Lucene Highlighter with stemming analyzer

Question

I am using Lucene's Highlighter class to highlight fragments of matched search results and it works well. I would like to switch from searching with the StandardAnalyzer to the EnglishAnalyzer, which will perform stemming of terms.

The search results are good, but now the highlighter doesn't always find a match. Here's an example of what I'm looking at:

document field text 1: Everyone likes goats.

document field text 2: I have a goat that eats everything.

Using the EnglishAnalyzer and searching for "goat", both documents are matched, but the highlighter is only able to find a matched fragment from document 2. Is there a way to have the highlighter return data for both documents?

I understand that the characters are different for the tokens, but the same tokens are still there, so it seems reasonable for it to just highlight whatever token is present at that location.

If it helps, this is using Lucene 3.5.

Deguzman

Answer 1

I found a solution to this problem. I changed from using the Highlighter class to using the FastVectorHighlighter. It looks like I'll pick up some speed improvements too (at the expense of storage of term vector data). For the benefit of anyone coming across this question later, here's a unit test showing how this all works together:

package com.sample.index;

import org.apache.lucene.analysis.Analyzer;
import org.apache.lucene.analysis.en.EnglishAnalyzer;
import org.apache.lucene.document.Document;
import org.apache.lucene.document.Field;
import org.apache.lucene.index.IndexReader;
import org.apache.lucene.index.IndexWriter;
import org.apache.lucene.index.IndexWriterConfig;
import org.apache.lucene.queryParser.ParseException;
import org.apache.lucene.queryParser.QueryParser;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.ScoreDoc;
import org.apache.lucene.search.TopDocs;
import org.apache.lucene.search.vectorhighlight.*;
import org.apache.lucene.store.RAMDirectory;
import org.apache.lucene.util.Version;
import org.junit.Before;
import org.junit.Test;

import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

import static junit.framework.Assert.assertEquals;

public class TestIndexStuff {
    public static final String FIELD_NORMAL = "normal";
    public static final String[] PRE_TAGS = new String[]{"["};
    public static final String[] POST_TAGS = new String[]{"]"};
    private IndexSearcher searcher;
    private Analyzer analyzer = new EnglishAnalyzer(Version.LUCENE_35);

    @Before
    public void init() throws IOException {
        RAMDirectory idx = new RAMDirectory();
        IndexWriterConfig config = new IndexWriterConfig(Version.LUCENE_35, analyzer);

        IndexWriter writer = new IndexWriter(idx, config);
        addDocs(writer);
        writer.close();

        searcher = new IndexSearcher(IndexReader.open(idx));
    }

    private void addDocs(IndexWriter writer) throws IOException {
        for (String text : new String[] {
              "Pretty much everyone likes goats.",
              "I have a goat that eats everything.", 
              "goats goats goats goats goats"}) {
            Document doc = new Document();
            doc.add(new Field(FIELD_NORMAL, text, Field.Store.YES, 
                    Field.Index.ANALYZED, Field.TermVector.WITH_POSITIONS_OFFSETS));
            writer.addDocument(doc);
        }
    }

    private FastVectorHighlighter makeHighlighter() {
        FragListBuilder fragListBuilder = new SimpleFragListBuilder(200);
        FragmentsBuilder fragmentBuilder = new SimpleFragmentsBuilder(PRE_TAGS, POST_TAGS);
        return new FastVectorHighlighter(true, true, fragListBuilder, fragmentBuilder);
    }

    @Test
    public void highlight() throws ParseException, IOException {
        Query query = new QueryParser(Version.LUCENE_35, FIELD_NORMAL, analyzer)
                    .parse("goat");
        FastVectorHighlighter highlighter = makeHighlighter();
        FieldQuery fieldQuery = highlighter.getFieldQuery(query);

        TopDocs topDocs = searcher.search(query, 10);
        List<String> fragments = new ArrayList<String>();
        for (ScoreDoc scoreDoc : topDocs.scoreDocs) {
            fragments.add(highlighter.getBestFragment(fieldQuery, searcher.getIndexReader(), 
                    scoreDoc.doc, FIELD_NORMAL, 10000));
        }

        assertEquals(3, fragments.size());
        assertEquals("[goats] [goats] [goats] [goats] [goats]", fragments.get(0).trim());
        assertEquals("Pretty much everyone likes [goats].", fragments.get(1).trim());
        assertEquals("I have a [goat] that eats everything.", fragments.get(2).trim());
    }
}

Deguzman