# 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.ml.param.shared import *
from pyspark import keyword_only
from pyspark.ml.util import JavaMLReadable, JavaMLWritable
from pyspark.ml.wrapper import JavaTransformer, JavaEstimator, JavaModel
from pyspark.ml.common import inherit_doc
from mmlspark.Utils import *
[docs]@inherit_doc
class PartitionSample(ComplexParamsMixin, JavaMLReadable, JavaMLWritable, JavaTransformer):
"""
Sampling mode. The options are:
- AssignToPartition
- RandomSample
- Head
The default is RandomSample.
Relevant parameters for the different modes are:
- When the mode is AssignToPartition:
- seed - the seed for random partition assignment.
- numParts - the number of partitions. The Default is 10.
- newColName - the name of the partition column.
The default is "Partition".
- When the mode is RandomSample:
- mode - Absolute or Percentage
- count - the number of rows to assign to each partition when
Absolute
- percent - the percentage per partition when Percentage
- When the mode is Head:
- count - the number of rows
Args:
count (long): Number of rows to return (default: 1000)
mode (str): AssignToPartition, RandomSample, or Head (default: RandomSample)
newColName (str): Name of the partition column (default: Partition)
numParts (int): Number of partitions (default: 10)
percent (double): Percent of rows to return (default: 0.01)
rsMode (str): Absolute or Percentage (default: Percentage)
seed (long): Seed for random operations (default: -1)
"""
@keyword_only
def __init__(self, count=1000, mode="RandomSample", newColName="Partition", numParts=10, percent=0.01, rsMode="Percentage", seed=-1):
super(PartitionSample, self).__init__()
self._java_obj = self._new_java_obj("com.microsoft.ml.spark.PartitionSample")
self.count = Param(self, "count", "count: Number of rows to return (default: 1000)")
self._setDefault(count=1000)
self.mode = Param(self, "mode", "mode: AssignToPartition, RandomSample, or Head (default: RandomSample)")
self._setDefault(mode="RandomSample")
self.newColName = Param(self, "newColName", "newColName: Name of the partition column (default: Partition)")
self._setDefault(newColName="Partition")
self.numParts = Param(self, "numParts", "numParts: Number of partitions (default: 10)")
self._setDefault(numParts=10)
self.percent = Param(self, "percent", "percent: Percent of rows to return (default: 0.01)")
self._setDefault(percent=0.01)
self.rsMode = Param(self, "rsMode", "rsMode: Absolute or Percentage (default: Percentage)")
self._setDefault(rsMode="Percentage")
self.seed = Param(self, "seed", "seed: Seed for random operations (default: -1)")
self._setDefault(seed=-1)
if hasattr(self, "_input_kwargs"):
kwargs = self._input_kwargs
else:
kwargs = self.__init__._input_kwargs
self.setParams(**kwargs)
[docs] @keyword_only
def setParams(self, count=1000, mode="RandomSample", newColName="Partition", numParts=10, percent=0.01, rsMode="Percentage", seed=-1):
"""
Set the (keyword only) parameters
Args:
count (long): Number of rows to return (default: 1000)
mode (str): AssignToPartition, RandomSample, or Head (default: RandomSample)
newColName (str): Name of the partition column (default: Partition)
numParts (int): Number of partitions (default: 10)
percent (double): Percent of rows to return (default: 0.01)
rsMode (str): Absolute or Percentage (default: Percentage)
seed (long): Seed for random operations (default: -1)
"""
if hasattr(self, "_input_kwargs"):
kwargs = self._input_kwargs
else:
kwargs = self.__init__._input_kwargs
return self._set(**kwargs)
[docs] def setCount(self, value):
"""
Args:
count (long): Number of rows to return (default: 1000)
"""
self._set(count=value)
return self
[docs] def getCount(self):
"""
Returns:
long: Number of rows to return (default: 1000)
"""
return self.getOrDefault(self.count)
[docs] def setMode(self, value):
"""
Args:
mode (str): AssignToPartition, RandomSample, or Head (default: RandomSample)
"""
self._set(mode=value)
return self
[docs] def getMode(self):
"""
Returns:
str: AssignToPartition, RandomSample, or Head (default: RandomSample)
"""
return self.getOrDefault(self.mode)
[docs] def setNewColName(self, value):
"""
Args:
newColName (str): Name of the partition column (default: Partition)
"""
self._set(newColName=value)
return self
[docs] def getNewColName(self):
"""
Returns:
str: Name of the partition column (default: Partition)
"""
return self.getOrDefault(self.newColName)
[docs] def setNumParts(self, value):
"""
Args:
numParts (int): Number of partitions (default: 10)
"""
self._set(numParts=value)
return self
[docs] def getNumParts(self):
"""
Returns:
int: Number of partitions (default: 10)
"""
return self.getOrDefault(self.numParts)
[docs] def setPercent(self, value):
"""
Args:
percent (double): Percent of rows to return (default: 0.01)
"""
self._set(percent=value)
return self
[docs] def getPercent(self):
"""
Returns:
double: Percent of rows to return (default: 0.01)
"""
return self.getOrDefault(self.percent)
[docs] def setRsMode(self, value):
"""
Args:
rsMode (str): Absolute or Percentage (default: Percentage)
"""
self._set(rsMode=value)
return self
[docs] def getRsMode(self):
"""
Returns:
str: Absolute or Percentage (default: Percentage)
"""
return self.getOrDefault(self.rsMode)
[docs] def setSeed(self, value):
"""
Args:
seed (long): Seed for random operations (default: -1)
"""
self._set(seed=value)
return self
[docs] def getSeed(self):
"""
Returns:
long: Seed for random operations (default: -1)
"""
return self.getOrDefault(self.seed)
[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.ml.spark.PartitionSample"
@staticmethod
def _from_java(java_stage):
module_name=PartitionSample.__module__
module_name=module_name.rsplit(".", 1)[0] + ".PartitionSample"
return from_java(java_stage, module_name)