Example usage for org.apache.lucene.queries.mlt MoreLikeThis setBoostFactor

List of usage examples for org.apache.lucene.queries.mlt MoreLikeThis setBoostFactor

  1. HOME
  2. Java
  3. org.apache.lucene
  4. org.apache.lucene.queries.mlt.*
  5. MoreLikeThis
  6. setBoostFactor

Introduction

In this page you can find the example usage for org.apache.lucene.queries.mlt MoreLikeThis setBoostFactor.

Prototype

public void setBoostFactor(float boostFactor)

Source Link

Document

Sets the boost factor to use when boosting terms

Usage

From source file:com.qwazr.search.query.MoreLikeThisQuery.java

License:Apache License

@Override
final public Query getQuery(QueryContext queryContext) throws IOException, ParseException {
    Objects.requireNonNull(doc_num, "The doc_num field is missing");
    final MoreLikeThis mlt = new MoreLikeThis(queryContext.indexSearcher.getIndexReader());
    if (is_boost != null)
        mlt.setBoost(is_boost);/*from   w ww  .j a  v  a2  s.  c om*/
    if (boost_factor != null)
        mlt.setBoostFactor(boost_factor);
    if (fieldnames != null)
        mlt.setFieldNames(fieldnames);
    if (max_doc_freq != null)
        mlt.setMaxDocFreq(max_doc_freq);
    if (max_doc_freq_pct != null)
        mlt.setMaxDocFreqPct(max_doc_freq_pct);
    if (max_num_tokens_parsed != null)
        mlt.setMaxNumTokensParsed(max_num_tokens_parsed);
    if (max_query_terms != null)
        mlt.setMaxQueryTerms(max_query_terms);
    if (max_word_len != null)
        mlt.setMaxWordLen(max_word_len);
    if (min_doc_freq != null)
        mlt.setMinDocFreq(min_doc_freq);
    if (min_term_freq != null)
        mlt.setMinTermFreq(min_term_freq);
    if (min_word_len != null)
        mlt.setMinWordLen(min_word_len);
    if (stop_words != null)
        mlt.setStopWords(stop_words);
    mlt.setAnalyzer(queryContext.analyzer);
    return mlt.like(doc_num);
}

From source file:org.apache.jackrabbit.oak.plugins.index.lucene.util.MoreLikeThisHelper.java

License:Apache License

public static Query getMoreLikeThis(IndexReader reader, Analyzer analyzer, String mltQueryString) {
    Query moreLikeThisQuery = null;
    MoreLikeThis mlt = new MoreLikeThis(reader);
    mlt.setAnalyzer(analyzer);/*from  w  w  w  . j a  va  2s  .c o  m*/
    try {
        String text = null;
        String[] fields = {};
        for (String param : mltQueryString.split("&")) {
            String[] keyValuePair = param.split("=");
            if (keyValuePair.length != 2 || keyValuePair[0] == null || keyValuePair[1] == null) {
                throw new RuntimeException("Unparsable native Lucene MLT query: " + mltQueryString);
            } else {
                if ("stream.body".equals(keyValuePair[0])) {
                    text = keyValuePair[1];
                } else if ("mlt.fl".equals(keyValuePair[0])) {
                    fields = keyValuePair[1].split(",");
                } else if ("mlt.mindf".equals(keyValuePair[0])) {
                    mlt.setMinDocFreq(Integer.parseInt(keyValuePair[1]));
                } else if ("mlt.mintf".equals(keyValuePair[0])) {
                    mlt.setMinTermFreq(Integer.parseInt(keyValuePair[1]));
                } else if ("mlt.boost".equals(keyValuePair[0])) {
                    mlt.setBoost(Boolean.parseBoolean(keyValuePair[1]));
                } else if ("mlt.qf".equals(keyValuePair[0])) {
                    mlt.setBoostFactor(Float.parseFloat(keyValuePair[1]));
                } else if ("mlt.maxdf".equals(keyValuePair[0])) {
                    mlt.setMaxDocFreq(Integer.parseInt(keyValuePair[1]));
                } else if ("mlt.maxdfp".equals(keyValuePair[0])) {
                    mlt.setMaxDocFreqPct(Integer.parseInt(keyValuePair[1]));
                } else if ("mlt.maxntp".equals(keyValuePair[0])) {
                    mlt.setMaxNumTokensParsed(Integer.parseInt(keyValuePair[1]));
                } else if ("mlt.maxqt".equals(keyValuePair[0])) {
                    mlt.setMaxQueryTerms(Integer.parseInt(keyValuePair[1]));
                } else if ("mlt.maxwl".equals(keyValuePair[0])) {
                    mlt.setMaxWordLen(Integer.parseInt(keyValuePair[1]));
                } else if ("mlt.minwl".equals(keyValuePair[0])) {
                    mlt.setMinWordLen(Integer.parseInt(keyValuePair[1]));
                }
            }
        }
        if (text != null) {
            if (FieldNames.PATH.equals(fields[0])) {
                IndexSearcher searcher = new IndexSearcher(reader);
                TermQuery q = new TermQuery(new Term(FieldNames.PATH, text));
                TopDocs top = searcher.search(q, 1);
                if (top.totalHits == 0) {
                    mlt.setFieldNames(fields);
                    moreLikeThisQuery = mlt.like(new StringReader(text), mlt.getFieldNames()[0]);
                } else {
                    ScoreDoc d = top.scoreDocs[0];
                    Document doc = reader.document(d.doc);
                    List<String> fieldNames = new ArrayList<String>();
                    for (IndexableField f : doc.getFields()) {
                        if (!FieldNames.PATH.equals(f.name())) {
                            fieldNames.add(f.name());
                        }
                    }
                    String[] docFields = fieldNames.toArray(new String[fieldNames.size()]);
                    mlt.setFieldNames(docFields);
                    moreLikeThisQuery = mlt.like(d.doc);
                }
            } else {
                mlt.setFieldNames(fields);
                moreLikeThisQuery = mlt.like(new StringReader(text), mlt.getFieldNames()[0]);
            }
        }
        return moreLikeThisQuery;
    } catch (Exception e) {
        throw new RuntimeException("could not handle MLT query " + mltQueryString);
    }
}

From source file:org.elasticsearch.common.lucene.search.MoreLikeThisQuery.java

License:Apache License

@Override
public Query rewrite(IndexReader reader) throws IOException {
    MoreLikeThis mlt = new MoreLikeThis(reader, similarity == null ? new DefaultSimilarity() : similarity);

    mlt.setFieldNames(moreLikeFields);/* ww w .j ava  2  s . co m*/
    mlt.setAnalyzer(analyzer);
    mlt.setMinTermFreq(minTermFrequency);
    mlt.setMinDocFreq(minDocFreq);
    mlt.setMaxDocFreq(maxDocFreq);
    mlt.setMaxQueryTerms(maxQueryTerms);
    mlt.setMinWordLen(minWordLen);
    mlt.setMaxWordLen(maxWordLen);
    mlt.setStopWords(stopWords);
    mlt.setBoost(boostTerms);
    mlt.setBoostFactor(boostTermsFactor);
    //LUCENE 4 UPGRADE this mapps the 3.6 behavior (only use the first field)
    BooleanQuery bq = (BooleanQuery) mlt.like(new FastStringReader(likeText), moreLikeFields[0]);
    BooleanClause[] clauses = bq.getClauses();

    bq.setMinimumNumberShouldMatch((int) (clauses.length * percentTermsToMatch));

    bq.setBoost(getBoost());
    return bq;
}

From source file:uk.gov.nationalarchives.discovery.taxonomy.common.service.impl.TSetBasedCategoriserServiceImpl.java

License:Mozilla Public License

/**
 * run More Like This process on a document by comparing its description to
 * the description of all items of the training set<br/>
 * currently we get a fixed number of the top results
 * //from ww w.  j a v a  2  s  .co  m
 * @param document
 *            document being tested
 * @return
 * @throws IOException
 */
public List<TSetBasedCategorisationResult> runMlt(Document document) {

    Map<String, TSetBasedCategorisationResult> result = null;
    IndexSearcher searcher = null;
    try {
        trainingSetSearcherManager.maybeRefresh();
        // Boolean wasRefreshed = trainingSetSearcherManager.maybeRefresh();
        // if (wasRefreshed) {
        // logger.debug(".runMlt: training set searcher had to be refreshed");
        // }
        searcher = trainingSetSearcherManager.acquire();

        // TODO TSETBASED refresh reader/searcher: Use readermanager and
        // refresh it?
        MoreLikeThis moreLikeThis = new MoreLikeThis(this.trainingSetIndexReader);
        moreLikeThis.setMinTermFreq(minTermFreq);
        moreLikeThis.setMinDocFreq(minDocFreq);
        moreLikeThis.setAnalyzer(this.trainingSetAnalyser);
        moreLikeThis.setFieldNames(fieldsToAnalyse.split(","));
        moreLikeThis.setBoost(true);

        BooleanQuery.Builder queryBuilder = new BooleanQuery.Builder();

        for (String fieldName : fieldsToAnalyse.split(",")) {
            String value = document.get(fieldName);
            if (value != null && !"null".equals(value)) {

                switch (InformationAssetViewFields.valueOf(fieldName)) {
                case DESCRIPTION:
                    moreLikeThis.setBoostFactor(descBoostingFactor);
                    break;
                case TITLE:
                    moreLikeThis.setBoostFactor(titleBoostingFactor);
                    break;
                case CONTEXTDESCRIPTION:
                    moreLikeThis.setBoostFactor(contextDescBoostingFactor);
                    break;
                default:
                case SUBJECTS:
                case CORPBODYS:
                case PERSON_FULLNAME:
                case PLACE_NAME:
                    moreLikeThis.setBoostFactor(1);
                    break;
                }
                Query query = moreLikeThis.like(fieldName, new StringReader(value));
                queryBuilder.add(query, Occur.SHOULD);
            }
        }
        BooleanQuery fullQuery = queryBuilder.build();

        TopDocs topDocs = searcher.search(fullQuery, this.maximumSimilarElements);
        logger.debug(".runMlt: found {} total hits, processed at maximum {} hits", topDocs.totalHits,
                this.maximumSimilarElements);

        result = new LinkedHashMap<String, TSetBasedCategorisationResult>();

        int size = 0;
        if (topDocs.totalHits <= this.maximumSimilarElements) {
            size = topDocs.totalHits - 1;
        } else {
            size = this.maximumSimilarElements - 1;
        }

        for (int i = 0; i < size; i++) {
            ScoreDoc scoreDoc = topDocs.scoreDocs[i];
            Float currrentScore = scoreDoc.score;

            if (currrentScore < this.mimimumScoreForMlt) {
                break;
            }

            Document hitDoc = searcher.doc(scoreDoc.doc);
            String category = hitDoc.get(InformationAssetViewFields.TAXONOMY.toString());
            String docReference = hitDoc.get(InformationAssetViewFields.DOCREFERENCE.toString());
            logger.debug(".runMlt: found doc, category: {}, score: {}, docreference: {}", category,
                    currrentScore, docReference);

            TSetBasedCategorisationResult existingCategorisationResult = result.get(category);
            Float scoreToSet = currrentScore;
            Integer numberOfFoundDocuments = 1;
            // k nearest neighbour algorithm
            if (existingCategorisationResult != null) {
                scoreToSet += existingCategorisationResult.getScore();
                numberOfFoundDocuments += existingCategorisationResult.getNumberOfFoundDocuments();
            }
            result.put(category,
                    new TSetBasedCategorisationResult(category, scoreToSet, numberOfFoundDocuments));

        }

    } catch (IOException e) {
        throw new TaxonomyException(TaxonomyErrorType.LUCENE_IO_EXCEPTION, e);
    } finally {
        LuceneHelperTools.releaseSearcherManagerQuietly(trainingSetSearcherManager, searcher);
    }

    List<TSetBasedCategorisationResult> sortedResults = sortCategorisationResultsByScoreDescAndFilterByGlobalScore(
            new ArrayList<TSetBasedCategorisationResult>(result.values()));

    return sortedResults;
}