From d584a2b8ac57eff3bf230c760e5bda205c6ea747 Mon Sep 17 00:00:00 2001
From: Kai Jiang <jiangkai@gmail.com>
Date: Thu, 28 Apr 2016 14:19:11 -0700
Subject: [SPARK-12810][PYSPARK] PySpark CrossValidatorModel should support
 avgMetrics

## What changes were proposed in this pull request?
support avgMetrics in CrossValidatorModel with Python
## How was this patch tested?
Doctest and `test_save_load` in `pyspark/ml/test.py`
[JIRA](https://issues.apache.org/jira/browse/SPARK-12810)

Author: Kai Jiang <jiangkai@gmail.com>

Closes #12464 from vectorijk/spark-12810.
---
 python/pyspark/ml/tests.py  | 27 +++++++++++++++++++++++++++
 python/pyspark/ml/tuning.py | 18 ++++++++++--------
 2 files changed, 37 insertions(+), 8 deletions(-)

diff --git a/python/pyspark/ml/tests.py b/python/pyspark/ml/tests.py
index ebef656632..36cecd4682 100644
--- a/python/pyspark/ml/tests.py
+++ b/python/pyspark/ml/tests.py
@@ -466,6 +466,31 @@ class InducedErrorEstimator(Estimator, HasInducedError):
 
 class CrossValidatorTests(PySparkTestCase):
 
+    def test_copy(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)
+        cvCopied = cv.copy()
+        self.assertEqual(cv.getEstimator().uid, cvCopied.getEstimator().uid)
+
+        cvModel = cv.fit(dataset)
+        cvModelCopied = cvModel.copy()
+        for index in range(len(cvModel.avgMetrics)):
+            self.assertTrue(abs(cvModel.avgMetrics[index] - cvModelCopied.avgMetrics[index])
+                            < 0.0001)
+
     def test_fit_minimize_metric(self):
         sqlContext = SQLContext(self.sc)
         dataset = sqlContext.createDataFrame([
@@ -539,6 +564,8 @@ class CrossValidatorTests(PySparkTestCase):
         cvModel.save(cvModelPath)
         loadedModel = CrossValidatorModel.load(cvModelPath)
         self.assertEqual(loadedModel.bestModel.uid, cvModel.bestModel.uid)
+        for index in range(len(loadedModel.avgMetrics)):
+            self.assertTrue(abs(loadedModel.avgMetrics[index] - cvModel.avgMetrics[index]) < 0.0001)
 
 
 class TrainValidationSplitTests(PySparkTestCase):
diff --git a/python/pyspark/ml/tuning.py b/python/pyspark/ml/tuning.py
index b16628bc70..22f9680cab 100644
--- a/python/pyspark/ml/tuning.py
+++ b/python/pyspark/ml/tuning.py
@@ -248,7 +248,7 @@ class CrossValidator(Estimator, ValidatorParams, MLReadable, MLWritable):
         h = 1.0 / nFolds
         randCol = self.uid + "_rand"
         df = dataset.select("*", rand(seed).alias(randCol))
-        metrics = np.zeros(numModels)
+        metrics = [0.0] * numModels
         for i in range(nFolds):
             validateLB = i * h
             validateUB = (i + 1) * h
@@ -266,7 +266,7 @@ class CrossValidator(Estimator, ValidatorParams, MLReadable, MLWritable):
         else:
             bestIndex = np.argmin(metrics)
         bestModel = est.fit(dataset, epm[bestIndex])
-        return self._copyValues(CrossValidatorModel(bestModel))
+        return self._copyValues(CrossValidatorModel(bestModel, metrics))
 
     @since("1.4.0")
     def copy(self, extra=None):
@@ -346,10 +346,11 @@ class CrossValidatorModel(Model, ValidatorParams, MLReadable, MLWritable):
     .. versionadded:: 1.4.0
     """
 
-    def __init__(self, bestModel):
+    def __init__(self, bestModel, avgMetrics=[]):
         super(CrossValidatorModel, self).__init__()
         #: best model from cross validation
         self.bestModel = bestModel
+        self.avgMetrics = avgMetrics
 
     def _transform(self, dataset):
         return self.bestModel.transform(dataset)
@@ -367,7 +368,9 @@ class CrossValidatorModel(Model, ValidatorParams, MLReadable, MLWritable):
         """
         if extra is None:
             extra = dict()
-        return CrossValidatorModel(self.bestModel.copy(extra))
+        bestModel = self.bestModel.copy(extra)
+        avgMetrics = self.avgMetrics
+        return CrossValidatorModel(bestModel, avgMetrics)
 
     @since("2.0.0")
     def write(self):
@@ -394,9 +397,10 @@ class CrossValidatorModel(Model, ValidatorParams, MLReadable, MLWritable):
 
         # Load information from java_stage to the instance.
         bestModel = JavaParams._from_java(java_stage.bestModel())
+        avgMetrics = list(java_stage.avgMetrics())
         estimator, epms, evaluator = super(CrossValidatorModel, cls)._from_java_impl(java_stage)
         # Create a new instance of this stage.
-        py_stage = cls(bestModel=bestModel)\
+        py_stage = cls(bestModel=bestModel, avgMetrics=avgMetrics)\
             .setEstimator(estimator).setEstimatorParamMaps(epms).setEvaluator(evaluator)
         py_stage._resetUid(java_stage.uid())
         return py_stage
@@ -408,12 +412,10 @@ class CrossValidatorModel(Model, ValidatorParams, MLReadable, MLWritable):
         :return: Java object equivalent to this instance.
         """
 
-        sc = SparkContext._active_spark_context
-
         _java_obj = JavaParams._new_java_obj("org.apache.spark.ml.tuning.CrossValidatorModel",
                                              self.uid,
                                              self.bestModel._to_java(),
-                                             _py2java(sc, []))
+                                             self.avgMetrics)
         estimator, epms, evaluator = super(CrossValidatorModel, self)._to_java_impl()
 
         _java_obj.set("evaluator", evaluator)
-- 
cgit v1.2.3