Java Code Examples for org.apache.lucene.index.SortedSetDocValues#nextOrd()

private Optional<DocValues> createSortedSetDocValues(int docid, String field, DocValuesType dvType)
    throws IOException {
  SortedSetDocValues ssvalues = IndexUtils.getSortedSetDocvalues(reader, field);

  if (ssvalues.advanceExact(docid)) {
    List<BytesRef> values = new ArrayList<>();

    long ord;
    while ((ord = ssvalues.nextOrd()) != SortedSetDocValues.NO_MORE_ORDS) {
      values.add(BytesRef.deepCopyOf(ssvalues.lookupOrd(ord)));
    }

    DocValues dv = DocValues.of(
        dvType,
        values,
        Collections.emptyList()
    );
    return Optional.of(dv);
  }

  return Optional.empty();
}
 

/** accumulates per-segment multi-valued facet counts, mapping to global ordinal space on-the-fly */
static void accumMultiGeneric(int counts[], int startTermIndex, SortedSetDocValues si, DocIdSetIterator disi, int subIndex, OrdinalMap map) throws IOException {
  final LongValues ordMap = map == null ? null : map.getGlobalOrds(subIndex);
  int doc;
  while ((doc = disi.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
    if (si.advanceExact(doc)) {
      // strange do-while to collect the missing count (first ord is NO_MORE_ORDS)
      int term = (int) si.nextOrd();
      do {
        if (map != null) {
          term = (int) ordMap.get(term);
        }
        int arrIdx = term-startTermIndex;
        if (arrIdx>=0 && arrIdx<counts.length) counts[arrIdx]++;
      } while ((term = (int) si.nextOrd()) >= 0);
    } else if (startTermIndex == -1) {
      counts[0]++; // missing count
    }
  }
}
 

private void collectPerSeg(SortedSetDocValues multiDv, DocIdSetIterator disi, LongValues toGlobal) throws IOException {
  int segMax = (int)multiDv.getValueCount();
  final int[] counts = getCountArr( segMax );

  int doc;
  while ((doc = disi.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
    if (multiDv.advanceExact(doc)) {
      for(;;) {
        int segOrd = (int)multiDv.nextOrd();
        if (segOrd < 0) break;
        counts[segOrd]++;
      }
    }
  }

  for (int i=0; i<segMax; i++) {
    int segCount = counts[i];
    if (segCount > 0) {
      int slot = toGlobal == null ? (i) : (int) toGlobal.get(i);
      countAcc.incrementCount(slot, segCount);
    }
  }
}
 

/** "typical" multi-valued faceting: not too many unique values, no prefixing. maps to global ordinals as a separate step */
static void accumMultiSeg(int counts[], SortedSetDocValues si, DocIdSetIterator disi, int subIndex, OrdinalMap map) throws IOException {
  // First count in seg-ord space:
  final int segCounts[];
  if (map == null) {
    segCounts = counts;
  } else {
    segCounts = new int[1+(int)si.getValueCount()];
  }
  
  int doc;
  while ((doc = disi.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
    if (si.advanceExact(doc)) {
      int term = (int) si.nextOrd();
      do {
        segCounts[1+term]++;
      } while ((term = (int)si.nextOrd()) >= 0);
    } else {
      counts[0]++; // missing
    }
  }
  
  // migrate to global ords (if necessary)
  if (map != null) {
    migrateGlobal(counts, segCounts, subIndex, map);
  }
}
 

private void collectDocs(SortedSetDocValues multiDv, DocIdSetIterator disi, LongValues toGlobal) throws IOException {
  int doc;
  while ((doc = disi.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
    if (multiDv.advanceExact(doc)) {
      for(;;) {
        int segOrd = (int)multiDv.nextOrd();
        if (segOrd < 0) break;
        collect(doc, segOrd, toGlobal);
      }
    }
  }
}
 

private void collectCounts(SortedSetDocValues multiDv, DocIdSetIterator disi, LongValues toGlobal) throws IOException {
  int doc;
  while ((doc = disi.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
    if (multiDv.advanceExact(doc)) {
      for(;;) {
        int segOrd = (int)multiDv.nextOrd();
        if (segOrd < 0) break;
        int ord = (int)toGlobal.get(segOrd);
        countAcc.incrementCount(ord, 1);
      }
    }
  }
}
 

private void assertEquals(int maxDoc, SortedSetDocValues expected, SortedSetDocValues actual) throws Exception {
  // can be null for the segment if no docs actually had any SortedDocValues
  // in this case FC.getDocTermsOrds returns EMPTY
  if (actual == null) {
    assertEquals(expected.getValueCount(), 0);
    return;
  }
  assertEquals(expected.getValueCount(), actual.getValueCount());
  while (true) {
    int docID = expected.nextDoc();
    assertEquals(docID, actual.nextDoc());
    if (docID == NO_MORE_DOCS) {
      break;
    }
    long expectedOrd;
    while ((expectedOrd = expected.nextOrd()) != NO_MORE_ORDS) {
      assertEquals(expectedOrd, actual.nextOrd());
    }
    assertEquals(NO_MORE_ORDS, actual.nextOrd());
  }
  
  // compare ord dictionary
  for (long i = 0; i < expected.getValueCount(); i++) {
    final BytesRef expectedBytes = BytesRef.deepCopyOf(expected.lookupOrd(i));
    final BytesRef actualBytes = actual.lookupOrd(i);
    assertEquals(expectedBytes, actualBytes);
  }
  
  // compare termsenum
  assertEquals(expected.getValueCount(), expected.termsEnum(), actual.termsEnum());
}
 

@Override
public Weight createWeight(IndexSearcher searcher, ScoreMode scoreMode, float boost) throws IOException {
  return new ConstantScoreWeight(this, boost) {

    @Override
    public Scorer scorer(LeafReaderContext context) throws IOException {
      final SortedSetDocValues values = DocValues.getSortedSet(context.reader(), field);
      final LongBitSet bits = new LongBitSet(values.getValueCount());
      boolean matchesAtLeastOneTerm = false;
      TermIterator iterator = termData.iterator();
      for (BytesRef term = iterator.next(); term != null; term = iterator.next()) {
        final long ord = values.lookupTerm(term);
        if (ord >= 0) {
          matchesAtLeastOneTerm = true;
          bits.set(ord);
        }
      }
      if (matchesAtLeastOneTerm == false) {
        return null;
      }
      return new ConstantScoreScorer(this, score(), scoreMode, new TwoPhaseIterator(values) {

        @Override
        public boolean matches() throws IOException {
          for (long ord = values.nextOrd(); ord != SortedSetDocValues.NO_MORE_ORDS; ord = values.nextOrd()) {
            if (bits.get(ord)) {
              return true;
            }
          }
          return false;
        }

        @Override
        public float matchCost() {
          return 3; // lookup in a bitset
        }

      });
    }

    @Override
    public boolean isCacheable(LeafReaderContext ctx) {
      return DocValues.isCacheable(ctx, field);
    }

  };
}
 

private void accumIntervalsMulti(SortedSetDocValues ssdv,
                                 DocIdSetIterator disi, Bits bits) throws IOException {
  // First update the ordinals in the intervals for this segment
  for (FacetInterval interval : intervals) {
    interval.updateContext(ssdv);
  }

  int doc;
  while ((doc = disi.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
    if (bits != null && bits.get(doc) == false) {
      continue;
    }
    if (doc > ssdv.docID()) {
      ssdv.advance(doc);
    }
    if (doc == ssdv.docID()) {
      long currOrd;
      int currentInterval = 0;
      while ((currOrd = ssdv.nextOrd()) != SortedSetDocValues.NO_MORE_ORDS) {
        boolean evaluateNextInterval = true;
        while (evaluateNextInterval && currentInterval < intervals.length) {
          IntervalCompareResult result = intervals[currentInterval].includes(currOrd);
          switch (result) {
          case INCLUDED:
            /*
             * Increment the current interval and move to the next one using
             * the same value
             */
            intervals[currentInterval].incCount();
            currentInterval++;
            break;
          case LOWER_THAN_START:
            /*
             * None of the next intervals will match this value (all of them have 
             * higher start value). Move to the next value for this document. 
             */
            evaluateNextInterval = false;
            break;
          case GREATER_THAN_END:
            /*
             * Next interval may match this value
             */
            currentInterval++;
            break;
          }
        }
      }
    }
  }
}
 

@Override
public Weight createWeight(IndexSearcher searcher, ScoreMode scoreMode, float boost) throws IOException {
  if (! (searcher instanceof SolrIndexSearcher)) {
    log.debug("Falling back to JoinQueryWeight because searcher [{}] is not the required SolrIndexSearcher", searcher);
    return super.createWeight(searcher, scoreMode, boost);
  }

  final SolrIndexSearcher solrSearcher = (SolrIndexSearcher) searcher;
  final JoinQueryWeight weight = new JoinQueryWeight(solrSearcher, ScoreMode.COMPLETE_NO_SCORES, 1.0f);
  final SolrIndexSearcher fromSearcher = weight.fromSearcher;
  final SolrIndexSearcher toSearcher = weight.toSearcher;

  try {
    final SortedSetDocValues topLevelFromDocValues = validateAndFetchDocValues(fromSearcher, fromField, "from");
    final SortedSetDocValues topLevelToDocValues = validateAndFetchDocValues(toSearcher, toField, "to");
    if (topLevelFromDocValues.getValueCount() == 0 || topLevelToDocValues.getValueCount() == 0) {
      return createNoMatchesWeight(boost);
    }

    final LongBitSet fromOrdBitSet = findFieldOrdinalsMatchingQuery(q, fromField, fromSearcher, topLevelFromDocValues);
    final LongBitSet toOrdBitSet = new LongBitSet(topLevelToDocValues.getValueCount());
    final BitsetBounds toBitsetBounds = convertFromOrdinalsIntoToField(fromOrdBitSet, topLevelFromDocValues, toOrdBitSet, topLevelToDocValues);

    final boolean toMultivalued = toSearcher.getSchema().getFieldOrNull(toField).multiValued();
    return new ConstantScoreWeight(this, boost) {
      public Scorer scorer(LeafReaderContext context) throws IOException {
        if (toBitsetBounds.lower == BitsetBounds.NO_MATCHES) {
          return null;
        }

        final DocIdSetIterator toApproximation = (toMultivalued) ? context.reader().getSortedSetDocValues(toField) :
            context.reader().getSortedDocValues(toField);
        if (toApproximation == null) {
          return null;
        }

        final int docBase = context.docBase;
        return new ConstantScoreScorer(this, this.score(), scoreMode, new TwoPhaseIterator(toApproximation) {
          public boolean matches() throws IOException {
            final boolean hasDoc = topLevelToDocValues.advanceExact(docBase + approximation.docID());
            if (hasDoc) {
              for (long ord = topLevelToDocValues.nextOrd(); ord != -1L; ord = topLevelToDocValues.nextOrd()) {
                if (toOrdBitSet.get(ord)) {
                  return true;
                }
              }
            }
            return false;
          }

          public float matchCost() {
            return 10.0F;
          }
        });

      }

      public boolean isCacheable(LeafReaderContext ctx) {
        return false;
      }
    };
  } catch (IOException e) {
    throw new RuntimeException(e);
  }
}
 

@Override
public SortedSetDocValues getSortedSetDocValues(String field) throws IOException {
  final SortedSetDocValues sortedSetDocValues = in.getSortedSetDocValues(field);
  if (sortedSetDocValues == null) {
    return null;
  }
  return new SortedSetDocValues() {

    private boolean _access;

    @Override
    public void setDocument(int docID) {
      try {
        if (_access = _accessControl.hasAccess(ReadType.SORTED_SET_DOC_VALUE, docID)) {
          sortedSetDocValues.setDocument(docID);
        }
      } catch (IOException e) {
        throw new RuntimeException(e);
      }
    }

    @Override
    public long nextOrd() {
      if (_access) {
        return sortedSetDocValues.nextOrd();
      }
      return NO_MORE_ORDS;
    }

    @Override
    public void lookupOrd(long ord, BytesRef result) {
      if (_access) {
        sortedSetDocValues.lookupOrd(ord, result);
      } else {
        result.bytes = BinaryDocValues.MISSING;
        result.length = 0;
        result.offset = 0;
      }
    }

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