# Copyright (C) Microsoft Corporation. All rights reserved.
# Licensed under the MIT License. See LICENSE in the project root for information.
import sys
if sys.version >= '3':
basestring = str
from pyspark import keyword_only
from pyspark.ml.common import inherit_doc
from pyspark.ml.evaluation import JavaEvaluator
from pyspark.ml.param import Param, Params, TypeConverters
from pyspark.ml.param.shared import HasLabelCol, HasPredictionCol
from pyspark.ml.util import JavaMLReadable, JavaMLWritable
[docs]@inherit_doc
class RankingEvaluator(JavaEvaluator, HasLabelCol, HasPredictionCol, JavaMLReadable, JavaMLWritable):
metricName = Param(Params._dummy(), "metricName",
"""metric name in evaluation - one of:
map -
ndcgAt -
mapk -
recallAtK -
diversityAtK -
maxDiversity - """,
typeConverter=TypeConverters.toString)
# todo: Should not need this, but not sure how to remove
labelCol = Param(Params._dummy(), "labelCol",
"""labelCol""",
typeConverter=TypeConverters.toString)
rawPredictionCol = Param(Params._dummy(), "rawPredictionCol",
"""rawPredictionCol""",
typeConverter=TypeConverters.toString)
k = Param(Params._dummy(), "k",
"""k""",
typeConverter=TypeConverters.toInt)
nItems = Param(Params._dummy(), "nItems",
"""number of items""",
typeConverter=TypeConverters.toInt)
@keyword_only
def __init__(self, rawPredictionCol="rawPrediction", labelCol="label", metricName="ndcgAt", k=3):
"""
__init__(self, rawPredictionCol="rawPrediction", labelCol="label", \
metricName="ndcgAt")
"""
super(RankingEvaluator, self).__init__()
self._java_obj = self._new_java_obj(
"com.microsoft.ml.spark.RankingEvaluator", self.uid)
self._setDefault(metricName="ndcgAt", k=3)
kwargs = self._input_kwargs
self._set(**kwargs)
[docs] def setRawPredictionCol(self, value):
"""
Sets the value of :py:attr:`rawPredictionCol`.
"""
return self._set(rawPredictionCol=value)
[docs] def getRawPredictionCol(self):
"""
Gets the value of rawPredictionCol or its default value.
"""
return self.getOrDefault(self.rawPredictionCol)
[docs] def setMetricName(self, value):
"""
Sets the value of :py:attr:`metricName`.
"""
return self._set(metricName=value)
[docs] def getMetricName(self):
"""
Gets the value of metricName or its default value.
"""
return self.getOrDefault(self.metricName)
[docs] def setK(self, value):
"""
Sets the value of :py:attr:`metricName`.
"""
return self._set(k=value)
[docs] def getK(self):
"""
Gets the value of metricName or its default value.
"""
return self.getOrDefault(self.k)
[docs] def setNumberItems(self, value):
"""
Sets the value of :py:attr:`nItems`.
"""
return self._set(nItems=value)
[docs] @keyword_only
def setParams(self, rawPredictionCol="rawPrediction", labelCol="label", metricName="ndcgAt", k=3):
"""
setParams(self, rawPredictionCol="rawPrediction", labelCol="label", \
metricName="ndcgAt")
Sets params for binary classification evaluator.
"""
kwargs = self._input_kwargs
return self._set(**kwargs)
[docs] def getMetricsMap(self, dataframe):
return self._call_java("getMetricsMap", dataframe)