Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/10152#discussion_r50961556
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/feature/Word2Vec.scala ---
@@ -289,17 +301,28 @@ class Word2Vec extends Serializable with Logging {
val expTable = sc.broadcast(createExpTable())
val bcVocab = sc.broadcast(vocab)
val bcVocabHash = sc.broadcast(vocabHash)
-
- val sentences: RDD[Array[Int]] = words.mapPartitions { iter =>
+ // each partition is a collection of sentences, will be translated
into arrays of Index integer
+ val sentences: RDD[Array[Int]] = dataset.mapPartitions { sentenceIter
=>
new Iterator[Array[Int]] {
- def hasNext: Boolean = iter.hasNext
+ var wordIter: Iterator[String] = null
+
+ def hasNext: Boolean = sentenceIter.hasNext || (wordIter != null
&& wordIter.hasNext)
def next(): Array[Int] = {
val sentence = ArrayBuilder.make[Int]
var sentenceLength = 0
- while (iter.hasNext && sentenceLength < MAX_SENTENCE_LENGTH) {
- val word = bcVocabHash.value.get(iter.next())
- word match {
+ // do translation of each word into its index in the vocabulary,
--- End diff --
Good point; I'm not sure if it was for efficiency or simply to copy the C++
implementation. Hm, I don't suppose we have benchmarks here. I'm OK with
keeping the longer more efficient implementation. but now that input sentence
boundaries matter, this can still be simplified. This part can still be used:
```
// Each input sentence will produce 1 or more Array[Int], so
flatMapPartitions
dataset.flatMapPartitions { sentenceIter =>
// Each sentence will map to 1 or more Array[Int], so map
sentenceIter.map { sentence =>
...
}
}
```
That is you can still `flatMapPartitions` over `Iterator`s for each
sentence to join them without trying to manage both the `Iterator` over
sentences and words.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
