Examples with SimpleHTMLFormatter org.apache.lucene.search.highlight.SimpleHTMLFormatter used on opensource projects

Example 1 with SimpleHTMLFormatter

use of org.apache.lucene.search.highlight.SimpleHTMLFormatter in project elasticsearch by elastic.

the class PlainHighlighter method highlight.

@Override
public HighlightField highlight(HighlighterContext highlighterContext) {
    SearchContextHighlight.Field field = highlighterContext.field;
    SearchContext context = highlighterContext.context;
    FetchSubPhase.HitContext hitContext = highlighterContext.hitContext;
    FieldMapper mapper = highlighterContext.mapper;
    Encoder encoder = field.fieldOptions().encoder().equals("html") ? HighlightUtils.Encoders.HTML : HighlightUtils.Encoders.DEFAULT;
    if (!hitContext.cache().containsKey(CACHE_KEY)) {
        Map<FieldMapper, org.apache.lucene.search.highlight.Highlighter> mappers = new HashMap<>();
        hitContext.cache().put(CACHE_KEY, mappers);
    }
    @SuppressWarnings("unchecked") Map<FieldMapper, org.apache.lucene.search.highlight.Highlighter> cache = (Map<FieldMapper, org.apache.lucene.search.highlight.Highlighter>) hitContext.cache().get(CACHE_KEY);
    org.apache.lucene.search.highlight.Highlighter entry = cache.get(mapper);
    if (entry == null) {
        QueryScorer queryScorer = new CustomQueryScorer(highlighterContext.query, field.fieldOptions().requireFieldMatch() ? mapper.fieldType().name() : null);
        queryScorer.setExpandMultiTermQuery(true);
        Fragmenter fragmenter;
        if (field.fieldOptions().numberOfFragments() == 0) {
            fragmenter = new NullFragmenter();
        } else if (field.fieldOptions().fragmenter() == null) {
            fragmenter = new SimpleSpanFragmenter(queryScorer, field.fieldOptions().fragmentCharSize());
        } else if ("simple".equals(field.fieldOptions().fragmenter())) {
            fragmenter = new SimpleFragmenter(field.fieldOptions().fragmentCharSize());
        } else if ("span".equals(field.fieldOptions().fragmenter())) {
            fragmenter = new SimpleSpanFragmenter(queryScorer, field.fieldOptions().fragmentCharSize());
        } else {
            throw new IllegalArgumentException("unknown fragmenter option [" + field.fieldOptions().fragmenter() + "] for the field [" + highlighterContext.fieldName + "]");
        }
        Formatter formatter = new SimpleHTMLFormatter(field.fieldOptions().preTags()[0], field.fieldOptions().postTags()[0]);
        entry = new org.apache.lucene.search.highlight.Highlighter(formatter, encoder, queryScorer);
        entry.setTextFragmenter(fragmenter);
        // always highlight across all data
        entry.setMaxDocCharsToAnalyze(Integer.MAX_VALUE);
        cache.put(mapper, entry);
    }
    // a HACK to make highlighter do highlighting, even though its using the single frag list builder
    int numberOfFragments = field.fieldOptions().numberOfFragments() == 0 ? 1 : field.fieldOptions().numberOfFragments();
    ArrayList<TextFragment> fragsList = new ArrayList<>();
    List<Object> textsToHighlight;
    Analyzer analyzer = context.mapperService().documentMapper(hitContext.hit().getType()).mappers().indexAnalyzer();
    try {
        textsToHighlight = HighlightUtils.loadFieldValues(field, mapper, context, hitContext);
        for (Object textToHighlight : textsToHighlight) {
            String text;
            if (textToHighlight instanceof BytesRef) {
                text = mapper.fieldType().valueForDisplay(textToHighlight).toString();
            } else {
                text = textToHighlight.toString();
            }
            try (TokenStream tokenStream = analyzer.tokenStream(mapper.fieldType().name(), text)) {
                if (!tokenStream.hasAttribute(CharTermAttribute.class) || !tokenStream.hasAttribute(OffsetAttribute.class)) {
                    // can't perform highlighting if the stream has no terms (binary token stream) or no offsets
                    continue;
                }
                TextFragment[] bestTextFragments = entry.getBestTextFragments(tokenStream, text, false, numberOfFragments);
                for (TextFragment bestTextFragment : bestTextFragments) {
                    if (bestTextFragment != null && bestTextFragment.getScore() > 0) {
                        fragsList.add(bestTextFragment);
                    }
                }
            }
        }
    } catch (Exception e) {
        if (ExceptionsHelper.unwrap(e, BytesRefHash.MaxBytesLengthExceededException.class) != null) {
            // the plain highlighter will parse the source and try to analyze it.
            return null;
        } else {
            throw new FetchPhaseExecutionException(context, "Failed to highlight field [" + highlighterContext.fieldName + "]", e);
        }
    }
    if (field.fieldOptions().scoreOrdered()) {
        CollectionUtil.introSort(fragsList, new Comparator<TextFragment>() {

            @Override
            public int compare(TextFragment o1, TextFragment o2) {
                return Math.round(o2.getScore() - o1.getScore());
            }
        });
    }
    String[] fragments;
    // number_of_fragments is set to 0 but we have a multivalued field
    if (field.fieldOptions().numberOfFragments() == 0 && textsToHighlight.size() > 1 && fragsList.size() > 0) {
        fragments = new String[fragsList.size()];
        for (int i = 0; i < fragsList.size(); i++) {
            fragments[i] = fragsList.get(i).toString();
        }
    } else {
        // refine numberOfFragments if needed
        numberOfFragments = fragsList.size() < numberOfFragments ? fragsList.size() : numberOfFragments;
        fragments = new String[numberOfFragments];
        for (int i = 0; i < fragments.length; i++) {
            fragments[i] = fragsList.get(i).toString();
        }
    }
    if (fragments.length > 0) {
        return new HighlightField(highlighterContext.fieldName, Text.convertFromStringArray(fragments));
    }
    int noMatchSize = highlighterContext.field.fieldOptions().noMatchSize();
    if (noMatchSize > 0 && textsToHighlight.size() > 0) {
        // Pull an excerpt from the beginning of the string but make sure to split the string on a term boundary.
        String fieldContents = textsToHighlight.get(0).toString();
        int end;
        try {
            end = findGoodEndForNoHighlightExcerpt(noMatchSize, analyzer, mapper.fieldType().name(), fieldContents);
        } catch (Exception e) {
            throw new FetchPhaseExecutionException(context, "Failed to highlight field [" + highlighterContext.fieldName + "]", e);
        }
        if (end > 0) {
            return new HighlightField(highlighterContext.fieldName, new Text[] { new Text(fieldContents.substring(0, end)) });
        }
    }
    return null;
}

Example 2 with SimpleHTMLFormatter

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

the class HighlightCustomQueryTest method highlightField.

/**
   * This method intended for use with
   * <tt>testHighlightingWithDefaultField()</tt>
   */
private String highlightField(Query query, String fieldName, String text) throws IOException, InvalidTokenOffsetsException {
    TokenStream tokenStream = new MockAnalyzer(random(), MockTokenizer.SIMPLE, true, MockTokenFilter.ENGLISH_STOPSET).tokenStream(fieldName, text);
    // Assuming "<B>", "</B>" used to highlight
    SimpleHTMLFormatter formatter = new SimpleHTMLFormatter();
    MyQueryScorer scorer = new MyQueryScorer(query, fieldName, FIELD_NAME);
    Highlighter highlighter = new Highlighter(formatter, scorer);
    highlighter.setTextFragmenter(new SimpleFragmenter(Integer.MAX_VALUE));
    String rv = highlighter.getBestFragments(tokenStream, text, 1, "(FIELD TEXT TRUNCATED)");
    return rv.length() == 0 ? text : rv;
}

Example 3 with SimpleHTMLFormatter

use of org.apache.lucene.search.highlight.SimpleHTMLFormatter in project OpenOLAT by OpenOLAT.

the class SearchResultsImpl method doHighlight.

/**
 * Highlight (bold,color) query words in result-document. Set HighlightResult for content or description.
 * @param query
 * @param analyzer
 * @param doc
 * @param resultDocument
 * @throws IOException
 */
private void doHighlight(Query query, Analyzer analyzer, Document doc, ResultDocument resultDocument) throws IOException {
    Highlighter highlighter = new Highlighter(new SimpleHTMLFormatter(HIGHLIGHT_PRE_TAG, HIGHLIGHT_POST_TAG), new SimpleHTMLEncoder(), new QueryScorer(query));
    // Get 3 best fragments of content and seperate with a "..."
    try {
        // highlight content
        String content = doc.get(AbstractOlatDocument.CONTENT_FIELD_NAME);
        TokenStream tokenStream = analyzer.tokenStream(AbstractOlatDocument.CONTENT_FIELD_NAME, new StringReader(content));
        String highlightResult = highlighter.getBestFragments(tokenStream, content, 3, HIGHLIGHT_SEPARATOR);
        // if no highlightResult is in content => look in description
        if (highlightResult.length() == 0) {
            String description = doc.get(AbstractOlatDocument.DESCRIPTION_FIELD_NAME);
            tokenStream = analyzer.tokenStream(AbstractOlatDocument.DESCRIPTION_FIELD_NAME, new StringReader(description));
            highlightResult = highlighter.getBestFragments(tokenStream, description, 3, HIGHLIGHT_SEPARATOR);
            resultDocument.setHighlightingDescription(true);
        }
        resultDocument.setHighlightResult(highlightResult);
        // highlight title
        String title = doc.get(AbstractOlatDocument.TITLE_FIELD_NAME);
        title = title.trim();
        if (title.length() > 128) {
            title = FilterFactory.getHtmlTagAndDescapingFilter().filter(title);
            title = Formatter.truncate(title, 128);
        }
        tokenStream = analyzer.tokenStream(AbstractOlatDocument.TITLE_FIELD_NAME, new StringReader(title));
        String highlightTitle = highlighter.getBestFragments(tokenStream, title, 3, " ");
        resultDocument.setHighlightTitle(highlightTitle);
    } catch (InvalidTokenOffsetsException e) {
        log.warn("", e);
    }
}

Example 4 with SimpleHTMLFormatter

use of org.apache.lucene.search.highlight.SimpleHTMLFormatter in project ansj_seg by NLPchina.

the class HeightLightTest method toHighlighter.

/**
 * 高亮设置
 *
 * @param query
 * @param doc
 * @return
 */
private static String toHighlighter(Analyzer analyzer, Query query, Document doc) {
    String field = "text";
    try {
        SimpleHTMLFormatter simpleHtmlFormatter = new SimpleHTMLFormatter("<font color=\"red\">", "</font>");
        Highlighter highlighter = new Highlighter(simpleHtmlFormatter, new QueryScorer(query));
        TokenStream tokenStream1 = indexAnalyzer.tokenStream("text", new StringReader(doc.get(field)));
        String highlighterStr = highlighter.getBestFragment(tokenStream1, doc.get(field));
        return highlighterStr == null ? doc.get(field) : highlighterStr;
    } catch (IOException e) {
        // TODO Auto-generated catch block
        e.printStackTrace();
    } catch (InvalidTokenOffsetsException e) {
        // TODO Auto-generated catch block
        e.printStackTrace();
    }
    return null;
}

Example 5 with SimpleHTMLFormatter

use of org.apache.lucene.search.highlight.SimpleHTMLFormatter in project ansj_seg by NLPchina.

the class IndexAndTest method toHighlighter.

/**
 * 高亮设置
 *
 * @param query
 * @param doc
 * @param field
 * @return
 */
private String toHighlighter(Analyzer analyzer, Query query, Document doc) {
    String field = "text";
    try {
        SimpleHTMLFormatter simpleHtmlFormatter = new SimpleHTMLFormatter("<font color=\"red\">", "</font>");
        Highlighter highlighter = new Highlighter(simpleHtmlFormatter, new QueryScorer(query));
        TokenStream tokenStream1 = analyzer.tokenStream("text", new StringReader(doc.get(field)));
        String highlighterStr = highlighter.getBestFragment(tokenStream1, doc.get(field));
        return highlighterStr == null ? doc.get(field) : highlighterStr;
    } catch (IOException e) {
        // TODO Auto-generated catch block
        e.printStackTrace();
    } catch (InvalidTokenOffsetsException e) {
        // TODO Auto-generated catch block
        e.printStackTrace();
    }
    return null;
}