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
176
177
178
179
180
181
182
183
|
/*
* 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.r
import org.apache.hadoop.fs.Path
import org.json4s._
import org.json4s.JsonDSL._
import org.json4s.jackson.JsonMethods._
import org.apache.spark.ml.{Pipeline, PipelineModel}
import org.apache.spark.ml.classification.{LogisticRegression, LogisticRegressionModel}
import org.apache.spark.ml.feature.{IndexToString, RFormula}
import org.apache.spark.ml.linalg.Vector
import org.apache.spark.ml.r.RWrapperUtils._
import org.apache.spark.ml.util._
import org.apache.spark.sql.{DataFrame, Dataset}
private[r] class LogisticRegressionWrapper private (
val pipeline: PipelineModel,
val features: Array[String],
val labels: Array[String]) extends MLWritable {
import LogisticRegressionWrapper._
private val lrModel: LogisticRegressionModel =
pipeline.stages(1).asInstanceOf[LogisticRegressionModel]
val rFeatures: Array[String] = if (lrModel.getFitIntercept) {
Array("(Intercept)") ++ features
} else {
features
}
val rCoefficients: Array[Double] = {
val numRows = lrModel.coefficientMatrix.numRows
val numCols = lrModel.coefficientMatrix.numCols
val numColsWithIntercept = if (lrModel.getFitIntercept) numCols + 1 else numCols
val coefficients: Array[Double] = new Array[Double](numRows * numColsWithIntercept)
val coefficientVectors: Seq[Vector] = lrModel.coefficientMatrix.rowIter.toSeq
var i = 0
if (lrModel.getFitIntercept) {
while (i < numRows) {
coefficients(i * numColsWithIntercept) = lrModel.interceptVector(i)
System.arraycopy(coefficientVectors(i).toArray, 0,
coefficients, i * numColsWithIntercept + 1, numCols)
i += 1
}
} else {
while (i < numRows) {
System.arraycopy(coefficientVectors(i).toArray, 0,
coefficients, i * numColsWithIntercept, numCols)
i += 1
}
}
coefficients
}
def transform(dataset: Dataset[_]): DataFrame = {
pipeline.transform(dataset)
.drop(PREDICTED_LABEL_INDEX_COL)
.drop(lrModel.getFeaturesCol)
.drop(lrModel.getLabelCol)
}
override def write: MLWriter = new LogisticRegressionWrapper.LogisticRegressionWrapperWriter(this)
}
private[r] object LogisticRegressionWrapper
extends MLReadable[LogisticRegressionWrapper] {
val PREDICTED_LABEL_INDEX_COL = "pred_label_idx"
val PREDICTED_LABEL_COL = "prediction"
def fit( // scalastyle:ignore
data: DataFrame,
formula: String,
regParam: Double,
elasticNetParam: Double,
maxIter: Int,
tol: Double,
family: String,
standardization: Boolean,
thresholds: Array[Double],
weightCol: String,
probabilityCol: String
): LogisticRegressionWrapper = {
val rFormula = new RFormula()
.setFormula(formula)
.setForceIndexLabel(true)
checkDataColumns(rFormula, data)
val rFormulaModel = rFormula.fit(data)
val fitIntercept = rFormula.hasIntercept
// get labels and feature names from output schema
val (features, labels) = getFeaturesAndLabels(rFormulaModel, data)
// assemble and fit the pipeline
val lr = new LogisticRegression()
.setRegParam(regParam)
.setElasticNetParam(elasticNetParam)
.setMaxIter(maxIter)
.setTol(tol)
.setFitIntercept(fitIntercept)
.setFamily(family)
.setStandardization(standardization)
.setWeightCol(weightCol)
.setFeaturesCol(rFormula.getFeaturesCol)
.setLabelCol(rFormula.getLabelCol)
.setProbabilityCol(probabilityCol)
.setPredictionCol(PREDICTED_LABEL_INDEX_COL)
if (thresholds.length > 1) {
lr.setThresholds(thresholds)
} else {
lr.setThreshold(thresholds(0))
}
val idxToStr = new IndexToString()
.setInputCol(PREDICTED_LABEL_INDEX_COL)
.setOutputCol(PREDICTED_LABEL_COL)
.setLabels(labels)
val pipeline = new Pipeline()
.setStages(Array(rFormulaModel, lr, idxToStr))
.fit(data)
new LogisticRegressionWrapper(pipeline, features, labels)
}
override def read: MLReader[LogisticRegressionWrapper] = new LogisticRegressionWrapperReader
override def load(path: String): LogisticRegressionWrapper = super.load(path)
class LogisticRegressionWrapperWriter(instance: LogisticRegressionWrapper) extends MLWriter {
override protected def saveImpl(path: String): Unit = {
val rMetadataPath = new Path(path, "rMetadata").toString
val pipelinePath = new Path(path, "pipeline").toString
val rMetadata = ("class" -> instance.getClass.getName) ~
("features" -> instance.features.toSeq) ~
("labels" -> instance.labels.toSeq)
val rMetadataJson: String = compact(render(rMetadata))
sc.parallelize(Seq(rMetadataJson), 1).saveAsTextFile(rMetadataPath)
instance.pipeline.save(pipelinePath)
}
}
class LogisticRegressionWrapperReader extends MLReader[LogisticRegressionWrapper] {
override def load(path: String): LogisticRegressionWrapper = {
implicit val format = DefaultFormats
val rMetadataPath = new Path(path, "rMetadata").toString
val pipelinePath = new Path(path, "pipeline").toString
val rMetadataStr = sc.textFile(rMetadataPath, 1).first()
val rMetadata = parse(rMetadataStr)
val features = (rMetadata \ "features").extract[Array[String]]
val labels = (rMetadata \ "labels").extract[Array[String]]
val pipeline = PipelineModel.load(pipelinePath)
new LogisticRegressionWrapper(pipeline, features, labels)
}
}
}
|
