Module org.apache.lucene.analysis.common
Package org.apache.lucene.analysis.commongrams

Class CommonGramsFilter

java.lang.Object
org.apache.lucene.util.AttributeSource
org.apache.lucene.analysis.TokenStream
org.apache.lucene.analysis.TokenFilter
org.apache.lucene.analysis.commongrams.CommonGramsFilter
All Implemented Interfaces:
Closeable, AutoCloseable, Unwrappable<TokenStream>
public final class CommonGramsFilter extends TokenFilter
Construct bigrams for frequently occurring terms while indexing. Single terms are still indexed too, with bigrams overlaid. This is achieved through the use of PositionIncrementAttribute.setPositionIncrement(int). Bigrams have a type of GRAM_TYPE Example:
  • input:"the quick brown fox"
  • output:|"the","the-quick"|"brown"|"fox"|
  • "the-quick" has a position increment of 0 so it is in the same position as "the" "the-quick" has a term.type() of "gram"

  • Field Details

  • Constructor Details

    • CommonGramsFilter

      public CommonGramsFilter(TokenStream input, CharArraySet commonWords)
      Construct a token stream filtering the given input using a Set of common words to create bigrams. Outputs both unigrams with position increment and bigrams with position increment 0 type=gram where one or both of the words in a potential bigram are in the set of common words .
      Parameters:
      input - TokenStream input in filter chain
      commonWords - The set of common words.

  • Method Details

    • incrementToken

      public boolean incrementToken() throws IOException
      Inserts bigrams for common words into a token stream. For each input token, output the token. If the token and/or the following token are in the list of common words also output a bigram with position increment 0 and type="gram"

      TODO:Consider adding an option to not emit unigram stopwords as in CDL XTF BigramStopFilter, CommonGramsQueryFilter would need to be changed to work with this.

      TODO: Consider optimizing for the case of three commongrams i.e "man of the year" normally produces 3 bigrams: "man-of", "of-the", "the-year" but with proper management of positions we could eliminate the middle bigram "of-the"and save a disk seek and a whole set of position lookups.

      Specified by:
      incrementToken in class TokenStream
      Returns:
      false for end of stream; true otherwise
      Throws:
      IOException

    • reset

      public void reset() throws IOException
      Description copied from class: TokenFilter
      This method is called by a consumer before it begins consumption using TokenStream.incrementToken().

      Resets this stream to a clean state. Stateful implementations must implement this method so that they can be reused, just as if they had been created fresh.

      If you override this method, always call super.reset(), otherwise some internal state will not be correctly reset (e.g., Tokenizer will throw IllegalStateException on further usage).

      NOTE: The default implementation chains the call to the input TokenStream, so be sure to call super.reset() when overriding this method.

      Overrides:
      reset in class TokenFilter
      Throws:
      IOException

    • isCommon

      private boolean isCommon()
      Determines if the current token is a common term
      Returns:
      true if the current token is a common term, false otherwise

    • saveTermBuffer

      private void saveTermBuffer()
      Saves this information to form the left part of a gram

    • gramToken

      private void gramToken()
      Constructs a compound token.