java.lang.Object
org.apache.lucene.analysis.hunspell.WordFormGenerator

public class WordFormGenerator extends Object
A utility class used for generating possible word forms by adding affixes to stems (getAllWordForms(String, String, Runnable)), and suggesting stems and flags to generate the given set of words (compress(List, Set, Runnable)).
  • Field Details

  • Constructor Details

    • WordFormGenerator

      public WordFormGenerator(Dictionary dictionary)
  • Method Details

    • fillAffixMap

      private void fillAffixMap(FST<IntsRef> fst, AffixKind kind)
    • toString

      private String toString(AffixKind kind, IntsRef input)
    • condition

      private AffixCondition condition(int affixId)
    • strip

      private String strip(int affixId)
    • getAllWordForms

      public List<AffixedWord> getAllWordForms(String root, Runnable checkCanceled)
      Generate all word forms for all dictionary entries with the given root word. The result order is stable but not specified. This is equivalent to "unmunch" from the "hunspell-tools" package.
      Parameters:
      checkCanceled - an object that's periodically called, allowing to interrupt the generation by throwing an exception
    • getAllWordForms

      public List<AffixedWord> getAllWordForms(String stem, String flags, Runnable checkCanceled)
      Generate all word forms for the given root pretending it has the given flags (in the same format as the dictionary uses). The result order is stable but not specified. This is equivalent to "unmunch" from the "hunspell-tools" package.
      Parameters:
      checkCanceled - an object that's periodically called, allowing to interrupt the generation by throwing an exception
    • getAllWordForms

      private List<AffixedWord> getAllWordForms(DictEntry entry, char[] encodedFlags, Runnable checkCanceled)
    • sortAndDeduplicate

      private static char[] sortAndDeduplicate(char[] flags)
    • deduplicate

      private static char[] deduplicate(char[] flags)
    • canStemToOriginal

      protected boolean canStemToOriginal(AffixedWord derived)
      A sanity-check that the word form generated by affixation in getAllWordForms(String, String, Runnable) is indeed accepted by the spell-checker and analyzed to be the form of the original dictionary entry. This can be overridden for cases where such check is unnecessary or can be done more efficiently.
    • isForbiddenWord

      private boolean isForbiddenWord(char[] chars, int offset, int length)
    • expand

      private List<AffixedWord> expand(AffixedWord stem, char[] flags, Runnable checkCanceled)
    • shouldConsiderAtAll

      private boolean shouldConsiderAtAll(char[] flags)
    • updateFlags

      private char[] updateFlags(char[] flags, char toRemove, char[] toAppend)
    • appendFlags

      private char[] appendFlags(WordFormGenerator.AffixEntry affix)
    • generateAllSimpleWords

      public void generateAllSimpleWords(Consumer<AffixedWord> consumer, Runnable checkCanceled)
      Traverse the whole dictionary and derive all word forms via affixation (as in getAllWordForms(String, String, Runnable)) for each of the entries. The iteration order is undefined. Only "simple" words are returned, no compounding flags are processed. Upper- and title-case variations are not returned, even if the spellchecker accepts them.
      Parameters:
      consumer - the object that receives each derived word form
      checkCanceled - an object that's periodically called, allowing to interrupt the traversal and generation by throwing an exception
    • compress

      public EntrySuggestion compress(List<String> words, Set<String> forbidden, Runnable checkCanceled)
      Given a list of words, try to produce a smaller set of dictionary entries (with some flags) that would generate these words. This is equivalent to "munch" from the "hunspell-tools" package. The algorithm tries to minimize the number of the dictionary entries to add or change, the number of flags involved, and the number of non-requested additionally generated words. All the mentioned words are in the dictionary format and case: no ICONV/OCONV/IGNORE conversions are applied.
      Parameters:
      words - the list of words to generate
      forbidden - the set of words to avoid generating
      checkCanceled - an object that's periodically called, allowing to interrupt the generation by throwing an exception
      Returns:
      the information about suggested dictionary entries and overgenerated words, or null if the algorithm couldn't generate anything
    • isCompatibleWithPreviousAffixes

      private boolean isCompatibleWithPreviousAffixes(AffixedWord stem, AffixKind kind, char flag)