@Deprecated
public abstract class Builder
extends java.lang.Object
Modifier and Type | Field and Description |
---|---|
protected int |
batchSize
Deprecated.
|
protected VocabCache |
cache
Deprecated.
|
protected boolean |
cleanup
Deprecated.
|
protected DocumentIterator |
docIter
Deprecated.
|
protected InvertedIndex |
index
Deprecated.
|
protected java.util.List<java.lang.String> |
labels
Deprecated.
|
protected int |
minWordFrequency
Deprecated.
|
protected double |
sample
Deprecated.
|
protected SentenceIterator |
sentenceIterator
Deprecated.
|
protected boolean |
stem
Deprecated.
|
protected java.util.List<java.lang.String> |
stopWords
Deprecated.
|
protected TokenizerFactory |
tokenizerFactory
Deprecated.
|
Constructor and Description |
---|
Builder()
Deprecated.
|
Modifier and Type | Method and Description |
---|---|
Builder |
batchSize(int batchSize)
Deprecated.
|
abstract TextVectorizer |
build()
Deprecated.
|
Builder |
cache(VocabCache cache)
Deprecated.
|
Builder |
cleanup(boolean cleanup)
Deprecated.
|
Builder |
index(InvertedIndex index)
Deprecated.
|
Builder |
iterate(DocumentIterator docIter)
Deprecated.
|
Builder |
iterate(SentenceIterator sentenceIterator)
Deprecated.
|
Builder |
labels(java.util.List<java.lang.String> labels)
Deprecated.
|
Builder |
minWords(int minWordFrequency)
Deprecated.
|
Builder |
sample(double sample)
Deprecated.
|
Builder |
stem(boolean stem)
Deprecated.
|
Builder |
stopWords(java.util.List<java.lang.String> stopWords)
Deprecated.
|
Builder |
tokenize(TokenizerFactory tokenizerFactory)
Deprecated.
|
protected VocabCache cache
protected TokenizerFactory tokenizerFactory
protected java.util.List<java.lang.String> stopWords
protected int minWordFrequency
protected DocumentIterator docIter
protected SentenceIterator sentenceIterator
protected java.util.List<java.lang.String> labels
protected InvertedIndex index
protected int batchSize
protected double sample
protected boolean stem
protected boolean cleanup
public Builder cleanup(boolean cleanup)
public Builder stem(boolean stem)
public Builder sample(double sample)
public Builder batchSize(int batchSize)
public Builder index(InvertedIndex index)
public Builder labels(java.util.List<java.lang.String> labels)
public Builder cache(VocabCache cache)
public Builder tokenize(TokenizerFactory tokenizerFactory)
public Builder stopWords(java.util.List<java.lang.String> stopWords)
public Builder minWords(int minWordFrequency)
public Builder iterate(DocumentIterator docIter)
public Builder iterate(SentenceIterator sentenceIterator)
public abstract TextVectorizer build()