Examples with Text org.elasticsearch.common.text.Text used on opensource projects

Example 26 with Text

use of org.elasticsearch.common.text.Text in project elasticsearch by elastic.

the class HighlightField method writeTo.

@Override
public void writeTo(StreamOutput out) throws IOException {
    out.writeString(name);
    if (fragments == null) {
        out.writeBoolean(false);
    } else {
        out.writeBoolean(true);
        out.writeVInt(fragments.length);
        for (Text fragment : fragments) {
            out.writeText(fragment);
        }
    }
}

Example 27 with Text

use of org.elasticsearch.common.text.Text in project elasticsearch by elastic.

the class PhraseSuggester method innerExecute.

/*
     * More Ideas:
     *   - add ability to find whitespace problems -> we can build a poor mans decompounder with our index based on a automaton?
     *   - add ability to build different error models maybe based on a confusion matrix?
     *   - try to combine a token with its subsequent token to find / detect word splits (optional)
     *      - for this to work we need some way to defined the position length of a candidate
     *   - phonetic filters could be interesting here too for candidate selection
     */
@Override
public Suggestion<? extends Entry<? extends Option>> innerExecute(String name, PhraseSuggestionContext suggestion, IndexSearcher searcher, CharsRefBuilder spare) throws IOException {
    double realWordErrorLikelihood = suggestion.realworldErrorLikelyhood();
    final PhraseSuggestion response = new PhraseSuggestion(name, suggestion.getSize());
    final IndexReader indexReader = searcher.getIndexReader();
    List<PhraseSuggestionContext.DirectCandidateGenerator> generators = suggestion.generators();
    final int numGenerators = generators.size();
    final List<CandidateGenerator> gens = new ArrayList<>(generators.size());
    for (int i = 0; i < numGenerators; i++) {
        PhraseSuggestionContext.DirectCandidateGenerator generator = generators.get(i);
        DirectSpellChecker directSpellChecker = generator.createDirectSpellChecker();
        Terms terms = MultiFields.getTerms(indexReader, generator.field());
        if (terms != null) {
            gens.add(new DirectCandidateGenerator(directSpellChecker, generator.field(), generator.suggestMode(), indexReader, realWordErrorLikelihood, generator.size(), generator.preFilter(), generator.postFilter(), terms));
        }
    }
    final String suggestField = suggestion.getField();
    final Terms suggestTerms = MultiFields.getTerms(indexReader, suggestField);
    if (gens.size() > 0 && suggestTerms != null) {
        final NoisyChannelSpellChecker checker = new NoisyChannelSpellChecker(realWordErrorLikelihood, suggestion.getRequireUnigram(), suggestion.getTokenLimit());
        final BytesRef separator = suggestion.separator();
        WordScorer wordScorer = suggestion.model().newScorer(indexReader, suggestTerms, suggestField, realWordErrorLikelihood, separator);
        Result checkerResult;
        try (TokenStream stream = checker.tokenStream(suggestion.getAnalyzer(), suggestion.getText(), spare, suggestion.getField())) {
            checkerResult = checker.getCorrections(stream, new MultiCandidateGeneratorWrapper(suggestion.getShardSize(), gens.toArray(new CandidateGenerator[gens.size()])), suggestion.maxErrors(), suggestion.getShardSize(), wordScorer, suggestion.confidence(), suggestion.gramSize());
        }
        PhraseSuggestion.Entry resultEntry = buildResultEntry(suggestion, spare, checkerResult.cutoffScore);
        response.addTerm(resultEntry);
        final BytesRefBuilder byteSpare = new BytesRefBuilder();
        final Function<Map<String, Object>, ExecutableScript> collateScript = suggestion.getCollateQueryScript();
        final boolean collatePrune = (collateScript != null) && suggestion.collatePrune();
        for (int i = 0; i < checkerResult.corrections.length; i++) {
            Correction correction = checkerResult.corrections[i];
            spare.copyUTF8Bytes(correction.join(SEPARATOR, byteSpare, null, null));
            boolean collateMatch = true;
            if (collateScript != null) {
                // Checks if the template query collateScript yields any documents
                // from the index for a correction, collateMatch is updated
                final Map<String, Object> vars = suggestion.getCollateScriptParams();
                vars.put(SUGGESTION_TEMPLATE_VAR_NAME, spare.toString());
                QueryShardContext shardContext = suggestion.getShardContext();
                final ExecutableScript executable = collateScript.apply(vars);
                final BytesReference querySource = (BytesReference) executable.run();
                try (XContentParser parser = XContentFactory.xContent(querySource).createParser(shardContext.getXContentRegistry(), querySource)) {
                    QueryBuilder innerQueryBuilder = shardContext.newParseContext(parser).parseInnerQueryBuilder();
                    final ParsedQuery parsedQuery = shardContext.toQuery(innerQueryBuilder);
                    collateMatch = Lucene.exists(searcher, parsedQuery.query());
                }
            }
            if (!collateMatch && !collatePrune) {
                continue;
            }
            Text phrase = new Text(spare.toString());
            Text highlighted = null;
            if (suggestion.getPreTag() != null) {
                spare.copyUTF8Bytes(correction.join(SEPARATOR, byteSpare, suggestion.getPreTag(), suggestion.getPostTag()));
                highlighted = new Text(spare.toString());
            }
            if (collatePrune) {
                resultEntry.addOption(new Suggestion.Entry.Option(phrase, highlighted, (float) (correction.score), collateMatch));
            } else {
                resultEntry.addOption(new Suggestion.Entry.Option(phrase, highlighted, (float) (correction.score)));
            }
        }
    } else {
        response.addTerm(buildResultEntry(suggestion, spare, Double.MIN_VALUE));
    }
    return response;
}

Example 28 with Text

use of org.elasticsearch.common.text.Text in project elasticsearch by elastic.

the class ExpandSearchPhaseTests method testFailOneItemFailsEntirePhase.

public void testFailOneItemFailsEntirePhase() throws IOException {
    AtomicBoolean executedMultiSearch = new AtomicBoolean(false);
    SearchHits collapsedHits = new SearchHits(new SearchHit[] { new SearchHit(2, "ID", new Text("type"), Collections.emptyMap()), new SearchHit(3, "ID", new Text("type"), Collections.emptyMap()) }, 1, 1.0F);
    MockSearchPhaseContext mockSearchPhaseContext = new MockSearchPhaseContext(1);
    String collapseValue = randomBoolean() ? null : "boom";
    mockSearchPhaseContext.getRequest().source(new SearchSourceBuilder().collapse(new CollapseBuilder("someField").setInnerHits(new InnerHitBuilder().setName("foobarbaz"))));
    mockSearchPhaseContext.searchTransport = new SearchTransportService(Settings.builder().put("search.remote.connect", false).build(), null, null) {

        @Override
        void sendExecuteMultiSearch(MultiSearchRequest request, SearchTask task, ActionListener<MultiSearchResponse> listener) {
            assertTrue(executedMultiSearch.compareAndSet(false, true));
            InternalSearchResponse internalSearchResponse = new InternalSearchResponse(collapsedHits, null, null, null, false, null, 1);
            SearchResponse response = mockSearchPhaseContext.buildSearchResponse(internalSearchResponse, null);
            listener.onResponse(new MultiSearchResponse(new MultiSearchResponse.Item[] { new MultiSearchResponse.Item(null, new RuntimeException("boom")), new MultiSearchResponse.Item(response, null) }));
        }
    };
    SearchHits hits = new SearchHits(new SearchHit[] { new SearchHit(1, "ID", new Text("type"), Collections.singletonMap("someField", new SearchHitField("someField", Collections.singletonList(collapseValue)))), new SearchHit(2, "ID2", new Text("type"), Collections.singletonMap("someField", new SearchHitField("someField", Collections.singletonList(collapseValue)))) }, 1, 1.0F);
    InternalSearchResponse internalSearchResponse = new InternalSearchResponse(hits, null, null, null, false, null, 1);
    SearchResponse response = mockSearchPhaseContext.buildSearchResponse(internalSearchResponse, null);
    AtomicReference<SearchResponse> reference = new AtomicReference<>();
    ExpandSearchPhase phase = new ExpandSearchPhase(mockSearchPhaseContext, response, r -> new SearchPhase("test") {

        @Override
        public void run() throws IOException {
            reference.set(r);
        }
    });
    phase.run();
    assertThat(mockSearchPhaseContext.phaseFailure.get(), Matchers.instanceOf(RuntimeException.class));
    assertEquals("boom", mockSearchPhaseContext.phaseFailure.get().getMessage());
    assertNotNull(mockSearchPhaseContext.phaseFailure.get());
    assertNull(reference.get());
    assertEquals(0, mockSearchPhaseContext.phasesExecuted.get());
}

Example 29 with Text

use of org.elasticsearch.common.text.Text in project elasticsearch by elastic.

the class ExpandSearchPhaseTests method testSkipPhase.

public void testSkipPhase() throws IOException {
    MockSearchPhaseContext mockSearchPhaseContext = new MockSearchPhaseContext(1);
    mockSearchPhaseContext.searchTransport = new SearchTransportService(Settings.builder().put("search.remote.connect", false).build(), null, null) {

        @Override
        void sendExecuteMultiSearch(MultiSearchRequest request, SearchTask task, ActionListener<MultiSearchResponse> listener) {
            fail("no collapsing here");
        }
    };
    SearchHits hits = new SearchHits(new SearchHit[] { new SearchHit(1, "ID", new Text("type"), Collections.singletonMap("someField", new SearchHitField("someField", Collections.singletonList(null)))), new SearchHit(2, "ID2", new Text("type"), Collections.singletonMap("someField", new SearchHitField("someField", Collections.singletonList(null)))) }, 1, 1.0F);
    InternalSearchResponse internalSearchResponse = new InternalSearchResponse(hits, null, null, null, false, null, 1);
    SearchResponse response = mockSearchPhaseContext.buildSearchResponse(internalSearchResponse, null);
    AtomicReference<SearchResponse> reference = new AtomicReference<>();
    ExpandSearchPhase phase = new ExpandSearchPhase(mockSearchPhaseContext, response, r -> new SearchPhase("test") {

        @Override
        public void run() throws IOException {
            reference.set(r);
        }
    });
    phase.run();
    mockSearchPhaseContext.assertNoFailure();
    assertNotNull(reference.get());
    assertEquals(1, mockSearchPhaseContext.phasesExecuted.get());
}

Example 30 with Text

use of org.elasticsearch.common.text.Text in project elasticsearch by elastic.

the class SearchPhaseControllerTests method generateQueryResults.

private AtomicArray<QuerySearchResultProvider> generateQueryResults(int nShards, List<CompletionSuggestion> suggestions, int searchHitsSize, boolean useConstantScore) {
    AtomicArray<QuerySearchResultProvider> queryResults = new AtomicArray<>(nShards);
    for (int shardIndex = 0; shardIndex < nShards; shardIndex++) {
        QuerySearchResult querySearchResult = new QuerySearchResult(shardIndex, new SearchShardTarget("", new Index("", ""), shardIndex));
        TopDocs topDocs = new TopDocs(0, new ScoreDoc[0], 0);
        if (searchHitsSize > 0) {
            int nDocs = randomIntBetween(0, searchHitsSize);
            ScoreDoc[] scoreDocs = new ScoreDoc[nDocs];
            float maxScore = 0F;
            for (int i = 0; i < nDocs; i++) {
                float score = useConstantScore ? 1.0F : Math.abs(randomFloat());
                scoreDocs[i] = new ScoreDoc(i, score);
                if (score > maxScore) {
                    maxScore = score;
                }
            }
            topDocs = new TopDocs(scoreDocs.length, scoreDocs, maxScore);
        }
        List<CompletionSuggestion> shardSuggestion = new ArrayList<>();
        for (CompletionSuggestion completionSuggestion : suggestions) {
            CompletionSuggestion suggestion = new CompletionSuggestion(completionSuggestion.getName(), completionSuggestion.getSize());
            final CompletionSuggestion.Entry completionEntry = new CompletionSuggestion.Entry(new Text(""), 0, 5);
            suggestion.addTerm(completionEntry);
            int optionSize = randomIntBetween(1, suggestion.getSize());
            float maxScore = randomIntBetween(suggestion.getSize(), (int) Float.MAX_VALUE);
            for (int i = 0; i < optionSize; i++) {
                completionEntry.addOption(new CompletionSuggestion.Entry.Option(i, new Text(""), maxScore, Collections.emptyMap()));
                float dec = randomIntBetween(0, optionSize);
                if (dec <= maxScore) {
                    maxScore -= dec;
                }
            }
            suggestion.setShardIndex(shardIndex);
            shardSuggestion.add(suggestion);
        }
        querySearchResult.topDocs(topDocs, null);
        querySearchResult.size(searchHitsSize);
        querySearchResult.suggest(new Suggest(new ArrayList<>(shardSuggestion)));
        queryResults.set(shardIndex, querySearchResult);
    }
    return queryResults;
}