIntercept="false"))
|
2016-01-06 13:43:03 -05:00
|
|
|
|
|
|
|
|
2015-01-28 20:14:23 -05:00
|
|
|
class PipelineTests(PySparkTestCase):
|
|
|
|
|
|
|
|
def test_pipeline(self):
|
|
|
|
dataset = MockDataset()
|
|
|
|
estimator0 = MockEstimator()
|
|
|
|
transformer1 = MockTransformer()
|
|
|
|
estimator2 = MockEstimator()
|
|
|
|
transformer3 = MockTransformer()
|
2015-05-18 15:02:18 -04:00
|
|
|
pipeline = Pipeline(stages=[estimator0, transformer1, estimator2, transformer3])
|
2015-01-28 20:14:23 -05:00
|
|
|
pipeline_model = pipeline.fit(dataset, {estimator0.fake: 0, transformer1.fake: 1})
|
2015-05-18 15:02:18 -04:00
|
|
|
model0, transformer1, model2, transformer3 = pipeline_model.stages
|
2015-01-28 20:14:23 -05:00
|
|
|
self.assertEqual(0, model0.dataset_index)
|
2015-05-18 15:02:18 -04:00
|
|
|
self.assertEqual(0, model0.getFake())
|
2015-01-28 20:14:23 -05:00
|
|
|
self.assertEqual(1, transformer1.dataset_index)
|
2015-05-18 15:02:18 -04:00
|
|
|
self.assertEqual(1, transformer1.getFake())
|
|
|
|
self.assertEqual(2, dataset.index)
|
|
|
|
self.assertIsNone(model2.dataset_index, "The last model shouldn't be called in fit.")
|
|
|
|
self.assertIsNone(transformer3.dataset_index,
|
|
|
|
"The last transformer shouldn't be called in fit.")
|
2015-01-28 20:14:23 -05:00
|
|
|
dataset = pipeline_model.transform(dataset)
|
|
|
|
self.assertEqual(2, model0.dataset_index)
|
|
|
|
self.assertEqual(3, transformer1.dataset_index)
|
|
|
|
self.assertEqual(4, model2.dataset_index)
|
|
|
|
self.assertEqual(5, transformer3.dataset_index)
|
|
|
|
self.assertEqual(6, dataset.index)
|
|
|
|
|
|
|
|
|
2015-05-20 18:16:12 -04:00
|
|
|
class TestParams(HasMaxIter, HasInputCol, HasSeed):
|
2015-04-16 02:49:42 -04:00
|
|
|
"""
|
2015-05-20 18:16:12 -04:00
|
|
|
A subclass of Params mixed with HasMaxIter, HasInputCol and HasSeed.
|
2015-04-16 02:49:42 -04:00
|
|
|
"""
|
2015-05-20 18:16:12 -04:00
|
|
|
@keyword_only
|
|
|
|
def __init__(self, seed=None):
|
2015-04-16 02:49:42 -04:00
|
|
|
super(TestParams, self).__init__()
|
|
|
|
self._setDefault(maxIter=10)
|
2015-05-20 18:16:12 -04:00
|
|
|
kwargs = self.__init__._input_kwargs
|
|
|
|
self.setParams(**kwargs)
|
|
|
|
|
|
|
|
@keyword_only
|
|
|
|
def setParams(self, seed=None):
|
|
|
|
"""
|
|
|
|
setParams(self, seed=None)
|
|
|
|
Sets params for this test.
|
|
|
|
"""
|
|
|
|
kwargs = self.setParams._input_kwargs
|
|
|
|
return self._set(**kwargs)
|
|
|
|
|
|
|
|
|
|
|
|
class OtherTestParams(HasMaxIter, HasInputCol, HasSeed):
|
|
|
|
"""
|
|
|
|
A subclass of Params mixed with HasMaxIter, HasInputCol and HasSeed.
|
|
|
|
"""
|
|
|
|
@keyword_only
|
|
|
|
def __init__(self, seed=None):
|
|
|
|
super(OtherTestParams, self).__init__()
|
|
|
|
self._setDefault(maxIter=10)
|
|
|
|
kwargs = self.__init__._input_kwargs
|
|
|
|
self.setParams(**kwargs)
|
|
|
|
|
|
|
|
@keyword_only
|
|
|
|
def setParams(self, seed=None):
|
|
|
|
"""
|
|
|
|
setParams(self, seed=None)
|
|
|
|
Sets params for this test.
|
|
|
|
"""
|
|
|
|
kwargs = self.setParams._input_kwargs
|
|
|
|
return self._set(**kwargs)
|
2015-04-16 02:49:42 -04:00
|
|
|
|
|
|
|
|
|
|
|
class ParamTests(PySparkTestCase):
|
|
|
|
|
2016-01-26 18:53:48 -05:00
|
|
|
def test_copy_new_parent(self):
|
|
|
|
testParams = TestParams()
|
|
|
|
# Copying an instantiated param should fail
|
|
|
|
with self.assertRaises(ValueError):
|
|
|
|
testParams.maxIter._copy_new_parent(testParams)
|
|
|
|
# Copying a dummy param should succeed
|
|
|
|
TestParams.maxIter._copy_new_parent(testParams)
|
|
|
|
maxIter = testParams.maxIter
|
|
|
|
self.assertEqual(maxIter.name, "maxIter")
|
|
|
|
self.assertEqual(maxIter.doc, "max number of iterations (>= 0).")
|
|
|
|
self.assertTrue(maxIter.parent == testParams.uid)
|
|
|
|
|
2015-04-16 02:49:42 -04:00
|
|
|
def test_param(self):
|
|
|
|
testParams = TestParams()
|
|
|
|
maxIter = testParams.maxIter
|
|
|
|
self.assertEqual(maxIter.name, "maxIter")
|
2015-10-20 19:51:32 -04:00
|
|
|
self.assertEqual(maxIter.doc, "max number of iterations (>= 0).")
|
2015-05-18 15:02:18 -04:00
|
|
|
self.assertTrue(maxIter.parent == testParams.uid)
|
2015-04-16 02:49:42 -04:00
|
|
|
|
2016-02-11 19:42:44 -05:00
|
|
|
def test_hasparam(self):
|
|
|
|
testParams = TestParams()
|
|
|
|
self.assertTrue(all([testParams.hasParam(p.name) for p in testParams.params]))
|
|
|
|
self.assertFalse(testParams.hasParam("notAParameter"))
|
|
|
|
|
2015-04-16 02:49:42 -04:00
|
|
|
def test_params(self):
|
|
|
|
testParams = TestParams()
|
|
|
|
maxIter = testParams.maxIter
|
|
|
|
inputCol = testParams.inputCol
|
2015-05-20 18:16:12 -04:00
|
|
|
seed = testParams.seed
|
2015-04-16 02:49:42 -04:00
|
|
|
|
|
|
|
params = testParams.params
|
2015-05-20 18:16:12 -04:00
|
|
|
self.assertEqual(params, [inputCol, maxIter, seed])
|
2015-04-16 02:49:42 -04:00
|
|
|
|
2016-02-11 19:42:44 -05:00
|
|
|
self.assertTrue(testParams.hasParam(maxIter.name))
|
2015-04-16 02:49:42 -04:00
|
|
|
self.assertTrue(testParams.hasDefault(maxIter))
|
|
|
|
self.assertFalse(testParams.isSet(maxIter))
|
|
|
|
self.assertTrue(testParams.isDefined(maxIter))
|
|
|
|
self.assertEqual(testParams.getMaxIter(), 10)
|
|
|
|
testParams.setMaxIter(100)
|
|
|
|
self.assertTrue(testParams.isSet(maxIter))
|
2015-09-18 12:53:52 -04:00
|
|
|
self.assertEqual(testParams.getMaxIter(), 100)
|
2015-04-16 02:49:42 -04:00
|
|
|
|
2016-02-11 19:42:44 -05:00
|
|
|
self.assertTrue(testParams.hasParam(inputCol.name))
|
2015-04-16 02:49:42 -04:00
|
|
|
self.assertFalse(testParams.hasDefault(inputCol))
|
|
|
|
self.assertFalse(testParams.isSet(inputCol))
|
|
|
|
self.assertFalse(testParams.isDefined(inputCol))
|
|
|
|
with self.assertRaises(KeyError):
|
|
|
|
testParams.getInputCol()
|
|
|
|
|
2015-05-20 18:16:12 -04:00
|
|
|
# Since the default is normally random, set it to a known number for debug str
|
|
|
|
testParams._setDefault(seed=41)
|
|
|
|
testParams.setSeed(43)
|
|
|
|
|
2015-09-18 12:53:52 -04:00
|
|
|
self.assertEqual(
|
2015-04-16 02:49:42 -04:00
|
|
|
testParams.explainParams(),
|
2015-10-20 19:51:32 -04:00
|
|
|
"\n".join(["inputCol: input column name. (undefined)",
|
|
|
|
"maxIter: max number of iterations (>= 0). (default: 10, current: 100)",
|
|
|
|
"seed: random seed. (default: 41, current: 43)"]))
|
2015-05-20 18:16:12 -04:00
|
|
|
|
2016-02-11 18:53:45 -05:00
|
|
|
def test_kmeans_param(self):
|
|
|
|
algo = KMeans()
|
|
|
|
self.assertEqual(algo.getInitMode(), "k-means||")
|
|
|
|
algo.setK(10)
|
|
|
|
self.assertEqual(algo.getK(), 10)
|
|
|
|
algo.setInitSteps(10)
|
|
|
|
self.assertEqual(algo.getInitSteps(), 10)
|
|
|
|
|
2015-05-20 18:16:12 -04:00
|
|
|
def test_hasseed(self):
|
|
|
|
noSeedSpecd = TestParams()
|
|
|
|
withSeedSpecd = TestParams(seed=42)
|
|
|
|
other = OtherTestParams()
|
|
|
|
# Check that we no longer use 42 as the magic number
|
|
|
|
self.assertNotEqual(noSeedSpecd.getSeed(), 42)
|
|
|
|
origSeed = noSeedSpecd.getSeed()
|
|
|
|
# Check that we only compute the seed once
|
|
|
|
self.assertEqual(noSeedSpecd.getSeed(), origSeed)
|
|
|
|
# Check that a specified seed is honored
|
|
|
|
self.assertEqual(withSeedSpecd.getSeed(), 42)
|
|
|
|
# Check that a different class has a different seed
|
|
|
|
self.assertNotEqual(other.getSeed(), noSeedSpecd.getSeed())
|
2015-04-16 02:49:42 -04:00
|
|
|
|
2016-03-08 20:34:25 -05:00
|
|
|
def test_param_property_error(self):
|
|
|
|
param_store = HasThrowableProperty()
|
|
|
|
self.assertRaises(RuntimeError, lambda: param_store.test_property)
|
|
|
|
params = param_store.params # should not invoke the property 'test_property'
|
|
|
|
self.assertEqual(len(params), 1)
|
|
|
|
|
2015-04-16 02:49:42 -04:00
|
|
|
|
2015-05-18 15:02:18 -04:00
|
|
|
class FeatureTests(PySparkTestCase):
|
|
|
|
|
|
|
|
def test_binarizer(self):
|
|
|
|
b0 = Binarizer()
|
|
|
|
self.assertListEqual(b0.params, [b0.inputCol, b0.outputCol, b0.threshold])
|
|
|
|
self.assertTrue(all([~b0.isSet(p) for p in b0.params]))
|
|
|
|
self.assertTrue(b0.hasDefault(b0.threshold))
|
|
|
|
self.assertEqual(b0.getThreshold(), 0.0)
|
|
|
|
b0.setParams(inputCol="input", outputCol="output").setThreshold(1.0)
|
|
|
|
self.assertTrue(all([b0.isSet(p) for p in b0.params]))
|
|
|
|
self.assertEqual(b0.getThreshold(), 1.0)
|
|
|
|
self.assertEqual(b0.getInputCol(), "input")
|
|
|
|
self.assertEqual(b0.getOutputCol(), "output")
|
|
|
|
|
|
|
|
b0c = b0.copy({b0.threshold: 2.0})
|
|
|
|
self.assertEqual(b0c.uid, b0.uid)
|
|
|
|
self.assertListEqual(b0c.params, b0.params)
|
|
|
|
self.assertEqual(b0c.getThreshold(), 2.0)
|
|
|
|
|
|
|
|
b1 = Binarizer(threshold=2.0, inputCol="input", outputCol="output")
|
|
|
|
self.assertNotEqual(b1.uid, b0.uid)
|
|
|
|
self.assertEqual(b1.getThreshold(), 2.0)
|
|
|
|
self.assertEqual(b1.getInputCol(), "input")
|
|
|
|
self.assertEqual(b1.getOutputCol(), "output")
|
|
|
|
|
|
|
|
def test_idf(self):
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
dataset = sqlContext.createDataFrame([
|
|
|
|
(DenseVector([1.0, 2.0]),),
|
|
|
|
(DenseVector([0.0, 1.0]),),
|
|
|
|
(DenseVector([3.0, 0.2]),)], ["tf"])
|
|
|
|
idf0 = IDF(inputCol="tf")
|
|
|
|
self.assertListEqual(idf0.params, [idf0.inputCol, idf0.minDocFreq, idf0.outputCol])
|
|
|
|
idf0m = idf0.fit(dataset, {idf0.outputCol: "idf"})
|
|
|
|
self.assertEqual(idf0m.uid, idf0.uid,
|
|
|
|
"Model should inherit the UID from its parent estimator.")
|
|
|
|
output = idf0m.transform(dataset)
|
|
|
|
self.assertIsNotNone(output.head().idf)
|
|
|
|
|
2015-06-29 21:40:30 -04:00
|
|
|
def test_ngram(self):
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
dataset = sqlContext.createDataFrame([
|
2015-09-01 13:48:57 -04:00
|
|
|
Row(input=["a", "b", "c", "d", "e"])])
|
2015-06-29 21:40:30 -04:00
|
|
|
ngram0 = NGram(n=4, inputCol="input", outputCol="output")
|
|
|
|
self.assertEqual(ngram0.getN(), 4)
|
|
|
|
self.assertEqual(ngram0.getInputCol(), "input")
|
|
|
|
self.assertEqual(ngram0.getOutputCol(), "output")
|
|
|
|
transformedDF = ngram0.transform(dataset)
|
2015-09-18 12:53:52 -04:00
|
|
|
self.assertEqual(transformedDF.head().output, ["a b c d", "b c d e"])
|
2015-06-29 21:40:30 -04:00
|
|
|
|
2015-09-01 13:48:57 -04:00
|
|
|
def test_stopwordsremover(self):
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
dataset = sqlContext.createDataFrame([Row(input=["a", "panda"])])
|
|
|
|
stopWordRemover = StopWordsRemover(inputCol="input", outputCol="output")
|
|
|
|
# Default
|
2015-09-18 12:53:52 -04:00
|
|
|
self.assertEqual(stopWordRemover.getInputCol(), "input")
|
2015-09-01 13:48:57 -04:00
|
|
|
transformedDF = stopWordRemover.transform(dataset)
|
2015-09-18 12:53:52 -04:00
|
|
|
self.assertEqual(transformedDF.head().output, ["panda"])
|
2015-09-01 13:48:57 -04:00
|
|
|
# Custom
|
|
|
|
stopwords = ["panda"]
|
|
|
|
stopWordRemover.setStopWords(stopwords)
|
2015-09-18 12:53:52 -04:00
|
|
|
self.assertEqual(stopWordRemover.getInputCol(), "input")
|
|
|
|
self.assertEqual(stopWordRemover.getStopWords(), stopwords)
|
2015-09-01 13:48:57 -04:00
|
|
|
transformedDF = stopWordRemover.transform(dataset)
|
2015-09-18 12:53:52 -04:00
|
|
|
self.assertEqual(transformedDF.head().output, ["a"])
|
2015-09-01 13:48:57 -04:00
|
|
|
|
2015-05-18 15:02:18 -04:00
|
|
|
|
2015-08-28 02:59:30 -04:00
|
|
|
class HasInducedError(Params):
|
|
|
|
|
|
|
|
def __init__(self):
|
|
|
|
super(HasInducedError, self).__init__()
|
|
|
|
self.inducedError = Param(self, "inducedError",
|
|
|
|
"Uniformly-distributed error added to feature")
|
|
|
|
|
|
|
|
def getInducedError(self):
|
|
|
|
return self.getOrDefault(self.inducedError)
|
|
|
|
|
|
|
|
|
|
|
|
class InducedErrorModel(Model, HasInducedError):
|
|
|
|
|
|
|
|
def __init__(self):
|
|
|
|
super(InducedErrorModel, self).__init__()
|
|
|
|
|
|
|
|
def _transform(self, dataset):
|
|
|
|
return dataset.withColumn("prediction",
|
|
|
|
dataset.feature + (rand(0) * self.getInducedError()))
|
|
|
|
|
|
|
|
|
|
|
|
class InducedErrorEstimator(Estimator, HasInducedError):
|
|
|
|
|
|
|
|
def __init__(self, inducedError=1.0):
|
|
|
|
super(InducedErrorEstimator, self).__init__()
|
|
|
|
self._set(inducedError=inducedError)
|
|
|
|
|
|
|
|
def _fit(self, dataset):
|
|
|
|
model = InducedErrorModel()
|
|
|
|
self._copyValues(model)
|
|
|
|
return model
|
|
|
|
|
|
|
|
|
|
|
|
class CrossValidatorTests(PySparkTestCase):
|
|
|
|
|
|
|
|
def test_fit_minimize_metric(self):
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
dataset = sqlContext.createDataFrame([
|
|
|
|
(10, 10.0),
|
|
|
|
(50, 50.0),
|
|
|
|
(100, 100.0),
|
|
|
|
(500, 500.0)] * 10,
|
|
|
|
["feature", "label"])
|
|
|
|
|
|
|
|
iee = InducedErrorEstimator()
|
|
|
|
evaluator = RegressionEvaluator(metricName="rmse")
|
|
|
|
|
|
|
|
grid = (ParamGridBuilder()
|
|
|
|
.addGrid(iee.inducedError, [100.0, 0.0, 10000.0])
|
|
|
|
.build())
|
|
|
|
cv = CrossValidator(estimator=iee, estimatorParamMaps=grid, evaluator=evaluator)
|
|
|
|
cvModel = cv.fit(dataset)
|
|
|
|
bestModel = cvModel.bestModel
|
|
|
|
bestModelMetric = evaluator.evaluate(bestModel.transform(dataset))
|
|
|
|
|
|
|
|
self.assertEqual(0.0, bestModel.getOrDefault('inducedError'),
|
|
|
|
"Best model should have zero induced error")
|
|
|
|
self.assertEqual(0.0, bestModelMetric, "Best model has RMSE of 0")
|
|
|
|
|
|
|
|
def test_fit_maximize_metric(self):
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
dataset = sqlContext.createDataFrame([
|
|
|
|
(10, 10.0),
|
|
|
|
(50, 50.0),
|
|
|
|
(100, 100.0),
|
|
|
|
(500, 500.0)] * 10,
|
|
|
|
["feature", "label"])
|
|
|
|
|
|
|
|
iee = InducedErrorEstimator()
|
|
|
|
evaluator = RegressionEvaluator(metricName="r2")
|
|
|
|
|
|
|
|
grid = (ParamGridBuilder()
|
|
|
|
.addGrid(iee.inducedError, [100.0, 0.0, 10000.0])
|
|
|
|
.build())
|
|
|
|
cv = CrossValidator(estimator=iee, estimatorParamMaps=grid, evaluator=evaluator)
|
|
|
|
cvModel = cv.fit(dataset)
|
|
|
|
bestModel = cvModel.bestModel
|
|
|
|
bestModelMetric = evaluator.evaluate(bestModel.transform(dataset))
|
|
|
|
|
|
|
|
self.assertEqual(0.0, bestModel.getOrDefault('inducedError'),
|
|
|
|
"Best model should have zero induced error")
|
|
|
|
self.assertEqual(1.0, bestModelMetric, "Best model has R-squared of 1")
|
2016-03-03 12:50:05 -05:00
|
|
|
|
|
|
|
|
|
|
|
class TrainValidationSplitTests(PySparkTestCase):
|
|
|
|
|
|
|
|
def test_fit_minimize_metric(self):
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
dataset = sqlContext.createDataFrame([
|
|
|
|
(10, 10.0),
|
|
|
|
(50, 50.0),
|
|
|
|
(100, 100.0),
|
|
|
|
(500, 500.0)] * 10,
|
|
|
|
["feature", "label"])
|
|
|
|
|
|
|
|
iee = InducedErrorEstimator()
|
|
|
|
evaluator = RegressionEvaluator(metricName="rmse")
|
|
|
|
|
|
|
|
grid = (ParamGridBuilder()
|
|
|
|
.addGrid(iee.inducedError, [100.0, 0.0, 10000.0])
|
|
|
|
.build())
|
|
|
|
tvs = TrainValidationSplit(estimator=iee, estimatorParamMaps=grid, evaluator=evaluator)
|
|
|
|
tvsModel = tvs.fit(dataset)
|
|
|
|
bestModel = tvsModel.bestModel
|
|
|
|
bestModelMetric = evaluator.evaluate(bestModel.transform(dataset))
|
|
|
|
|
|
|
|
self.assertEqual(0.0, bestModel.getOrDefault('inducedError'),
|
|
|
|
"Best model should have zero induced error")
|
|
|
|
self.assertEqual(0.0, bestModelMetric, "Best model has RMSE of 0")
|
|
|
|
|
|
|
|
def test_fit_maximize_metric(self):
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
dataset = sqlContext.createDataFrame([
|
|
|
|
(10, 10.0),
|
|
|
|
(50, 50.0),
|
|
|
|
(100, 100.0),
|
|
|
|
(500, 500.0)] * 10,
|
|
|
|
["feature", "label"])
|
|
|
|
|
|
|
|
iee = InducedErrorEstimator()
|
|
|
|
evaluator = RegressionEvaluator(metricName="r2")
|
|
|
|
|
|
|
|
grid = (ParamGridBuilder()
|
|
|
|
.addGrid(iee.inducedError, [100.0, 0.0, 10000.0])
|
|
|
|
.build())
|
|
|
|
tvs = TrainValidationSplit(estimator=iee, estimatorParamMaps=grid, evaluator=evaluator)
|
|
|
|
tvsModel = tvs.fit(dataset)
|
|
|
|
bestModel = tvsModel.bestModel
|
|
|
|
bestModelMetric = evaluator.evaluate(bestModel.transform(dataset))
|
|
|
|
|
|
|
|
self.assertEqual(0.0, bestModel.getOrDefault('inducedError'),
|
|
|
|
"Best model should have zero induced error")
|
|
|
|
self.assertEqual(1.0, bestModelMetric, "Best model has R-squared of 1")
|
2015-08-28 02:59:30 -04:00
|
|
|
|
|
|
|
|
2016-01-29 12:22:24 -05:00
|
|
|
class PersistenceTest(PySparkTestCase):
|
|
|
|
|
|
|
|
def test_linear_regression(self):
|
|
|
|
lr = LinearRegression(maxIter=1)
|
|
|
|
path = tempfile.mkdtemp()
|
|
|
|
lr_path = path + "/lr"
|
|
|
|
lr.save(lr_path)
|
|
|
|
lr2 = LinearRegression.load(lr_path)
|
|
|
|
self.assertEqual(lr2.uid, lr2.maxIter.parent,
|
|
|
|
"Loaded LinearRegression instance uid (%s) did not match Param's uid (%s)"
|
|
|
|
% (lr2.uid, lr2.maxIter.parent))
|
|
|
|
self.assertEqual(lr._defaultParamMap[lr.maxIter], lr2._defaultParamMap[lr2.maxIter],
|
|
|
|
"Loaded LinearRegression instance default params did not match " +
|
|
|
|
"original defaults")
|
|
|
|
try:
|
|
|
|
rmtree(path)
|
|
|
|
except OSError:
|
|
|
|
pass
|
|
|
|
|
2016-03-16 17:21:42 -04:00
|
|
|
def test_logistic_regression(self):
|
|
|
|
lr = LogisticRegression(maxIter=1)
|
|
|
|
path = tempfile.mkdtemp()
|
|
|
|
lr_path = path + "/logreg"
|
|
|
|
lr.save(lr_path)
|
|
|
|
lr2 = LogisticRegression.load(lr_path)
|
|
|
|
self.assertEqual(lr2.uid, lr2.maxIter.parent,
|
|
|
|
"Loaded LogisticRegression instance uid (%s) "
|
|
|
|
"did not match Param's uid (%s)"
|
|
|
|
% (lr2.uid, lr2.maxIter.parent))
|
|
|
|
self.assertEqual(lr._defaultParamMap[lr.maxIter], lr2._defaultParamMap[lr2.maxIter],
|
|
|
|
"Loaded LogisticRegression instance default params did not match " +
|
|
|
|
"original defaults")
|
|
|
|
try:
|
|
|
|
rmtree(path)
|
|
|
|
except OSError:
|
|
|
|
pass
|
|
|
|
|
2016-03-22 15:11:23 -04:00
|
|
|
def _compare_pipelines(self, m1, m2):
|
|
|
|
"""
|
|
|
|
Compare 2 ML types, asserting that they are equivalent.
|
|
|
|
This currently supports:
|
|
|
|
- basic types
|
|
|
|
- Pipeline, PipelineModel
|
|
|
|
This checks:
|
|
|
|
- uid
|
|
|
|
- type
|
|
|
|
- Param values and parents
|
|
|
|
"""
|
|
|
|
self.assertEqual(m1.uid, m2.uid)
|
|
|
|
self.assertEqual(type(m1), type(m2))
|
|
|
|
if isinstance(m1, JavaWrapper):
|
|
|
|
self.assertEqual(len(m1.params), len(m2.params))
|
|
|
|
for p in m1.params:
|
|
|
|
self.assertEqual(m1.getOrDefault(p), m2.getOrDefault(p))
|
|
|
|
self.assertEqual(p.parent, m2.getParam(p.name).parent)
|
|
|
|
elif isinstance(m1, Pipeline):
|
|
|
|
self.assertEqual(len(m1.getStages()), len(m2.getStages()))
|
|
|
|
for s1, s2 in zip(m1.getStages(), m2.getStages()):
|
|
|
|
self._compare_pipelines(s1, s2)
|
|
|
|
elif isinstance(m1, PipelineModel):
|
|
|
|
self.assertEqual(len(m1.stages), len(m2.stages))
|
|
|
|
for s1, s2 in zip(m1.stages, m2.stages):
|
|
|
|
self._compare_pipelines(s1, s2)
|
|
|
|
else:
|
|
|
|
raise RuntimeError("_compare_pipelines does not yet support type: %s" % type(m1))
|
|
|
|
|
2016-03-16 16:49:40 -04:00
|
|
|
def test_pipeline_persistence(self):
|
2016-03-22 15:11:23 -04:00
|
|
|
"""
|
|
|
|
Pipeline[HashingTF, PCA]
|
|
|
|
"""
|
2016-03-16 16:49:40 -04:00
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
temp_path = tempfile.mkdtemp()
|
|
|
|
|
|
|
|
try:
|
|
|
|
df = sqlContext.createDataFrame([(["a", "b", "c"],), (["c", "d", "e"],)], ["words"])
|
|
|
|
tf = HashingTF(numFeatures=10, inputCol="words", outputCol="features")
|
|
|
|
pca = PCA(k=2, inputCol="features", outputCol="pca_features")
|
|
|
|
pl = Pipeline(stages=[tf, pca])
|
|
|
|
model = pl.fit(df)
|
2016-03-22 15:11:23 -04:00
|
|
|
|
2016-03-16 16:49:40 -04:00
|
|
|
pipeline_path = temp_path + "/pipeline"
|
|
|
|
pl.save(pipeline_path)
|
|
|
|
loaded_pipeline = Pipeline.load(pipeline_path)
|
2016-03-22 15:11:23 -04:00
|
|
|
self._compare_pipelines(pl, loaded_pipeline)
|
|
|
|
|
|
|
|
model_path = temp_path + "/pipeline-model"
|
|
|
|
model.save(model_path)
|
|
|
|
loaded_model = PipelineModel.load(model_path)
|
|
|
|
self._compare_pipelines(model, loaded_model)
|
|
|
|
finally:
|
|
|
|
try:
|
|
|
|
rmtree(temp_path)
|
|
|
|
except OSError:
|
|
|
|
pass
|
2016-03-16 16:49:40 -04:00
|
|
|
|
2016-03-22 15:11:23 -04:00
|
|
|
def test_nested_pipeline_persistence(self):
|
|
|
|
"""
|
|
|
|
Pipeline[HashingTF, Pipeline[PCA]]
|
|
|
|
"""
|
|
|
|
sqlContext = SQLContext(self.sc)
|
|
|
|
temp_path = tempfile.mkdtemp()
|
|
|
|
|
|
|
|
try:
|
|
|
|
df = sqlContext.createDataFrame([(["a", "b", "c"],), (["c", "d", "e"],)], ["words"])
|
|
|
|
tf = HashingTF(numFeatures=10, inputCol="words", outputCol="features")
|
|
|
|
pca = PCA(k=2, inputCol="features", outputCol="pca_features")
|
|
|
|
p0 = Pipeline(stages=[pca])
|
|
|
|
pl = Pipeline(stages=[tf, p0])
|
|
|
|
model = pl.fit(df)
|
2016-03-16 16:49:40 -04:00
|
|
|
|
2016-03-22 15:11:23 -04:00
|
|
|
pipeline_path = temp_path + "/pipeline"
|
|
|
|
pl.save(pipeline_path)
|
|
|
|
loaded_pipeline = Pipeline.load(pipeline_path)
|
|
|
|
self._compare_pipelines(pl, loaded_pipeline)
|
2016-03-16 16:49:40 -04:00
|
|
|
|
|
|
|
model_path = temp_path + "/pipeline-model"
|
|
|
|
model.save(model_path)
|
|
|
|
loaded_model = PipelineModel.load(model_path)
|
2016-03-22 15:11:23 -04:00
|
|
|
self._compare_pipelines(model, loaded_model)
|
2016-03-16 16:49:40 -04:00
|
|
|
finally:
|
|
|
|
try:
|
|
|
|
rmtree(temp_path)
|
|
|
|
except OSError:
|
|
|
|
pass
|
|
|
|
|
2016-01-29 12:22:24 -05:00
|
|
|
|
2016-03-08 20:34:25 -05:00
|
|
|
class HasThrowableProperty(Params):
|
|
|
|
|
|
|
|
def __init__(self):
|
|
|
|
super(HasThrowableProperty, self).__init__()
|
|
|
|
self.p = Param(self, "none", "empty param")
|
|
|
|
|
|
|
|
@property
|
|
|
|
def test_property(self):
|
|
|
|
raise RuntimeError("Test property to raise error when invoked")
|
|
|
|
|
|
|
|
|
2015-01-28 20:14:23 -05:00
|
|
|
if __name__ == "__main__":
|
2016-01-20 14:11:10 -05:00
|
|
|
from pyspark.ml.tests import *
|
2015-10-22 18:27:11 -04:00
|
|
|
if xmlrunner:
|
|
|
|
unittest.main(testRunner=xmlrunner.XMLTestRunner(output='target/test-reports'))
|
|
|
|
else:
|
|
|
|
unittest.main()
|