Example usage for org.apache.lucene.search.spans SpanQuery rewrite

List of usage examples for org.apache.lucene.search.spans SpanQuery rewrite

  1. HOME
  2. Java
  3. org.apache.lucene
  4. org.apache.lucene.search.spans.*
  5. SpanQuery
  6. rewrite

Introduction

In this page you can find the example usage for org.apache.lucene.search.spans SpanQuery rewrite.

Prototype

public Query rewrite(IndexReader reader) throws IOException

Source Link

Document

Expert: called to re-write queries into primitive queries.

Usage

From source file:edu.mit.ll.vizlinc.highlight.WeightedSpanTermExtractor.java

License:Apache License

/**
 * Fills a <code>Map</code> with <@link WeightedSpanTerm>s using the terms from the supplied <code>SpanQuery</code>.
 * /*  w  ww .j a  v  a  2 s .com*/
 * @param terms
 *          Map to place created WeightedSpanTerms in
 * @param spanQuery
 *          SpanQuery to extract Terms from
 * @throws IOException
 */
private void extractWeightedSpanTerms(Map<String, WeightedSpanTerm> terms, SpanQuery spanQuery)
        throws IOException {
    Set<String> fieldNames;

    if (fieldName == null) {
        fieldNames = new HashSet<String>();
        collectSpanQueryFields(spanQuery, fieldNames);
    } else {
        fieldNames = new HashSet<String>(1);
        fieldNames.add(fieldName);
    }
    // To support the use of the default field name
    if (defaultField != null) {
        fieldNames.add(defaultField);
    }

    Map<String, SpanQuery> queries = new HashMap<String, SpanQuery>();

    Set<Term> nonWeightedTerms = new HashSet<Term>();
    final boolean mustRewriteQuery = mustRewriteQuery(spanQuery);
    if (mustRewriteQuery) {
        for (final String field : fieldNames) {
            final SpanQuery rewrittenQuery = (SpanQuery) spanQuery.rewrite(getReaderForField(field));
            queries.put(field, rewrittenQuery);
            rewrittenQuery.extractTerms(nonWeightedTerms);
        }
    } else {
        spanQuery.extractTerms(nonWeightedTerms);
    }

    List<PositionSpan> spanPositions = new ArrayList<PositionSpan>();

    for (final String field : fieldNames) {

        IndexReader reader = getReaderForField(field);
        final Spans spans;
        if (mustRewriteQuery) {
            spans = queries.get(field).getSpans(reader);
        } else {
            spans = spanQuery.getSpans(reader);
        }

        // collect span positions
        while (spans.next()) {
            spanPositions.add(new PositionSpan(spans.start(), spans.end() - 1));
        }

    }

    if (spanPositions.size() == 0) {
        // no spans found
        return;
    }

    for (final Term queryTerm : nonWeightedTerms) {

        if (fieldNameComparator(queryTerm.field())) {
            WeightedSpanTerm weightedSpanTerm = terms.get(queryTerm.text());

            if (weightedSpanTerm == null) {
                weightedSpanTerm = new WeightedSpanTerm(spanQuery.getBoost(), queryTerm.text());
                weightedSpanTerm.addPositionSpans(spanPositions);
                weightedSpanTerm.positionSensitive = true;
                terms.put(queryTerm.text(), weightedSpanTerm);
            } else {
                if (spanPositions.size() > 0) {
                    weightedSpanTerm.addPositionSpans(spanPositions);
                }
            }
        }
    }
}

From source file:nl.inl.blacklab.search.Hits.java

License:Apache License

/**
 * Construct an empty Hits object//  w  w  w  .  j a va  2 s.c om
 *
 * @param searcher
 *            the searcher object
 * @param concordanceFieldPropName
 *            field to use by default when finding concordances
 * @param sourceQuery
 *            the query to execute to get the hits
 * @throws TooManyClauses if the query is overly broad (expands to too many terms)
 */
public Hits(Searcher searcher, String concordanceFieldPropName, SpanQuery sourceQuery) throws TooManyClauses {
    this(searcher, concordanceFieldPropName);

    try {
        DirectoryReader reader = searcher == null ? null : searcher.getIndexReader();
        spanQuery = (SpanQuery) sourceQuery.rewrite(reader);
        termContexts = new HashMap<Term, TermContext>();
        Set<Term> terms = new HashSet<Term>();
        spanQuery.extractTerms(terms);
        etiquette = new ThreadPriority();
        for (Term term : terms) {
            try {
                etiquette.behave();
            } catch (InterruptedException e) {
                // Taking too long, break it off.
                // Not a very graceful way to do it... but at least it won't
                // be stuck forever.
                Thread.currentThread().interrupt(); // client can check this
                throw new RuntimeException("Query matches too many terms; aborted.");
            }
            termContexts.put(term, TermContext.build(reader.getContext(), term, true));
        }

        currentSourceSpans = null;
        atomicReaderContexts = reader == null ? null : reader.leaves();
        atomicReaderContextIndex = -1;
        hitQueryContext = new HitQueryContext(); // to keep track of captured groups, etc.
        //sourceSpans = BLSpansWrapper.optWrap(spanQuery.getSpans(srw != null ? srw.getContext() : null, srw != null ? srw.getLiveDocs() : null, termContexts));
    } catch (IOException e) {
        throw new RuntimeException(e);
    }

    sourceSpansFullyRead = false;
}

From source file:nl.inl.blacklab.search.HitsImpl.java

License:Apache License

/**
 * Construct a Hits object from a SpanQuery.
 *
 * @param searcher//from  w w w.j  av a 2  s  .c  om
 *            the searcher object
 * @param concordanceFieldPropName
 *            field to use by default when finding concordances
 * @param sourceQuery
 *            the query to execute to get the hits
 * @throws TooManyClauses if the query is overly broad (expands to too many terms)
 */
HitsImpl(Searcher searcher, SpanQuery sourceQuery) throws TooManyClauses {
    this(searcher, (List<Hit>) null);
    try {
        IndexReader reader = searcher.getIndexReader();
        spanQuery = (SpanQuery) sourceQuery.rewrite(reader);
        termContexts = new HashMap<>();
        Set<Term> terms = new HashSet<>();
        extractTermsFromSpanQuery(terms);
        etiquette = new ThreadPriority();
        for (Term term : terms) {
            try {
                etiquette.behave();
            } catch (InterruptedException e) {
                // Taking too long, break it off.
                // Not a very graceful way to do it... but at least it won't
                // be stuck forever.
                Thread.currentThread().interrupt(); // client can check this
                throw new RuntimeException("Query matches too many terms; aborted.");
            }
            termContexts.put(term, TermContext.build(reader.getContext(), term));
        }

        currentSourceSpans = null;
        atomicReaderContexts = reader == null ? null : reader.leaves();
        atomicReaderContextIndex = -1;
        //sourceSpans = BLSpansWrapper.optWrap(spanQuery.getSpans(srw != null ? srw.getContext() : null, srw != null ? srw.getLiveDocs() : null, termContexts));
    } catch (IOException e) {
        throw new RuntimeException(e);
    }

    sourceSpansFullyRead = false;
}

From source file:nl.inl.blacklab.search.lucene.BLSpanOrQuery.java

License:Apache License

@Override
public Query rewrite(IndexReader reader) throws IOException {
    BLSpanOrQuery clone = null;/*  w  w w . ja v  a  2 s  . com*/
    for (int i = 0; i < clauses.size(); i++) {
        SpanQuery c = clauses.get(i);
        SpanQuery query = (SpanQuery) c.rewrite(reader);
        if (query != c) { // clause rewrote: must clone
            if (clone == null)
                clone = this.clone();
            clone.clauses.set(i, query);
        }
    }
    if (clone != null) {
        return clone; // some clauses rewrote
    }
    return this; // no clauses rewrote
}

From source file:nl.inl.blacklab.search.lucene.SpanQueryAndNot.java

License:Apache License

@Override
public Query rewrite(IndexReader reader) throws IOException {
    SpanQueryAndNot clone = null;// w w  w  . ja v  a  2  s.  c  o m

    for (int i = 0; i < clauses.length; i++) {
        SpanQuery c = clauses[i];
        SpanQuery query = (SpanQuery) c.rewrite(reader);
        if (query != c) { // clause rewrote: must clone
            if (clone == null)
                clone = (SpanQueryAndNot) clone();
            clone.clauses[i] = query;
        }
    }
    if (clone != null) {
        return clone; // some clauses rewrote
    }
    return this; // no clauses rewrote
}

From source file:nl.inl.blacklab.search.lucene.SpanQueryBase.java

License:Apache License

@Override
public Query rewrite(IndexReader reader) throws IOException {
    SpanQueryBase clone = null;// ww  w. j  ava  2  s. co m
    for (int i = 0; i < clauses.length; i++) {
        SpanQuery c = clauses[i];
        SpanQuery query = c == null ? null : (SpanQuery) c.rewrite(reader);
        if (query != c) {
            // clause rewritten: must clone
            if (clone == null)
                clone = (SpanQueryBase) clone();
            clone.clauses[i] = query;
        }
    }
    if (clone != null) {
        return clone; // some clauses rewritten
    }
    return this; // no clauses rewritten
}

From source file:nl.inl.blacklab.search.lucene.SpanQueryDocLevelAndNot.java

License:Apache License

@Override
public Query rewrite(IndexReader reader) throws IOException {
    SpanQueryDocLevelAndNot clone = null;

    for (int i = 0; i < clauses.length; i++) {
        SpanQuery c = clauses[i];
        SpanQuery query = (SpanQuery) c.rewrite(reader);
        if (query != c) { // clause rewrote: must clone
            if (clone == null)
                clone = (SpanQueryDocLevelAndNot) clone();
            clone.clauses[i] = query;//from  w  ww .ja  v a  2 s  . c  o  m
        }
    }
    if (clone != null) {
        return clone; // some clauses rewrote
    }
    return this; // no clauses rewrote
}

From source file:org.apache.solr.analysis.entity.TestEntityFilterFactory.java

License:Apache License

private long countSpans(SpanQuery q) throws Exception {
    List<AtomicReaderContext> ctxs = reader.leaves();
    assert (ctxs.size() == 1);
    AtomicReaderContext ctx = ctxs.get(0);
    q = (SpanQuery) q.rewrite(ctx.reader());
    Spans spans = q.getSpans(ctx, null, new HashMap<Term, TermContext>());

    long i = 0;//from   w ww  . j  a v  a2s.  com
    while (spans.next()) {
        i++;
    }
    return i;
}

From source file:org.apache.solr.analysis.entity.TestEntityFilterFactory.java

License:Apache License

private long countDocs(SpanQuery q) throws Exception {
    OpenBitSet docs = new OpenBitSet();
    List<AtomicReaderContext> ctxs = reader.leaves();
    assert (ctxs.size() == 1);
    AtomicReaderContext ctx = ctxs.get(0);
    IndexReaderContext parentCtx = reader.getContext();
    q = (SpanQuery) q.rewrite(ctx.reader());

    Set<Term> qTerms = new HashSet<Term>();
    q.extractTerms(qTerms);//w w  w  .  j a v a2 s. c om
    Map<Term, TermContext> termContexts = new HashMap<Term, TermContext>();

    for (Term t : qTerms) {
        TermContext c = TermContext.build(parentCtx, t);
        termContexts.put(t, c);
    }

    Spans spans = q.getSpans(ctx, null, termContexts);

    while (spans.next()) {
        docs.set(spans.doc());
    }
    long spanDocHits = docs.cardinality();
    // double check with a regular searcher
    TotalHitCountCollector coll = new TotalHitCountCollector();
    searcher.search(q, coll);
    assertEquals(coll.getTotalHits(), spanDocHits);
    return spanDocHits;

}

From source file:org.tallison.lucene.queryparser.spans.SQPTestBase.java

License:Apache License

long countSpans(String field, Query q) throws Exception {
    List<LeafReaderContext> ctxs = reader.leaves();

    assert (ctxs.size() == 1);
    LeafReaderContext leafReaderContext = ctxs.get(0);
    SpanQuery sq = convert(field, q);
    sq = (SpanQuery) sq.rewrite(reader);
    SpanWeight sw = sq.createWeight(searcher, ScoreMode.COMPLETE_NO_SCORES, 1.0f);

    final Spans spans = sw.getSpans(leafReaderContext, SpanWeight.Postings.POSITIONS);

    long i = 0;/* w w w .  j a va 2 s .c  om*/
    if (spans != null) {
        while (spans.nextDoc() != Spans.NO_MORE_DOCS) {
            while (spans.nextStartPosition() != Spans.NO_MORE_POSITIONS) {
                i++;
            }
        }
    }
    return i;
}