public OpenBitSetDISI TermToBitSet(string term, IndexReader indexReader)
{
var facetQuery = new TermQuery(new Term(this.Field, term));
var facetQueryFilter = new CachingWrapperFilter(new QueryWrapperFilter(facetQuery));
var bitSet = new OpenBitSetDISI(facetQueryFilter.GetDocIdSet(indexReader).Iterator(), indexReader.MaxDoc());
return bitSet;
}
private List<FacetMatch> FindMatchesInQuery(Facet facet, Filter query, Filter filter, IndexReader indexReader)
{
var matches = facet.Values.Select(value =>
{
var bitsQuery = new OpenBitSetDISI(query.GetDocIdSet(indexReader).Iterator(), indexReader.MaxDoc());
bitsQuery.And(value.Item2);
if (filter != null)
{
//TODO: Remove this hard coded value (1000)
var bitsFilter = new OpenBitSetDISI(filter.GetDocIdSet(indexReader).Iterator(), 1000);
bitsQuery.And(bitsFilter);
}
var count = bitsQuery.Cardinality();
return new FacetMatch() { Count = count, Value = value.Item1, Id = facet.Id };
}).ToList();
return matches;
}
private void doChain(OpenBitSetDISI result, int logic, DocIdSet dis)
{
if (dis is OpenBitSet)
{
// optimized case for OpenBitSets
switch (logic)
{
case OR:
result.Or((OpenBitSet)dis);
break;
case AND:
result.And((OpenBitSet)dis);
break;
case ANDNOT:
result.AndNot((OpenBitSet)dis);
break;
case XOR:
result.Xor((OpenBitSet)dis);
break;
default:
doChain(result, DEFAULT, dis);
break;
}
}
else
{
DocIdSetIterator disi;
if (dis == null)
{
disi = DocIdSet.EMPTY_DOCIDSET.Iterator();
}
else
{
disi = dis.Iterator();
if (disi == null)
{
disi = DocIdSet.EMPTY_DOCIDSET.Iterator();
}
}
switch (logic)
{
case OR:
result.InPlaceOr(disi);
break;
case AND:
result.InPlaceAnd(disi);
break;
case ANDNOT:
result.InPlaceNot(disi);
break;
case XOR:
result.InPlaceXor(disi);
break;
default:
doChain(result, DEFAULT, dis);
break;
}
}
}
/** Provide a SortedVIntList when it is definitely
* smaller than an OpenBitSet
* @deprecated Either use CachingWrapperFilter, or
* switch to a different DocIdSet implementation yourself.
* This method will be removed in Lucene 4.0
**/
protected DocIdSet finalResult(OpenBitSetDISI result, int maxDocs)
{
return result;
}
private OpenBitSetDISI initialResult(IndexReader reader, int logic, int[] index)
{
OpenBitSetDISI result;
/**
* First AND operation takes place against a completely false
* bitset and will always return zero results.
*/
if (logic == AND)
{
result = new OpenBitSetDISI(getDISI(chain[index[0]], reader), reader.MaxDoc);
++index[0];
}
else if (logic == ANDNOT)
{
result = new OpenBitSetDISI(getDISI(chain[index[0]], reader), reader.MaxDoc);
result.Flip(0, reader.MaxDoc); // NOTE: may set bits for deleted docs.
++index[0];
}
else
{
result = new OpenBitSetDISI(reader.MaxDoc);
}
return result;
}
public ISearchBits GetBits()
{
var query = CreateQuery();
IndexSearcher searcher;
try {
searcher = new IndexSearcher(_directory, true);
}
catch {
// index might not exist if it has been rebuilt
Logger.Information("Attempt to read a none existing index");
return null;
}
try {
var filter = new QueryWrapperFilter(query);
var bits = filter.GetDocIdSet(searcher.GetIndexReader());
var disi = new OpenBitSetDISI(bits.Iterator(), searcher.MaxDoc());
return new SearchBits(disi);
}
finally {
searcher.Close();
}
}
private IEnumerable<FacetMatch> FindMatchesInQuery(Query baseQueryWithoutFacetDrilldown, IList<FacetFieldInfo> allFacetFieldInfos, FacetFieldInfo facetFieldInfoToCalculateFor)
{
var calculations = 0;
var queryFilter = new CachingWrapperFilter(new QueryWrapperFilter(CreateFacetedQuery(baseQueryWithoutFacetDrilldown, allFacetFieldInfos, facetFieldInfoToCalculateFor.FieldName)));
var bitsQueryWithoutFacetDrilldown = new OpenBitSetDISI(queryFilter.GetDocIdSet(IndexReader).Iterator(), IndexReader.MaxDoc);
var baseQueryWithoutFacetDrilldownCopy = new OpenBitSetDISI(bitsQueryWithoutFacetDrilldown.Bits.Length)
{
Bits = new long[bitsQueryWithoutFacetDrilldown.Bits.Length]
};
var calculatedFacetCounts = new ResultCollection(facetFieldInfoToCalculateFor);
foreach (var facetValueBitSet in GetOrCreateFacetBitSet(facetFieldInfoToCalculateFor.FieldName).FacetValueBitSetList)
{
var isSelected = calculatedFacetCounts.IsSelected(facetValueBitSet.Value);
if (!isSelected && facetValueBitSet.Count < calculatedFacetCounts.MinCountForNonSelected) //Impossible to get a better result
{
if (calculatedFacetCounts.HaveEnoughResults)
break;
}
bitsQueryWithoutFacetDrilldown.Bits.CopyTo(baseQueryWithoutFacetDrilldownCopy.Bits, 0);
baseQueryWithoutFacetDrilldownCopy.NumWords = bitsQueryWithoutFacetDrilldown.NumWords;
var bitset = facetValueBitSet.Bitset ?? CalculateOpenBitSetDisi(facetFieldInfoToCalculateFor.FieldName, facetValueBitSet.Value);
baseQueryWithoutFacetDrilldownCopy.And(bitset);
var count = baseQueryWithoutFacetDrilldownCopy.Cardinality();
if (count == 0)
continue;
var match = new FacetMatch
{
Count = count,
Value = facetValueBitSet.Value,
FacetFieldName = facetFieldInfoToCalculateFor.FieldName
};
calculations++;
if (isSelected)
calculatedFacetCounts.AddToSelected(match);
else
calculatedFacetCounts.AddToNonSelected(match);
}
return calculatedFacetCounts.GetList();
}
/// <summary>
/// Get the id set for the filter.
/// </summary>
/// <param name="reader">The reader.</param>
/// <returns>The filter set to use.</returns>
public override DocIdSet GetDocIdSet(IndexReader reader)
{
OpenBitSetDISI res = null;
if (shouldFilters != null)
{
for (int i = 0; i < shouldFilters.Count; i++)
{
if (res == null)
{
res = new OpenBitSetDISI(GetDISI(shouldFilters, i, reader), reader.MaxDoc());
}
else
{
DocIdSet dis = ((Filter)shouldFilters[i]).GetDocIdSet(reader);
if (dis is OpenBitSet)
{
// optimized case for OpenBitSets
res.Or((OpenBitSet)dis);
}
else
{
res.InPlaceOr(GetDISI(shouldFilters, i, reader));
}
}
}
}
if (notFilters != null)
{
for (int i = 0; i < notFilters.Count; i++)
{
if (res == null)
{
res = new OpenBitSetDISI(GetDISI(notFilters, i, reader), reader.MaxDoc());
res.Flip(0, reader.MaxDoc()); // NOTE: may set bits on deleted docs
}
else
{
DocIdSet dis = ((Filter)notFilters[i]).GetDocIdSet(reader);
if (dis is OpenBitSet)
{
// optimized case for OpenBitSets
res.AndNot((OpenBitSet)dis);
}
else
{
res.InPlaceNot(GetDISI(notFilters, i, reader));
}
}
}
}
if (mustFilters != null)
{
for (int i = 0; i < mustFilters.Count; i++)
{
if (res == null)
{
res = new OpenBitSetDISI(GetDISI(mustFilters, i, reader), reader.MaxDoc());
}
else
{
DocIdSet dis = ((Filter)mustFilters[i]).GetDocIdSet(reader);
if (dis is OpenBitSet)
{
// optimized case for OpenBitSets
res.And((OpenBitSet)dis);
}
else
{
res.InPlaceAnd(GetDISI(mustFilters, i, reader));
}
}
}
}
if (res != null)
return FinalResult(res, reader.MaxDoc());
else
{
//TODO: 2.- change return DocIdSet.EMPTY_DOCIDSET;
return null;
}
}
// TODO: in 3.0, instead of removing this deprecated
// method, make it a no-op and mark it final
/** Provide a SortedVIntList when it is definitely smaller
* than an OpenBitSet.
* @deprecated Either use CachingWrapperFilter, or
* switch to a different DocIdSet implementation yourself. */
protected DocIdSet FinalResult(OpenBitSetDISI result, int maxDocs)
{
return (result.Cardinality() < (maxDocs / 9))
? (DocIdSet)new SortedVIntList(result)
: (DocIdSet)result;
}
public SimpleFacetedSearch(IndexReader reader, string[] groupByFields)
{
this._Reader = reader;
List<FieldValuesBitSets> fieldValuesBitSets = new List<FieldValuesBitSets>();
//STEP 1
//f1 = A, B
//f2 = I, J
//f3 = 1, 2, 3
int maxFacets = 1;
List<List<string>> inputToCP = new List<List<string>>();
foreach (string field in groupByFields)
{
FieldValuesBitSets f = new FieldValuesBitSets(reader, field);
maxFacets *= f.FieldValueBitSetPair.Count;
if (maxFacets > MAX_FACETS) throw new Exception("Facet count exceeded " + MAX_FACETS);
fieldValuesBitSets.Add(f);
inputToCP.Add(f.FieldValueBitSetPair.Keys.ToList());
}
//STEP 2
// comb1: A I 1
// comb2: A I 2 etc.
var cp = inputToCP.CartesianProduct();
//SETP 3
//create a single BitSet for each combination
//BitSet1: A AND I AND 1
//BitSet2: A AND I AND 2 etc.
//and remove impossible comb's (for ex, B J 3) from list.
Parallel.ForEach(cp, combinations =>
{
OpenBitSetDISI bitSet = new OpenBitSetDISI(_Reader.MaxDoc());
bitSet.Set(0, bitSet.Size());
List<string> comb = combinations.ToList();
for (int j = 0; j < comb.Count; j++)
{
bitSet.And(fieldValuesBitSets[j].FieldValueBitSetPair[comb[j]]);
}
//STEP 3
if (bitSet.Cardinality() > 0)
{
lock(_Groups)
_Groups.Add(new KeyValuePair<List<string>, OpenBitSetDISI>(comb, bitSet));
}
});
//Now _Groups has 7 rows (as <List<string>, BitSet> pairs)
}
public IEnumerable<IHit> Query(int pageIndex, int pageSize, out int totalCount, out IEnumerable<FacetGroup> facetedResults)
{
totalCount = 0;
facetedResults = null;
if (searchPaths == null || searchPaths.Count <= 0)
searchPaths.AddRange(indexPaths.Values.Select(o => o.Path));
List<LuceneHit> results = new List<LuceneHit>();
List<IndexSearcher> subSearchs = new List<IndexSearcher>();
searchPaths.ForEach(o => subSearchs.Add(new IndexSearcher(FSDirectory.Open(o))));
if (facetFields != null && facetFields.Count > 0)
{
var facetGroups = new List<FacetGroup>();
var mainQueryFilter = new CachingWrapperFilter(new QueryWrapperFilter(query));
MultiReader readers = new MultiReader(subSearchs.Select(o => o.IndexReader).ToArray());
foreach (var facetField in facetFields)
{
FacetGroup fg = new FacetGroup();
fg.FieldName = facetFieldNameProvider.GetMapName(TypeName, facetField);
var items = new List<FacetItem>();
var allDistinctField = FieldCache_Fields.DEFAULT.GetStrings(readers, facetField).Distinct().ToArray();
int totalHits = 0;
Parallel.ForEach(allDistinctField, fieldValue =>
{
//foreach (var fieldValue in allDistinctField)
//{
var facetQuery = new TermQuery(new Term(facetField, fieldValue));
var facetQueryFilter = new CachingWrapperFilter(new QueryWrapperFilter(facetQuery));
var bs = new OpenBitSetDISI(facetQueryFilter.GetDocIdSet(readers).Iterator(), readers.MaxDoc);
bs.InPlaceAnd(mainQueryFilter.GetDocIdSet(readers).Iterator());
int count = (Int32)bs.Cardinality();
FacetItem item = new FacetItem();
item.GroupValue = fieldValue;
item.Count = count;
items.Add(item);
totalHits += count;
}
);
fg.FacetItems = items.OrderByDescending(o => o.Count);
fg.TotalHits = totalHits;
facetGroups.Add(fg);
}
facetedResults = facetGroups.OrderBy(o => o.FieldName);
}
ParallelMultiSearcher searcher = new ParallelMultiSearcher(subSearchs.ToArray());
Sort sort = null;
if (sortFields != null && sortFields.Count > 0)
{
sort = new Sort(sortFields.ToArray());
}
int maxDoc = searcher.MaxDoc;
int startIndex = 0;
if (pageIndex >= 0 && pageSize > 0)
{
startIndex = pageIndex * pageSize;
maxDoc = pageSize * (pageIndex + 1);
}
var docs = sort == null ? searcher.Search(query, null, maxDoc) : searcher.Search(query, null, maxDoc, sort);
totalCount = docs.TotalHits;
int endIndex = docs.TotalHits - startIndex;
for (int i = startIndex; i < endIndex; i++)
{
LuceneHit h = new LuceneHit(TypeName, DocumentBuilder, searcher.Doc(docs.ScoreDocs[i].Doc));
results.Add(h);
}
return results;
}
private static long CalculateFacetCount(DocIdSet baseBitSet, DocIdSet filterDocSet)
{
var baseDisi = new OpenBitSetDISI(baseBitSet.Iterator(), 25000);
var filterIterator = filterDocSet.Iterator();
baseDisi.InPlaceAnd(filterIterator);
var total = baseDisi.Cardinality();
return total;
}
public OpenBitSetDISI GetBitSetFromFilter(Filter filter, IndexReader indexReader)
{
var bitSet = new OpenBitSetDISI(filter.GetDocIdSet(indexReader).Iterator(), indexReader.MaxDoc());
return bitSet;
}
public void AddValue(string value, OpenBitSetDISI matchingDocuments, Filter facetFilter)
{
this.Values.Add(new Tuple<string, OpenBitSetDISI, Filter>(value, matchingDocuments, facetFilter));
}