Examples with NodeSet org.exist.dom.persistent.NodeSet used on opensource projects

Example 16 with NodeSet

use of org.exist.dom.persistent.NodeSet in project exist by eXist-db.

the class NGramIndexWorker method search.

public NodeSet search(final int contextId, final DocumentSet docs, final List<QName> qnames, final String query, final String ngram, final XQueryContext context, final NodeSet contextSet, final int axis) throws XPathException {
    final List<QName> searchQnames;
    if (qnames == null || qnames.isEmpty()) {
        searchQnames = getDefinedIndexes(context.getBroker(), docs);
    } else {
        searchQnames = qnames;
    }
    final NodeSet result = new ExtArrayNodeSet(docs.getDocumentCount(), 250);
    for (final Iterator<Collection> iter = docs.getCollectionIterator(); iter.hasNext(); ) {
        final int collectionId = iter.next().getId();
        for (final QName qname : searchQnames) {
            final NGramQNameKey key = new NGramQNameKey(collectionId, qname, index.getBrokerPool().getSymbols(), query);
            try (final ManagedLock<ReentrantLock> dbLock = lockManager.acquireBtreeReadLock(index.db.getLockName())) {
                final SearchCallback cb = new SearchCallback(contextId, query, ngram, docs, contextSet, context, result, axis == NodeSet.ANCESTOR);
                final int op = query.codePointCount(0, query.length()) < getN() ? IndexQuery.TRUNC_RIGHT : IndexQuery.EQ;
                index.db.query(new IndexQuery(op, key), cb);
            } catch (final LockException e) {
                LOG.warn("Failed to acquire lock for '{}'", FileUtils.fileName(index.db.getFile()), e);
            } catch (final IOException | BTreeException e) {
                LOG.error("{} in '{}'", e.getMessage(), FileUtils.fileName(index.db.getFile()), e);
            }
        }
    }
    // ensure result is ready to use
    result.iterate();
    return result;
}

Example 17 with NodeSet

use of org.exist.dom.persistent.NodeSet in project exist by eXist-db.

the class NGramSearch method eval.

@Override
public Sequence eval(Sequence contextSequence, Item contextItem) throws XPathException {
    if (contextItem != null)
        contextSequence = contextItem.toSequence();
    NodeSet result;
    if (preselectResult == null) {
        Sequence input = getArgument(0).eval(contextSequence, contextItem);
        if (input.isEmpty())
            result = NodeSet.EMPTY_SET;
        else {
            long start = System.currentTimeMillis();
            NodeSet inNodes = input.toNodeSet();
            DocumentSet docs = inNodes.getDocumentSet();
            NGramIndexWorker index = (NGramIndexWorker) context.getBroker().getIndexController().getWorkerByIndexId(NGramIndex.ID);
            // Alternate design
            // NGramIndexWorker index =
            // (NGramIndexWorker)context.getBroker().getBrokerPool().getIndexManager().getIndexById(NGramIndex.ID).getWorker();
            String key = getArgument(1).eval(contextSequence, contextItem).getStringValue();
            List<QName> qnames = null;
            if (contextQName != null) {
                qnames = new ArrayList<>(1);
                qnames.add(contextQName);
            }
            result = processMatches(index, docs, qnames, key, inNodes, NodeSet.ANCESTOR);
            if (context.getProfiler().traceFunctions()) {
                // report index use
                context.getProfiler().traceIndexUsage(context, "ngram", this, PerformanceStats.BASIC_INDEX, System.currentTimeMillis() - start);
            }
        }
    } else {
        contextStep.setPreloadedData(contextSequence.getDocumentSet(), preselectResult);
        result = getArgument(0).eval(contextSequence).toNodeSet();
    }
    return result;
}

Example 18 with NodeSet

use of org.exist.dom.persistent.NodeSet in project exist by eXist-db.

the class NGramSearch method fixedStringSearch.

public NodeSet fixedStringSearch(final NGramIndexWorker index, final DocumentSet docs, final List<QName> qnames, final String query, final NodeSet nodeSet, final int axis) throws XPathException {
    String[] ngrams = NGramSearch.getDistinctNGrams(query, index.getN());
    // Nothing to search for? The find nothing.
    if (ngrams.length == 0)
        return new EmptyNodeSet();
    String firstNgramm = ngrams[0];
    LOG.trace("First NGRAM: {}", firstNgramm);
    NodeSet result = index.search(getExpressionId(), docs, qnames, firstNgramm, firstNgramm, context, nodeSet, axis);
    for (int i = 1; i < ngrams.length; i++) {
        String ngram = ngrams[i];
        int len = ngram.codePointCount(0, ngram.length());
        int fillSize = index.getN() - len;
        String filledNgram = ngram;
        // ngrams lead to a considerable performance loss.
        if (fillSize > 0) {
            String filler = ngrams[i - 1];
            StringBuilder buf = new StringBuilder();
            int pos = filler.offsetByCodePoints(0, len);
            for (int j = 0; j < fillSize; j++) {
                int codepoint = filler.codePointAt(pos);
                pos += Character.charCount(codepoint);
                buf.appendCodePoint(codepoint);
            }
            buf.append(ngram);
            filledNgram = buf.toString();
            LOG.debug("Filled: {}", filledNgram);
        }
        NodeSet nodes = index.search(getExpressionId(), docs, qnames, filledNgram, ngram, context, nodeSet, axis);
        final NodeSet nodesContainingFirstINgrams = result;
        result = NodeSets.transformNodes(nodes, proxy -> Optional.ofNullable(nodesContainingFirstINgrams.get(proxy)).map(before -> getContinuousMatches(before, proxy)).orElse(null));
    }
    return result;
}

Example 19 with NodeSet

use of org.exist.dom.persistent.NodeSet in project exist by eXist-db.

the class NGramSearch method processMatches.

private NodeSet processMatches(NGramIndexWorker index, DocumentSet docs, List<QName> qnames, String query, NodeSet nodeSet, int axis) throws XPathException {
    EvaluatableExpression parsedQuery = null;
    if (getLocalName().equals("wildcard-contains"))
        parsedQuery = parseQuery(query);
    else
        parsedQuery = new FixedString(this, query);
    LOG.debug("Parsed Query: {}", parsedQuery);
    NodeSet result = parsedQuery.eval(index, docs, qnames, nodeSet, axis, this.getExpressionId());
    if (getLocalName().startsWith("starts-with"))
        result = NodeSets.getNodesMatchingAtStart(result, getExpressionId());
    else if (getLocalName().startsWith("ends-with"))
        result = NodeSets.getNodesMatchingAtEnd(result, getExpressionId());
    result = NodeSets.transformNodes(result, proxy -> NodeProxies.transformOwnMatches(proxy, Match::filterOutOverlappingOffsets, getExpressionId()));
    return result;
}

Example 20 with NodeSet

use of org.exist.dom.persistent.NodeSet in project exist by eXist-db.

the class Query method eval.

public Sequence eval(Sequence contextSequence, Item contextItem) throws XPathException {
    if (contextItem != null)
        contextSequence = contextItem.toSequence();
    if (contextSequence != null && !contextSequence.isPersistentSet())
        // in-memory docs won't have an index
        return Sequence.EMPTY_SEQUENCE;
    NodeSet result;
    if (preselectResult == null) {
        long start = System.currentTimeMillis();
        Sequence input = getArgument(0).eval(contextSequence);
        if (!(input instanceof VirtualNodeSet) && input.isEmpty())
            result = NodeSet.EMPTY_SET;
        else {
            NodeSet inNodes = input.toNodeSet();
            DocumentSet docs = inNodes.getDocumentSet();
            LuceneIndexWorker index = (LuceneIndexWorker) context.getBroker().getIndexController().getWorkerByIndexId(LuceneIndex.ID);
            Item key = getKey(contextSequence, contextItem);
            List<QName> qnames = null;
            if (contextQName != null) {
                qnames = new ArrayList<>(1);
                qnames.add(contextQName);
            }
            QueryOptions options = parseOptions(this, contextSequence, contextItem, 3);
            try {
                if (key != null && Type.subTypeOf(key.getType(), Type.ELEMENT)) {
                    final Element queryXML = (Element) ((NodeValue) key).getNode();
                    result = index.query(getExpressionId(), docs, inNodes, qnames, queryXML, NodeSet.ANCESTOR, options);
                } else {
                    final String query = key == null ? null : key.getStringValue();
                    result = index.query(getExpressionId(), docs, inNodes, qnames, query, NodeSet.ANCESTOR, options);
                }
            } catch (IOException | org.apache.lucene.queryparser.classic.ParseException e) {
                throw new XPathException(this, e.getMessage());
            }
        }
        if (context.getProfiler().traceFunctions()) {
            context.getProfiler().traceIndexUsage(context, "lucene", this, PerformanceStats.BASIC_INDEX, System.currentTimeMillis() - start);
        }
    } else {
        // DW: contextSequence can be null
        contextStep.setPreloadedData(contextSequence.getDocumentSet(), preselectResult);
        result = getArgument(0).eval(contextSequence).toNodeSet();
    }
    return result;
}