[SPARK-19826][ML][PYTHON]add spark.ml Python API for PIC

mllib/src/main/scala/org/apache/spark/ml/clustering/PowerIterationClustering.scala

@@ -97,13 +97,15 @@ private[clustering] trait PowerIterationClusteringParams extends Params with Has
   def getNeighborsCol: String = $(neighborsCol)
 
   /**
-   * Param for the name of the input column for neighbors in the adjacency list representation.
+   * Param for the name of the input column for non-negative weights (similarities) of edges
+   * between the vertex in `idCol` and each neighbor in `neighborsCol`.
    * Default: "similarities"
    * @group param
    */
   @Since("2.4.0")
   val similaritiesCol = new Param[String](this, "similaritiesCol",
-    "Name of the input column for neighbors in the adjacency list representation.",
+    "Name of the input column for non-negative weights (similarities) of edges between the " +
+    "vertex in `idCol` and each neighbor in `neighborsCol`.",
     (value: String) => value.nonEmpty)
 
   setDefault(similaritiesCol, "similarities")

python/pyspark/ml/clustering.py

@@ -19,14 +19,14 @@
 
 from pyspark import since, keyword_only
 from pyspark.ml.util import *
-from pyspark.ml.wrapper import JavaEstimator, JavaModel, JavaWrapper
+from pyspark.ml.wrapper import JavaEstimator, JavaModel, JavaParams, JavaTransformer, JavaWrapper
 from pyspark.ml.param.shared import *
 from pyspark.ml.common import inherit_doc
 
 __all__ = ['BisectingKMeans', 'BisectingKMeansModel', 'BisectingKMeansSummary',
            'KMeans', 'KMeansModel',
            'GaussianMixture', 'GaussianMixtureModel', 'GaussianMixtureSummary',
-           'LDA', 'LDAModel', 'LocalLDAModel', 'DistributedLDAModel']
+           'LDA', 'LDAModel', 'LocalLDAModel', 'DistributedLDAModel', 'PowerIterationClustering']
 
 
 class ClusteringSummary(JavaWrapper):
@@ -836,7 +836,7 @@ class LDA(JavaEstimator, HasFeaturesCol, HasMaxIter, HasSeed, HasCheckpointInter
 
     Terminology:
 
-     - "term" = "word": an el
+     - "term" = "word": an element of the vocabulary
      - "token": instance of a term appearing in a document
      - "topic": multinomial distribution over terms representing some concept
      - "document": one piece of text, corresponding to one row in the input data
@@ -938,7 +938,7 @@ def __init__(self, featuresCol="features", maxIter=20, seed=None, checkpointInte
                   k=10, optimizer="online", learningOffset=1024.0, learningDecay=0.51,\
                   subsamplingRate=0.05, optimizeDocConcentration=True,\
                   docConcentration=None, topicConcentration=None,\
-                  topicDistributionCol="topicDistribution", keepLastCheckpoint=True):
+                  topicDistributionCol="topicDistribution", keepLastCheckpoint=True)
         """
         super(LDA, self).__init__()
         self._java_obj = self._new_java_obj("org.apache.spark.ml.clustering.LDA", self.uid)
@@ -967,7 +967,7 @@ def setParams(self, featuresCol="features", maxIter=20, seed=None, checkpointInt
                   k=10, optimizer="online", learningOffset=1024.0, learningDecay=0.51,\
                   subsamplingRate=0.05, optimizeDocConcentration=True,\
                   docConcentration=None, topicConcentration=None,\
-                  topicDistributionCol="topicDistribution", keepLastCheckpoint=True):
+                  topicDistributionCol="topicDistribution", keepLastCheckpoint=True)
 
         Sets params for LDA.
         """
@@ -1156,6 +1156,204 @@ def getKeepLastCheckpoint(self):
         return self.getOrDefault(self.keepLastCheckpoint)
 
 
+@inherit_doc
+class PowerIterationClustering(HasMaxIter, HasPredictionCol, JavaTransformer, JavaParams,
+                               JavaMLReadable, JavaMLWritable):
+    """
+    .. note:: Experimental
+    Power Iteration Clustering (PIC), a scalable graph clustering algorithm developed by
+    <a href=http://www.icml2010.org/papers/387.pdf>Lin and Cohen</a>. From the abstract:
+    PIC finds a very low-dimensional embedding of a dataset using truncated power
+    iteration on a normalized pair-wise similarity matrix of the data.
+
+    PIC takes an affinity matrix between items (or vertices) as input.  An affinity matrix
+    is a symmetric matrix whose entries are non-negative similarities between items.
+    PIC takes this matrix (or graph) as an adjacency matrix.  Specifically, each input row
+    includes:
+
+     - :py:class:`idCol`: vertex ID
+     - :py:class:`neighborsCol`: neighbors of vertex in :py:class:`idCol`
+     - :py:class:`similaritiesCol`: non-negative weights (similarities) of edges between the
+        vertex in :py:class:`idCol` and each neighbor in :py:class:`neighborsCol`
+
+    PIC returns a cluster assignment for each input vertex.  It appends a new column
+    :py:class:`predictionCol` containing the cluster assignment in :py:class:`[0,k)` for
+    each row (vertex).
+
+    Notes:
+
+     - [[PowerIterationClustering]] is a transformer with an expensive [[transform]] operation.
+        Transform runs the iterative PIC algorithm to cluster the whole input dataset.
+     - Input validation: This validates that similarities are non-negative but does NOT validate
+        that the input matrix is symmetric.
+
+    @see <a href=http://en.wikipedia.org/wiki/Spectral_clustering>
+    Spectral clustering (Wikipedia)</a>
+
+    >>> from pyspark.sql.types import ArrayType, DoubleType, LongType, StructField, StructType
+    >>> similarities = [((long)(1), [0], [0.5]), ((long)(2), [0, 1], [0.7,0.5]), \
+                        ((long)(3), [0, 1, 2], [0.9, 0.7, 0.5]), \
+                        ((long)(4), [0, 1, 2, 3], [1.1, 0.9, 0.7,0.5]), \
+                        ((long)(5), [0, 1, 2, 3, 4], [1.3, 1.1, 0.9, 0.7,0.5])]
+    >>> rdd = sc.parallelize(similarities, 2)
+    >>> schema = StructType([StructField("id", LongType(), False), \
+                 StructField("neighbors", ArrayType(LongType(), False), True), \
+                 StructField("similarities", ArrayType(DoubleType(), False), True)])
+    >>> df = spark.createDataFrame(rdd, schema)
+    >>> pic = PowerIterationClustering()
+    >>> result = pic.setK(2).setMaxIter(10).transform(df)
+    >>> predictions = sorted(set([(i[0], i[1]) for i in result.select(result.id, result.prediction)
+    ...     .collect()]), key=lambda x: x[0])
+    >>> predictions[0]
+    (1, 1)
+    >>> predictions[1]
+    (2, 1)
+    >>> predictions[2]
+    (3, 0)
+    >>> predictions[3]
+    (4, 0)
+    >>> predictions[4]
+    (5, 0)
+    >>> pic_path = temp_path + "/pic"
+    >>> pic.save(pic_path)
+    >>> pic2 = PowerIterationClustering.load(pic_path)
+    >>> pic2.getK()
+    2
+    >>> pic2.getMaxIter()
+    10
+    >>> pic3 = PowerIterationClustering(k=4, initMode="degree")
+    >>> pic3.getIdCol()
+    'id'
+    >>> pic3.getK()
+    4
+    >>> pic3.getMaxIter()
+    20
+    >>> pic3.getInitMode()
+    'degree'
+
+    .. versionadded:: 2.4.0
+    """
+
+    k = Param(Params._dummy(), "k",
+              "The number of clusters to create. Must be > 1.",
+              typeConverter=TypeConverters.toInt)
+    initMode = Param(Params._dummy(), "initMode",
+                     "The initialization algorithm. This can be either " +
+                     "'random' to use a random vector as vertex properties, or 'degree' to use " +
+                     "a normalized sum of similarities with other vertices.  Supported options: " +
+                     "'random' and 'degree'.",
+                     typeConverter=TypeConverters.toString)
+    idCol = Param(Params._dummy(), "idCol",
+                  "Name of the input column for vertex IDs.",
+                  typeConverter=TypeConverters.toString)
+    neighborsCol = Param(Params._dummy(), "neighborsCol",
+                         "Name of the input column for neighbors in the adjacency list " +
+                         "representation.",
+                         typeConverter=TypeConverters.toString)
+    similaritiesCol = Param(Params._dummy(), "similaritiesCol",
+                            "Name of the input column for non-negative weights (similarities) " +
+                            "of edges between the vertex in `idCol` and each neighbor in " +
+                            "`neighborsCol`",
+                            typeConverter=TypeConverters.toString)
+
+    @keyword_only
+    def __init__(self, predictionCol="prediction", k=2, maxIter=20, initMode="random",
+                 idCol="id", neighborsCol="neighbors", similaritiesCol="similarities"):
+        """
+        __init__(self, predictionCol="prediction", k=2, maxIter=20, initMode="random",\
+                 idCol="id", neighborsCol="neighbors", similaritiesCol="similarities")
+        """
+        super(PowerIterationClustering, self).__init__()
+        self._java_obj = self._new_java_obj(
+            "org.apache.spark.ml.clustering.PowerIterationClustering", self.uid)
+        self._setDefault(k=2, maxIter=20, initMode="random", idCol="id", neighborsCol="neighbors",
+                         similaritiesCol="similarities")
+        kwargs = self._input_kwargs
+        self.setParams(**kwargs)
+
+    @keyword_only
+    @since("2.4.0")
+    def setParams(self, predictionCol="prediction", k=2, maxIter=20, initMode="random",
+                  idCol="id", neighborsCol="neighbors", similaritiesCol="similarities"):
+        """
+        setParams(self, predictionCol="prediction", k=2, maxIter=20, initMode="random",\
+                  idCol="id", neighborsCol="neighbors", similaritiesCol="similarities")
+        Sets params for PowerIterationClustering.
+        """
+        kwargs = self._input_kwargs
+        return self._set(**kwargs)
+
+    @since("2.4.0")
+    def setK(self, value):
+        """
+        Sets the value of :py:attr:`k`.
+        """
+        return self._set(k=value)
+
+    @since("2.4.0")
+    def getK(self):
+        """
+        Gets the value of :py:attr:`k`.
+        """
+        return self.getOrDefault(self.k)
+
+    @since("2.4.0")
+    def setInitMode(self, value):
+        """
+        Sets the value of :py:attr:`initMode`.
+        """
+        return self._set(initMode=value)
+
+    @since("2.4.0")
+    def getInitMode(self):
+        """
+        Gets the value of `initMode`
+        """
+        return self.getOrDefault(self.initMode)
+
+    @since("2.4.0")
+    def setIdCol(self, value):
+        """
+        Sets the value of :py:attr:`idCol`.
+        """
+        return self._set(idCol=value)
+
+    @since("2.4.0")
+    def getIdCol(self):
+        """
+        Gets the value of :py:attr:`idCol`.
+        """
+        return self.getOrDefault(self.idCol)
+
+    @since("2.4.0")
+    def setNeighborsCol(self, value):
+        """
+        Sets the value of :py:attr:`neighborsCol.
+        """
+        return self._set(neighborsCol=value)
+
+    @since("2.4.0")
+    def getNeighborsCol(self):
+        """
+        Gets the value of :py:attr:`neighborsCol`.
+        """
+        return self.getOrDefault(self.neighborsCol)
+
+    @since("2.4.0")
+    def setSimilaritiesCol(self, value):
+        """
+        Sets the value of :py:attr:`similaritiesCol`.
+        """
+        return self._set(similaritiesCol=value)
+
+    @since("2.4.0")
+    def getSimilaritiesCol(self):
+        """
+        Gets the value of :py:attr:`similaritiesCol`.
+        """
+        return self.getOrDefault(self.binary)
+
+
 if __name__ == "__main__":
     import doctest
     import pyspark.ml.clustering

python/pyspark/ml/tests.py

@@ -1873,6 +1873,53 @@ def test_kmeans_cosine_distance(self):
         self.assertTrue(result[4].prediction == result[5].prediction)
 
 
+class PowerIterationClustering(SparkSessionTestCase):
+
+    def test_power_iteration_clustering(self):
+        from pyspark.sql.types import ArrayType, DoubleType, LongType, StructField, StructType
+        from pyspark.ml.clustering import PowerIterationClustering
+        import math
+
+        def genCircle(r, n):
+            points = []
+            for i in range(0, n):
+                theta = 2.0 * math.pi * i / n
+                points.append((r * math.cos(theta), r * math.sin(theta)))
+            return points
+
+        def sim(x, y):
+            dist = (x[0] - y[0]) * (x[0] - y[0]) + (x[1] - y[1]) * (x[1] - y[1])
+            return math.exp(-dist / 2.0)
+
+        r1 = 1.0
+        n1 = 10
+        r2 = 4.0
+        n2 = 40
+        n = n1 + n2
+        points = genCircle(r1, n1) + genCircle(r2, n2)
+        similarities = []
+        for i in range(1, n):
+            neighbor = []
+            weight = []
+            for j in range(i):
+                neighbor.append((long)(j))
+                weight.append(sim(points[i], points[j]))
+            similarities.append([(long)(i), neighbor, weight])
+        rdd = self.sc.parallelize(similarities, 2)
+        schema = StructType([StructField("id", LongType(), False),
+                            StructField("neighbors", ArrayType(LongType(), False), True),
+                            StructField("similarities", ArrayType(DoubleType(), False), True)])
+        df = self.spark.createDataFrame(rdd, schema)
+        pic = PowerIterationClustering()
+        result = pic.setK(2).setMaxIter(40).transform(df)
+        predictions = sorted(set([(i[0], i[1]) for i in result.select(result.id,
+                             result.prediction).collect()]), key=lambda x: x[0])
+        for i in range(0, 8):
+            self.assertEqual(predictions[i], (i+1, 1))
+        for i in range(9, 48):
+            self.assertEqual(predictions[i], (i+1, 0))
+
+
 class OneVsRestTests(SparkSessionTestCase):
 
     def test_copy(self):