forked from apache/spark
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Vectors.scala
559 lines (492 loc) · 15.9 KB
/
Vectors.scala
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
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
/*
* 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.mllib.linalg
import java.util
import java.lang.{Double => JavaDouble, Integer => JavaInteger, Iterable => JavaIterable}
import scala.annotation.varargs
import scala.collection.JavaConverters._
import breeze.linalg.{DenseVector => BDV, SparseVector => BSV, Vector => BV}
import org.apache.spark.SparkException
import org.apache.spark.annotation.DeveloperApi
import org.apache.spark.mllib.util.NumericParser
import org.apache.spark.sql.Row
import org.apache.spark.sql.catalyst.expressions.GenericMutableRow
import org.apache.spark.sql.types._
/**
* Represents a numeric vector, whose index type is Int and value type is Double.
*
* Note: Users should not implement this interface.
*/
@SQLUserDefinedType(udt = classOf[VectorUDT])
sealed trait Vector extends Serializable {
/**
* Size of the vector.
*/
def size: Int
/**
* Converts the instance to a double array.
*/
def toArray: Array[Double]
override def equals(other: Any): Boolean = {
other match {
case v2: Vector => {
if (this.size != v2.size) return false
(this, v2) match {
case (s1: SparseVector, s2: SparseVector) =>
Vectors.equals(s1.indices, s1.values, s2.indices, s2.values)
case (s1: SparseVector, d1: DenseVector) =>
Vectors.equals(s1.indices, s1.values, 0 until d1.size, d1.values)
case (d1: DenseVector, s1: SparseVector) =>
Vectors.equals(0 until d1.size, d1.values, s1.indices, s1.values)
case (_, _) => util.Arrays.equals(this.toArray, v2.toArray)
}
}
case _ => false
}
}
override def hashCode(): Int = {
var result: Int = size + 31
this.foreachActive { case (index, value) =>
// ignore explict 0 for comparison between sparse and dense
if (value != 0) {
result = 31 * result + index
// refer to {@link java.util.Arrays.equals} for hash algorithm
val bits = java.lang.Double.doubleToLongBits(value)
result = 31 * result + (bits ^ (bits >>> 32)).toInt
}
}
result
}
/**
* Converts the instance to a breeze vector.
*/
private[mllib] def toBreeze: BV[Double]
/**
* Gets the value of the ith element.
* @param i index
*/
def apply(i: Int): Double = toBreeze(i)
/**
* Makes a deep copy of this vector.
*/
def copy: Vector = {
throw new NotImplementedError(s"copy is not implemented for ${this.getClass}.")
}
/**
* Applies a function `f` to all the active elements of dense and sparse vector.
*
* @param f the function takes two parameters where the first parameter is the index of
* the vector with type `Int`, and the second parameter is the corresponding value
* with type `Double`.
*/
private[spark] def foreachActive(f: (Int, Double) => Unit)
}
/**
* :: DeveloperApi ::
*
* User-defined type for [[Vector]] which allows easy interaction with SQL
* via [[org.apache.spark.sql.DataFrame]].
*
* NOTE: This is currently private[spark] but will be made public later once it is stabilized.
*/
@DeveloperApi
private[spark] class VectorUDT extends UserDefinedType[Vector] {
override def sqlType: StructType = {
// type: 0 = sparse, 1 = dense
// We only use "values" for dense vectors, and "size", "indices", and "values" for sparse
// vectors. The "values" field is nullable because we might want to add binary vectors later,
// which uses "size" and "indices", but not "values".
StructType(Seq(
StructField("type", ByteType, nullable = false),
StructField("size", IntegerType, nullable = true),
StructField("indices", ArrayType(IntegerType, containsNull = false), nullable = true),
StructField("values", ArrayType(DoubleType, containsNull = false), nullable = true)))
}
override def serialize(obj: Any): Row = {
val row = new GenericMutableRow(4)
obj match {
case SparseVector(size, indices, values) =>
row.setByte(0, 0)
row.setInt(1, size)
row.update(2, indices.toSeq)
row.update(3, values.toSeq)
case DenseVector(values) =>
row.setByte(0, 1)
row.setNullAt(1)
row.setNullAt(2)
row.update(3, values.toSeq)
}
row
}
override def deserialize(datum: Any): Vector = {
datum match {
// TODO: something wrong with UDT serialization
case v: Vector =>
v
case row: Row =>
require(row.length == 4,
s"VectorUDT.deserialize given row with length ${row.length} but requires length == 4")
val tpe = row.getByte(0)
tpe match {
case 0 =>
val size = row.getInt(1)
val indices = row.getAs[Iterable[Int]](2).toArray
val values = row.getAs[Iterable[Double]](3).toArray
new SparseVector(size, indices, values)
case 1 =>
val values = row.getAs[Iterable[Double]](3).toArray
new DenseVector(values)
}
}
}
override def pyUDT: String = "pyspark.mllib.linalg.VectorUDT"
override def userClass: Class[Vector] = classOf[Vector]
override def equals(o: Any): Boolean = {
o match {
case v: VectorUDT => true
case _ => false
}
}
override def hashCode: Int = 7919
override def typeName: String = "vector"
private[spark] override def asNullable: VectorUDT = this
}
/**
* Factory methods for [[org.apache.spark.mllib.linalg.Vector]].
* We don't use the name `Vector` because Scala imports
* [[scala.collection.immutable.Vector]] by default.
*/
object Vectors {
/**
* Creates a dense vector from its values.
*/
@varargs
def dense(firstValue: Double, otherValues: Double*): Vector =
new DenseVector((firstValue +: otherValues).toArray)
// A dummy implicit is used to avoid signature collision with the one generated by @varargs.
/**
* Creates a dense vector from a double array.
*/
def dense(values: Array[Double]): Vector = new DenseVector(values)
/**
* Creates a sparse vector providing its index array and value array.
*
* @param size vector size.
* @param indices index array, must be strictly increasing.
* @param values value array, must have the same length as indices.
*/
def sparse(size: Int, indices: Array[Int], values: Array[Double]): Vector =
new SparseVector(size, indices, values)
/**
* Creates a sparse vector using unordered (index, value) pairs.
*
* @param size vector size.
* @param elements vector elements in (index, value) pairs.
*/
def sparse(size: Int, elements: Seq[(Int, Double)]): Vector = {
require(size > 0)
val (indices, values) = elements.sortBy(_._1).unzip
var prev = -1
indices.foreach { i =>
require(prev < i, s"Found duplicate indices: $i.")
prev = i
}
require(prev < size)
new SparseVector(size, indices.toArray, values.toArray)
}
/**
* Creates a sparse vector using unordered (index, value) pairs in a Java friendly way.
*
* @param size vector size.
* @param elements vector elements in (index, value) pairs.
*/
def sparse(size: Int, elements: JavaIterable[(JavaInteger, JavaDouble)]): Vector = {
sparse(size, elements.asScala.map { case (i, x) =>
(i.intValue(), x.doubleValue())
}.toSeq)
}
/**
* Creates a vector of all zeros.
*
* @param size vector size
* @return a zero vector
*/
def zeros(size: Int): Vector = {
new DenseVector(new Array[Double](size))
}
/**
* Parses a string resulted from [[Vector.toString]] into a [[Vector]].
*/
def parse(s: String): Vector = {
parseNumeric(NumericParser.parse(s))
}
private[mllib] def parseNumeric(any: Any): Vector = {
any match {
case values: Array[Double] =>
Vectors.dense(values)
case Seq(size: Double, indices: Array[Double], values: Array[Double]) =>
Vectors.sparse(size.toInt, indices.map(_.toInt), values)
case other =>
throw new SparkException(s"Cannot parse $other.")
}
}
/**
* Creates a vector instance from a breeze vector.
*/
private[mllib] def fromBreeze(breezeVector: BV[Double]): Vector = {
breezeVector match {
case v: BDV[Double] =>
if (v.offset == 0 && v.stride == 1 && v.length == v.data.length) {
new DenseVector(v.data)
} else {
new DenseVector(v.toArray) // Can't use underlying array directly, so make a new one
}
case v: BSV[Double] =>
if (v.index.length == v.used) {
new SparseVector(v.length, v.index, v.data)
} else {
new SparseVector(v.length, v.index.slice(0, v.used), v.data.slice(0, v.used))
}
case v: BV[_] =>
sys.error("Unsupported Breeze vector type: " + v.getClass.getName)
}
}
/**
* Returns the p-norm of this vector.
* @param vector input vector.
* @param p norm.
* @return norm in L^p^ space.
*/
def norm(vector: Vector, p: Double): Double = {
require(p >= 1.0)
val values = vector match {
case DenseVector(vs) => vs
case SparseVector(n, ids, vs) => vs
case v => throw new IllegalArgumentException("Do not support vector type " + v.getClass)
}
val size = values.size
if (p == 1) {
var sum = 0.0
var i = 0
while (i < size) {
sum += math.abs(values(i))
i += 1
}
sum
} else if (p == 2) {
var sum = 0.0
var i = 0
while (i < size) {
sum += values(i) * values(i)
i += 1
}
math.sqrt(sum)
} else if (p == Double.PositiveInfinity) {
var max = 0.0
var i = 0
while (i < size) {
val value = math.abs(values(i))
if (value > max) max = value
i += 1
}
max
} else {
var sum = 0.0
var i = 0
while (i < size) {
sum += math.pow(math.abs(values(i)), p)
i += 1
}
math.pow(sum, 1.0 / p)
}
}
/**
* Returns the squared distance between two Vectors.
* @param v1 first Vector.
* @param v2 second Vector.
* @return squared distance between two Vectors.
*/
def sqdist(v1: Vector, v2: Vector): Double = {
require(v1.size == v2.size, "vector dimension mismatch")
var squaredDistance = 0.0
(v1, v2) match {
case (v1: SparseVector, v2: SparseVector) =>
val v1Values = v1.values
val v1Indices = v1.indices
val v2Values = v2.values
val v2Indices = v2.indices
val nnzv1 = v1Indices.size
val nnzv2 = v2Indices.size
var kv1 = 0
var kv2 = 0
while (kv1 < nnzv1 || kv2 < nnzv2) {
var score = 0.0
if (kv2 >= nnzv2 || (kv1 < nnzv1 && v1Indices(kv1) < v2Indices(kv2))) {
score = v1Values(kv1)
kv1 += 1
} else if (kv1 >= nnzv1 || (kv2 < nnzv2 && v2Indices(kv2) < v1Indices(kv1))) {
score = v2Values(kv2)
kv2 += 1
} else {
score = v1Values(kv1) - v2Values(kv2)
kv1 += 1
kv2 += 1
}
squaredDistance += score * score
}
case (v1: SparseVector, v2: DenseVector) =>
squaredDistance = sqdist(v1, v2)
case (v1: DenseVector, v2: SparseVector) =>
squaredDistance = sqdist(v2, v1)
case (DenseVector(vv1), DenseVector(vv2)) =>
var kv = 0
val sz = vv1.size
while (kv < sz) {
val score = vv1(kv) - vv2(kv)
squaredDistance += score * score
kv += 1
}
case _ =>
throw new IllegalArgumentException("Do not support vector type " + v1.getClass +
" and " + v2.getClass)
}
squaredDistance
}
/**
* Returns the squared distance between DenseVector and SparseVector.
*/
private[mllib] def sqdist(v1: SparseVector, v2: DenseVector): Double = {
var kv1 = 0
var kv2 = 0
val indices = v1.indices
var squaredDistance = 0.0
val nnzv1 = indices.size
val nnzv2 = v2.size
var iv1 = if (nnzv1 > 0) indices(kv1) else -1
while (kv2 < nnzv2) {
var score = 0.0
if (kv2 != iv1) {
score = v2(kv2)
} else {
score = v1.values(kv1) - v2(kv2)
if (kv1 < nnzv1 - 1) {
kv1 += 1
iv1 = indices(kv1)
}
}
squaredDistance += score * score
kv2 += 1
}
squaredDistance
}
/**
* Check equality between sparse/dense vectors
*/
private[mllib] def equals(
v1Indices: IndexedSeq[Int],
v1Values: Array[Double],
v2Indices: IndexedSeq[Int],
v2Values: Array[Double]): Boolean = {
val v1Size = v1Values.size
val v2Size = v2Values.size
var k1 = 0
var k2 = 0
var allEqual = true
while (allEqual) {
while (k1 < v1Size && v1Values(k1) == 0) k1 += 1
while (k2 < v2Size && v2Values(k2) == 0) k2 += 1
if (k1 >= v1Size || k2 >= v2Size) {
return k1 >= v1Size && k2 >= v2Size // check end alignment
}
allEqual = v1Indices(k1) == v2Indices(k2) && v1Values(k1) == v2Values(k2)
k1 += 1
k2 += 1
}
allEqual
}
}
/**
* A dense vector represented by a value array.
*/
@SQLUserDefinedType(udt = classOf[VectorUDT])
class DenseVector(val values: Array[Double]) extends Vector {
override def size: Int = values.length
override def toString: String = values.mkString("[", ",", "]")
override def toArray: Array[Double] = values
private[mllib] override def toBreeze: BV[Double] = new BDV[Double](values)
override def apply(i: Int): Double = values(i)
override def copy: DenseVector = {
new DenseVector(values.clone())
}
private[spark] override def foreachActive(f: (Int, Double) => Unit) = {
var i = 0
val localValuesSize = values.size
val localValues = values
while (i < localValuesSize) {
f(i, localValues(i))
i += 1
}
}
}
object DenseVector {
/** Extracts the value array from a dense vector. */
def unapply(dv: DenseVector): Option[Array[Double]] = Some(dv.values)
}
/**
* A sparse vector represented by an index array and an value array.
*
* @param size size of the vector.
* @param indices index array, assume to be strictly increasing.
* @param values value array, must have the same length as the index array.
*/
@SQLUserDefinedType(udt = classOf[VectorUDT])
class SparseVector(
override val size: Int,
val indices: Array[Int],
val values: Array[Double]) extends Vector {
require(indices.length == values.length)
override def toString: String =
"(%s,%s,%s)".format(size, indices.mkString("[", ",", "]"), values.mkString("[", ",", "]"))
override def toArray: Array[Double] = {
val data = new Array[Double](size)
var i = 0
val nnz = indices.length
while (i < nnz) {
data(indices(i)) = values(i)
i += 1
}
data
}
override def copy: SparseVector = {
new SparseVector(size, indices.clone(), values.clone())
}
private[mllib] override def toBreeze: BV[Double] = new BSV[Double](indices, values, size)
private[spark] override def foreachActive(f: (Int, Double) => Unit) = {
var i = 0
val localValuesSize = values.size
val localIndices = indices
val localValues = values
while (i < localValuesSize) {
f(localIndices(i), localValues(i))
i += 1
}
}
}
object SparseVector {
def unapply(sv: SparseVector): Option[(Int, Array[Int], Array[Double])] =
Some((sv.size, sv.indices, sv.values))
}