internal SegmentResult(int[] counts, int total, TermsEnum tenum, int startFacetOrd, int endFacetOrd)
: base(counts, total - counts[0], counts[0], endFacetOrd + 1)
{
this.tenum = tenum;
this.mergePos = startFacetOrd == -1 ? 1 : startFacetOrd + 1;
if (mergePos < maxTermPos)
{
Debug.Assert(tenum != null);
tenum.SeekExact(startFacetOrd == -1 ? 0 : startFacetOrd);
mergeTerm = tenum.Term();
}
}
public virtual void TestRandom()
{
// token -> docID -> tokens
IDictionary <string, IDictionary <int?, IList <Token> > > actualTokens = new Dictionary <string, IDictionary <int?, IList <Token> > >();
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random, dir, iwc);
int numDocs = AtLeast(20);
//final int numDocs = AtLeast(5);
FieldType ft = new FieldType(TextField.TYPE_NOT_STORED);
// TODO: randomize what IndexOptions we use; also test
// changing this up in one IW buffered segment...:
ft.IndexOptions = IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS;
if (Random.NextBoolean())
{
ft.StoreTermVectors = true;
ft.StoreTermVectorOffsets = Random.NextBoolean();
ft.StoreTermVectorPositions = Random.NextBoolean();
}
for (int docCount = 0; docCount < numDocs; docCount++)
{
Document doc = new Document();
doc.Add(new Int32Field("id", docCount, Field.Store.NO));
IList <Token> tokens = new List <Token>();
int numTokens = AtLeast(100);
//final int numTokens = AtLeast(20);
int pos = -1;
int offset = 0;
//System.out.println("doc id=" + docCount);
for (int tokenCount = 0; tokenCount < numTokens; tokenCount++)
{
string text;
if (Random.NextBoolean())
{
text = "a";
}
else if (Random.NextBoolean())
{
text = "b";
}
else if (Random.NextBoolean())
{
text = "c";
}
else
{
text = "d";
}
int posIncr = Random.NextBoolean() ? 1 : Random.Next(5);
if (tokenCount == 0 && posIncr == 0)
{
posIncr = 1;
}
int offIncr = Random.NextBoolean() ? 0 : Random.Next(5);
int tokenOffset = Random.Next(5);
Token token = MakeToken(text, posIncr, offset + offIncr, offset + offIncr + tokenOffset);
if (!actualTokens.TryGetValue(text, out IDictionary <int?, IList <Token> > postingsByDoc))
{
actualTokens[text] = postingsByDoc = new Dictionary <int?, IList <Token> >();
}
if (!postingsByDoc.TryGetValue(docCount, out IList <Token> postings))
{
postingsByDoc[docCount] = postings = new List <Token>();
}
postings.Add(token);
tokens.Add(token);
pos += posIncr;
// stuff abs position into type:
token.Type = "" + pos;
offset += offIncr + tokenOffset;
//System.out.println(" " + token + " posIncr=" + token.getPositionIncrement() + " pos=" + pos + " off=" + token.StartOffset + "/" + token.EndOffset + " (freq=" + postingsByDoc.Get(docCount).Size() + ")");
}
doc.Add(new Field("content", new CannedTokenStream(tokens.ToArray()), ft));
w.AddDocument(doc);
}
DirectoryReader r = w.GetReader();
w.Dispose();
string[] terms = new string[] { "a", "b", "c", "d" };
foreach (AtomicReaderContext ctx in r.Leaves)
{
// TODO: improve this
AtomicReader sub = (AtomicReader)ctx.Reader;
//System.out.println("\nsub=" + sub);
TermsEnum termsEnum = sub.Fields.GetTerms("content").GetEnumerator();
DocsEnum docs = null;
DocsAndPositionsEnum docsAndPositions = null;
DocsAndPositionsEnum docsAndPositionsAndOffsets = null;
FieldCache.Int32s docIDToID = FieldCache.DEFAULT.GetInt32s(sub, "id", false);
foreach (string term in terms)
{
//System.out.println(" term=" + term);
if (termsEnum.SeekExact(new BytesRef(term)))
{
docs = termsEnum.Docs(null, docs);
Assert.IsNotNull(docs);
int doc;
//System.out.println(" doc/freq");
while ((doc = docs.NextDoc()) != DocIdSetIterator.NO_MORE_DOCS)
{
IList <Token> expected = actualTokens[term][docIDToID.Get(doc)];
//System.out.println(" doc=" + docIDToID.Get(doc) + " docID=" + doc + " " + expected.Size() + " freq");
Assert.IsNotNull(expected);
Assert.AreEqual(expected.Count, docs.Freq);
}
// explicitly exclude offsets here
docsAndPositions = termsEnum.DocsAndPositions(null, docsAndPositions, DocsAndPositionsFlags.PAYLOADS);
Assert.IsNotNull(docsAndPositions);
//System.out.println(" doc/freq/pos");
while ((doc = docsAndPositions.NextDoc()) != DocIdSetIterator.NO_MORE_DOCS)
{
IList <Token> expected = actualTokens[term][docIDToID.Get(doc)];
//System.out.println(" doc=" + docIDToID.Get(doc) + " " + expected.Size() + " freq");
Assert.IsNotNull(expected);
Assert.AreEqual(expected.Count, docsAndPositions.Freq);
foreach (Token token in expected)
{
int pos = Convert.ToInt32(token.Type);
//System.out.println(" pos=" + pos);
Assert.AreEqual(pos, docsAndPositions.NextPosition());
}
}
docsAndPositionsAndOffsets = termsEnum.DocsAndPositions(null, docsAndPositions);
Assert.IsNotNull(docsAndPositionsAndOffsets);
//System.out.println(" doc/freq/pos/offs");
while ((doc = docsAndPositionsAndOffsets.NextDoc()) != DocIdSetIterator.NO_MORE_DOCS)
{
IList <Token> expected = actualTokens[term][docIDToID.Get(doc)];
//System.out.println(" doc=" + docIDToID.Get(doc) + " " + expected.Size() + " freq");
Assert.IsNotNull(expected);
Assert.AreEqual(expected.Count, docsAndPositionsAndOffsets.Freq);
foreach (Token token in expected)
{
int pos = Convert.ToInt32(token.Type);
//System.out.println(" pos=" + pos);
Assert.AreEqual(pos, docsAndPositionsAndOffsets.NextPosition());
Assert.AreEqual(token.StartOffset, docsAndPositionsAndOffsets.StartOffset);
Assert.AreEqual(token.EndOffset, docsAndPositionsAndOffsets.EndOffset);
}
}
}
}
// TODO: test advance:
}
r.Dispose();
dir.Dispose();
}
private void Verify(AtomicReader r, int[][] idToOrds, BytesRef[] termsArray, BytesRef prefixRef)
{
DocTermOrds dto = new DocTermOrds(r, r.LiveDocs, "field", prefixRef, int.MaxValue, TestUtil.NextInt32(Random, 2, 10));
FieldCache.Int32s docIDToID = FieldCache.DEFAULT.GetInt32s(r, "id", false);
/*
* for(int docID=0;docID<subR.MaxDoc;docID++) {
* System.out.println(" docID=" + docID + " id=" + docIDToID[docID]);
* }
*/
if (Verbose)
{
Console.WriteLine("TEST: verify prefix=" + (prefixRef == null ? "null" : prefixRef.Utf8ToString()));
Console.WriteLine("TEST: all TERMS:");
TermsEnum allTE = MultiFields.GetTerms(r, "field").GetEnumerator();
int ord = 0;
while (allTE.MoveNext())
{
Console.WriteLine(" ord=" + (ord++) + " term=" + allTE.Term.Utf8ToString());
}
}
//final TermsEnum te = subR.Fields.Terms("field").iterator();
TermsEnum te = dto.GetOrdTermsEnum(r);
if (dto.NumTerms == 0)
{
if (prefixRef == null)
{
Assert.IsNull(MultiFields.GetTerms(r, "field"));
}
else
{
Terms terms = MultiFields.GetTerms(r, "field");
if (terms != null)
{
TermsEnum termsEnum = terms.GetEnumerator();
TermsEnum.SeekStatus result = termsEnum.SeekCeil(prefixRef);
if (result != TermsEnum.SeekStatus.END)
{
Assert.IsFalse(StringHelper.StartsWith(termsEnum.Term, prefixRef), "term=" + termsEnum.Term.Utf8ToString() + " matches prefix=" + prefixRef.Utf8ToString());
}
else
{
// ok
}
}
else
{
// ok
}
}
return;
}
if (Verbose)
{
Console.WriteLine("TEST: TERMS:");
te.SeekExact(0);
while (true)
{
Console.WriteLine(" ord=" + te.Ord + " term=" + te.Term.Utf8ToString());
if (!te.MoveNext())
{
break;
}
}
}
SortedSetDocValues iter = dto.GetIterator(r);
for (int docID = 0; docID < r.MaxDoc; docID++)
{
if (Verbose)
{
Console.WriteLine("TEST: docID=" + docID + " of " + r.MaxDoc + " (id=" + docIDToID.Get(docID) + ")");
}
iter.SetDocument(docID);
int[] answers = idToOrds[docIDToID.Get(docID)];
int upto = 0;
long ord;
while ((ord = iter.NextOrd()) != SortedSetDocValues.NO_MORE_ORDS)
{
te.SeekExact(ord);
BytesRef expected = termsArray[answers[upto++]];
if (Verbose)
{
Console.WriteLine(" exp=" + expected.Utf8ToString() + " actual=" + te.Term.Utf8ToString());
}
Assert.AreEqual(expected, te.Term, "expected=" + expected.Utf8ToString() + " actual=" + te.Term.Utf8ToString() + " ord=" + ord);
}
Assert.AreEqual(answers.Length, upto);
}
}
public virtual void TestSortedTermsEnum()
{
Directory directory = NewDirectory();
Analyzer analyzer = new MockAnalyzer(Random());
IndexWriterConfig iwconfig = NewIndexWriterConfig(TEST_VERSION_CURRENT, analyzer);
iwconfig.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter iwriter = new RandomIndexWriter(Random(), directory, iwconfig);
Document doc = new Document();
doc.Add(new StringField("field", "hello", Field.Store.NO));
iwriter.AddDocument(doc);
doc = new Document();
doc.Add(new StringField("field", "world", Field.Store.NO));
iwriter.AddDocument(doc);
doc = new Document();
doc.Add(new StringField("field", "beer", Field.Store.NO));
iwriter.AddDocument(doc);
iwriter.ForceMerge(1);
DirectoryReader ireader = iwriter.Reader;
iwriter.Dispose();
AtomicReader ar = GetOnlySegmentReader(ireader);
SortedSetDocValues dv = FieldCache.DEFAULT.GetDocTermOrds(ar, "field");
Assert.AreEqual(3, dv.ValueCount);
TermsEnum termsEnum = dv.GetTermsEnum();
// next()
Assert.AreEqual("beer", termsEnum.Next().Utf8ToString());
Assert.AreEqual(0, termsEnum.Ord);
Assert.AreEqual("hello", termsEnum.Next().Utf8ToString());
Assert.AreEqual(1, termsEnum.Ord);
Assert.AreEqual("world", termsEnum.Next().Utf8ToString());
Assert.AreEqual(2, termsEnum.Ord);
// seekCeil()
Assert.AreEqual(SeekStatus.NOT_FOUND, termsEnum.SeekCeil(new BytesRef("ha!")));
Assert.AreEqual("hello", termsEnum.Term.Utf8ToString());
Assert.AreEqual(1, termsEnum.Ord);
Assert.AreEqual(SeekStatus.FOUND, termsEnum.SeekCeil(new BytesRef("beer")));
Assert.AreEqual("beer", termsEnum.Term.Utf8ToString());
Assert.AreEqual(0, termsEnum.Ord);
Assert.AreEqual(SeekStatus.END, termsEnum.SeekCeil(new BytesRef("zzz")));
// seekExact()
Assert.IsTrue(termsEnum.SeekExact(new BytesRef("beer")));
Assert.AreEqual("beer", termsEnum.Term.Utf8ToString());
Assert.AreEqual(0, termsEnum.Ord);
Assert.IsTrue(termsEnum.SeekExact(new BytesRef("hello")));
Assert.AreEqual("hello", termsEnum.Term.Utf8ToString());
Assert.AreEqual(1, termsEnum.Ord);
Assert.IsTrue(termsEnum.SeekExact(new BytesRef("world")));
Assert.AreEqual("world", termsEnum.Term.Utf8ToString());
Assert.AreEqual(2, termsEnum.Ord);
Assert.IsFalse(termsEnum.SeekExact(new BytesRef("bogus")));
// seek(ord)
termsEnum.SeekExact(0);
Assert.AreEqual("beer", termsEnum.Term.Utf8ToString());
Assert.AreEqual(0, termsEnum.Ord);
termsEnum.SeekExact(1);
Assert.AreEqual("hello", termsEnum.Term.Utf8ToString());
Assert.AreEqual(1, termsEnum.Ord);
termsEnum.SeekExact(2);
Assert.AreEqual("world", termsEnum.Term.Utf8ToString());
Assert.AreEqual(2, termsEnum.Ord);
ireader.Dispose();
directory.Dispose();
}
/// <summary>
/// Returns the term (<see cref="BytesRef"/>) corresponding to
/// the provided ordinal.
/// </summary>
public virtual BytesRef LookupTerm(TermsEnum termsEnum, int ord)
{
termsEnum.SeekExact(ord);
return(termsEnum.Term);
}
private void AssertEquals(long numOrds, TermsEnum expected, TermsEnum actual)
{
BytesRef @ref;
// sequential next() through all terms
while ((@ref = expected.Next()) != null)
{
Assert.AreEqual(@ref, actual.Next());
Assert.AreEqual(expected.Ord(), actual.Ord());
Assert.AreEqual(expected.Term(), actual.Term());
}
Assert.IsNull(actual.Next());
// sequential seekExact(ord) through all terms
for (long i = 0; i < numOrds; i++)
{
expected.SeekExact(i);
actual.SeekExact(i);
Assert.AreEqual(expected.Ord(), actual.Ord());
Assert.AreEqual(expected.Term(), actual.Term());
}
// sequential seekExact(BytesRef) through all terms
for (long i = 0; i < numOrds; i++)
{
expected.SeekExact(i);
Assert.IsTrue(actual.SeekExact(expected.Term()));
Assert.AreEqual(expected.Ord(), actual.Ord());
Assert.AreEqual(expected.Term(), actual.Term());
}
// sequential seekCeil(BytesRef) through all terms
for (long i = 0; i < numOrds; i++)
{
expected.SeekExact(i);
Assert.AreEqual(SeekStatus.FOUND, actual.SeekCeil(expected.Term()));
Assert.AreEqual(expected.Ord(), actual.Ord());
Assert.AreEqual(expected.Term(), actual.Term());
}
// random seekExact(ord)
for (long i = 0; i < numOrds; i++)
{
long randomOrd = TestUtil.NextLong(Random(), 0, numOrds - 1);
expected.SeekExact(randomOrd);
actual.SeekExact(randomOrd);
Assert.AreEqual(expected.Ord(), actual.Ord());
Assert.AreEqual(expected.Term(), actual.Term());
}
// random seekExact(BytesRef)
for (long i = 0; i < numOrds; i++)
{
long randomOrd = TestUtil.NextLong(Random(), 0, numOrds - 1);
expected.SeekExact(randomOrd);
actual.SeekExact(expected.Term());
Assert.AreEqual(expected.Ord(), actual.Ord());
Assert.AreEqual(expected.Term(), actual.Term());
}
// random seekCeil(BytesRef)
for (long i = 0; i < numOrds; i++)
{
BytesRef target = new BytesRef(TestUtil.RandomUnicodeString(Random()));
SeekStatus expectedStatus = expected.SeekCeil(target);
Assert.AreEqual(expectedStatus, actual.SeekCeil(target));
if (expectedStatus != SeekStatus.END)
{
Assert.AreEqual(expected.Ord(), actual.Ord());
Assert.AreEqual(expected.Term(), actual.Term());
}
}
}
// sugar
private bool SeekExact(TermsEnum te, string term)
{
return te.SeekExact(new BytesRef(term));
}
public override void SeekExact(long ord)
{
m_input.SeekExact(ord);
}
public override void SeekExact(long ord)
{
@in.SeekExact(ord);
}
// sugar
private bool SeekExact(TermsEnum te, string term)
{
return(te.SeekExact(new BytesRef(term)));
}
private void TestRandomSeeks(IndexReader r, params string[] validTermStrings)
{
BytesRef[] validTerms = new BytesRef[validTermStrings.Length];
for (int termIDX = 0; termIDX < validTermStrings.Length; termIDX++)
{
validTerms[termIDX] = new BytesRef(validTermStrings[termIDX]);
}
Array.Sort(validTerms);
if (Verbose)
{
Console.WriteLine("TEST: " + validTerms.Length + " terms:");
foreach (BytesRef t in validTerms)
{
Console.WriteLine(" " + t.Utf8ToString() + " " + t);
}
}
TermsEnum te = MultiFields.GetTerms(r, FIELD).GetEnumerator();
int END_LOC = -validTerms.Length - 1;
IList <TermAndState> termStates = new List <TermAndState>();
for (int iter = 0; iter < 100 * RandomMultiplier; iter++)
{
BytesRef t;
int loc;
TermState termState;
if (Random.Next(6) == 4)
{
// pick term that doens't exist:
t = GetNonExistTerm(validTerms);
termState = null;
if (Verbose)
{
Console.WriteLine("\nTEST: invalid term=" + t.Utf8ToString());
}
loc = Array.BinarySearch(validTerms, t);
}
else if (termStates.Count != 0 && Random.Next(4) == 1)
{
TermAndState ts = termStates[Random.Next(termStates.Count)];
t = ts.Term;
loc = Array.BinarySearch(validTerms, t);
Assert.IsTrue(loc >= 0);
termState = ts.State;
if (Verbose)
{
Console.WriteLine("\nTEST: valid termState term=" + t.Utf8ToString());
}
}
else
{
// pick valid term
loc = Random.Next(validTerms.Length);
t = BytesRef.DeepCopyOf(validTerms[loc]);
termState = null;
if (Verbose)
{
Console.WriteLine("\nTEST: valid term=" + t.Utf8ToString());
}
}
// seekCeil or seekExact:
bool doSeekExact = Random.NextBoolean();
if (termState != null)
{
if (Verbose)
{
Console.WriteLine(" seekExact termState");
}
te.SeekExact(t, termState);
}
else if (doSeekExact)
{
if (Verbose)
{
Console.WriteLine(" seekExact");
}
Assert.AreEqual(loc >= 0, te.SeekExact(t));
}
else
{
if (Verbose)
{
Console.WriteLine(" seekCeil");
}
TermsEnum.SeekStatus result = te.SeekCeil(t);
if (Verbose)
{
Console.WriteLine(" got " + result);
}
if (loc >= 0)
{
Assert.AreEqual(TermsEnum.SeekStatus.FOUND, result);
}
else if (loc == END_LOC)
{
Assert.AreEqual(TermsEnum.SeekStatus.END, result);
}
else
{
if (Debugging.AssertsEnabled)
{
Debugging.Assert(loc >= -validTerms.Length);
}
Assert.AreEqual(TermsEnum.SeekStatus.NOT_FOUND, result);
}
}
if (loc >= 0)
{
Assert.AreEqual(t, te.Term);
}
else if (doSeekExact)
{
// TermsEnum is unpositioned if seekExact returns false
continue;
}
else if (loc == END_LOC)
{
continue;
}
else
{
loc = -loc - 1;
Assert.AreEqual(validTerms[loc], te.Term);
}
// Do a bunch of next's after the seek
int numNext = Random.Next(validTerms.Length);
for (int nextCount = 0; nextCount < numNext; nextCount++)
{
if (Verbose)
{
Console.WriteLine("\nTEST: next loc=" + loc + " of " + validTerms.Length);
}
bool moved = te.MoveNext();
//BytesRef t2 = te.Term;
loc++;
if (loc == validTerms.Length)
{
//Assert.IsNull(t2); // LUCENENET specific - accessing the Term after MoveNext() returns false results in an assertion failure
Assert.IsFalse(moved);
break;
}
else
{
Assert.AreEqual(validTerms[loc], te.Term);
if (Random.Next(40) == 17 && termStates.Count < 100)
{
termStates.Add(new TermAndState(validTerms[loc], te.GetTermState()));
}
}
}
}
}
public virtual void Test()
{
Random random = new Random(Random.Next());
LineFileDocs docs = new LineFileDocs(random, DefaultCodecSupportsDocValues);
Directory d = NewDirectory();
MockAnalyzer analyzer = new MockAnalyzer(LuceneTestCase.Random);
analyzer.MaxTokenLength = TestUtil.NextInt32(LuceneTestCase.Random, 1, IndexWriter.MAX_TERM_LENGTH);
RandomIndexWriter w = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
LuceneTestCase.Random, d, analyzer);
int numDocs = AtLeast(10);
for (int docCount = 0; docCount < numDocs; docCount++)
{
w.AddDocument(docs.NextDoc());
}
IndexReader r = w.GetReader();
w.Dispose();
List <BytesRef> terms = new List <BytesRef>();
TermsEnum termsEnum = MultiFields.GetTerms(r, "body").GetEnumerator();
while (termsEnum.MoveNext())
{
terms.Add(BytesRef.DeepCopyOf(termsEnum.Term));
}
if (Verbose)
{
Console.WriteLine("TEST: " + terms.Count + " terms");
}
int upto = -1;
int iters = AtLeast(200);
for (int iter = 0; iter < iters; iter++)
{
bool isEnd;
if (upto != -1 && LuceneTestCase.Random.NextBoolean())
{
// next
if (Verbose)
{
Console.WriteLine("TEST: iter next");
}
isEnd = termsEnum.MoveNext() == false;
upto++;
if (isEnd)
{
if (Verbose)
{
Console.WriteLine(" end");
}
Assert.AreEqual(upto, terms.Count);
upto = -1;
}
else
{
if (Verbose)
{
Console.WriteLine(" got term=" + termsEnum.Term.Utf8ToString() + " expected=" + terms[upto].Utf8ToString());
}
Assert.IsTrue(upto < terms.Count);
Assert.AreEqual(terms[upto], termsEnum.Term);
}
}
else
{
BytesRef target;
string exists;
if (LuceneTestCase.Random.NextBoolean())
{
// likely fake term
if (LuceneTestCase.Random.NextBoolean())
{
target = new BytesRef(TestUtil.RandomSimpleString(LuceneTestCase.Random));
}
else
{
target = new BytesRef(TestUtil.RandomRealisticUnicodeString(LuceneTestCase.Random));
}
exists = "likely not";
}
else
{
// real term
target = terms[LuceneTestCase.Random.Next(terms.Count)];
exists = "yes";
}
upto = terms.BinarySearch(target);
if (LuceneTestCase.Random.NextBoolean())
{
if (Verbose)
{
Console.WriteLine("TEST: iter seekCeil target=" + target.Utf8ToString() + " exists=" + exists);
}
// seekCeil
TermsEnum.SeekStatus status = termsEnum.SeekCeil(target);
if (Verbose)
{
Console.WriteLine(" got " + status);
}
if (upto < 0)
{
upto = -(upto + 1);
if (upto >= terms.Count)
{
Assert.AreEqual(TermsEnum.SeekStatus.END, status);
upto = -1;
}
else
{
Assert.AreEqual(TermsEnum.SeekStatus.NOT_FOUND, status);
Assert.AreEqual(terms[upto], termsEnum.Term);
}
}
else
{
Assert.AreEqual(TermsEnum.SeekStatus.FOUND, status);
Assert.AreEqual(terms[upto], termsEnum.Term);
}
}
else
{
if (Verbose)
{
Console.WriteLine("TEST: iter seekExact target=" + target.Utf8ToString() + " exists=" + exists);
}
// seekExact
bool result = termsEnum.SeekExact(target);
if (Verbose)
{
Console.WriteLine(" got " + result);
}
if (upto < 0)
{
Assert.IsFalse(result);
upto = -1;
}
else
{
Assert.IsTrue(result);
Assert.AreEqual(target, termsEnum.Term);
}
}
}
}
r.Dispose();
d.Dispose();
docs.Dispose();
}
// DocValues updates
private void ApplyDocValuesUpdates <T1>(IEnumerable <T1> updates, ReadersAndUpdates rld, SegmentReader reader, DocValuesFieldUpdates.Container dvUpdatesContainer) where T1 : DocValuesUpdate
{
lock (this)
{
Fields fields = reader.Fields;
if (fields == null)
{
// this reader has no postings
return;
}
// TODO: we can process the updates per DV field, from last to first so that
// if multiple terms affect same document for the same field, we add an update
// only once (that of the last term). To do that, we can keep a bitset which
// marks which documents have already been updated. So e.g. if term T1
// updates doc 7, and then we process term T2 and it updates doc 7 as well,
// we don't apply the update since we know T1 came last and therefore wins
// the update.
// We can also use that bitset as 'liveDocs' to pass to TermEnum.docs(), so
// that these documents aren't even returned.
string currentField = null;
TermsEnum termsEnum = null;
DocsEnum docs = null;
//System.out.println(Thread.currentThread().getName() + " numericDVUpdate reader=" + reader);
foreach (DocValuesUpdate update in updates)
{
Term term = update.term;
int limit = update.docIDUpto;
// TODO: we traverse the terms in update order (not term order) so that we
// apply the updates in the correct order, i.e. if two terms udpate the
// same document, the last one that came in wins, irrespective of the
// terms lexical order.
// we can apply the updates in terms order if we keep an updatesGen (and
// increment it with every update) and attach it to each NumericUpdate. Note
// that we cannot rely only on docIDUpto because an app may send two updates
// which will get same docIDUpto, yet will still need to respect the order
// those updates arrived.
if (!string.Equals(term.Field, currentField, StringComparison.Ordinal))
{
// if we change the code to process updates in terms order, enable this assert
// assert currentField == null || currentField.CompareToOrdinal(term.Field) < 0;
currentField = term.Field;
Terms terms = fields.GetTerms(currentField);
if (terms != null)
{
termsEnum = terms.GetIterator(termsEnum);
}
else
{
termsEnum = null;
continue; // no terms in that field
}
}
if (termsEnum == null)
{
continue;
}
// System.out.println(" term=" + term);
if (termsEnum.SeekExact(term.Bytes))
{
// we don't need term frequencies for this
DocsEnum docsEnum = termsEnum.Docs(rld.LiveDocs, docs, DocsFlags.NONE);
//System.out.println("BDS: got docsEnum=" + docsEnum);
DocValuesFieldUpdates dvUpdates = dvUpdatesContainer.GetUpdates(update.field, update.type);
if (dvUpdates == null)
{
dvUpdates = dvUpdatesContainer.NewUpdates(update.field, update.type, reader.MaxDoc);
}
int doc;
while ((doc = docsEnum.NextDoc()) != DocIdSetIterator.NO_MORE_DOCS)
{
//System.out.println(Thread.currentThread().getName() + " numericDVUpdate term=" + term + " doc=" + docID);
if (doc >= limit)
{
break; // no more docs that can be updated for this term
}
dvUpdates.Add(doc, update.value);
}
}
}
}
}
// Delete by Term
private long ApplyTermDeletes(IEnumerable <Term> termsIter, ReadersAndUpdates rld, SegmentReader reader)
{
lock (this)
{
long delCount = 0;
Fields fields = reader.Fields;
if (fields == null)
{
// this reader has no postings
return(0);
}
TermsEnum termsEnum = null;
string currentField = null;
DocsEnum docs = null;
Debug.Assert(CheckDeleteTerm(null));
bool any = false;
//System.out.println(Thread.currentThread().getName() + " del terms reader=" + reader);
foreach (Term term in termsIter)
{
// Since we visit terms sorted, we gain performance
// by re-using the same TermsEnum and seeking only
// forwards
if (!string.Equals(term.Field, currentField, StringComparison.Ordinal))
{
Debug.Assert(currentField == null || currentField.CompareToOrdinal(term.Field) < 0);
currentField = term.Field;
Terms terms = fields.GetTerms(currentField);
if (terms != null)
{
termsEnum = terms.GetIterator(termsEnum);
}
else
{
termsEnum = null;
}
}
if (termsEnum == null)
{
continue;
}
Debug.Assert(CheckDeleteTerm(term));
// System.out.println(" term=" + term);
if (termsEnum.SeekExact(term.Bytes))
{
// we don't need term frequencies for this
DocsEnum docsEnum = termsEnum.Docs(rld.LiveDocs, docs, DocsFlags.NONE);
//System.out.println("BDS: got docsEnum=" + docsEnum);
if (docsEnum != null)
{
while (true)
{
int docID = docsEnum.NextDoc();
//System.out.println(Thread.currentThread().getName() + " del term=" + term + " doc=" + docID);
if (docID == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
if (!any)
{
rld.InitWritableLiveDocs();
any = true;
}
// NOTE: there is no limit check on the docID
// when deleting by Term (unlike by Query)
// because on flush we apply all Term deletes to
// each segment. So all Term deleting here is
// against prior segments:
if (rld.Delete(docID))
{
delCount++;
}
}
}
}
}
return(delCount);
}
}
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);
}
}
}
}
}
public virtual void _run()
{
for (int iter = 0; iter < NUM_TEST_ITER; iter++)
{
FieldData field = Fields[Random().Next(Fields.Length)];
TermsEnum termsEnum = TermsDict.Terms(field.FieldInfo.Name).Iterator(null);
if (Si.Codec is Lucene3xCodec)
{
// code below expects unicode sort order
continue;
}
int upto = 0;
// Test straight enum of the terms:
while (true)
{
BytesRef term = termsEnum.Next();
if (term == null)
{
break;
}
BytesRef expected = new BytesRef(field.Terms[upto++].Text2);
Assert.IsTrue(expected.BytesEquals(term), "expected=" + expected + " vs actual " + term);
}
Assert.AreEqual(upto, field.Terms.Length);
// Test random seek:
TermData term2 = field.Terms[Random().Next(field.Terms.Length)];
TermsEnum.SeekStatus status = termsEnum.SeekCeil(new BytesRef(term2.Text2));
Assert.AreEqual(status, TermsEnum.SeekStatus.FOUND);
Assert.AreEqual(term2.Docs.Length, termsEnum.DocFreq());
if (field.OmitTF)
{
this.VerifyDocs(term2.Docs, term2.Positions, TestUtil.Docs(Random(), termsEnum, null, null, DocsEnum.FLAG_NONE), false);
}
else
{
this.VerifyDocs(term2.Docs, term2.Positions, termsEnum.DocsAndPositions(null, null), true);
}
// Test random seek by ord:
int idx = Random().Next(field.Terms.Length);
term2 = field.Terms[idx];
bool success = false;
try
{
termsEnum.SeekExact(idx);
success = true;
}
catch (System.NotSupportedException uoe)
{
// ok -- skip it
}
if (success)
{
Assert.AreEqual(status, TermsEnum.SeekStatus.FOUND);
Assert.IsTrue(termsEnum.Term().BytesEquals(new BytesRef(term2.Text2)));
Assert.AreEqual(term2.Docs.Length, termsEnum.DocFreq());
if (field.OmitTF)
{
this.VerifyDocs(term2.Docs, term2.Positions, TestUtil.Docs(Random(), termsEnum, null, null, DocsEnum.FLAG_NONE), false);
}
else
{
this.VerifyDocs(term2.Docs, term2.Positions, termsEnum.DocsAndPositions(null, null), true);
}
}
// Test seek to non-existent terms:
if (VERBOSE)
{
Console.WriteLine("TEST: seek non-exist terms");
}
for (int i = 0; i < 100; i++)
{
string text2 = TestUtil.RandomUnicodeString(Random()) + ".";
status = termsEnum.SeekCeil(new BytesRef(text2));
Assert.IsTrue(status == TermsEnum.SeekStatus.NOT_FOUND || status == TermsEnum.SeekStatus.END);
}
// Seek to each term, backwards:
if (VERBOSE)
{
Console.WriteLine("TEST: seek terms backwards");
}
for (int i = field.Terms.Length - 1; i >= 0; i--)
{
Assert.AreEqual(TermsEnum.SeekStatus.FOUND, termsEnum.SeekCeil(new BytesRef(field.Terms[i].Text2)), Thread.CurrentThread.Name + ": field=" + field.FieldInfo.Name + " term=" + field.Terms[i].Text2);
Assert.AreEqual(field.Terms[i].Docs.Length, termsEnum.DocFreq());
}
// Seek to each term by ord, backwards
for (int i = field.Terms.Length - 1; i >= 0; i--)
{
try
{
termsEnum.SeekExact(i);
Assert.AreEqual(field.Terms[i].Docs.Length, termsEnum.DocFreq());
Assert.IsTrue(termsEnum.Term().BytesEquals(new BytesRef(field.Terms[i].Text2)));
}
catch (System.NotSupportedException uoe)
{
}
}
// Seek to non-existent empty-string term
status = termsEnum.SeekCeil(new BytesRef(""));
Assert.IsNotNull(status);
//Assert.AreEqual(TermsEnum.SeekStatus.NOT_FOUND, status);
// Make sure we're now pointing to first term
Assert.IsTrue(termsEnum.Term().BytesEquals(new BytesRef(field.Terms[0].Text2)));
// Test docs enum
termsEnum.SeekCeil(new BytesRef(""));
upto = 0;
do
{
term2 = field.Terms[upto];
if (Random().Next(3) == 1)
{
DocsEnum docs;
DocsEnum docsAndFreqs;
DocsAndPositionsEnum postings;
if (!field.OmitTF)
{
postings = termsEnum.DocsAndPositions(null, null);
if (postings != null)
{
docs = docsAndFreqs = postings;
}
else
{
docs = docsAndFreqs = TestUtil.Docs(Random(), termsEnum, null, null, DocsEnum.FLAG_FREQS);
}
}
else
{
postings = null;
docsAndFreqs = null;
docs = TestUtil.Docs(Random(), termsEnum, null, null, DocsEnum.FLAG_NONE);
}
Assert.IsNotNull(docs);
int upto2 = -1;
bool ended = false;
while (upto2 < term2.Docs.Length - 1)
{
// Maybe skip:
int left = term2.Docs.Length - upto2;
int doc;
if (Random().Next(3) == 1 && left >= 1)
{
int inc = 1 + Random().Next(left - 1);
upto2 += inc;
if (Random().Next(2) == 1)
{
doc = docs.Advance(term2.Docs[upto2]);
Assert.AreEqual(term2.Docs[upto2], doc);
}
else
{
doc = docs.Advance(1 + term2.Docs[upto2]);
if (doc == DocIdSetIterator.NO_MORE_DOCS)
{
// skipped past last doc
Debug.Assert(upto2 == term2.Docs.Length - 1);
ended = true;
break;
}
else
{
// skipped to next doc
Debug.Assert(upto2 < term2.Docs.Length - 1);
if (doc >= term2.Docs[1 + upto2])
{
upto2++;
}
}
}
}
else
{
doc = docs.NextDoc();
Assert.IsTrue(doc != -1);
upto2++;
}
Assert.AreEqual(term2.Docs[upto2], doc);
if (!field.OmitTF)
{
Assert.AreEqual(term2.Positions[upto2].Length, postings.Freq());
if (Random().Next(2) == 1)
{
this.VerifyPositions(term2.Positions[upto2], postings);
}
}
}
if (!ended)
{
Assert.AreEqual(DocIdSetIterator.NO_MORE_DOCS, docs.NextDoc());
}
}
upto++;
} while (termsEnum.Next() != null);
Assert.AreEqual(upto, field.Terms.Length);
}
}
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 virtual void AssertEquals(RandomTokenStream tk, FieldType ft, Terms terms)
{
Assert.AreEqual(1, terms.DocCount);
int termCount = new JCG.HashSet <string>(tk.terms).Count;
Assert.AreEqual((long)termCount, terms.Count); // LUCENENET specific - cast required because types don't match (xUnit checks this)
Assert.AreEqual((long)termCount, terms.SumDocFreq); // LUCENENET specific - cast required because types don't match (xUnit checks this)
Assert.AreEqual(ft.StoreTermVectorPositions, terms.HasPositions);
Assert.AreEqual(ft.StoreTermVectorOffsets, terms.HasOffsets);
Assert.AreEqual(ft.StoreTermVectorPayloads && tk.HasPayloads(), terms.HasPayloads);
ISet <BytesRef> uniqueTerms = new JCG.HashSet <BytesRef>();
foreach (string term in tk.freqs.Keys)
{
uniqueTerms.Add(new BytesRef(term));
}
BytesRef[] sortedTerms = uniqueTerms.ToArray(/*new BytesRef[0]*/);
Array.Sort(sortedTerms, terms.Comparer);
TermsEnum termsEnum = terms.GetEnumerator(Random.NextBoolean() ? null : this.termsEnum.Value);
this.termsEnum.Value = termsEnum;
for (int i = 0; i < sortedTerms.Length; ++i)
{
Assert.IsTrue(termsEnum.MoveNext());
Assert.AreEqual(sortedTerms[i], termsEnum.Term);
Assert.AreEqual(1, termsEnum.DocFreq);
FixedBitSet bits = new FixedBitSet(1);
DocsEnum docsEnum = termsEnum.Docs(bits, Random.NextBoolean() ? null : this.docsEnum.Value);
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.NextDoc());
bits.Set(0);
docsEnum = termsEnum.Docs(Random.NextBoolean() ? bits : null, Random.NextBoolean() ? null : docsEnum);
Assert.IsNotNull(docsEnum);
Assert.AreEqual(0, docsEnum.NextDoc());
Assert.AreEqual(0, docsEnum.DocID);
Assert.AreEqual(tk.freqs[termsEnum.Term.Utf8ToString()], docsEnum.Freq);
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.NextDoc());
this.docsEnum.Value = docsEnum;
bits.Clear(0);
DocsAndPositionsEnum docsAndPositionsEnum = termsEnum.DocsAndPositions(bits, Random.NextBoolean() ? null : this.docsAndPositionsEnum.Value);
Assert.AreEqual(ft.StoreTermVectorOffsets || ft.StoreTermVectorPositions, docsAndPositionsEnum != null);
if (docsAndPositionsEnum != null)
{
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsAndPositionsEnum.NextDoc());
}
bits.Set(0);
docsAndPositionsEnum = termsEnum.DocsAndPositions(Random.NextBoolean() ? bits : null, Random.NextBoolean() ? null : docsAndPositionsEnum);
Assert.AreEqual(ft.StoreTermVectorOffsets || ft.StoreTermVectorPositions, docsAndPositionsEnum != null);
if (terms.HasPositions || terms.HasOffsets)
{
Assert.AreEqual(0, docsAndPositionsEnum.NextDoc());
int freq = docsAndPositionsEnum.Freq;
Assert.AreEqual(tk.freqs[termsEnum.Term.Utf8ToString()], freq);
if (docsAndPositionsEnum != null)
{
for (int k = 0; k < freq; ++k)
{
int position = docsAndPositionsEnum.NextPosition();
ISet <int> indexes;
if (terms.HasPositions)
{
indexes = tk.positionToTerms[position];
Assert.IsNotNull(indexes);
}
else
{
indexes = tk.startOffsetToTerms[docsAndPositionsEnum.StartOffset];
Assert.IsNotNull(indexes);
}
if (terms.HasPositions)
{
bool foundPosition = false;
foreach (int index in indexes)
{
if (tk.termBytes[index].Equals(termsEnum.Term) && tk.positions[index] == position)
{
foundPosition = true;
break;
}
}
Assert.IsTrue(foundPosition);
}
if (terms.HasOffsets)
{
bool foundOffset = false;
foreach (int index in indexes)
{
if (tk.termBytes[index].Equals(termsEnum.Term) && tk.startOffsets[index] == docsAndPositionsEnum.StartOffset && tk.endOffsets[index] == docsAndPositionsEnum.EndOffset)
{
foundOffset = true;
break;
}
}
Assert.IsTrue(foundOffset);
}
if (terms.HasPayloads)
{
bool foundPayload = false;
foreach (int index in indexes)
{
if (tk.termBytes[index].Equals(termsEnum.Term) && Equals(tk.payloads[index], docsAndPositionsEnum.GetPayload()))
{
foundPayload = true;
break;
}
}
Assert.IsTrue(foundPayload);
}
}
try
{
docsAndPositionsEnum.NextPosition();
Assert.Fail();
}
catch (Exception e) when(e.IsException())
{
// ok
}
catch (Exception e) when(e.IsAssertionError())
{
// ok
}
}
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsAndPositionsEnum.NextDoc());
}
this.docsAndPositionsEnum.Value = docsAndPositionsEnum;
}
Assert.IsFalse(termsEnum.MoveNext());
for (int i = 0; i < 5; ++i)
{
if (Random.NextBoolean())
{
Assert.IsTrue(termsEnum.SeekExact(RandomPicks.RandomFrom(Random, tk.termBytes)));
}
else
{
Assert.AreEqual(SeekStatus.FOUND, termsEnum.SeekCeil(RandomPicks.RandomFrom(Random, tk.termBytes)));
}
}
}
/// <exception cref="System.IO.IOException"></exception>
internal SegmentResult(int[] counts, int total, int missingCountIndex, TermsEnum
tenum, int startFacetOrd, int endFacetOrd) : base(counts, total - counts[missingCountIndex
], counts[missingCountIndex], endFacetOrd == missingCountIndex + 1 ? missingCountIndex
: endFacetOrd)
{
this.tenum = tenum;
this.mergePos = startFacetOrd;
if (tenum != null)
{
tenum.SeekExact(mergePos);
mergeTerm = tenum.Term();
}
}
protected internal virtual void AssertEquals(RandomTokenStream tk, FieldType ft, Terms terms)
{
Assert.AreEqual(1, terms.DocCount);
int termCount = (new HashSet <string>(Arrays.AsList(tk.Terms))).Count;
Assert.AreEqual(termCount, terms.Count);
Assert.AreEqual(termCount, terms.SumDocFreq);
Assert.AreEqual(ft.StoreTermVectorPositions, terms.HasPositions);
Assert.AreEqual(ft.StoreTermVectorOffsets, terms.HasOffsets);
Assert.AreEqual(ft.StoreTermVectorPayloads && tk.HasPayloads(), terms.HasPayloads);
HashSet <BytesRef> uniqueTerms = new HashSet <BytesRef>();
foreach (string term in tk.Freqs.Keys)
{
uniqueTerms.Add(new BytesRef(term));
}
BytesRef[] sortedTerms = uniqueTerms.ToArray(/*new BytesRef[0]*/);
Array.Sort(sortedTerms, terms.Comparer);
TermsEnum termsEnum = terms.GetIterator(Random().NextBoolean() ? null : this.termsEnum.Value);
this.termsEnum.Value = termsEnum;
for (int i = 0; i < sortedTerms.Length; ++i)
{
BytesRef nextTerm = termsEnum.Next();
Assert.AreEqual(sortedTerms[i], nextTerm);
Assert.AreEqual(sortedTerms[i], termsEnum.Term);
Assert.AreEqual(1, termsEnum.DocFreq);
FixedBitSet bits = new FixedBitSet(1);
DocsEnum docsEnum = termsEnum.Docs(bits, Random().NextBoolean() ? null : this.docsEnum.Value);
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.NextDoc());
bits.Set(0);
docsEnum = termsEnum.Docs(Random().NextBoolean() ? bits : null, Random().NextBoolean() ? null : docsEnum);
Assert.IsNotNull(docsEnum);
Assert.AreEqual(0, docsEnum.NextDoc());
Assert.AreEqual(0, docsEnum.DocID);
Assert.AreEqual(tk.Freqs[termsEnum.Term.Utf8ToString()], (int?)docsEnum.Freq);
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.NextDoc());
this.docsEnum.Value = docsEnum;
bits.Clear(0);
DocsAndPositionsEnum docsAndPositionsEnum = termsEnum.DocsAndPositions(bits, Random().NextBoolean() ? null : this.docsAndPositionsEnum.Value);
Assert.AreEqual(ft.StoreTermVectorOffsets || ft.StoreTermVectorPositions, docsAndPositionsEnum != null);
if (docsAndPositionsEnum != null)
{
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsAndPositionsEnum.NextDoc());
}
bits.Set(0);
docsAndPositionsEnum = termsEnum.DocsAndPositions(Random().NextBoolean() ? bits : null, Random().NextBoolean() ? null : docsAndPositionsEnum);
Assert.AreEqual(ft.StoreTermVectorOffsets || ft.StoreTermVectorPositions, docsAndPositionsEnum != null);
if (terms.HasPositions || terms.HasOffsets)
{
Assert.AreEqual(0, docsAndPositionsEnum.NextDoc());
int freq = docsAndPositionsEnum.Freq;
Assert.AreEqual(tk.Freqs[termsEnum.Term.Utf8ToString()], (int?)freq);
if (docsAndPositionsEnum != null)
{
for (int k = 0; k < freq; ++k)
{
int position = docsAndPositionsEnum.NextPosition();
ISet <int?> indexes;
if (terms.HasPositions)
{
indexes = tk.PositionToTerms[position];
Assert.IsNotNull(indexes);
}
else
{
indexes = tk.StartOffsetToTerms[docsAndPositionsEnum.StartOffset];
Assert.IsNotNull(indexes);
}
if (terms.HasPositions)
{
bool foundPosition = false;
foreach (int index in indexes)
{
if (tk.TermBytes[index].Equals(termsEnum.Term) && tk.Positions[index] == position)
{
foundPosition = true;
break;
}
}
Assert.IsTrue(foundPosition);
}
if (terms.HasOffsets)
{
bool foundOffset = false;
foreach (int index in indexes)
{
if (tk.TermBytes[index].Equals(termsEnum.Term) && tk.StartOffsets[index] == docsAndPositionsEnum.StartOffset && tk.EndOffsets[index] == docsAndPositionsEnum.EndOffset)
{
foundOffset = true;
break;
}
}
Assert.IsTrue(foundOffset);
}
if (terms.HasPayloads)
{
bool foundPayload = false;
foreach (int index in indexes)
{
if (tk.TermBytes[index].Equals(termsEnum.Term) && Equals(tk.Payloads[index], docsAndPositionsEnum.GetPayload()))
{
foundPayload = true;
break;
}
}
Assert.IsTrue(foundPayload);
}
}
try
{
docsAndPositionsEnum.NextPosition();
Assert.Fail();
}
#pragma warning disable 168
catch (Exception e)
#pragma warning restore 168
{
// ok
}
}
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsAndPositionsEnum.NextDoc());
}
this.docsAndPositionsEnum.Value = docsAndPositionsEnum;
}
Assert.IsNull(termsEnum.Next());
for (int i = 0; i < 5; ++i)
{
if (Random().NextBoolean())
{
Assert.IsTrue(termsEnum.SeekExact(RandomInts.RandomFrom(Random(), tk.TermBytes)));
}
else
{
Assert.AreEqual(SeekStatus.FOUND, termsEnum.SeekCeil(RandomInts.RandomFrom(Random(), tk.TermBytes)));
}
}
}
/// <summary>
/// Returns the term (<seealso cref="BytesRef"/>) corresponding to
/// the provided ordinal.
/// </summary>
public virtual BytesRef LookupTerm(TermsEnum termsEnum, int ord)
{
termsEnum.SeekExact(ord);
return termsEnum.Term();
}
public virtual void VerifyEquals(DirectoryReader r1, DirectoryReader r2, string idField)
{
if (VERBOSE)
{
Console.WriteLine("\nr1 docs:");
PrintDocs(r1);
Console.WriteLine("\nr2 docs:");
PrintDocs(r2);
}
if (r1.NumDocs != r2.NumDocs)
{
Debug.Assert(false, "r1.NumDocs=" + r1.NumDocs + " vs r2.NumDocs=" + r2.NumDocs);
}
bool hasDeletes = !(r1.MaxDoc == r2.MaxDoc && r1.NumDocs == r1.MaxDoc);
int[] r2r1 = new int[r2.MaxDoc]; // r2 id to r1 id mapping
// create mapping from id2 space to id2 based on idField
Fields f1 = MultiFields.GetFields(r1);
if (f1 == null)
{
// make sure r2 is empty
Assert.IsNull(MultiFields.GetFields(r2));
return;
}
Terms terms1 = f1.GetTerms(idField);
if (terms1 == null)
{
Assert.IsTrue(MultiFields.GetFields(r2) == null || MultiFields.GetFields(r2).GetTerms(idField) == null);
return;
}
TermsEnum termsEnum = terms1.GetIterator(null);
IBits liveDocs1 = MultiFields.GetLiveDocs(r1);
IBits liveDocs2 = MultiFields.GetLiveDocs(r2);
Fields fields = MultiFields.GetFields(r2);
if (fields == null)
{
// make sure r1 is in fact empty (eg has only all
// deleted docs):
IBits liveDocs = MultiFields.GetLiveDocs(r1);
DocsEnum docs = null;
while (termsEnum.Next() != null)
{
docs = TestUtil.Docs(Random(), termsEnum, liveDocs, docs, DocsFlags.NONE);
while (docs.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
Assert.Fail("r1 is not empty but r2 is");
}
}
return;
}
Terms terms2 = fields.GetTerms(idField);
TermsEnum termsEnum2 = terms2.GetIterator(null);
DocsEnum termDocs1 = null;
DocsEnum termDocs2 = null;
while (true)
{
BytesRef term = termsEnum.Next();
//System.out.println("TEST: match id term=" + term);
if (term == null)
{
break;
}
termDocs1 = TestUtil.Docs(Random(), termsEnum, liveDocs1, termDocs1, DocsFlags.NONE);
if (termsEnum2.SeekExact(term))
{
termDocs2 = TestUtil.Docs(Random(), termsEnum2, liveDocs2, termDocs2, DocsFlags.NONE);
}
else
{
termDocs2 = null;
}
if (termDocs1.NextDoc() == DocIdSetIterator.NO_MORE_DOCS)
{
// this doc is deleted and wasn't replaced
Assert.IsTrue(termDocs2 == null || termDocs2.NextDoc() == DocIdSetIterator.NO_MORE_DOCS);
continue;
}
int id1 = termDocs1.DocID;
Assert.AreEqual(DocIdSetIterator.NO_MORE_DOCS, termDocs1.NextDoc());
Assert.IsTrue(termDocs2.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
int id2 = termDocs2.DocID;
Assert.AreEqual(DocIdSetIterator.NO_MORE_DOCS, termDocs2.NextDoc());
r2r1[id2] = id1;
// verify stored fields are equivalent
try
{
VerifyEquals(r1.Document(id1), r2.Document(id2));
}
catch (Exception t)
{
Console.WriteLine("FAILED id=" + term + " id1=" + id1 + " id2=" + id2 + " term=" + term);
Console.WriteLine(" d1=" + r1.Document(id1));
Console.WriteLine(" d2=" + r2.Document(id2));
throw t;
}
try
{
// verify term vectors are equivalent
VerifyEquals(r1.GetTermVectors(id1), r2.GetTermVectors(id2));
}
catch (Exception e)
{
Console.WriteLine("FAILED id=" + term + " id1=" + id1 + " id2=" + id2);
Fields tv1 = r1.GetTermVectors(id1);
Console.WriteLine(" d1=" + tv1);
if (tv1 != null)
{
DocsAndPositionsEnum dpEnum = null;
DocsEnum dEnum = null;
foreach (string field in tv1)
{
Console.WriteLine(" " + field + ":");
Terms terms3 = tv1.GetTerms(field);
Assert.IsNotNull(terms3);
TermsEnum termsEnum3 = terms3.GetIterator(null);
BytesRef term2;
while ((term2 = termsEnum3.Next()) != null)
{
Console.WriteLine(" " + term2.Utf8ToString() + ": freq=" + termsEnum3.TotalTermFreq);
dpEnum = termsEnum3.DocsAndPositions(null, dpEnum);
if (dpEnum != null)
{
Assert.IsTrue(dpEnum.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
int freq = dpEnum.Freq;
Console.WriteLine(" doc=" + dpEnum.DocID + " freq=" + freq);
for (int posUpto = 0; posUpto < freq; posUpto++)
{
Console.WriteLine(" pos=" + dpEnum.NextPosition());
}
}
else
{
dEnum = TestUtil.Docs(Random(), termsEnum3, null, dEnum, DocsFlags.FREQS);
Assert.IsNotNull(dEnum);
Assert.IsTrue(dEnum.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
int freq = dEnum.Freq;
Console.WriteLine(" doc=" + dEnum.DocID + " freq=" + freq);
}
}
}
}
Fields tv2 = r2.GetTermVectors(id2);
Console.WriteLine(" d2=" + tv2);
if (tv2 != null)
{
DocsAndPositionsEnum dpEnum = null;
DocsEnum dEnum = null;
foreach (string field in tv2)
{
Console.WriteLine(" " + field + ":");
Terms terms3 = tv2.GetTerms(field);
Assert.IsNotNull(terms3);
TermsEnum termsEnum3 = terms3.GetIterator(null);
BytesRef term2;
while ((term2 = termsEnum3.Next()) != null)
{
Console.WriteLine(" " + term2.Utf8ToString() + ": freq=" + termsEnum3.TotalTermFreq);
dpEnum = termsEnum3.DocsAndPositions(null, dpEnum);
if (dpEnum != null)
{
Assert.IsTrue(dpEnum.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
int freq = dpEnum.Freq;
Console.WriteLine(" doc=" + dpEnum.DocID + " freq=" + freq);
for (int posUpto = 0; posUpto < freq; posUpto++)
{
Console.WriteLine(" pos=" + dpEnum.NextPosition());
}
}
else
{
dEnum = TestUtil.Docs(Random(), termsEnum3, null, dEnum, DocsFlags.FREQS);
Assert.IsNotNull(dEnum);
Assert.IsTrue(dEnum.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
int freq = dEnum.Freq;
Console.WriteLine(" doc=" + dEnum.DocID + " freq=" + freq);
}
}
}
}
throw e;
}
}
//System.out.println("TEST: done match id");
// Verify postings
//System.out.println("TEST: create te1");
Fields fields1 = MultiFields.GetFields(r1);
IEnumerator <string> fields1Enum = fields1.GetEnumerator();
Fields fields2 = MultiFields.GetFields(r2);
IEnumerator <string> fields2Enum = fields2.GetEnumerator();
string field1 = null, field2 = null;
TermsEnum termsEnum1 = null;
termsEnum2 = null;
DocsEnum docs1 = null, docs2 = null;
// pack both doc and freq into single element for easy sorting
long[] info1 = new long[r1.NumDocs];
long[] info2 = new long[r2.NumDocs];
for (; ;)
{
BytesRef term1 = null, term2 = null;
// iterate until we get some docs
int len1;
for (; ;)
{
len1 = 0;
if (termsEnum1 == null)
{
if (!fields1Enum.MoveNext())
{
break;
}
field1 = fields1Enum.Current;
Terms terms = fields1.GetTerms(field1);
if (terms == null)
{
continue;
}
termsEnum1 = terms.GetIterator(null);
}
term1 = termsEnum1.Next();
if (term1 == null)
{
// no more terms in this field
termsEnum1 = null;
continue;
}
//System.out.println("TEST: term1=" + term1);
docs1 = TestUtil.Docs(Random(), termsEnum1, liveDocs1, docs1, DocsFlags.FREQS);
while (docs1.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
int d = docs1.DocID;
int f = docs1.Freq;
info1[len1] = (((long)d) << 32) | (uint)f;
len1++;
}
if (len1 > 0)
{
break;
}
}
// iterate until we get some docs
int len2;
for (; ;)
{
len2 = 0;
if (termsEnum2 == null)
{
if (!fields2Enum.MoveNext())
{
break;
}
field2 = fields2Enum.Current;
Terms terms = fields2.GetTerms(field2);
if (terms == null)
{
continue;
}
termsEnum2 = terms.GetIterator(null);
}
term2 = termsEnum2.Next();
if (term2 == null)
{
// no more terms in this field
termsEnum2 = null;
continue;
}
//System.out.println("TEST: term1=" + term1);
docs2 = TestUtil.Docs(Random(), termsEnum2, liveDocs2, docs2, DocsFlags.FREQS);
while (docs2.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
int d = r2r1[docs2.DocID];
int f = docs2.Freq;
info2[len2] = (((long)d) << 32) | (uint)f;
len2++;
}
if (len2 > 0)
{
break;
}
}
Assert.AreEqual(len1, len2);
if (len1 == 0) // no more terms
{
break;
}
Assert.AreEqual(field1, field2);
Assert.IsTrue(term1.BytesEquals(term2));
if (!hasDeletes)
{
Assert.AreEqual(termsEnum1.DocFreq, termsEnum2.DocFreq);
}
Assert.AreEqual(term1, term2, "len1=" + len1 + " len2=" + len2 + " deletes?=" + hasDeletes);
// sort info2 to get it into ascending docid
Array.Sort(info2, 0, len2);
// now compare
for (int i = 0; i < len1; i++)
{
Assert.AreEqual(info1[i], info2[i], "i=" + i + " len=" + len1 + " d1=" + ((long)((ulong)info1[i] >> 32)) + " f1=" + (info1[i] & int.MaxValue) + " d2=" + ((long)((ulong)info2[i] >> 32)) + " f2=" + (info2[i] & int.MaxValue) + " field=" + field1 + " term=" + term1.Utf8ToString());
}
}
}
