From a0411aebee7c134f0426f0c2b2cb4c1c7856a291 Mon Sep 17 00:00:00 2001 From: Yanbo Liang Date: Thu, 9 Apr 2015 15:10:10 -0700 Subject: [PATCH] [SPARK-6264] [MLLIB] Support FPGrowth algorithm in Python API Support FPGrowth algorithm in Python API. Should we remove "Experimental" which were marked for FPGrowth and FPGrowthModel in Scala? jkbradley Author: Yanbo Liang Closes #5213 from yanboliang/spark-6264 and squashes the following commits: ed62ead [Yanbo Liang] trigger jenkins 8ce0359 [Yanbo Liang] fix docstring style 544c725 [Yanbo Liang] address comments a2d7cf7 [Yanbo Liang] add doc for FPGrowth.train() dcf7d73 [Yanbo Liang] add python doc b18fd07 [Yanbo Liang] trigger jenkins 2c951b8 [Yanbo Liang] fix typos 7f62c8f [Yanbo Liang] add fpm to __init__.py b96206a [Yanbo Liang] Support FPGrowth algorithm in Python API --- .../api/python/FPGrowthModelWrapper.scala | 33 ++++++++ .../mllib/api/python/PythonMLLibAPI.scala | 23 +++++- python/docs/pyspark.mllib.rst | 7 ++ python/pyspark/mllib/__init__.py | 2 +- python/pyspark/mllib/fpm.py | 81 +++++++++++++++++++ python/run-tests | 1 + 6 files changed, 143 insertions(+), 4 deletions(-) create mode 100644 mllib/src/main/scala/org/apache/spark/mllib/api/python/FPGrowthModelWrapper.scala create mode 100644 python/pyspark/mllib/fpm.py diff --git a/mllib/src/main/scala/org/apache/spark/mllib/api/python/FPGrowthModelWrapper.scala b/mllib/src/main/scala/org/apache/spark/mllib/api/python/FPGrowthModelWrapper.scala new file mode 100644 index 0000000000000..ee933f4cfcafd --- /dev/null +++ b/mllib/src/main/scala/org/apache/spark/mllib/api/python/FPGrowthModelWrapper.scala @@ -0,0 +1,33 @@ +/* + * 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. + */ + +package org.apache.spark.mllib.api.python + +import org.apache.spark.api.java.JavaRDD +import org.apache.spark.mllib.fpm.{FPGrowth, FPGrowthModel} +import org.apache.spark.rdd.RDD + +/** + * A Wrapper of FPGrowthModel to provide helper method for Python + */ +private[python] class FPGrowthModelWrapper(model: FPGrowthModel[Any]) + extends FPGrowthModel(model.freqItemsets) { + + def getFreqItemsets: RDD[Array[Any]] = { + SerDe.fromTuple2RDD(model.freqItemsets.map(x => (x.javaItems, x.freq))) + } +} diff --git a/mllib/src/main/scala/org/apache/spark/mllib/api/python/PythonMLLibAPI.scala b/mllib/src/main/scala/org/apache/spark/mllib/api/python/PythonMLLibAPI.scala index 6c386cacfb7ca..1faa3def0e042 100644 --- a/mllib/src/main/scala/org/apache/spark/mllib/api/python/PythonMLLibAPI.scala +++ b/mllib/src/main/scala/org/apache/spark/mllib/api/python/PythonMLLibAPI.scala @@ -34,6 +34,7 @@ import org.apache.spark.api.python.SerDeUtil import org.apache.spark.mllib.classification._ import org.apache.spark.mllib.clustering._ import org.apache.spark.mllib.feature._ +import org.apache.spark.mllib.fpm.{FPGrowth, FPGrowthModel} import org.apache.spark.mllib.linalg._ import org.apache.spark.mllib.optimization._ import org.apache.spark.mllib.random.{RandomRDDs => RG} @@ -358,9 +359,7 @@ private[python] class PythonMLLibAPI extends Serializable { val model = new GaussianMixtureModel(weight, gaussians) model.predictSoft(data) } - - - + /** * Java stub for Python mllib ALS.train(). This stub returns a handle * to the Java object instead of the content of the Java object. Extra care @@ -420,6 +419,24 @@ private[python] class PythonMLLibAPI extends Serializable { new MatrixFactorizationModelWrapper(model) } + /** + * Java stub for Python mllib FPGrowth.train(). This stub returns a handle + * to the Java object instead of the content of the Java object. Extra care + * needs to be taken in the Python code to ensure it gets freed on exit; see + * the Py4J documentation. + */ + def trainFPGrowthModel( + data: JavaRDD[java.lang.Iterable[Any]], + minSupport: Double, + numPartitions: Int): FPGrowthModel[Any] = { + val fpg = new FPGrowth() + .setMinSupport(minSupport) + .setNumPartitions(numPartitions) + + val model = fpg.run(data.rdd.map(_.asScala.toArray)) + new FPGrowthModelWrapper(model) + } + /** * Java stub for Normalizer.transform() */ diff --git a/python/docs/pyspark.mllib.rst b/python/docs/pyspark.mllib.rst index 15101470afc07..26ece4c2c389a 100644 --- a/python/docs/pyspark.mllib.rst +++ b/python/docs/pyspark.mllib.rst @@ -31,6 +31,13 @@ pyspark.mllib.feature module :undoc-members: :show-inheritance: +pyspark.mllib.fpm module +------------------------ + +.. automodule:: pyspark.mllib.fpm + :members: + :undoc-members: + pyspark.mllib.linalg module --------------------------- diff --git a/python/pyspark/mllib/__init__.py b/python/pyspark/mllib/__init__.py index 6449800d9c120..f2ef573fe9f6f 100644 --- a/python/pyspark/mllib/__init__.py +++ b/python/pyspark/mllib/__init__.py @@ -25,7 +25,7 @@ if numpy.version.version < '1.4': raise Exception("MLlib requires NumPy 1.4+") -__all__ = ['classification', 'clustering', 'feature', 'linalg', 'random', +__all__ = ['classification', 'clustering', 'feature', 'fpm', 'linalg', 'random', 'recommendation', 'regression', 'stat', 'tree', 'util'] import sys diff --git a/python/pyspark/mllib/fpm.py b/python/pyspark/mllib/fpm.py new file mode 100644 index 0000000000000..3aa6d79d7093c --- /dev/null +++ b/python/pyspark/mllib/fpm.py @@ -0,0 +1,81 @@ +# +# 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 JavaModelWrapper, callMLlibFunc, inherit_doc + +__all__ = ['FPGrowth', 'FPGrowthModel'] + + +@inherit_doc +class FPGrowthModel(JavaModelWrapper): + + """ + .. note:: Experimental + + A FP-Growth model for mining frequent itemsets + using the Parallel FP-Growth algorithm. + + >>> data = [["a", "b", "c"], ["a", "b", "d", "e"], ["a", "c", "e"], ["a", "c", "f"]] + >>> rdd = sc.parallelize(data, 2) + >>> model = FPGrowth.train(rdd, 0.6, 2) + >>> sorted(model.freqItemsets().collect()) + [([u'a'], 4), ([u'c'], 3), ([u'c', u'a'], 3)] + """ + + def freqItemsets(self): + """ + Get the frequent itemsets of this model + """ + return self.call("getFreqItemsets") + + +class FPGrowth(object): + """ + .. note:: Experimental + + A Parallel FP-growth algorithm to mine frequent itemsets. + """ + + @classmethod + def train(cls, data, minSupport=0.3, numPartitions=-1): + """ + Computes an FP-Growth model that contains frequent itemsets. + :param data: The input data set, each element + contains a transaction. + :param minSupport: The minimal support level + (default: `0.3`). + :param numPartitions: The number of partitions used by parallel + FP-growth (default: same as input data). + """ + model = callMLlibFunc("trainFPGrowthModel", data, float(minSupport), int(numPartitions)) + return FPGrowthModel(model) + + +def _test(): + import doctest + import pyspark.mllib.fpm + globs = pyspark.mllib.fpm.__dict__.copy() + globs['sc'] = SparkContext('local[4]', 'PythonTest') + (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/run-tests b/python/run-tests index b7630c356cfae..f569a56fb7a9a 100755 --- a/python/run-tests +++ b/python/run-tests @@ -77,6 +77,7 @@ function run_mllib_tests() { run_test "pyspark/mllib/clustering.py" run_test "pyspark/mllib/evaluation.py" run_test "pyspark/mllib/feature.py" + run_test "pyspark/mllib/fpm.py" run_test "pyspark/mllib/linalg.py" run_test "pyspark/mllib/rand.py" run_test "pyspark/mllib/recommendation.py"