Source code for synapse.ml.nn.ConditionalKNNModel

# Copyright (C) Microsoft Corporation. All rights reserved.
# Licensed under the MIT License. See LICENSE in project root for information.


import sys
if sys.version >= '3':
    basestring = str

from pyspark import SparkContext, SQLContext
from pyspark.sql import DataFrame
from pyspark.ml.param.shared import *
from pyspark import keyword_only
from pyspark.ml.util import JavaMLReadable, JavaMLWritable
from synapse.ml.core.serialize.java_params_patch import *
from pyspark.ml.wrapper import JavaTransformer, JavaEstimator, JavaModel
from pyspark.ml.evaluation import JavaEvaluator
from pyspark.ml.common import inherit_doc
from synapse.ml.core.schema.Utils import *
from pyspark.ml.param import TypeConverters
from synapse.ml.core.schema.TypeConversionUtils import generateTypeConverter, complexTypeConverter


[docs]@inherit_doc class ConditionalKNNModel(ComplexParamsMixin, JavaMLReadable, JavaMLWritable, JavaModel): """ Args: ballTree (object): the ballTree model used for perfoming queries conditionerCol (object): column holding identifiers for features that will be returned when queried featuresCol (object): The name of the features column k (int): number of matches to return labelCol (object): The name of the label column leafSize (int): max size of the leaves of the tree outputCol (object): The name of the output column valuesCol (object): column holding values for each feature (key) that will be returned when queried """ ballTree = Param(Params._dummy(), "ballTree", "the ballTree model used for perfoming queries") conditionerCol = Param(Params._dummy(), "conditionerCol", "column holding identifiers for features that will be returned when queried") featuresCol = Param(Params._dummy(), "featuresCol", "The name of the features column") k = Param(Params._dummy(), "k", "number of matches to return", typeConverter=TypeConverters.toInt) labelCol = Param(Params._dummy(), "labelCol", "The name of the label column") leafSize = Param(Params._dummy(), "leafSize", "max size of the leaves of the tree", typeConverter=TypeConverters.toInt) outputCol = Param(Params._dummy(), "outputCol", "The name of the output column") valuesCol = Param(Params._dummy(), "valuesCol", "column holding values for each feature (key) that will be returned when queried") @keyword_only def __init__( self, java_obj=None, ballTree=None, conditionerCol=None, featuresCol=None, k=None, labelCol=None, leafSize=None, outputCol=None, valuesCol=None ): super(ConditionalKNNModel, self).__init__() if java_obj is None: self._java_obj = self._new_java_obj("com.microsoft.azure.synapse.ml.nn.ConditionalKNNModel", self.uid) else: self._java_obj = java_obj if hasattr(self, "_input_kwargs"): kwargs = self._input_kwargs else: kwargs = self.__init__._input_kwargs if java_obj is None: for k,v in kwargs.items(): if v is not None: getattr(self, "set" + k[0].upper() + k[1:])(v)
[docs] @keyword_only def setParams( self, ballTree=None, conditionerCol=None, featuresCol=None, k=None, labelCol=None, leafSize=None, outputCol=None, valuesCol=None ): """ Set the (keyword only) parameters """ if hasattr(self, "_input_kwargs"): kwargs = self._input_kwargs else: kwargs = self.__init__._input_kwargs return self._set(**kwargs)
[docs] @classmethod def read(cls): """ Returns an MLReader instance for this class. """ return JavaMMLReader(cls)
[docs] @staticmethod def getJavaPackage(): """ Returns package name String. """ return "com.microsoft.azure.synapse.ml.nn.ConditionalKNNModel"
@staticmethod def _from_java(java_stage): module_name=ConditionalKNNModel.__module__ module_name=module_name.rsplit(".", 1)[0] + ".ConditionalKNNModel" return from_java(java_stage, module_name)
[docs] def setBallTree(self, value): """ Args: ballTree: the ballTree model used for perfoming queries """ self._set(ballTree=value) return self
[docs] def setConditionerCol(self, value): """ Args: conditionerCol: column holding identifiers for features that will be returned when queried """ self._set(conditionerCol=value) return self
[docs] def setFeaturesCol(self, value): """ Args: featuresCol: The name of the features column """ self._set(featuresCol=value) return self
[docs] def setK(self, value): """ Args: k: number of matches to return """ self._set(k=value) return self
[docs] def setLabelCol(self, value): """ Args: labelCol: The name of the label column """ self._set(labelCol=value) return self
[docs] def setLeafSize(self, value): """ Args: leafSize: max size of the leaves of the tree """ self._set(leafSize=value) return self
[docs] def setOutputCol(self, value): """ Args: outputCol: The name of the output column """ self._set(outputCol=value) return self
[docs] def setValuesCol(self, value): """ Args: valuesCol: column holding values for each feature (key) that will be returned when queried """ self._set(valuesCol=value) return self
[docs] def getBallTree(self): """ Returns: ballTree: the ballTree model used for perfoming queries """ return self.getOrDefault(self.ballTree)
[docs] def getConditionerCol(self): """ Returns: conditionerCol: column holding identifiers for features that will be returned when queried """ return self.getOrDefault(self.conditionerCol)
[docs] def getFeaturesCol(self): """ Returns: featuresCol: The name of the features column """ return self.getOrDefault(self.featuresCol)
[docs] def getK(self): """ Returns: k: number of matches to return """ return self.getOrDefault(self.k)
[docs] def getLabelCol(self): """ Returns: labelCol: The name of the label column """ return self.getOrDefault(self.labelCol)
[docs] def getLeafSize(self): """ Returns: leafSize: max size of the leaves of the tree """ return self.getOrDefault(self.leafSize)
[docs] def getOutputCol(self): """ Returns: outputCol: The name of the output column """ return self.getOrDefault(self.outputCol)
[docs] def getValuesCol(self): """ Returns: valuesCol: column holding values for each feature (key) that will be returned when queried """ return self.getOrDefault(self.valuesCol)