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

Example 61 with TopDocs

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

the class SpatialPrefixTreeTest method testBadPrefixTreePrune.

/**
   * A PrefixTree pruning optimization gone bad, applicable when optimize=true.
   * See <a href="https://issues.apache.org/jira/browse/LUCENE-4770">LUCENE-4770</a>.
   */
@Test
public void testBadPrefixTreePrune() throws Exception {
    trie = new QuadPrefixTree(ctx, 12);
    TermQueryPrefixTreeStrategy strategy = new TermQueryPrefixTreeStrategy(trie, "geo");
    Document doc = new Document();
    doc.add(new TextField("id", "1", Store.YES));
    Shape area = ctx.makeRectangle(-122.82, -122.78, 48.54, 48.56);
    Field[] fields = strategy.createIndexableFields(area, 0.025);
    for (Field field : fields) {
        doc.add(field);
    }
    addDocument(doc);
    Point upperleft = ctx.makePoint(-122.88, 48.54);
    Point lowerright = ctx.makePoint(-122.82, 48.62);
    Query query = strategy.makeQuery(new SpatialArgs(SpatialOperation.Intersects, ctx.makeRectangle(upperleft, lowerright)));
    commit();
    TopDocs search = indexSearcher.search(query, 10);
    ScoreDoc[] scoreDocs = search.scoreDocs;
    for (ScoreDoc scoreDoc : scoreDocs) {
        System.out.println(indexSearcher.doc(scoreDoc.doc));
    }
    assertEquals(1, search.totalHits);
}

Example 62 with TopDocs

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

the class TestHierarchicalDocBuilder method assertSearch.

private void assertSearch(Query query, String field, String... values) throws IOException {
    /* The limit of search queue is doubled to catch the error in case when for some reason there are more docs than expected  */
    SolrIndexSearcher searcher = req.getSearcher();
    TopDocs result = searcher.search(query, values.length * 2);
    assertEquals(values.length, result.totalHits);
    List<String> actualValues = new ArrayList<String>();
    for (int index = 0; index < values.length; ++index) {
        Document doc = searcher.doc(result.scoreDocs[index].doc);
        actualValues.add(doc.get(field));
    }
    for (String expectedValue : values) {
        boolean removed = actualValues.remove(expectedValue);
        if (!removed) {
            fail("Search result does not contain expected values");
        }
    }
}

Example 63 with TopDocs

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

the class SimpleMLTQParser method parse.

public Query parse() {
    String defaultField = req.getSchema().getUniqueKeyField().getName();
    String uniqueValue = localParams.get(QueryParsing.V);
    String[] qf = localParams.getParams("qf");
    SolrIndexSearcher searcher = req.getSearcher();
    Query docIdQuery = createIdQuery(defaultField, uniqueValue);
    Map<String, Float> boostFields = new HashMap<>();
    try {
        TopDocs td = searcher.search(docIdQuery, 1);
        if (td.totalHits != 1)
            throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, "Error completing MLT request. Could not fetch " + "document with id [" + uniqueValue + "]");
        ScoreDoc[] scoreDocs = td.scoreDocs;
        MoreLikeThis mlt = new MoreLikeThis(req.getSearcher().getIndexReader());
        mlt.setMinTermFreq(localParams.getInt("mintf", MoreLikeThis.DEFAULT_MIN_TERM_FREQ));
        mlt.setMinDocFreq(localParams.getInt("mindf", MoreLikeThis.DEFAULT_MIN_DOC_FREQ));
        mlt.setMinWordLen(localParams.getInt("minwl", MoreLikeThis.DEFAULT_MIN_WORD_LENGTH));
        mlt.setMaxWordLen(localParams.getInt("maxwl", MoreLikeThis.DEFAULT_MAX_WORD_LENGTH));
        mlt.setMaxQueryTerms(localParams.getInt("maxqt", MoreLikeThis.DEFAULT_MAX_QUERY_TERMS));
        mlt.setMaxNumTokensParsed(localParams.getInt("maxntp", MoreLikeThis.DEFAULT_MAX_NUM_TOKENS_PARSED));
        mlt.setMaxDocFreq(localParams.getInt("maxdf", MoreLikeThis.DEFAULT_MAX_DOC_FREQ));
        Boolean boost = localParams.getBool("boost", false);
        mlt.setBoost(boost);
        String[] fieldNames;
        if (qf != null) {
            ArrayList<String> fields = new ArrayList<>();
            for (String fieldName : qf) {
                if (!StringUtils.isEmpty(fieldName)) {
                    String[] strings = splitList.split(fieldName);
                    for (String string : strings) {
                        if (!StringUtils.isEmpty(string)) {
                            fields.add(string);
                        }
                    }
                }
            }
            // Parse field names and boosts from the fields
            boostFields = SolrPluginUtils.parseFieldBoosts(fields.toArray(new String[0]));
            fieldNames = boostFields.keySet().toArray(new String[0]);
        } else {
            Map<String, SchemaField> fieldDefinitions = req.getSearcher().getSchema().getFields();
            ArrayList<String> fields = new ArrayList();
            for (String fieldName : fieldDefinitions.keySet()) {
                if (fieldDefinitions.get(fieldName).indexed() && fieldDefinitions.get(fieldName).stored())
                    if (fieldDefinitions.get(fieldName).getType().getNumberType() == null)
                        fields.add(fieldName);
            }
            fieldNames = fields.toArray(new String[0]);
        }
        if (fieldNames.length < 1) {
            throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, "MoreLikeThis requires at least one similarity field: qf");
        }
        mlt.setFieldNames(fieldNames);
        mlt.setAnalyzer(req.getSchema().getIndexAnalyzer());
        Query rawMLTQuery = mlt.like(scoreDocs[0].doc);
        BooleanQuery boostedMLTQuery = (BooleanQuery) rawMLTQuery;
        if (boost && boostFields.size() > 0) {
            BooleanQuery.Builder newQ = new BooleanQuery.Builder();
            newQ.setMinimumNumberShouldMatch(boostedMLTQuery.getMinimumNumberShouldMatch());
            for (BooleanClause clause : boostedMLTQuery) {
                Query q = clause.getQuery();
                float originalBoost = 1f;
                if (q instanceof BoostQuery) {
                    BoostQuery bq = (BoostQuery) q;
                    q = bq.getQuery();
                    originalBoost = bq.getBoost();
                }
                Float fieldBoost = boostFields.get(((TermQuery) q).getTerm().field());
                q = ((fieldBoost != null) ? new BoostQuery(q, fieldBoost * originalBoost) : clause.getQuery());
                newQ.add(q, clause.getOccur());
            }
            boostedMLTQuery = newQ.build();
        }
        // exclude current document from results
        BooleanQuery.Builder realMLTQuery = new BooleanQuery.Builder();
        realMLTQuery.add(boostedMLTQuery, BooleanClause.Occur.MUST);
        realMLTQuery.add(docIdQuery, BooleanClause.Occur.MUST_NOT);
        return realMLTQuery.build();
    } catch (IOException e) {
        throw new SolrException(SolrException.ErrorCode.BAD_REQUEST, "Error completing MLT request" + e.getMessage());
    }
}

Example 64 with TopDocs

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

the class TopGroupsResultTransformer method transformToNative.

/**
   * {@inheritDoc}
   */
@Override
public Map<String, ?> transformToNative(NamedList<NamedList> shardResponse, Sort groupSort, Sort withinGroupSort, String shard) {
    Map<String, Object> result = new HashMap<>();
    final IndexSchema schema = rb.req.getSearcher().getSchema();
    for (Map.Entry<String, NamedList> entry : shardResponse) {
        String key = entry.getKey();
        NamedList commandResult = entry.getValue();
        Integer totalGroupedHitCount = (Integer) commandResult.get("totalGroupedHitCount");
        Integer totalHits = (Integer) commandResult.get("totalHits");
        if (totalHits != null) {
            Integer matches = (Integer) commandResult.get("matches");
            Float maxScore = (Float) commandResult.get("maxScore");
            if (maxScore == null) {
                maxScore = Float.NaN;
            }
            @SuppressWarnings("unchecked") List<NamedList<Object>> documents = (List<NamedList<Object>>) commandResult.get("documents");
            ScoreDoc[] scoreDocs = transformToNativeShardDoc(documents, groupSort, shard, schema);
            final TopDocs topDocs;
            if (withinGroupSort.equals(Sort.RELEVANCE)) {
                topDocs = new TopDocs(totalHits, scoreDocs, maxScore);
            } else {
                topDocs = new TopFieldDocs(totalHits, scoreDocs, withinGroupSort.getSort(), maxScore);
            }
            result.put(key, new QueryCommandResult(topDocs, matches));
            continue;
        }
        Integer totalHitCount = (Integer) commandResult.get("totalHitCount");
        List<GroupDocs<BytesRef>> groupDocs = new ArrayList<>();
        for (int i = 2; i < commandResult.size(); i++) {
            String groupValue = commandResult.getName(i);
            @SuppressWarnings("unchecked") NamedList<Object> groupResult = (NamedList<Object>) commandResult.getVal(i);
            Integer totalGroupHits = (Integer) groupResult.get("totalHits");
            Float maxScore = (Float) groupResult.get("maxScore");
            if (maxScore == null) {
                maxScore = Float.NaN;
            }
            @SuppressWarnings("unchecked") List<NamedList<Object>> documents = (List<NamedList<Object>>) groupResult.get("documents");
            ScoreDoc[] scoreDocs = transformToNativeShardDoc(documents, withinGroupSort, shard, schema);
            BytesRef groupValueRef = groupValue != null ? new BytesRef(groupValue) : null;
            groupDocs.add(new GroupDocs<>(Float.NaN, maxScore, totalGroupHits, scoreDocs, groupValueRef, null));
        }
        @SuppressWarnings("unchecked") GroupDocs<BytesRef>[] groupDocsArr = groupDocs.toArray(new GroupDocs[groupDocs.size()]);
        TopGroups<BytesRef> topGroups = new TopGroups<>(groupSort.getSort(), withinGroupSort.getSort(), totalHitCount, totalGroupedHitCount, groupDocsArr, Float.NaN);
        result.put(key, topGroups);
    }
    return result;
}

Example 65 with TopDocs

use of org.apache.lucene.search.TopDocs in project zm-mailbox by Zimbra.

the class RemoteMailQueue method search0.

private void search0(SearchResult result, IndexReader indexReader, Query query, int offset, int limit) throws IOException {
    if (ZimbraLog.rmgmt.isDebugEnabled()) {
        ZimbraLog.rmgmt.debug("searching query=" + query + " offset=" + offset + " limit=" + limit + " " + this);
    }
    Searcher searcher = null;
    try {
        searcher = new IndexSearcher(indexReader);
        TopDocs topDocs = searcher.search(query, (Filter) null, limit);
        ScoreDoc[] hits = topDocs.scoreDocs;
        if (offset < hits.length) {
            int n;
            if (limit <= 0) {
                n = hits.length;
            } else {
                n = Math.min(offset + limit, hits.length);
            }
            for (int i = offset; i < n; i++) {
                Document doc = searcher.doc(hits[i].doc);
                Map<QueueAttr, String> qitem = docToQueueItem(doc);
                result.qitems.add(qitem);
            }
        }
        result.hits = hits.length;
    } finally {
        if (searcher != null) {
            searcher.close();
        }
    }
}