Examples with TermsEnum org.apache.lucene.index.TermsEnum used on opensource projects

Example 61 with TermsEnum

use of org.apache.lucene.index.TermsEnum in project lucene-solr by apache.

the class TestMultiPhraseQuery method testPhrasePrefix.

public void testPhrasePrefix() throws IOException {
    Directory indexStore = newDirectory();
    RandomIndexWriter writer = new RandomIndexWriter(random(), indexStore);
    add("blueberry pie", writer);
    add("blueberry strudel", writer);
    add("blueberry pizza", writer);
    add("blueberry chewing gum", writer);
    add("bluebird pizza", writer);
    add("bluebird foobar pizza", writer);
    add("piccadilly circus", writer);
    IndexReader reader = writer.getReader();
    IndexSearcher searcher = newSearcher(reader);
    // search for "blueberry pi*":
    MultiPhraseQuery.Builder query1builder = new MultiPhraseQuery.Builder();
    // search for "strawberry pi*":
    MultiPhraseQuery.Builder query2builder = new MultiPhraseQuery.Builder();
    query1builder.add(new Term("body", "blueberry"));
    query2builder.add(new Term("body", "strawberry"));
    LinkedList<Term> termsWithPrefix = new LinkedList<>();
    // this TermEnum gives "piccadilly", "pie" and "pizza".
    String prefix = "pi";
    TermsEnum te = MultiFields.getFields(reader).terms("body").iterator();
    te.seekCeil(new BytesRef(prefix));
    do {
        String s = te.term().utf8ToString();
        if (s.startsWith(prefix)) {
            termsWithPrefix.add(new Term("body", s));
        } else {
            break;
        }
    } while (te.next() != null);
    query1builder.add(termsWithPrefix.toArray(new Term[0]));
    MultiPhraseQuery query1 = query1builder.build();
    assertEquals("body:\"blueberry (piccadilly pie pizza)\"", query1.toString());
    query2builder.add(termsWithPrefix.toArray(new Term[0]));
    MultiPhraseQuery query2 = query2builder.build();
    assertEquals("body:\"strawberry (piccadilly pie pizza)\"", query2.toString());
    ScoreDoc[] result;
    result = searcher.search(query1, 1000).scoreDocs;
    assertEquals(2, result.length);
    result = searcher.search(query2, 1000).scoreDocs;
    assertEquals(0, result.length);
    // search for "blue* pizza":
    MultiPhraseQuery.Builder query3builder = new MultiPhraseQuery.Builder();
    termsWithPrefix.clear();
    prefix = "blue";
    te.seekCeil(new BytesRef(prefix));
    do {
        if (te.term().utf8ToString().startsWith(prefix)) {
            termsWithPrefix.add(new Term("body", te.term().utf8ToString()));
        }
    } while (te.next() != null);
    query3builder.add(termsWithPrefix.toArray(new Term[0]));
    query3builder.add(new Term("body", "pizza"));
    MultiPhraseQuery query3 = query3builder.build();
    result = searcher.search(query3, 1000).scoreDocs;
    // blueberry pizza, bluebird pizza
    assertEquals(2, result.length);
    assertEquals("body:\"(blueberry bluebird) pizza\"", query3.toString());
    // test slop:
    query3builder.setSlop(1);
    query3 = query3builder.build();
    result = searcher.search(query3, 1000).scoreDocs;
    // just make sure no exc:
    searcher.explain(query3, 0);
    // blueberry pizza, bluebird pizza, bluebird
    assertEquals(3, result.length);
    // foobar pizza
    MultiPhraseQuery.Builder query4builder = new MultiPhraseQuery.Builder();
    expectThrows(IllegalArgumentException.class, () -> {
        query4builder.add(new Term("field1", "foo"));
        query4builder.add(new Term("field2", "foobar"));
    });
    writer.close();
    reader.close();
    indexStore.close();
}

Example 62 with TermsEnum

use of org.apache.lucene.index.TermsEnum in project lucene-solr by apache.

the class SimpleTextDocValuesWriter method addSortedField.

@Override
public void addSortedField(FieldInfo field, DocValuesProducer valuesProducer) throws IOException {
    assert fieldSeen(field.name);
    assert field.getDocValuesType() == DocValuesType.SORTED;
    writeFieldEntry(field, DocValuesType.SORTED);
    int valueCount = 0;
    int maxLength = -1;
    TermsEnum terms = valuesProducer.getSorted(field).termsEnum();
    for (BytesRef value = terms.next(); value != null; value = terms.next()) {
        maxLength = Math.max(maxLength, value.length);
        valueCount++;
    }
    // write numValues
    SimpleTextUtil.write(data, NUMVALUES);
    SimpleTextUtil.write(data, Integer.toString(valueCount), scratch);
    SimpleTextUtil.writeNewline(data);
    // write maxLength
    SimpleTextUtil.write(data, MAXLENGTH);
    SimpleTextUtil.write(data, Integer.toString(maxLength), scratch);
    SimpleTextUtil.writeNewline(data);
    int maxBytesLength = Integer.toString(maxLength).length();
    StringBuilder sb = new StringBuilder();
    for (int i = 0; i < maxBytesLength; i++) {
        sb.append('0');
    }
    // write our pattern for encoding lengths
    SimpleTextUtil.write(data, PATTERN);
    SimpleTextUtil.write(data, sb.toString(), scratch);
    SimpleTextUtil.writeNewline(data);
    final DecimalFormat encoder = new DecimalFormat(sb.toString(), new DecimalFormatSymbols(Locale.ROOT));
    int maxOrdBytes = Long.toString(valueCount + 1L).length();
    sb.setLength(0);
    for (int i = 0; i < maxOrdBytes; i++) {
        sb.append('0');
    }
    // write our pattern for ords
    SimpleTextUtil.write(data, ORDPATTERN);
    SimpleTextUtil.write(data, sb.toString(), scratch);
    SimpleTextUtil.writeNewline(data);
    final DecimalFormat ordEncoder = new DecimalFormat(sb.toString(), new DecimalFormatSymbols(Locale.ROOT));
    // for asserts:
    int valuesSeen = 0;
    terms = valuesProducer.getSorted(field).termsEnum();
    for (BytesRef value = terms.next(); value != null; value = terms.next()) {
        // write length
        SimpleTextUtil.write(data, LENGTH);
        SimpleTextUtil.write(data, encoder.format(value.length), scratch);
        SimpleTextUtil.writeNewline(data);
        // write bytes -- don't use SimpleText.write
        // because it escapes:
        data.writeBytes(value.bytes, value.offset, value.length);
        // pad to fit
        for (int i = value.length; i < maxLength; i++) {
            data.writeByte((byte) ' ');
        }
        SimpleTextUtil.writeNewline(data);
        valuesSeen++;
        assert valuesSeen <= valueCount;
    }
    assert valuesSeen == valueCount;
    SortedDocValues values = valuesProducer.getSorted(field);
    for (int i = 0; i < numDocs; ++i) {
        if (values.docID() < i) {
            values.nextDoc();
            assert values.docID() >= i;
        }
        int ord = -1;
        if (values.docID() == i) {
            ord = values.ordValue();
        }
        SimpleTextUtil.write(data, ordEncoder.format(ord + 1L), scratch);
        SimpleTextUtil.writeNewline(data);
    }
}

Example 63 with TermsEnum

use of org.apache.lucene.index.TermsEnum in project lucene-solr by apache.

the class SimpleTextDocValuesWriter method addSortedSetField.

@Override
public void addSortedSetField(FieldInfo field, DocValuesProducer valuesProducer) throws IOException {
    assert fieldSeen(field.name);
    assert field.getDocValuesType() == DocValuesType.SORTED_SET;
    writeFieldEntry(field, DocValuesType.SORTED_SET);
    long valueCount = 0;
    int maxLength = 0;
    TermsEnum terms = valuesProducer.getSortedSet(field).termsEnum();
    for (BytesRef value = terms.next(); value != null; value = terms.next()) {
        maxLength = Math.max(maxLength, value.length);
        valueCount++;
    }
    // write numValues
    SimpleTextUtil.write(data, NUMVALUES);
    SimpleTextUtil.write(data, Long.toString(valueCount), scratch);
    SimpleTextUtil.writeNewline(data);
    // write maxLength
    SimpleTextUtil.write(data, MAXLENGTH);
    SimpleTextUtil.write(data, Integer.toString(maxLength), scratch);
    SimpleTextUtil.writeNewline(data);
    int maxBytesLength = Integer.toString(maxLength).length();
    StringBuilder sb = new StringBuilder();
    for (int i = 0; i < maxBytesLength; i++) {
        sb.append('0');
    }
    // write our pattern for encoding lengths
    SimpleTextUtil.write(data, PATTERN);
    SimpleTextUtil.write(data, sb.toString(), scratch);
    SimpleTextUtil.writeNewline(data);
    final DecimalFormat encoder = new DecimalFormat(sb.toString(), new DecimalFormatSymbols(Locale.ROOT));
    // compute ord pattern: this is funny, we encode all values for all docs to find the maximum length
    int maxOrdListLength = 0;
    StringBuilder sb2 = new StringBuilder();
    SortedSetDocValues values = valuesProducer.getSortedSet(field);
    for (int doc = values.nextDoc(); doc != DocIdSetIterator.NO_MORE_DOCS; doc = values.nextDoc()) {
        sb2.setLength(0);
        for (long ord = values.nextOrd(); ord != SortedSetDocValues.NO_MORE_ORDS; ord = values.nextOrd()) {
            if (sb2.length() > 0) {
                sb2.append(",");
            }
            sb2.append(Long.toString(ord));
        }
        maxOrdListLength = Math.max(maxOrdListLength, sb2.length());
    }
    sb2.setLength(0);
    for (int i = 0; i < maxOrdListLength; i++) {
        sb2.append('X');
    }
    // write our pattern for ord lists
    SimpleTextUtil.write(data, ORDPATTERN);
    SimpleTextUtil.write(data, sb2.toString(), scratch);
    SimpleTextUtil.writeNewline(data);
    // for asserts:
    long valuesSeen = 0;
    terms = valuesProducer.getSortedSet(field).termsEnum();
    for (BytesRef value = terms.next(); value != null; value = terms.next()) {
        // write length
        SimpleTextUtil.write(data, LENGTH);
        SimpleTextUtil.write(data, encoder.format(value.length), scratch);
        SimpleTextUtil.writeNewline(data);
        // write bytes -- don't use SimpleText.write
        // because it escapes:
        data.writeBytes(value.bytes, value.offset, value.length);
        // pad to fit
        for (int i = value.length; i < maxLength; i++) {
            data.writeByte((byte) ' ');
        }
        SimpleTextUtil.writeNewline(data);
        valuesSeen++;
        assert valuesSeen <= valueCount;
    }
    assert valuesSeen == valueCount;
    values = valuesProducer.getSortedSet(field);
    // write the ords for each doc comma-separated
    for (int i = 0; i < numDocs; ++i) {
        if (values.docID() < i) {
            values.nextDoc();
            assert values.docID() >= i;
        }
        sb2.setLength(0);
        if (values.docID() == i) {
            for (long ord = values.nextOrd(); ord != SortedSetDocValues.NO_MORE_ORDS; ord = values.nextOrd()) {
                if (sb2.length() > 0) {
                    sb2.append(",");
                }
                sb2.append(Long.toString(ord));
            }
        }
        // now pad to fit: these are numbers so spaces work well. reader calls trim()
        int numPadding = maxOrdListLength - sb2.length();
        for (int j = 0; j < numPadding; j++) {
            sb2.append(' ');
        }
        SimpleTextUtil.write(data, sb2.toString(), scratch);
        SimpleTextUtil.writeNewline(data);
    }
}

Example 64 with TermsEnum

use of org.apache.lucene.index.TermsEnum in project lucene-solr by apache.

the class DocValuesConsumer method mergeSortedSetField.

/**
   * Merges the sortedset docvalues from <code>toMerge</code>.
   * <p>
   * The default implementation calls {@link #addSortedSetField}, passing
   * an Iterable that merges ordinals and values and filters deleted documents .
   */
public void mergeSortedSetField(FieldInfo mergeFieldInfo, final MergeState mergeState) throws IOException {
    List<SortedSetDocValues> toMerge = new ArrayList<>();
    for (int i = 0; i < mergeState.docValuesProducers.length; i++) {
        SortedSetDocValues values = null;
        DocValuesProducer docValuesProducer = mergeState.docValuesProducers[i];
        if (docValuesProducer != null) {
            FieldInfo fieldInfo = mergeState.fieldInfos[i].fieldInfo(mergeFieldInfo.name);
            if (fieldInfo != null && fieldInfo.getDocValuesType() == DocValuesType.SORTED_SET) {
                values = docValuesProducer.getSortedSet(fieldInfo);
            }
        }
        if (values == null) {
            values = DocValues.emptySortedSet();
        }
        toMerge.add(values);
    }
    // step 1: iterate thru each sub and mark terms still in use
    TermsEnum[] liveTerms = new TermsEnum[toMerge.size()];
    long[] weights = new long[liveTerms.length];
    for (int sub = 0; sub < liveTerms.length; sub++) {
        SortedSetDocValues dv = toMerge.get(sub);
        Bits liveDocs = mergeState.liveDocs[sub];
        if (liveDocs == null) {
            liveTerms[sub] = dv.termsEnum();
            weights[sub] = dv.getValueCount();
        } else {
            LongBitSet bitset = new LongBitSet(dv.getValueCount());
            int docID;
            while ((docID = dv.nextDoc()) != NO_MORE_DOCS) {
                if (liveDocs.get(docID)) {
                    long ord;
                    while ((ord = dv.nextOrd()) != SortedSetDocValues.NO_MORE_ORDS) {
                        bitset.set(ord);
                    }
                }
            }
            liveTerms[sub] = new BitsFilteredTermsEnum(dv.termsEnum(), bitset);
            weights[sub] = bitset.cardinality();
        }
    }
    // step 2: create ordinal map (this conceptually does the "merging")
    final OrdinalMap map = OrdinalMap.build(null, liveTerms, weights, PackedInts.COMPACT);
    // step 3: add field
    addSortedSetField(mergeFieldInfo, new EmptyDocValuesProducer() {

        @Override
        public SortedSetDocValues getSortedSet(FieldInfo fieldInfo) throws IOException {
            if (fieldInfo != mergeFieldInfo) {
                throw new IllegalArgumentException("wrong FieldInfo");
            }
            // We must make new iterators + DocIDMerger for each iterator:
            List<SortedSetDocValuesSub> subs = new ArrayList<>();
            long cost = 0;
            for (int i = 0; i < mergeState.docValuesProducers.length; i++) {
                SortedSetDocValues values = null;
                DocValuesProducer docValuesProducer = mergeState.docValuesProducers[i];
                if (docValuesProducer != null) {
                    FieldInfo readerFieldInfo = mergeState.fieldInfos[i].fieldInfo(mergeFieldInfo.name);
                    if (readerFieldInfo != null && readerFieldInfo.getDocValuesType() == DocValuesType.SORTED_SET) {
                        values = docValuesProducer.getSortedSet(readerFieldInfo);
                    }
                }
                if (values == null) {
                    values = DocValues.emptySortedSet();
                }
                cost += values.cost();
                subs.add(new SortedSetDocValuesSub(mergeState.docMaps[i], values, map.getGlobalOrds(i)));
            }
            final DocIDMerger<SortedSetDocValuesSub> docIDMerger = DocIDMerger.of(subs, mergeState.needsIndexSort);
            final long finalCost = cost;
            return new SortedSetDocValues() {

                private int docID = -1;

                private SortedSetDocValuesSub currentSub;

                @Override
                public int docID() {
                    return docID;
                }

                @Override
                public int nextDoc() throws IOException {
                    currentSub = docIDMerger.next();
                    if (currentSub == null) {
                        docID = NO_MORE_DOCS;
                    } else {
                        docID = currentSub.mappedDocID;
                    }
                    return docID;
                }

                @Override
                public int advance(int target) throws IOException {
                    throw new UnsupportedOperationException();
                }

                @Override
                public boolean advanceExact(int target) throws IOException {
                    throw new UnsupportedOperationException();
                }

                @Override
                public long nextOrd() throws IOException {
                    long subOrd = currentSub.values.nextOrd();
                    if (subOrd == NO_MORE_ORDS) {
                        return NO_MORE_ORDS;
                    }
                    return currentSub.map.get(subOrd);
                }

                @Override
                public long cost() {
                    return finalCost;
                }

                @Override
                public BytesRef lookupOrd(long ord) throws IOException {
                    int segmentNumber = map.getFirstSegmentNumber(ord);
                    long segmentOrd = map.getFirstSegmentOrd(ord);
                    return toMerge.get(segmentNumber).lookupOrd(segmentOrd);
                }

                @Override
                public long getValueCount() {
                    return map.getValueCount();
                }
            };
        }
    });
}

Example 65 with TermsEnum

use of org.apache.lucene.index.TermsEnum in project lucene-solr by apache.

the class TestOrdsBlockTree method testBasic.

public void testBasic() throws Exception {
    Directory dir = newDirectory();
    RandomIndexWriter w = new RandomIndexWriter(random(), dir);
    Document doc = new Document();
    doc.add(newTextField("field", "a b c", Field.Store.NO));
    w.addDocument(doc);
    IndexReader r = w.getReader();
    TermsEnum te = MultiFields.getTerms(r, "field").iterator();
    // Test next()
    assertEquals(new BytesRef("a"), te.next());
    assertEquals(0L, te.ord());
    assertEquals(new BytesRef("b"), te.next());
    assertEquals(1L, te.ord());
    assertEquals(new BytesRef("c"), te.next());
    assertEquals(2L, te.ord());
    assertNull(te.next());
    // Test seekExact by term
    assertTrue(te.seekExact(new BytesRef("b")));
    assertEquals(1, te.ord());
    assertTrue(te.seekExact(new BytesRef("a")));
    assertEquals(0, te.ord());
    assertTrue(te.seekExact(new BytesRef("c")));
    assertEquals(2, te.ord());
    // Test seekExact by ord
    te.seekExact(1);
    assertEquals(new BytesRef("b"), te.term());
    te.seekExact(0);
    assertEquals(new BytesRef("a"), te.term());
    te.seekExact(2);
    assertEquals(new BytesRef("c"), te.term());
    r.close();
    w.close();
    dir.close();
}