# Copyright (C) Microsoft Corporation. All rights reserved.
# Licensed under the MIT License. See LICENSE in project root for information.
from mmlspark.vw._VowpalWabbitContextualBandit import _VowpalWabbitContextualBandit, _VowpalWabbitContextualBanditModel
from pyspark.ml.common import inherit_doc
from pyspark import SparkContext, SQLContext
from pyspark.sql import DataFrame
from pyspark.ml.wrapper import JavaWrapper
from pyspark.ml.common import _py2java, _java2py
[docs]def to_java_params(sc, model, pyParamMap):
paramMap = JavaWrapper._new_java_obj("org.apache.spark.ml.param.ParamMap")
for param, value in pyParamMap.items():
java_param = model._java_obj.getParam(param.name)
java_value = _py2java(sc, value)
paramMap.put([java_param.w(java_value)])
return paramMap
[docs]@inherit_doc
class VowpalWabbitContextualBandit(_VowpalWabbitContextualBandit):
def _create_model(self, java_model):
model = VowpalWabbitContextualBanditModel()
model._java_obj = java_model
model._transfer_params_from_java()
return model
[docs] def setInitialModel(self, model):
"""
Initialize the estimator with a previously trained model.
"""
self._java_obj.setInitialModel(model._java_obj.getModel())
[docs] def parallelFit(self, dataset, param_maps):
sc = SparkContext._active_spark_context
self._transfer_params_to_java()
javaParamMaps = [to_java_params(sc, self, x) for x in param_maps]
javaModels = self._java_obj.parallelFit(dataset._jdf, javaParamMaps)
return [self._copyValues(self._create_model(x)) for x in javaModels]
[docs]@inherit_doc
class VowpalWabbitContextualBanditModel(_VowpalWabbitContextualBanditModel):
[docs] def saveNativeModel(self, filename):
"""
Save the native model to a local or WASB remote location.
"""
self._java_obj.saveNativeModel(filename)
[docs] def getNativeModel(self):
"""
Get the binary native VW model.
"""
return self._java_obj.getModel()
[docs] def getReadableModel(self):
return self._java_obj.getReadableModel()
[docs] def getPerformanceStatistics(self):
ctx = SparkContext._active_spark_context
sql_ctx = SQLContext.getOrCreate(ctx)
return DataFrame(self._java_obj.getPerformanceStatistics(), sql_ctx)