org.languagetool.tagging.disambiguation

Interface Disambiguator

All Known Implementing Classes:

AbstractDisambiguator, ArabicHybridDisambiguator, CatalanHybridDisambiguator, DemoDisambiguator, EnglishHybridDisambiguator, FrenchHybridDisambiguator, GalicianHybridDisambiguator, GermanRuleDisambiguator, IrishHybridDisambiguator, ItalianRuleDisambiguator, MultiWordChunker, MultiWordChunker2, NoopDisambiguator, PolishHybridDisambiguator, PortugueseHybridDisambiguator, RussianHybridDisambiguator, SpanishHybridDisambiguator, SwedishHybridDisambiguator, UkrainianHybridDisambiguator, XmlRuleDisambiguator

public interface Disambiguator

Disambiguator interface. Particular implementations are language-dependent.

The POS tagger might assign multiple tags to the token. The goal is to filter out the incorrect tags and leave ideally only one per token.

Method Summary

Modifier and Type	Method and Description
AnalyzedSentence	disambiguate(AnalyzedSentence input) If possible, filters out the wrong POS tags.
default AnalyzedSentence	disambiguate(AnalyzedSentence input, JLanguageTool.CheckCancelledCallback checkCanceled) The same as disambiguate(AnalyzedSentence), but may call checkCanceled (if it's non-null) to allow for better interruptibility.
AnalyzedSentence	preDisambiguate(AnalyzedSentence input) If possible, filters out the wrong POS tags.

Method Detail

preDisambiguate

AnalyzedSentence preDisambiguate(AnalyzedSentence input)

If possible, filters out the wrong POS tags. This code will run before disambiguation rules from xml are called. This allows to have some initial disambiguation logic in Java.

Parameters:

input - The sentence with already tagged words. The words are expected to have multiple tags.

Returns:

Analyzed sentence, where each word has only one (possibly the most correct) tag.

Since:

3.7

disambiguate

AnalyzedSentence disambiguate(AnalyzedSentence input)
                       throws IOException

If possible, filters out the wrong POS tags.

Parameters:

input - The sentence with already tagged words. The words are expected to have multiple tags.

Returns:

Analyzed sentence, where each word has only one (possibly the most correct) tag.

Throws:

IOException

disambiguate

default AnalyzedSentence disambiguate(AnalyzedSentence input,
                                      @Nullable
                                      JLanguageTool.CheckCancelledCallback checkCanceled)
                               throws IOException

The same as disambiguate(AnalyzedSentence), but may call checkCanceled (if it's non-null) to allow for better interruptibility.

Throws:

IOException