Search in sources :

Example 1 with FilterCollector

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

the class TestSort method testSort.

public void testSort() throws Exception {
    Directory dir = new RAMDirectory();
    Field f = new StringField("f", "0", Field.Store.NO);
    Field f2 = new StringField("f2", "0", Field.Store.NO);
    for (int iterCnt = 0; iterCnt < iter; iterCnt++) {
        IndexWriter iw = new IndexWriter(dir, new IndexWriterConfig(new SimpleAnalyzer()).setOpenMode(IndexWriterConfig.OpenMode.CREATE));
        final MyDoc[] mydocs = new MyDoc[ndocs];
        int v1EmptyPercent = 50;
        int v2EmptyPercent = 50;
        int commitCountdown = commitCount;
        for (int i = 0; i < ndocs; i++) {
            MyDoc mydoc = new MyDoc();
            mydoc.doc = i;
            mydocs[i] = mydoc;
            Document document = new Document();
            if (r.nextInt(100) < v1EmptyPercent) {
                mydoc.val = Integer.toString(r.nextInt(maxval));
                f.setStringValue(mydoc.val);
                document.add(f);
            }
            if (r.nextInt(100) < v2EmptyPercent) {
                mydoc.val2 = Integer.toString(r.nextInt(maxval));
                f2.setStringValue(mydoc.val2);
                document.add(f2);
            }
            iw.addDocument(document);
            if (--commitCountdown <= 0) {
                commitCountdown = commitCount;
                iw.commit();
            }
        }
        iw.close();
        Map<String, UninvertingReader.Type> mapping = new HashMap<>();
        mapping.put("f", UninvertingReader.Type.SORTED);
        mapping.put("f2", UninvertingReader.Type.SORTED);
        DirectoryReader reader = UninvertingReader.wrap(DirectoryReader.open(dir), mapping);
        IndexSearcher searcher = new IndexSearcher(reader);
        // System.out.println("segments="+searcher.getIndexReader().getSequentialSubReaders().length);
        assertTrue(reader.leaves().size() > 1);
        for (int i = 0; i < qiter; i++) {
            Filter filt = new Filter() {

                @Override
                public DocIdSet getDocIdSet(LeafReaderContext context, Bits acceptDocs) {
                    return BitsFilteredDocIdSet.wrap(randSet(context.reader().maxDoc()), acceptDocs);
                }

                @Override
                public String toString(String field) {
                    return "TestSortFilter";
                }

                @Override
                public boolean equals(Object other) {
                    return other == this;
                }

                @Override
                public int hashCode() {
                    return System.identityHashCode(this);
                }
            };
            int top = r.nextInt((ndocs >> 3) + 1) + 1;
            final boolean luceneSort = r.nextBoolean();
            final boolean sortMissingLast = !luceneSort && r.nextBoolean();
            final boolean sortMissingFirst = !luceneSort && !sortMissingLast;
            final boolean reverse = r.nextBoolean();
            List<SortField> sfields = new ArrayList<>();
            final boolean secondary = r.nextBoolean();
            final boolean luceneSort2 = r.nextBoolean();
            final boolean sortMissingLast2 = !luceneSort2 && r.nextBoolean();
            final boolean sortMissingFirst2 = !luceneSort2 && !sortMissingLast2;
            final boolean reverse2 = r.nextBoolean();
            if (r.nextBoolean())
                sfields.add(new SortField(null, SortField.Type.SCORE));
            // hit both use-cases of sort-missing-last
            sfields.add(Sorting.getStringSortField("f", reverse, sortMissingLast, sortMissingFirst));
            if (secondary) {
                sfields.add(Sorting.getStringSortField("f2", reverse2, sortMissingLast2, sortMissingFirst2));
            }
            if (r.nextBoolean())
                sfields.add(new SortField(null, SortField.Type.SCORE));
            Sort sort = new Sort(sfields.toArray(new SortField[sfields.size()]));
            final String nullRep = luceneSort || sortMissingFirst && !reverse || sortMissingLast && reverse ? "" : "zzz";
            final String nullRep2 = luceneSort2 || sortMissingFirst2 && !reverse2 || sortMissingLast2 && reverse2 ? "" : "zzz";
            boolean trackScores = r.nextBoolean();
            boolean trackMaxScores = r.nextBoolean();
            boolean scoreInOrder = r.nextBoolean();
            final TopFieldCollector topCollector = TopFieldCollector.create(sort, top, true, trackScores, trackMaxScores);
            final List<MyDoc> collectedDocs = new ArrayList<>();
            // delegate and collect docs ourselves
            Collector myCollector = new FilterCollector(topCollector) {

                @Override
                public LeafCollector getLeafCollector(LeafReaderContext context) throws IOException {
                    final int docBase = context.docBase;
                    return new FilterLeafCollector(super.getLeafCollector(context)) {

                        @Override
                        public void collect(int doc) throws IOException {
                            super.collect(doc);
                            collectedDocs.add(mydocs[docBase + doc]);
                        }
                    };
                }
            };
            searcher.search(filt, myCollector);
            Collections.sort(collectedDocs, (o1, o2) -> {
                String v1 = o1.val == null ? nullRep : o1.val;
                String v2 = o2.val == null ? nullRep : o2.val;
                int cmp = v1.compareTo(v2);
                if (reverse)
                    cmp = -cmp;
                if (cmp != 0)
                    return cmp;
                if (secondary) {
                    v1 = o1.val2 == null ? nullRep2 : o1.val2;
                    v2 = o2.val2 == null ? nullRep2 : o2.val2;
                    cmp = v1.compareTo(v2);
                    if (reverse2)
                        cmp = -cmp;
                }
                cmp = cmp == 0 ? o1.doc - o2.doc : cmp;
                return cmp;
            });
            TopDocs topDocs = topCollector.topDocs();
            ScoreDoc[] sdocs = topDocs.scoreDocs;
            for (int j = 0; j < sdocs.length; j++) {
                int id = sdocs[j].doc;
                if (id != collectedDocs.get(j).doc) {
                    log.error("Error at pos " + j + "\n\tsortMissingFirst=" + sortMissingFirst + " sortMissingLast=" + sortMissingLast + " reverse=" + reverse + "\n\tEXPECTED=" + collectedDocs);
                }
                assertEquals(id, collectedDocs.get(j).doc);
            }
        }
        reader.close();
    }
    dir.close();
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) SimpleAnalyzer(org.apache.lucene.analysis.core.SimpleAnalyzer) HashMap(java.util.HashMap) ArrayList(java.util.ArrayList) SortField(org.apache.lucene.search.SortField) Document(org.apache.lucene.document.Document) ScoreDoc(org.apache.lucene.search.ScoreDoc) TopDocs(org.apache.lucene.search.TopDocs) StringField(org.apache.lucene.document.StringField) SchemaField(org.apache.solr.schema.SchemaField) SortField(org.apache.lucene.search.SortField) Field(org.apache.lucene.document.Field) LeafCollector(org.apache.lucene.search.LeafCollector) FilterLeafCollector(org.apache.lucene.search.FilterLeafCollector) FilterCollector(org.apache.lucene.search.FilterCollector) Collector(org.apache.lucene.search.Collector) TopFieldCollector(org.apache.lucene.search.TopFieldCollector) LeafReaderContext(org.apache.lucene.index.LeafReaderContext) Sort(org.apache.lucene.search.Sort) TopFieldCollector(org.apache.lucene.search.TopFieldCollector) RAMDirectory(org.apache.lucene.store.RAMDirectory) Directory(org.apache.lucene.store.Directory) DirectoryReader(org.apache.lucene.index.DirectoryReader) RAMDirectory(org.apache.lucene.store.RAMDirectory) Type(org.apache.lucene.search.SortField.Type) FilterCollector(org.apache.lucene.search.FilterCollector) IndexWriter(org.apache.lucene.index.IndexWriter) StringField(org.apache.lucene.document.StringField) FilterLeafCollector(org.apache.lucene.search.FilterLeafCollector) Bits(org.apache.lucene.util.Bits) IndexWriterConfig(org.apache.lucene.index.IndexWriterConfig)

Example 2 with FilterCollector

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

the class DrillSideways method search.

/**
   * Search, collecting hits with a {@link Collector}, and
   * computing drill down and sideways counts.
   */
public DrillSidewaysResult search(DrillDownQuery query, Collector hitCollector) throws IOException {
    Map<String, Integer> drillDownDims = query.getDims();
    FacetsCollector drillDownCollector = new FacetsCollector();
    if (drillDownDims.isEmpty()) {
        // There are no drill-down dims, so there is no
        // drill-sideways to compute:
        searcher.search(query, MultiCollector.wrap(hitCollector, drillDownCollector));
        return new DrillSidewaysResult(buildFacetsResult(drillDownCollector, null, null), null);
    }
    Query baseQuery = query.getBaseQuery();
    if (baseQuery == null) {
        // TODO: we could optimize this pure-browse case by
        // making a custom scorer instead:
        baseQuery = new MatchAllDocsQuery();
    }
    Query[] drillDownQueries = query.getDrillDownQueries();
    FacetsCollector[] drillSidewaysCollectors = new FacetsCollector[drillDownDims.size()];
    for (int i = 0; i < drillSidewaysCollectors.length; i++) {
        drillSidewaysCollectors[i] = new FacetsCollector();
    }
    DrillSidewaysQuery dsq = new DrillSidewaysQuery(baseQuery, drillDownCollector, drillSidewaysCollectors, drillDownQueries, scoreSubDocsAtOnce());
    if (hitCollector.needsScores() == false) {
        // this is a horrible hack in order to make sure IndexSearcher will not
        // attempt to cache the DrillSidewaysQuery
        hitCollector = new FilterCollector(hitCollector) {

            @Override
            public boolean needsScores() {
                return true;
            }
        };
    }
    searcher.search(dsq, hitCollector);
    return new DrillSidewaysResult(buildFacetsResult(drillDownCollector, drillSidewaysCollectors, drillDownDims.keySet().toArray(new String[drillDownDims.size()])), null);
}
Also used : Query(org.apache.lucene.search.Query) MatchAllDocsQuery(org.apache.lucene.search.MatchAllDocsQuery) MatchAllDocsQuery(org.apache.lucene.search.MatchAllDocsQuery) FilterCollector(org.apache.lucene.search.FilterCollector)

Aggregations

FilterCollector (org.apache.lucene.search.FilterCollector)2 ArrayList (java.util.ArrayList)1 HashMap (java.util.HashMap)1 SimpleAnalyzer (org.apache.lucene.analysis.core.SimpleAnalyzer)1 Document (org.apache.lucene.document.Document)1 Field (org.apache.lucene.document.Field)1 StringField (org.apache.lucene.document.StringField)1 DirectoryReader (org.apache.lucene.index.DirectoryReader)1 IndexWriter (org.apache.lucene.index.IndexWriter)1 IndexWriterConfig (org.apache.lucene.index.IndexWriterConfig)1 LeafReaderContext (org.apache.lucene.index.LeafReaderContext)1 Collector (org.apache.lucene.search.Collector)1 FilterLeafCollector (org.apache.lucene.search.FilterLeafCollector)1 IndexSearcher (org.apache.lucene.search.IndexSearcher)1 LeafCollector (org.apache.lucene.search.LeafCollector)1 MatchAllDocsQuery (org.apache.lucene.search.MatchAllDocsQuery)1 Query (org.apache.lucene.search.Query)1 ScoreDoc (org.apache.lucene.search.ScoreDoc)1 Sort (org.apache.lucene.search.Sort)1 SortField (org.apache.lucene.search.SortField)1