Examples with KeywordMatchingType edu.uci.ics.textdb.exp.keywordmatcher.KeywordMatchingType used on opensource projects

Example 1 with KeywordMatchingType

use of edu.uci.ics.textdb.exp.keywordmatcher.KeywordMatchingType in project textdb by TextDB.

the class DictionaryMatcherSourceOperator method getNextTuple.

/**
     * @about Gets the next matched tuple. <br>
     *        Returns the tuple with results in spanList. <br>
     * 
     *        Performs SCAN, KEYWORD_BASIC, or KEYWORD_PHRASE depends on the
     *        dictionary predicate. <br>
     * 
     *        DictionaryOperatorType.SCAN: <br>
     *        Scan the tuples using ScanSourceOperator. <br>
     *        For each tuple, loop through the dictionary and find results. <br>
     *        We assume the dictionary is smaller than the data at the source
     *        operator, we treat the data source as the outer relation to reduce
     *        the number of disk IOs. <br>
     * 
     *        DictionaryOperatorType.KEYWORD_BASIC, KEYWORD_PHRASE: <br>
     *        Use KeywordMatcher to find results. <br>
     * 
     *        KEYWORD_BASIC corresponds to KeywordOperatorType.BASIC, which
     *        performs keyword search on the document. The input query is
     *        tokenized. The order of the tokens doesn't matter. <br>
     * 
     *        KEYWORD_PHRASE corresponds to KeywordOperatorType.PHRASE, which
     *        performs phrase search on the document. The input query is
     *        tokenized. The order of the tokens does matter. Stopwords are
     *        treated as placeholders to indicate an arbitary token. <br>
     * 
     */
@Override
public Tuple getNextTuple() throws TextDBException {
    if (resultCursor >= limit + offset - 1) {
        return null;
    }
    if (predicate.getKeywordMatchingType() == KeywordMatchingType.PHRASE_INDEXBASED || predicate.getKeywordMatchingType() == KeywordMatchingType.CONJUNCTION_INDEXBASED) {
        while (true) {
            // If there's result from current keywordMatcher, return it.
            if ((sourceTuple = keywordSource.getNextTuple()) != null) {
                resultCursor++;
                if (resultCursor >= offset) {
                    return sourceTuple;
                }
                continue;
            }
            // return null if reach the end of dictionary.
            if ((currentDictionaryEntry = predicate.getDictionary().getNextEntry()) == null) {
                return null;
            }
            // Construct a new KeywordMatcher with the new dictionary entry.
            KeywordMatchingType keywordMatchingType;
            if (predicate.getKeywordMatchingType() == KeywordMatchingType.PHRASE_INDEXBASED) {
                keywordMatchingType = KeywordMatchingType.PHRASE_INDEXBASED;
            } else {
                keywordMatchingType = KeywordMatchingType.CONJUNCTION_INDEXBASED;
            }
            keywordSource.close();
            KeywordSourcePredicate keywordSourcePredicate = new KeywordSourcePredicate(currentDictionaryEntry, predicate.getAttributeNames(), predicate.getAnalyzerString(), keywordMatchingType, predicate.getTableName(), predicate.getSpanListName());
            keywordSource = new KeywordMatcherSourceOperator(keywordSourcePredicate);
            keywordSource.open();
        }
    } else // Substring matching (based on scan)
    {
        Tuple sourceTuple;
        Tuple resultTuple = null;
        while ((sourceTuple = indexSource.getNextTuple()) != null) {
            sourceTuple = DataflowUtils.getSpanTuple(sourceTuple.getFields(), new ArrayList<Span>(), outputSchema);
            resultTuple = computeMatchingResult(currentDictionaryEntry, sourceTuple);
            if (resultTuple != null) {
                resultCursor++;
            }
            if (resultTuple != null && resultCursor >= offset) {
                break;
            }
        }
        return resultTuple;
    }
}