From 05163057a1810f0a32b722e8c93e5435240636d9 Mon Sep 17 00:00:00 2001
From: Tor Myklebust <tmyklebu@gmail.com>
Date: Wed, 25 Dec 2013 00:08:05 -0500
Subject: [PATCH] Split the mllib bindings into a whole bunch of modules and
rename some things.
---
python/pyspark/__init__.py | 7 +-
python/pyspark/mllib/__init__.py | 46 +++++
python/pyspark/{mllib.py => mllib/_common.py} | 190 ++----------------
python/pyspark/mllib/classification.py | 86 ++++++++
python/pyspark/mllib/clustering.py | 79 ++++++++
python/pyspark/mllib/recommendation.py | 74 +++++++
python/pyspark/mllib/regression.py | 110 ++++++++++
7 files changed, 409 insertions(+), 183 deletions(-)
create mode 100644 python/pyspark/mllib/__init__.py
rename python/pyspark/{mllib.py => mllib/_common.py} (55%)
create mode 100644 python/pyspark/mllib/classification.py
create mode 100644 python/pyspark/mllib/clustering.py
create mode 100644 python/pyspark/mllib/recommendation.py
create mode 100644 python/pyspark/mllib/regression.py
diff --git a/python/pyspark/__init__.py b/python/pyspark/__init__.py
index 3d73d95909..1f35f6f939 100644
--- a/python/pyspark/__init__.py
+++ b/python/pyspark/__init__.py
@@ -42,11 +42,6 @@ from pyspark.context import SparkContext
from pyspark.rdd import RDD
from pyspark.files import SparkFiles
from pyspark.storagelevel import StorageLevel
-from pyspark.mllib import LinearRegressionModel, LassoModel, \
- RidgeRegressionModel, LogisticRegressionModel, SVMModel, KMeansModel, \
- ALSModel
-__all__ = ["SparkContext", "RDD", "SparkFiles", "StorageLevel",
- "LinearRegressionModel", "LassoModel", "RidgeRegressionModel",
- "LogisticRegressionModel", "SVMModel", "KMeansModel", "ALSModel"];
+__all__ = ["SparkContext", "RDD", "SparkFiles", "StorageLevel"]
diff --git a/python/pyspark/mllib/__init__.py b/python/pyspark/mllib/__init__.py
new file mode 100644
index 0000000000..6037a3aa63
--- /dev/null
+++ b/python/pyspark/mllib/__init__.py
@@ -0,0 +1,46 @@
+#
+# 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.
+#
+
+"""
+PySpark is the Python API for Spark.
+
+Public classes:
+
+ - L{SparkContext<pyspark.context.SparkContext>}
+ Main entry point for Spark functionality.
+ - L{RDD<pyspark.rdd.RDD>}
+ A Resilient Distributed Dataset (RDD), the basic abstraction in Spark.
+ - L{Broadcast<pyspark.broadcast.Broadcast>}
+ A broadcast variable that gets reused across tasks.
+ - L{Accumulator<pyspark.accumulators.Accumulator>}
+ An "add-only" shared variable that tasks can only add values to.
+ - L{SparkFiles<pyspark.files.SparkFiles>}
+ Access files shipped with jobs.
+ - L{StorageLevel<pyspark.storagelevel.StorageLevel>}
+ Finer-grained cache persistence levels.
+"""
+import sys
+import os
+sys.path.insert(0, os.path.join(os.environ["SPARK_HOME"], "python/lib/py4j0.7.egg"))
+
+from pyspark.mllib.regression import LinearRegressionModel, LassoModel, RidgeRegressionModel, LinearRegressionWithSGD, LassoWithSGD, RidgeRegressionWithSGD
+from pyspark.mllib.classification import LogisticRegressionModel, SVMModel, LogisticRegressionWithSGD, SVMWithSGD
+from pyspark.mllib.recommendation import MatrixFactorizationModel, ALS
+from pyspark.mllib.clustering import KMeansModel, KMeans
+
+
+__all__ = ["LinearRegressionModel", "LassoModel", "RidgeRegressionModel", "LinearRegressionWithSGD", "LassoWithSGD", "RidgeRegressionWithSGD", "LogisticRegressionModel", "SVMModel", "LogisticRegressionWithSGD", "SVMWithSGD", "MatrixFactorizationModel", "ALS", "KMeansModel", "KMeans"]
diff --git a/python/pyspark/mllib.py b/python/pyspark/mllib/_common.py
similarity index 55%
rename from python/pyspark/mllib.py
rename to python/pyspark/mllib/_common.py
index 46f368b1ec..e68bd8a9db 100644
--- a/python/pyspark/mllib.py
+++ b/python/pyspark/mllib/_common.py
@@ -15,7 +15,7 @@
# limitations under the License.
#
-from numpy import *
+from numpy import ndarray, copyto, float64, int64, int32, zeros, array_equal, array, dot, shape
from pyspark import SparkContext
# Double vector format:
@@ -137,7 +137,7 @@ def _linear_predictor_typecheck(x, coeffs):
pass
else:
raise RuntimeError("Got array of %d elements; wanted %d"
- % shape(x)[0] % shape(coeffs)[0])
+ % (shape(x)[0], shape(coeffs)[0]))
else:
raise RuntimeError("Bulk predict not yet supported.")
elif (type(x) == RDD):
@@ -145,6 +145,17 @@ def _linear_predictor_typecheck(x, coeffs):
else:
raise TypeError("Argument of type " + type(x) + " unsupported")
+def _get_unmangled_rdd(data, serializer):
+ dataBytes = data.map(serializer)
+ dataBytes._bypass_serializer = True
+ dataBytes.cache()
+ return dataBytes
+
+# Map a pickled Python RDD of numpy double vectors to a Java RDD of
+# _serialized_double_vectors
+def _get_unmangled_double_vector_rdd(data):
+ return _get_unmangled_rdd(data, _serialize_double_vector)
+
class LinearModel(object):
"""Something that has a vector of coefficients and an intercept."""
def __init__(self, coeff, intercept):
@@ -164,17 +175,6 @@ class LinearRegressionModelBase(LinearModel):
_linear_predictor_typecheck(x, self._coeff)
return dot(self._coeff, x) + self._intercept
-def _get_unmangled_rdd(data, serializer):
- dataBytes = data.map(serializer)
- dataBytes._bypass_serializer = True
- dataBytes.cache()
- return dataBytes
-
-# Map a pickled Python RDD of numpy double vectors to a Java RDD of
-# _serialized_double_vectors
-def _get_unmangled_double_vector_rdd(data):
- return _get_unmangled_rdd(data, _serialize_double_vector)
-
# If we weren't given initial weights, take a zero vector of the appropriate
# length.
def _get_initial_weights(initial_weights, data):
@@ -206,133 +206,6 @@ def _regression_train_wrapper(sc, train_func, klass, data, initial_weights):
+ type(ans[0]) + " which is not float")
return klass(_deserialize_double_vector(ans[0]), ans[1])
-class LinearRegressionModel(LinearRegressionModelBase):
- """A linear regression model derived from a least-squares fit.
-
- >>> data = array([0.0, 0.0, 1.0, 1.0, 3.0, 2.0, 2.0, 3.0]).reshape(4,2)
- >>> lrm = LinearRegressionModel.train(sc, sc.parallelize(data), initial_weights=array([1.0]))
- """
- @classmethod
- def train(cls, sc, data, iterations=100, step=1.0,
- mini_batch_fraction=1.0, initial_weights=None):
- """Train a linear regression model on the given data."""
- return _regression_train_wrapper(sc, lambda d, i:
- sc._jvm.PythonMLLibAPI().trainLinearRegressionModel(
- d._jrdd, iterations, step, mini_batch_fraction, i),
- LinearRegressionModel, data, initial_weights)
-
-class LassoModel(LinearRegressionModelBase):
- """A linear regression model derived from a least-squares fit with an
- l_1 penalty term.
-
- >>> data = array([0.0, 0.0, 1.0, 1.0, 3.0, 2.0, 2.0, 3.0]).reshape(4,2)
- >>> lrm = LassoModel.train(sc, sc.parallelize(data), initial_weights=array([1.0]))
- """
- @classmethod
- def train(cls, sc, data, iterations=100, step=1.0, reg_param=1.0,
- mini_batch_fraction=1.0, initial_weights=None):
- """Train a Lasso regression model on the given data."""
- return _regression_train_wrapper(sc, lambda d, i:
- sc._jvm.PythonMLLibAPI().trainLassoModel(d._jrdd,
- iterations, step, reg_param, mini_batch_fraction, i),
- LassoModel, data, initial_weights)
-
-class RidgeRegressionModel(LinearRegressionModelBase):
- """A linear regression model derived from a least-squares fit with an
- l_2 penalty term.
-
- >>> data = array([0.0, 0.0, 1.0, 1.0, 3.0, 2.0, 2.0, 3.0]).reshape(4,2)
- >>> lrm = RidgeRegressionModel.train(sc, sc.parallelize(data), initial_weights=array([1.0]))
- """
- @classmethod
- def train(cls, sc, data, iterations=100, step=1.0, reg_param=1.0,
- mini_batch_fraction=1.0, initial_weights=None):
- """Train a ridge regression model on the given data."""
- return _regression_train_wrapper(sc, lambda d, i:
- sc._jvm.PythonMLLibAPI().trainRidgeModel(d._jrdd,
- iterations, step, reg_param, mini_batch_fraction, i),
- RidgeRegressionModel, data, initial_weights)
-
-class LogisticRegressionModel(LinearModel):
- """A linear binary classification model derived from logistic regression.
-
- >>> data = array([0.0, 0.0, 1.0, 1.0, 1.0, 2.0, 1.0, 3.0]).reshape(4,2)
- >>> lrm = LogisticRegressionModel.train(sc, sc.parallelize(data))
- """
- def predict(self, x):
- _linear_predictor_typecheck(x, _coeff)
- margin = dot(x, _coeff) + intercept
- prob = 1/(1 + exp(-margin))
- return 1 if prob > 0.5 else 0
-
- @classmethod
- def train(cls, sc, data, iterations=100, step=1.0,
- mini_batch_fraction=1.0, initial_weights=None):
- """Train a logistic regression model on the given data."""
- return _regression_train_wrapper(sc, lambda d, i:
- sc._jvm.PythonMLLibAPI().trainLogisticRegressionModel(d._jrdd,
- iterations, step, mini_batch_fraction, i),
- LogisticRegressionModel, data, initial_weights)
-
-class SVMModel(LinearModel):
- """A support vector machine.
-
- >>> data = array([0.0, 0.0, 1.0, 1.0, 1.0, 2.0, 1.0, 3.0]).reshape(4,2)
- >>> svm = SVMModel.train(sc, sc.parallelize(data))
- """
- def predict(self, x):
- _linear_predictor_typecheck(x, _coeff)
- margin = dot(x, _coeff) + intercept
- return 1 if margin >= 0 else 0
- @classmethod
- def train(cls, sc, data, iterations=100, step=1.0, reg_param=1.0,
- mini_batch_fraction=1.0, initial_weights=None):
- """Train a support vector machine on the given data."""
- return _regression_train_wrapper(sc, lambda d, i:
- sc._jvm.PythonMLLibAPI().trainSVMModel(d._jrdd,
- iterations, step, reg_param, mini_batch_fraction, i),
- SVMModel, data, initial_weights)
-
-class KMeansModel(object):
- """A clustering model derived from the k-means method.
-
- >>> data = array([0.0,0.0, 1.0,1.0, 9.0,8.0, 8.0,9.0]).reshape(4,2)
- >>> clusters = KMeansModel.train(sc, sc.parallelize(data), 2, maxIterations=10, runs=30, initialization_mode="random")
- >>> clusters.predict(array([0.0, 0.0])) == clusters.predict(array([1.0, 1.0]))
- True
- >>> clusters.predict(array([8.0, 9.0])) == clusters.predict(array([9.0, 8.0]))
- True
- >>> clusters = KMeansModel.train(sc, sc.parallelize(data), 2)
- """
- def __init__(self, centers_):
- self.centers = centers_
-
- def predict(self, x):
- """Find the cluster to which x belongs in this model."""
- best = 0
- best_distance = 1e75
- for i in range(0, self.centers.shape[0]):
- diff = x - self.centers[i]
- distance = sqrt(dot(diff, diff))
- if distance < best_distance:
- best = i
- best_distance = distance
- return best
-
- @classmethod
- def train(cls, sc, data, k, maxIterations=100, runs=1,
- initialization_mode="k-means||"):
- """Train a k-means clustering model."""
- dataBytes = _get_unmangled_double_vector_rdd(data)
- ans = sc._jvm.PythonMLLibAPI().trainKMeansModel(dataBytes._jrdd,
- k, maxIterations, runs, initialization_mode)
- if len(ans) != 1:
- raise RuntimeError("JVM call result had unexpected length")
- elif type(ans[0]) != bytearray:
- raise RuntimeError("JVM call result had first element of type "
- + type(ans[0]) + " which is not bytearray")
- return KMeansModel(_deserialize_double_matrix(ans[0]))
-
def _serialize_rating(r):
ba = bytearray(16)
intpart = ndarray(shape=[2], buffer=ba, dtype=int32)
@@ -340,43 +213,6 @@ def _serialize_rating(r):
intpart[0], intpart[1], doublepart[0] = r
return ba
-class ALSModel(object):
- """A matrix factorisation model trained by regularized alternating
- least-squares.
-
- >>> r1 = (1, 1, 1.0)
- >>> r2 = (1, 2, 2.0)
- >>> r3 = (2, 1, 2.0)
- >>> ratings = sc.parallelize([r1, r2, r3])
- >>> model = ALSModel.trainImplicit(sc, ratings, 1)
- >>> model.predict(2,2) is not None
- True
- """
-
- def __init__(self, sc, java_model):
- self._context = sc
- self._java_model = java_model
-
- def __del__(self):
- self._context._gateway.detach(self._java_model)
-
- def predict(self, user, product):
- return self._java_model.predict(user, product)
-
- @classmethod
- def train(cls, sc, ratings, rank, iterations=5, lambda_=0.01, blocks=-1):
- ratingBytes = _get_unmangled_rdd(ratings, _serialize_rating)
- mod = sc._jvm.PythonMLLibAPI().trainALSModel(ratingBytes._jrdd,
- rank, iterations, lambda_, blocks)
- return ALSModel(sc, mod)
-
- @classmethod
- def trainImplicit(cls, sc, ratings, rank, iterations=5, lambda_=0.01, blocks=-1, alpha=0.01):
- ratingBytes = _get_unmangled_rdd(ratings, _serialize_rating)
- mod = sc._jvm.PythonMLLibAPI().trainImplicitALSModel(ratingBytes._jrdd,
- rank, iterations, lambda_, blocks, alpha)
- return ALSModel(sc, mod)
-
def _test():
import doctest
globs = globals().copy()
diff --git a/python/pyspark/mllib/classification.py b/python/pyspark/mllib/classification.py
new file mode 100644
index 0000000000..70de332d34
--- /dev/null
+++ b/python/pyspark/mllib/classification.py
@@ -0,0 +1,86 @@
+#
+# 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.
+#
+
+from numpy import array, dot, shape
+from pyspark import SparkContext
+from pyspark.mllib._common import \
+ _get_unmangled_rdd, _get_unmangled_double_vector_rdd, \
+ _serialize_double_matrix, _deserialize_double_matrix, \
+ _serialize_double_vector, _deserialize_double_vector, \
+ _get_initial_weights, _serialize_rating, _regression_train_wrapper, \
+ LinearModel, _linear_predictor_typecheck
+from math import exp, log
+
+class LogisticRegressionModel(LinearModel):
+ """A linear binary classification model derived from logistic regression.
+
+ >>> data = array([0.0, 0.0, 1.0, 1.0, 1.0, 2.0, 1.0, 3.0]).reshape(4,2)
+ >>> lrm = LogisticRegressionWithSGD.train(sc, sc.parallelize(data))
+ >>> lrm.predict(array([1.0])) != None
+ True
+ """
+ def predict(self, x):
+ _linear_predictor_typecheck(x, self._coeff)
+ margin = dot(x, self._coeff) + self._intercept
+ prob = 1/(1 + exp(-margin))
+ return 1 if prob > 0.5 else 0
+
+class LogisticRegressionWithSGD(object):
+ @classmethod
+ def train(cls, sc, data, iterations=100, step=1.0,
+ mini_batch_fraction=1.0, initial_weights=None):
+ """Train a logistic regression model on the given data."""
+ return _regression_train_wrapper(sc, lambda d, i:
+ sc._jvm.PythonMLLibAPI().trainLogisticRegressionModelWithSGD(d._jrdd,
+ iterations, step, mini_batch_fraction, i),
+ LogisticRegressionModel, data, initial_weights)
+
+class SVMModel(LinearModel):
+ """A support vector machine.
+
+ >>> data = array([0.0, 0.0, 1.0, 1.0, 1.0, 2.0, 1.0, 3.0]).reshape(4,2)
+ >>> svm = SVMWithSGD.train(sc, sc.parallelize(data))
+ >>> svm.predict(array([1.0])) != None
+ True
+ """
+ def predict(self, x):
+ _linear_predictor_typecheck(x, self._coeff)
+ margin = dot(x, self._coeff) + self._intercept
+ return 1 if margin >= 0 else 0
+
+class SVMWithSGD(object):
+ @classmethod
+ def train(cls, sc, data, iterations=100, step=1.0, reg_param=1.0,
+ mini_batch_fraction=1.0, initial_weights=None):
+ """Train a support vector machine on the given data."""
+ return _regression_train_wrapper(sc, lambda d, i:
+ sc._jvm.PythonMLLibAPI().trainSVMModelWithSGD(d._jrdd,
+ iterations, step, reg_param, mini_batch_fraction, i),
+ SVMModel, data, initial_weights)
+
+def _test():
+ import doctest
+ globs = globals().copy()
+ globs['sc'] = SparkContext('local[4]', 'PythonTest', batchSize=2)
+ (failure_count, test_count) = doctest.testmod(globs=globs,
+ optionflags=doctest.ELLIPSIS)
+ globs['sc'].stop()
+ if failure_count:
+ exit(-1)
+
+if __name__ == "__main__":
+ _test()
diff --git a/python/pyspark/mllib/clustering.py b/python/pyspark/mllib/clustering.py
new file mode 100644
index 0000000000..8cf20e591a
--- /dev/null
+++ b/python/pyspark/mllib/clustering.py
@@ -0,0 +1,79 @@
+#
+# 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.
+#
+
+from numpy import array, dot
+from math import sqrt
+from pyspark import SparkContext
+from pyspark.mllib._common import \
+ _get_unmangled_rdd, _get_unmangled_double_vector_rdd, \
+ _serialize_double_matrix, _deserialize_double_matrix, \
+ _serialize_double_vector, _deserialize_double_vector, \
+ _get_initial_weights, _serialize_rating, _regression_train_wrapper
+
+class KMeansModel(object):
+ """A clustering model derived from the k-means method.
+
+ >>> data = array([0.0,0.0, 1.0,1.0, 9.0,8.0, 8.0,9.0]).reshape(4,2)
+ >>> clusters = KMeans.train(sc, sc.parallelize(data), 2, maxIterations=10, runs=30, initialization_mode="random")
+ >>> clusters.predict(array([0.0, 0.0])) == clusters.predict(array([1.0, 1.0]))
+ True
+ >>> clusters.predict(array([8.0, 9.0])) == clusters.predict(array([9.0, 8.0]))
+ True
+ >>> clusters = KMeans.train(sc, sc.parallelize(data), 2)
+ """
+ def __init__(self, centers_):
+ self.centers = centers_
+
+ def predict(self, x):
+ """Find the cluster to which x belongs in this model."""
+ best = 0
+ best_distance = 1e75
+ for i in range(0, self.centers.shape[0]):
+ diff = x - self.centers[i]
+ distance = sqrt(dot(diff, diff))
+ if distance < best_distance:
+ best = i
+ best_distance = distance
+ return best
+
+class KMeans(object):
+ @classmethod
+ def train(cls, sc, data, k, maxIterations=100, runs=1,
+ initialization_mode="k-means||"):
+ """Train a k-means clustering model."""
+ dataBytes = _get_unmangled_double_vector_rdd(data)
+ ans = sc._jvm.PythonMLLibAPI().trainKMeansModel(dataBytes._jrdd,
+ k, maxIterations, runs, initialization_mode)
+ if len(ans) != 1:
+ raise RuntimeError("JVM call result had unexpected length")
+ elif type(ans[0]) != bytearray:
+ raise RuntimeError("JVM call result had first element of type "
+ + type(ans[0]) + " which is not bytearray")
+ return KMeansModel(_deserialize_double_matrix(ans[0]))
+
+def _test():
+ import doctest
+ globs = globals().copy()
+ globs['sc'] = SparkContext('local[4]', 'PythonTest', batchSize=2)
+ (failure_count, test_count) = doctest.testmod(globs=globs,
+ optionflags=doctest.ELLIPSIS)
+ globs['sc'].stop()
+ if failure_count:
+ exit(-1)
+
+if __name__ == "__main__":
+ _test()
diff --git a/python/pyspark/mllib/recommendation.py b/python/pyspark/mllib/recommendation.py
new file mode 100644
index 0000000000..14d06cba21
--- /dev/null
+++ b/python/pyspark/mllib/recommendation.py
@@ -0,0 +1,74 @@
+#
+# 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.
+#
+
+from pyspark import SparkContext
+from pyspark.mllib._common import \
+ _get_unmangled_rdd, _get_unmangled_double_vector_rdd, \
+ _serialize_double_matrix, _deserialize_double_matrix, \
+ _serialize_double_vector, _deserialize_double_vector, \
+ _get_initial_weights, _serialize_rating, _regression_train_wrapper
+
+class MatrixFactorizationModel(object):
+ """A matrix factorisation model trained by regularized alternating
+ least-squares.
+
+ >>> r1 = (1, 1, 1.0)
+ >>> r2 = (1, 2, 2.0)
+ >>> r3 = (2, 1, 2.0)
+ >>> ratings = sc.parallelize([r1, r2, r3])
+ >>> model = ALS.trainImplicit(sc, ratings, 1)
+ >>> model.predict(2,2) is not None
+ True
+ """
+
+ def __init__(self, sc, java_model):
+ self._context = sc
+ self._java_model = java_model
+
+ def __del__(self):
+ self._context._gateway.detach(self._java_model)
+
+ def predict(self, user, product):
+ return self._java_model.predict(user, product)
+
+class ALS(object):
+ @classmethod
+ def train(cls, sc, ratings, rank, iterations=5, lambda_=0.01, blocks=-1):
+ ratingBytes = _get_unmangled_rdd(ratings, _serialize_rating)
+ mod = sc._jvm.PythonMLLibAPI().trainALSModel(ratingBytes._jrdd,
+ rank, iterations, lambda_, blocks)
+ return MatrixFactorizationModel(sc, mod)
+
+ @classmethod
+ def trainImplicit(cls, sc, ratings, rank, iterations=5, lambda_=0.01, blocks=-1, alpha=0.01):
+ ratingBytes = _get_unmangled_rdd(ratings, _serialize_rating)
+ mod = sc._jvm.PythonMLLibAPI().trainImplicitALSModel(ratingBytes._jrdd,
+ rank, iterations, lambda_, blocks, alpha)
+ return MatrixFactorizationModel(sc, mod)
+
+def _test():
+ import doctest
+ globs = globals().copy()
+ globs['sc'] = SparkContext('local[4]', 'PythonTest', batchSize=2)
+ (failure_count, test_count) = doctest.testmod(globs=globs,
+ optionflags=doctest.ELLIPSIS)
+ globs['sc'].stop()
+ if failure_count:
+ exit(-1)
+
+if __name__ == "__main__":
+ _test()
diff --git a/python/pyspark/mllib/regression.py b/python/pyspark/mllib/regression.py
new file mode 100644
index 0000000000..a3a68b29e0
--- /dev/null
+++ b/python/pyspark/mllib/regression.py
@@ -0,0 +1,110 @@
+#
+# 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.
+#
+
+from numpy import array, dot
+from pyspark import SparkContext
+from pyspark.mllib._common import \
+ _get_unmangled_rdd, _get_unmangled_double_vector_rdd, \
+ _serialize_double_matrix, _deserialize_double_matrix, \
+ _serialize_double_vector, _deserialize_double_vector, \
+ _get_initial_weights, _serialize_rating, _regression_train_wrapper, \
+ _linear_predictor_typecheck
+
+class LinearModel(object):
+ """Something that has a vector of coefficients and an intercept."""
+ def __init__(self, coeff, intercept):
+ self._coeff = coeff
+ self._intercept = intercept
+
+class LinearRegressionModelBase(LinearModel):
+ """A linear regression model.
+
+ >>> lrmb = LinearRegressionModelBase(array([1.0, 2.0]), 0.1)
+ >>> abs(lrmb.predict(array([-1.03, 7.777])) - 14.624) < 1e-6
+ True
+ """
+ def predict(self, x):
+ """Predict the value of the dependent variable given a vector x"""
+ """containing values for the independent variables."""
+ _linear_predictor_typecheck(x, self._coeff)
+ return dot(self._coeff, x) + self._intercept
+
+class LinearRegressionModel(LinearRegressionModelBase):
+ """A linear regression model derived from a least-squares fit.
+
+ >>> data = array([0.0, 0.0, 1.0, 1.0, 3.0, 2.0, 2.0, 3.0]).reshape(4,2)
+ >>> lrm = LinearRegressionWithSGD.train(sc, sc.parallelize(data), initial_weights=array([1.0]))
+ """
+
+class LinearRegressionWithSGD(object):
+ @classmethod
+ def train(cls, sc, data, iterations=100, step=1.0,
+ mini_batch_fraction=1.0, initial_weights=None):
+ """Train a linear regression model on the given data."""
+ return _regression_train_wrapper(sc, lambda d, i:
+ sc._jvm.PythonMLLibAPI().trainLinearRegressionModelWithSGD(
+ d._jrdd, iterations, step, mini_batch_fraction, i),
+ LinearRegressionModel, data, initial_weights)
+
+class LassoModel(LinearRegressionModelBase):
+ """A linear regression model derived from a least-squares fit with an
+ l_1 penalty term.
+
+ >>> data = array([0.0, 0.0, 1.0, 1.0, 3.0, 2.0, 2.0, 3.0]).reshape(4,2)
+ >>> lrm = LassoWithSGD.train(sc, sc.parallelize(data), initial_weights=array([1.0]))
+ """
+
+class LassoWithSGD(object):
+ @classmethod
+ def train(cls, sc, data, iterations=100, step=1.0, reg_param=1.0,
+ mini_batch_fraction=1.0, initial_weights=None):
+ """Train a Lasso regression model on the given data."""
+ return _regression_train_wrapper(sc, lambda d, i:
+ sc._jvm.PythonMLLibAPI().trainLassoModelWithSGD(d._jrdd,
+ iterations, step, reg_param, mini_batch_fraction, i),
+ LassoModel, data, initial_weights)
+
+class RidgeRegressionModel(LinearRegressionModelBase):
+ """A linear regression model derived from a least-squares fit with an
+ l_2 penalty term.
+
+ >>> data = array([0.0, 0.0, 1.0, 1.0, 3.0, 2.0, 2.0, 3.0]).reshape(4,2)
+ >>> lrm = RidgeRegressionWithSGD.train(sc, sc.parallelize(data), initial_weights=array([1.0]))
+ """
+
+class RidgeRegressionWithSGD(object):
+ @classmethod
+ def train(cls, sc, data, iterations=100, step=1.0, reg_param=1.0,
+ mini_batch_fraction=1.0, initial_weights=None):
+ """Train a ridge regression model on the given data."""
+ return _regression_train_wrapper(sc, lambda d, i:
+ sc._jvm.PythonMLLibAPI().trainRidgeModelWithSGD(d._jrdd,
+ iterations, step, reg_param, mini_batch_fraction, i),
+ RidgeRegressionModel, data, initial_weights)
+
+def _test():
+ import doctest
+ globs = globals().copy()
+ globs['sc'] = SparkContext('local[4]', 'PythonTest', batchSize=2)
+ (failure_count, test_count) = doctest.testmod(globs=globs,
+ optionflags=doctest.ELLIPSIS)
+ globs['sc'].stop()
+ if failure_count:
+ exit(-1)
+
+if __name__ == "__main__":
+ _test()
--
GitLab