Package org.apache.lucene.search.highlight

package org.apache.lucene.search.highlight

The highlight package contains classes to provide "keyword in context" features typically used to highlight search terms in the text of results pages. The Highlighter class is the central component and can be used to extract the most interesting sections of a piece of text and highlight them, with the help of Fragmenter, fragment Scorer, and Formatter classes.

Example Usage

  //... Above, create documents with two fields, one with term vectors (tv) and one without (notv)
  IndexSearcher searcher = new IndexSearcher(directory);
  QueryParser parser = new QueryParser("notv", analyzer);
  Query query = parser.parse("million");

  TopDocs hits = searcher.search(query, 10);

  SimpleHTMLFormatter htmlFormatter = new SimpleHTMLFormatter();
  Highlighter highlighter = new Highlighter(htmlFormatter, new QueryScorer(query));
  for (int i = 0; i invalid input: '<' 10; i++) {
    int id = hits.scoreDocs[i].doc;
    Document doc = searcher.doc(id);
    String text = doc.get("notv");
    TokenStream tokenStream = TokenSources.getAnyTokenStream(searcher.getIndexReader(), id, "notv", analyzer);
    TextFragment[] frag = highlighter.getBestTextFragments(tokenStream, text, false, 10);//highlighter.getBestFragments(tokenStream, text, 3, "...");
    for (int j = 0; j invalid input: '<' frag.length; j++) {
      if ((frag[j] != null) invalid input: '&'invalid input: '&' (frag[j].getScore() > 0)) {
        System.out.println((frag[j].toString()));
      }
    }
    //Term vector
    text = doc.get("tv");
    tokenStream = TokenSources.getAnyTokenStream(searcher.getIndexReader(), hits.scoreDocs[i].doc, "tv", analyzer);
    frag = highlighter.getBestTextFragments(tokenStream, text, false, 10);
    for (int j = 0; j invalid input: '<' frag.length; j++) {
      if ((frag[j] != null) invalid input: '&'invalid input: '&' (frag[j].getScore() > 0)) {
        System.out.println((frag[j].toString()));
      }
    }
    System.out.println("-------------");
  }

New features 06/02/2005

This release adds options for encoding (thanks to Nicko Cadell). An "Encoder" implementation such as the new SimpleHTMLEncoder class can be passed to the highlighter to encode all those non-xhtml standard characters such as & into legal values. This simple class may not suffice for some languages - Commons Lang has an implementation that could be used: escapeHtml(String) in http://svn.apache.org/viewcvs.cgi/jakarta/commons/proper/lang/trunk/src/java/org/apache/commons/lang/StringEscapeUtils.java?rev=137958invalid input: '&view'=markup

New features 22/12/2004

This release adds some new capabilities:

1. Faster highlighting using Term vector support
2. New formatting options to use color intensity to show informational value
3. Options for better summarization by using term IDF scores to influence fragment selection

The highlighter takes a TokenStream as input. Until now these streams have typically been produced using an Analyzer but the new class TokenSources provides helper methods for obtaining TokenStreams from the new TermVector position support (see latest CVS version).

The new class GradientFormatter can use a scale of colors to highlight terms according to their score. A subtle use of color can help emphasise the reasons for matching (useful when doing "MoreLikeThis" queries and you want to see what the basis of the similarities are).

The QueryScorer class has a new constructor which can use an IndexReader to derive the IDF (inverse document frequency) for each term in order to influence the score. This is useful for helping to extracting the most significant sections of a document and in supplying scores used by the new GradientFormatter to color significant words more strongly. The QueryScorer.getMaxWeight method is useful when passed to the GradientFormatter constructor to define the top score which is associated with the top color.

Related Packages

Package	Description
org.apache.lucene.search	Code to search indices.

Class	Description
DefaultEncoder	Simple Encoder implementation that does not modify the output
Encoder	Encodes original text.
Formatter	Processes terms found in the original text, typically by applying some form of mark-up to highlight terms in HTML search results pages.
Fragmenter	Implements the policy for breaking text into multiple fragments for consideration by the Highlighter class.
GradientFormatter	Formats text with different color intensity depending on the score of the term.
Highlighter	Class used to markup highlighted terms found in the best sections of a text, using configurable Fragmenter, Scorer, Formatter, Encoder and tokenizers.
InvalidTokenOffsetsException	Exception thrown if TokenStream Tokens are incompatible with provided text
NullFragmenter	Fragmenter implementation which does not fragment the text.
OffsetLimitTokenFilter	This TokenFilter limits the number of tokens while indexing by adding up the current offset.
PositionSpan	Utility class to record Positions Spans
QueryScorer	Scorer implementation which scores text fragments by the number of unique query terms found.
QueryTermExtractor	Utility class used to extract the terms used in a query, plus any weights.
QueryTermScorer	Scorer implementation which scores text fragments by the number of unique query terms found.
Scorer	A Scorer is responsible for scoring a stream of tokens.
SimpleFragmenter	Fragmenter implementation which breaks text up into same-size fragments with no concerns over spotting sentence boundaries.
SimpleHTMLEncoder	Simple Encoder implementation to escape text for HTML output
SimpleHTMLFormatter	Simple Formatter implementation to highlight terms with a pre and post tag.
SimpleSpanFragmenter	Fragmenter implementation which breaks text up into same-size fragments but does not split up Spans.
SpanGradientFormatter	Formats text with different color intensity depending on the score of the term using the span tag.
TextFragment	Low-level class used to record information about a section of a document with a score.
TokenGroup	One, or several overlapping tokens, along with the score(s) and the scope of the original text
TokenSources	Hides implementation issues associated with obtaining a TokenStream for use with the higlighter - can obtain from TermFreqVectors with offsets and (optionally) positions or from Analyzer class reparsing the stored content.
TokenStreamFromTermPositionVector	TokenStream created from a term vector field.
WeightedSpanTerm	Lightweight class to hold term, weight, and positions used for scoring this term.
WeightedSpanTermExtractor	Class used to extract WeightedSpanTerms from a Query based on whether Terms from the Query are contained in a supplied TokenStream.
WeightedTerm	Lightweight class to hold term and a weight value used for scoring this term