1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
|
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.spark.ml.feature
import org.apache.hadoop.fs.Path
import org.apache.spark.annotation.{Experimental, Since}
import org.apache.spark.ml.{Estimator, Model}
import org.apache.spark.ml.param.{ParamMap, Params}
import org.apache.spark.ml.param.shared.{HasInputCol, HasOutputCol}
import org.apache.spark.ml.util._
import org.apache.spark.mllib.linalg.{Vector, Vectors, VectorUDT}
import org.apache.spark.mllib.stat.Statistics
import org.apache.spark.sql._
import org.apache.spark.sql.functions._
import org.apache.spark.sql.types.{StructField, StructType}
/**
* Params for [[MaxAbsScaler]] and [[MaxAbsScalerModel]].
*/
private[feature] trait MaxAbsScalerParams extends Params with HasInputCol with HasOutputCol {
/** Validates and transforms the input schema. */
protected def validateAndTransformSchema(schema: StructType): StructType = {
val inputType = schema($(inputCol)).dataType
require(inputType.isInstanceOf[VectorUDT],
s"Input column ${$(inputCol)} must be a vector column")
require(!schema.fieldNames.contains($(outputCol)),
s"Output column ${$(outputCol)} already exists.")
val outputFields = schema.fields :+ StructField($(outputCol), new VectorUDT, false)
StructType(outputFields)
}
}
/**
* :: Experimental ::
* Rescale each feature individually to range [-1, 1] by dividing through the largest maximum
* absolute value in each feature. It does not shift/center the data, and thus does not destroy
* any sparsity.
*/
@Experimental
class MaxAbsScaler @Since("2.0.0") (override val uid: String)
extends Estimator[MaxAbsScalerModel] with MaxAbsScalerParams with DefaultParamsWritable {
@Since("2.0.0")
def this() = this(Identifiable.randomUID("maxAbsScal"))
/** @group setParam */
def setInputCol(value: String): this.type = set(inputCol, value)
/** @group setParam */
def setOutputCol(value: String): this.type = set(outputCol, value)
override def fit(dataset: DataFrame): MaxAbsScalerModel = {
transformSchema(dataset.schema, logging = true)
val input = dataset.select($(inputCol)).rdd.map { case Row(v: Vector) => v }
val summary = Statistics.colStats(input)
val minVals = summary.min.toArray
val maxVals = summary.max.toArray
val n = minVals.length
val maxAbs = Array.tabulate(n) { i => math.max(math.abs(minVals(i)), math.abs(maxVals(i))) }
copyValues(new MaxAbsScalerModel(uid, Vectors.dense(maxAbs)).setParent(this))
}
override def transformSchema(schema: StructType): StructType = {
validateAndTransformSchema(schema)
}
override def copy(extra: ParamMap): MaxAbsScaler = defaultCopy(extra)
}
@Since("1.6.0")
object MaxAbsScaler extends DefaultParamsReadable[MaxAbsScaler] {
@Since("1.6.0")
override def load(path: String): MaxAbsScaler = super.load(path)
}
/**
* :: Experimental ::
* Model fitted by [[MaxAbsScaler]].
*
*/
@Experimental
class MaxAbsScalerModel private[ml] (
override val uid: String,
val maxAbs: Vector)
extends Model[MaxAbsScalerModel] with MaxAbsScalerParams with MLWritable {
import MaxAbsScalerModel._
/** @group setParam */
def setInputCol(value: String): this.type = set(inputCol, value)
/** @group setParam */
def setOutputCol(value: String): this.type = set(outputCol, value)
override def transform(dataset: DataFrame): DataFrame = {
transformSchema(dataset.schema, logging = true)
// TODO: this looks hack, we may have to handle sparse and dense vectors separately.
val maxAbsUnzero = Vectors.dense(maxAbs.toArray.map(x => if (x == 0) 1 else x))
val reScale = udf { (vector: Vector) =>
val brz = vector.toBreeze / maxAbsUnzero.toBreeze
Vectors.fromBreeze(brz)
}
dataset.withColumn($(outputCol), reScale(col($(inputCol))))
}
override def transformSchema(schema: StructType): StructType = {
validateAndTransformSchema(schema)
}
override def copy(extra: ParamMap): MaxAbsScalerModel = {
val copied = new MaxAbsScalerModel(uid, maxAbs)
copyValues(copied, extra).setParent(parent)
}
@Since("1.6.0")
override def write: MLWriter = new MaxAbsScalerModelWriter(this)
}
@Since("1.6.0")
object MaxAbsScalerModel extends MLReadable[MaxAbsScalerModel] {
private[MaxAbsScalerModel]
class MaxAbsScalerModelWriter(instance: MaxAbsScalerModel) extends MLWriter {
private case class Data(maxAbs: Vector)
override protected def saveImpl(path: String): Unit = {
DefaultParamsWriter.saveMetadata(instance, path, sc)
val data = new Data(instance.maxAbs)
val dataPath = new Path(path, "data").toString
sqlContext.createDataFrame(Seq(data)).repartition(1).write.parquet(dataPath)
}
}
private class MaxAbsScalerModelReader extends MLReader[MaxAbsScalerModel] {
private val className = classOf[MaxAbsScalerModel].getName
override def load(path: String): MaxAbsScalerModel = {
val metadata = DefaultParamsReader.loadMetadata(path, sc, className)
val dataPath = new Path(path, "data").toString
val Row(maxAbs: Vector) = sqlContext.read.parquet(dataPath)
.select("maxAbs")
.head()
val model = new MaxAbsScalerModel(metadata.uid, maxAbs)
DefaultParamsReader.getAndSetParams(model, metadata)
model
}
}
@Since("1.6.0")
override def read: MLReader[MaxAbsScalerModel] = new MaxAbsScalerModelReader
@Since("1.6.0")
override def load(path: String): MaxAbsScalerModel = super.load(path)
}
|
