internal RandomBits(int maxDoc, double pctLive, Random random)
{
bits = new FixedBitSet(maxDoc);
for (int i = 0; i < maxDoc; i++)
{
if (random.NextDouble() <= pctLive)
{
bits.Set(i);
}
}
}
protected virtual void CollectDocs(FixedBitSet bitSet)
{
//WARN: keep this specialization in sync
Debug.Assert(termsEnum != null);
docsEnum = termsEnum.Docs(acceptDocs, docsEnum, DocsEnum.FLAG_NONE);
int docid;
while ((docid = docsEnum.NextDoc()) != DocIdSetIterator.NO_MORE_DOCS)
{
bitSet.Set(docid);
}
}
public override DocIdSet GetDocIdSet(AtomicReaderContext context, Bits acceptDocs)
{
AtomicReader reader = context.AtomicReader;
FixedBitSet result = null; // lazy init if needed - no need to create a big bitset ahead of time
Fields fields = reader.Fields;
BytesRef spare = new BytesRef(this.termsBytes);
if (fields == null)
{
return result;
}
Terms terms = null;
TermsEnum termsEnum = null;
DocsEnum docs = null;
foreach (TermsAndField termsAndField in this.termsAndFields)
{
if ((terms = fields.Terms(termsAndField.field)) != null)
{
termsEnum = terms.Iterator(termsEnum); // this won't return null
for (int i = termsAndField.start; i < termsAndField.end; i++)
{
spare.Offset = offsets[i];
spare.Length = offsets[i + 1] - offsets[i];
if (termsEnum.SeekExact(spare))
{
docs = termsEnum.Docs(acceptDocs, docs, DocsEnum.FLAG_NONE); // no freq since we don't need them
if (result == null)
{
if (docs.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
result = new FixedBitSet(reader.MaxDoc);
// lazy init but don't do it in the hot loop since we could read many docs
result.Set(docs.DocID());
}
}
while (docs.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
result.Set(docs.DocID());
}
}
}
}
}
return result;
}
private FixedBitSet CorrectBits(AtomicReader reader, Bits acceptDocs)
{
FixedBitSet bits = new FixedBitSet(reader.MaxDoc); //assume all are INvalid
Terms terms = reader.Fields.Terms(fieldName);
if (terms == null)
{
return bits;
}
TermsEnum termsEnum = terms.Iterator(null);
DocsEnum docs = null;
while (true)
{
BytesRef currTerm = termsEnum.Next();
if (currTerm == null)
{
break;
}
else
{
docs = termsEnum.Docs(acceptDocs, docs, DocsEnum.FLAG_NONE);
int doc = docs.NextDoc();
if (doc != DocIdSetIterator.NO_MORE_DOCS)
{
if (keepMode == KeepMode.KM_USE_FIRST_OCCURRENCE)
{
bits.Set(doc);
}
else
{
int lastDoc = doc;
while (true)
{
lastDoc = doc;
doc = docs.NextDoc();
if (doc == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
}
bits.Set(lastDoc);
}
}
}
}
return bits;
}
private FixedBitSet FastBits(AtomicReader reader, Bits acceptDocs)
{
FixedBitSet bits = new FixedBitSet(reader.MaxDoc);
bits.Set(0, reader.MaxDoc); //assume all are valid
Terms terms = reader.Fields.Terms(fieldName);
if (terms == null)
{
return bits;
}
TermsEnum termsEnum = terms.Iterator(null);
DocsEnum docs = null;
while (true)
{
BytesRef currTerm = termsEnum.Next();
if (currTerm == null)
{
break;
}
else
{
if (termsEnum.DocFreq() > 1)
{
// unset potential duplicates
docs = termsEnum.Docs(acceptDocs, docs, DocsEnum.FLAG_NONE);
int doc = docs.NextDoc();
if (doc != DocIdSetIterator.NO_MORE_DOCS)
{
if (keepMode == KeepMode.KM_USE_FIRST_OCCURRENCE)
{
doc = docs.NextDoc();
}
}
int lastDoc = -1;
while (true)
{
lastDoc = doc;
bits.Clear(lastDoc);
doc = docs.NextDoc();
if (doc == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
}
if (keepMode == KeepMode.KM_USE_LAST_OCCURRENCE)
{
// restore the last bit
bits.Set(lastDoc);
}
}
}
}
return bits;
}
protected virtual void FillDocsAndScores(FixedBitSet matchingDocs, Bits acceptDocs,
TermsEnum termsEnum)
{
BytesRef spare = new BytesRef();
DocsEnum docsEnum = null;
for (int i = 0; i < outerInstance._terms.Size(); i++)
{
if (termsEnum.SeekExact(outerInstance._terms.Get(outerInstance._ords[i], spare)))
{
docsEnum = termsEnum.Docs(acceptDocs, docsEnum, FLAG_NONE);
float score = outerInstance._scores[outerInstance._ords[i]];
for (int doc = docsEnum.NextDoc();
doc != NO_MORE_DOCS;
doc = docsEnum.NextDoc())
{
matchingDocs.Set(doc);
// In the case the same doc is also related to a another doc, a score might be overwritten. I think this
// can only happen in a many-to-many relation
scores[doc] = score;
}
}
}
}
protected override void FillDocsAndScores(FixedBitSet matchingDocs, Bits acceptDocs,
TermsEnum termsEnum)
{
BytesRef spare = new BytesRef();
DocsEnum docsEnum = null;
for (int i = 0; i < outerInstance._terms.Size(); i++)
{
if (termsEnum.SeekExact(outerInstance._terms.Get(outerInstance._ords[i], spare)))
{
docsEnum = termsEnum.Docs(acceptDocs, docsEnum, FLAG_NONE);
float score = outerInstance._scores[outerInstance._ords[i]];
for (int doc = docsEnum.NextDoc();
doc != NO_MORE_DOCS;
doc = docsEnum.NextDoc())
{
// I prefer this:
/*if (scores[doc] < score) {
scores[doc] = score;
matchingDocs.set(doc);
}*/
// But this behaves the same as MVInnerScorer and only then the tests will pass:
if (!matchingDocs.Get(doc))
{
scores[doc] = score;
matchingDocs.Set(doc);
}
}
}
}
}
internal Bits RandomLiveDocs(int maxDoc)
{
if (Rarely())
{
if (Random().nextBoolean())
{
return null;
}
else
{
return new Bits_MatchNoBits(maxDoc);
}
}
FixedBitSet bits = new FixedBitSet(maxDoc);
int bitsSet = TestUtil.NextInt(Random(), 1, maxDoc - 1);
for (int i = 0; i < bitsSet; ++i)
{
while (true)
{
int index = Random().nextInt(maxDoc);
if (!bits.Get(index))
{
bits.Set(index);
break;
}
}
}
return bits;
}
