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

Example 96 with ScoreDoc

use of org.apache.lucene.search.ScoreDoc in project searchcode-server by boyter.

the class CodeSearcher method getByCodeId.

/**
     * Only used as fallback if getByRepoFileName fails for some reason due to what appears to be a lucene index bug
     * this should always work as the path used is sha1 and should be unique for anything the current codebase can
     * deal with
     */
public CodeResult getByCodeId(String codeId) {
    CodeResult codeResult = null;
    try {
        IndexReader reader = DirectoryReader.open(FSDirectory.open(Paths.get(this.INDEXPATH)));
        IndexSearcher searcher = new IndexSearcher(reader);
        Analyzer analyzer = new CodeAnalyzer();
        QueryParser parser = new QueryParser(CODEFIELD, analyzer);
        Query query = parser.parse(Values.CODEID + ":" + QueryParser.escape(codeId));
        Singleton.getLogger().info("Query to get by " + Values.CODEID + ":" + QueryParser.escape(codeId));
        TopDocs results = searcher.search(query, 1);
        ScoreDoc[] hits = results.scoreDocs;
        if (hits.length != 0) {
            Document doc = searcher.doc(hits[0].doc);
            String filepath = doc.get(Values.PATH);
            List<String> code = new ArrayList<>();
            try {
                code = Singleton.getHelpers().readFileLinesGuessEncoding(filepath, Singleton.getHelpers().tryParseInt(Properties.getProperties().getProperty(Values.MAXFILELINEDEPTH, Values.DEFAULTMAXFILELINEDEPTH), Values.DEFAULTMAXFILELINEDEPTH));
            } catch (Exception ex) {
                Singleton.getLogger().info("Indexed file appears to binary: " + filepath);
            }
            codeResult = new CodeResult(code, null);
            codeResult.setFilePath(filepath);
            codeResult.setCodePath(doc.get(Values.FILELOCATIONFILENAME));
            codeResult.setFileName(doc.get(Values.FILENAME));
            codeResult.setLanguageName(doc.get(Values.LANGUAGENAME));
            codeResult.setMd5hash(doc.get(Values.MD5HASH));
            codeResult.setCodeLines(doc.get(Values.CODELINES));
            codeResult.setDocumentId(hits[0].doc);
            codeResult.setRepoName(doc.get(Values.REPONAME));
            codeResult.setRepoLocation(doc.get(Values.REPOLOCATION));
            codeResult.setCodeOwner(doc.get(Values.CODEOWNER));
            codeResult.setCodeId(doc.get(Values.CODEID));
        }
        reader.close();
    } catch (Exception ex) {
        LOGGER.severe(" caught a " + ex.getClass() + "\n with message: " + ex.getMessage());
    }
    return codeResult;
}

Example 97 with ScoreDoc

use of org.apache.lucene.search.ScoreDoc in project Anserini by castorini.

the class TweetSearcherAPI method search.

@POST
@Path("search")
@Produces(MediaType.APPLICATION_JSON)
public List<SearchResult> search(SearchAPIQuery query) {
    try {
        Query q = new QueryParser(TweetStreamIndexer.StatusField.TEXT.name, TweetSearcher.ANALYZER).parse(query.getQuery());
        try {
            reader = DirectoryReader.open(TweetSearcher.indexWriter, true, true);
        } catch (IOException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
        IndexReader newReader = DirectoryReader.openIfChanged((DirectoryReader) reader, TweetSearcher.indexWriter, true);
        if (newReader != null) {
            reader.close();
            reader = newReader;
        }
        IndexSearcher searcher = new IndexSearcher(reader);
        int topN = query.getCount();
        TopScoreDocCollector collector = TopScoreDocCollector.create(topN);
        searcher.search(q, collector);
        ScoreDoc[] hits = collector.topDocs().scoreDocs;
        List<SearchResult> resultHits = new ArrayList<>();
        for (int i = 0; i < hits.length && i < topN; ++i) {
            int docId = hits[i].doc;
            Document d = searcher.doc(docId);
            resultHits.add(new SearchResult(String.valueOf(d.get(TweetStreamIndexer.StatusField.ID.name))));
        }
        return resultHits;
    } catch (Exception e) {
        e.printStackTrace();
        return new ArrayList<>();
    }
}

Example 98 with ScoreDoc

use of org.apache.lucene.search.ScoreDoc in project Anserini by castorini.

the class TweetServlet method doGet.

@Override
protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException {
    if (request.getRequestURI().equals("/search")) {
        response.setStatus(HttpServletResponse.SC_OK);
        response.setContentType("text/html");
        request.setCharacterEncoding("UTF-8");
        Query q;
        try {
            q = new QueryParser(StatusField.TEXT.name, TweetSearcher.ANALYZER).parse(request.getParameter("query"));
            try {
                reader = DirectoryReader.open(TweetSearcher.indexWriter, true, true);
            } catch (IOException e) {
                // TODO Auto-generated catch block
                e.printStackTrace();
            }
            IndexReader newReader = DirectoryReader.openIfChanged((DirectoryReader) reader, TweetSearcher.indexWriter, true);
            if (newReader != null) {
                reader.close();
                reader = newReader;
            }
            IndexSearcher searcher = new IndexSearcher(reader);
            int topN;
            if (request.getParameter("top") != null) {
                topN = Integer.parseInt(request.getParameter("top"));
            } else {
                // TODO configurable, default(parameter unspecified in url) topN = 20
                topN = 20;
            }
            TopScoreDocCollector collector = TopScoreDocCollector.create(topN);
            searcher.search(q, collector);
            ScoreDoc[] hits = collector.topDocs().scoreDocs;
            TweetHits tweetHits = new TweetHits(request.getParameter("query"), hits.length);
            for (int i = 0; i < hits.length; ++i) {
                int docId = hits[i].doc;
                Document d = searcher.doc(docId);
                tweetHits.addHit(i, String.valueOf(d.get(StatusField.ID.name)));
            }
            MustacheFactory mf = new DefaultMustacheFactory();
            Mustache mustache = mf.compile(MustacheTemplatePath);
            mustache.execute(response.getWriter(), tweetHits).flush();
        } catch (ParseException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
    } else {
        response.setStatus(HttpServletResponse.SC_NOT_FOUND);
    }
}

Example 99 with ScoreDoc

use of org.apache.lucene.search.ScoreDoc in project Anserini by castorini.

the class BaseFeatureExtractor method printFeatures.

/**
     * Prints feature vectors wrt to the qrels, one vector per qrel
     * @param out
     * @throws IOException
     */
public void printFeatures(PrintStream out) throws IOException {
    Map<String, RerankerContext> queryContextMap = buildRerankerContextMap();
    FeatureExtractors extractors = getExtractors();
    Bits liveDocs = MultiFields.getLiveDocs(reader);
    Set<String> fieldsToLoad = getFieldsToLoad();
    // We need to open a searcher
    IndexSearcher searcher = new IndexSearcher(reader);
    this.printHeader(out, extractors);
    // Iterate through all the qrels and for each document id we have for them
    LOG.debug("Processing queries");
    for (String qid : this.qrels.getQids()) {
        LOG.debug(String.format("Processing qid: %s", qid));
        // Get the map of documents
        RerankerContext context = queryContextMap.get(qid);
        for (Map.Entry<String, Integer> entry : this.qrels.getDocMap(qid).entrySet()) {
            String docId = entry.getKey();
            int qrelScore = entry.getValue();
            // We issue a specific query
            TopDocs topDocs = searcher.search(docIdQuery(docId), 1);
            if (topDocs.totalHits == 0) {
                LOG.warn(String.format("Document Id %s expected but not found in index, skipping...", docId));
                continue;
            }
            ScoreDoc hit = topDocs.scoreDocs[0];
            Document doc = reader.document(hit.doc, fieldsToLoad);
            //TODO factor for test
            Terms terms = reader.getTermVector(hit.doc, getTermVectorField());
            if (terms == null) {
                LOG.debug(String.format("No term vectors found for doc %s, qid %s", docId, qid));
                continue;
            }
            float[] featureValues = extractors.extractAll(doc, terms, context);
            writeFeatureVector(out, qid, qrelScore, docId, featureValues);
        }
        LOG.debug(String.format("Finished processing for qid: %s", qid));
        out.flush();
    }
}

Example 100 with ScoreDoc

use of org.apache.lucene.search.ScoreDoc in project ddf by codice.

the class GeoNamesQueryLuceneIndex method doGetNearestCities.

protected List<NearbyLocation> doGetNearestCities(final Shape shape, final int radiusInKm, final int maxResults, final Directory directory) throws GeoEntryQueryException {
    notNull(shape, "GeoNamesQueryLuceneIndex.doGetNearestCities(): argument 'shape' may not be null.");
    if (radiusInKm <= 0) {
        throw new IllegalArgumentException("GeoNamesQueryLuceneIndex.doGetNearestCities(): radiusInKm must be positive.");
    }
    if (maxResults <= 0) {
        throw new IllegalArgumentException("GeoNamesQueryLuceneIndex.doGetNearestCities(): maxResults must be positive.");
    }
    if (directory == null) {
        return Collections.emptyList();
    }
    try (final IndexReader indexReader = createIndexReader(directory)) {
        final IndexSearcher indexSearcher = createIndexSearcher(indexReader);
        final List<NearbyLocation> closestCities = new ArrayList<>();
        final Point center = shape.getCenter();
        final Query filter = createSpatialQuery(center, radiusInKm);
        // Query for all the documents in the index that are cities, then filter those
        // results for the ones that are in the search area.
        final BooleanQuery booleanQuery = new BooleanQuery.Builder().add(PPL_QUERY, BooleanClause.Occur.MUST).add(filter, BooleanClause.Occur.FILTER).build();
        final TopDocs topDocs = indexSearcher.search(booleanQuery, maxResults, SORT);
        if (topDocs.totalHits > 0) {
            for (ScoreDoc scoreDoc : topDocs.scoreDocs) {
                final double lat = Double.parseDouble(indexSearcher.doc(scoreDoc.doc).get(GeoNamesLuceneConstants.LATITUDE_FIELD));
                final double lon = Double.parseDouble(indexSearcher.doc(scoreDoc.doc).get(GeoNamesLuceneConstants.LONGITUDE_FIELD));
                final String name = indexSearcher.doc(scoreDoc.doc).get(GeoNamesLuceneConstants.NAME_FIELD);
                final NearbyLocation city = new NearbyLocationImpl(center, new PointImpl(lon, lat, SPATIAL_CONTEXT), name);
                closestCities.add(city);
            }
        }
        return closestCities;
    } catch (IOException e) {
        throw new GeoEntryQueryException("Error reading the index", e);
    }
}