Examples with TopDocs org.apache.lucene.search.TopDocs used on opensource projects

Example 66 with TopDocs

use of org.apache.lucene.search.TopDocs in project lucene-solr by apache.

the class TestNRTCachingDirectory method testNRTAndCommit.

public void testNRTAndCommit() throws Exception {
    Directory dir = newDirectory();
    NRTCachingDirectory cachedDir = new NRTCachingDirectory(dir, 2.0, 25.0);
    MockAnalyzer analyzer = new MockAnalyzer(random());
    analyzer.setMaxTokenLength(TestUtil.nextInt(random(), 1, IndexWriter.MAX_TERM_LENGTH));
    IndexWriterConfig conf = newIndexWriterConfig(analyzer);
    RandomIndexWriter w = new RandomIndexWriter(random(), cachedDir, conf);
    final LineFileDocs docs = new LineFileDocs(random());
    final int numDocs = TestUtil.nextInt(random(), 100, 400);
    if (VERBOSE) {
        System.out.println("TEST: numDocs=" + numDocs);
    }
    final List<BytesRef> ids = new ArrayList<>();
    DirectoryReader r = null;
    for (int docCount = 0; docCount < numDocs; docCount++) {
        final Document doc = docs.nextDoc();
        ids.add(new BytesRef(doc.get("docid")));
        w.addDocument(doc);
        if (random().nextInt(20) == 17) {
            if (r == null) {
                r = DirectoryReader.open(w.w);
            } else {
                final DirectoryReader r2 = DirectoryReader.openIfChanged(r);
                if (r2 != null) {
                    r.close();
                    r = r2;
                }
            }
            assertEquals(1 + docCount, r.numDocs());
            final IndexSearcher s = newSearcher(r);
            // Just make sure search can run; we can't assert
            // totHits since it could be 0
            TopDocs hits = s.search(new TermQuery(new Term("body", "the")), 10);
        // System.out.println("tot hits " + hits.totalHits);
        }
    }
    if (r != null) {
        r.close();
    }
    // Close should force cache to clear since all files are sync'd
    w.close();
    final String[] cachedFiles = cachedDir.listCachedFiles();
    for (String file : cachedFiles) {
        System.out.println("FAIL: cached file " + file + " remains after sync");
    }
    assertEquals(0, cachedFiles.length);
    r = DirectoryReader.open(dir);
    for (BytesRef id : ids) {
        assertEquals(1, r.docFreq(new Term("docid", id)));
    }
    r.close();
    cachedDir.close();
    docs.close();
}

Example 67 with TopDocs

use of org.apache.lucene.search.TopDocs in project lucene-solr by apache.

the class TestSimilarityBase method testHeartRanking.

/** Test whether all similarities return document 3 before documents 7 and 8. */
public void testHeartRanking() throws IOException {
    Query q = new TermQuery(new Term(FIELD_BODY, "heart"));
    for (SimilarityBase sim : sims) {
        searcher.setSimilarity(sim);
        TopDocs topDocs = searcher.search(q, 1000);
        assertEquals("Failed: " + sim.toString(), "2", reader.document(topDocs.scoreDocs[0].doc).get(FIELD_ID));
    }
}

Example 68 with TopDocs

use of org.apache.lucene.search.TopDocs in project lucene-solr by apache.

the class TokenSourcesTest method testOverlapWithOffset.

public void testOverlapWithOffset() throws IOException, InvalidTokenOffsetsException {
    final String TEXT = "the fox did not jump";
    final Directory directory = newDirectory();
    final IndexWriter indexWriter = new IndexWriter(directory, newIndexWriterConfig(null));
    try {
        final Document document = new Document();
        FieldType customType = new FieldType(TextField.TYPE_NOT_STORED);
        customType.setStoreTermVectors(true);
        // no positions!
        customType.setStoreTermVectorOffsets(true);
        document.add(new Field(FIELD, new OverlappingTokenStream(), customType));
        indexWriter.addDocument(document);
    } finally {
        indexWriter.close();
    }
    final IndexReader indexReader = DirectoryReader.open(directory);
    assertEquals(1, indexReader.numDocs());
    final IndexSearcher indexSearcher = newSearcher(indexReader);
    try {
        final DisjunctionMaxQuery query = new DisjunctionMaxQuery(Arrays.asList(new SpanTermQuery(new Term(FIELD, "{fox}")), new SpanTermQuery(new Term(FIELD, "fox"))), 1);
        // final Query phraseQuery = new SpanNearQuery(new SpanQuery[] {
        // new SpanTermQuery(new Term(FIELD, "{fox}")),
        // new SpanTermQuery(new Term(FIELD, "fox")) }, 0, true);
        TopDocs hits = indexSearcher.search(query, 1);
        assertEquals(1, hits.totalHits);
        final Highlighter highlighter = new Highlighter(new SimpleHTMLFormatter(), new SimpleHTMLEncoder(), new QueryScorer(query));
        final TokenStream tokenStream = TokenSources.getTermVectorTokenStreamOrNull(FIELD, indexReader.getTermVectors(0), -1);
        assertEquals("<B>the fox</B> did not jump", highlighter.getBestFragment(tokenStream, TEXT));
    } finally {
        indexReader.close();
        directory.close();
    }
}

Example 69 with TopDocs

use of org.apache.lucene.search.TopDocs in project lucene-solr by apache.

the class SynonymTokenizer method testCustomScoreQueryHighlight.

public void testCustomScoreQueryHighlight() throws Exception {
    TermQuery termQuery = new TermQuery(new Term(FIELD_NAME, "very"));
    CustomScoreQuery query = new CustomScoreQuery(termQuery);
    searcher = newSearcher(reader);
    TopDocs hits = searcher.search(query, 10, new Sort(SortField.FIELD_DOC, SortField.FIELD_SCORE));
    assertEquals(2, hits.totalHits);
    QueryScorer scorer = new QueryScorer(query, FIELD_NAME);
    Highlighter highlighter = new Highlighter(scorer);
    final int docId0 = hits.scoreDocs[0].doc;
    Document doc = searcher.doc(docId0);
    String storedField = doc.get(FIELD_NAME);
    TokenStream stream = getAnyTokenStream(FIELD_NAME, docId0);
    Fragmenter fragmenter = new SimpleSpanFragmenter(scorer);
    highlighter.setTextFragmenter(fragmenter);
    String fragment = highlighter.getBestFragment(stream, storedField);
    assertEquals("Hello this is a piece of text that is <B>very</B> long and contains too much preamble and the meat is really here which says kennedy has been shot", fragment);
}

Example 70 with TopDocs

use of org.apache.lucene.search.TopDocs in project lucene-solr by apache.

the class TokenSourcesTest method testOverlapWithOffsetExactPhrase.

public void testOverlapWithOffsetExactPhrase() throws IOException, InvalidTokenOffsetsException {
    final String TEXT = "the fox did not jump";
    final Directory directory = newDirectory();
    final IndexWriter indexWriter = new IndexWriter(directory, newIndexWriterConfig(null));
    try {
        final Document document = new Document();
        FieldType customType = new FieldType(TextField.TYPE_NOT_STORED);
        customType.setStoreTermVectors(true);
        // no positions!
        customType.setStoreTermVectorOffsets(true);
        document.add(new Field(FIELD, new OverlappingTokenStream(), customType));
        indexWriter.addDocument(document);
    } finally {
        indexWriter.close();
    }
    final IndexReader indexReader = DirectoryReader.open(directory);
    try {
        assertEquals(1, indexReader.numDocs());
        final IndexSearcher indexSearcher = newSearcher(indexReader);
        // final DisjunctionMaxQuery query = new DisjunctionMaxQuery(1);
        // query.add(new SpanTermQuery(new Term(FIELD, "{fox}")));
        // query.add(new SpanTermQuery(new Term(FIELD, "fox")));
        final Query phraseQuery = new SpanNearQuery(new SpanQuery[] { new SpanTermQuery(new Term(FIELD, "the")), new SpanTermQuery(new Term(FIELD, "fox")) }, 0, true);
        TopDocs hits = indexSearcher.search(phraseQuery, 1);
        assertEquals(1, hits.totalHits);
        final Highlighter highlighter = new Highlighter(new SimpleHTMLFormatter(), new SimpleHTMLEncoder(), new QueryScorer(phraseQuery));
        final TokenStream tokenStream = TokenSources.getTermVectorTokenStreamOrNull(FIELD, indexReader.getTermVectors(0), -1);
        assertEquals("<B>the fox</B> did not jump", highlighter.getBestFragment(tokenStream, TEXT));
    } finally {
        indexReader.close();
        directory.close();
    }
}