private void VerifyVector(TermsEnum vector, int num)
{
StringBuilder temp = new StringBuilder();
while (vector.Next() != null)
{
temp.Append(vector.Term().Utf8ToString());
}
if (!English.IntToEnglish(num).Trim().Equals(temp.ToString().Trim()))
{
Console.WriteLine("wrong term result");
}
}
public virtual void Test()
{
Directory dir = NewDirectory();
MockAnalyzer analyzer = new MockAnalyzer(Random());
analyzer.MaxTokenLength = TestUtil.NextInt(Random(), 1, IndexWriter.MAX_TERM_LENGTH);
RandomIndexWriter w = new RandomIndexWriter(Random(), dir, analyzer);
LineFileDocs docs = new LineFileDocs(Random(), DefaultCodecSupportsDocValues());
int charsToIndex = AtLeast(100000);
int charsIndexed = 0;
//System.out.println("bytesToIndex=" + charsToIndex);
while (charsIndexed < charsToIndex)
{
Document doc = docs.NextDoc();
charsIndexed += doc.Get("body").Length;
w.AddDocument(doc);
//System.out.println(" bytes=" + charsIndexed + " add: " + doc);
}
IndexReader r = w.Reader;
//System.out.println("numDocs=" + r.NumDocs);
w.Dispose();
IndexSearcher s = NewSearcher(r);
Terms terms = MultiFields.GetFields(r).Terms("body");
int termCount = 0;
TermsEnum termsEnum = terms.Iterator(null);
while (termsEnum.Next() != null)
{
termCount++;
}
Assert.IsTrue(termCount > 0);
// Target ~10 terms to search:
double chance = 10.0 / termCount;
termsEnum = terms.Iterator(termsEnum);
IDictionary <BytesRef, TopDocs> answers = new Dictionary <BytesRef, TopDocs>();
while (termsEnum.Next() != null)
{
if (Random().NextDouble() <= chance)
{
BytesRef term = BytesRef.DeepCopyOf(termsEnum.Term());
answers[term] = s.Search(new TermQuery(new Term("body", term)), 100);
}
}
if (answers.Count > 0)
{
CountdownEvent startingGun = new CountdownEvent(1);
int numThreads = TestUtil.NextInt(Random(), 2, 5);
ThreadClass[] threads = new ThreadClass[numThreads];
for (int threadID = 0; threadID < numThreads; threadID++)
{
ThreadClass thread = new ThreadAnonymousInnerClassHelper(this, s, answers, startingGun);
threads[threadID] = thread;
thread.Start();
}
startingGun.Signal();
foreach (ThreadClass thread in threads)
{
thread.Join();
}
}
r.Dispose();
dir.Dispose();
}
private void DuellReaders(CompositeReader other, AtomicReader memIndexReader)
{
AtomicReader competitor = SlowCompositeReaderWrapper.Wrap(other);
Fields memFields = memIndexReader.Fields;
foreach (string field in competitor.Fields)
{
Terms memTerms = memFields.Terms(field);
Terms iwTerms = memIndexReader.Terms(field);
if (iwTerms == null)
{
assertNull(memTerms);
}
else
{
NumericDocValues normValues = competitor.GetNormValues(field);
NumericDocValues memNormValues = memIndexReader.GetNormValues(field);
if (normValues != null)
{
// mem idx always computes norms on the fly
assertNotNull(memNormValues);
assertEquals(normValues.Get(0), memNormValues.Get(0));
}
assertNotNull(memTerms);
assertEquals(iwTerms.DocCount, memTerms.DocCount);
assertEquals(iwTerms.SumDocFreq, memTerms.SumDocFreq);
assertEquals(iwTerms.SumTotalTermFreq, memTerms.SumTotalTermFreq);
TermsEnum iwTermsIter = iwTerms.Iterator(null);
TermsEnum memTermsIter = memTerms.Iterator(null);
if (iwTerms.HasPositions())
{
bool offsets = iwTerms.HasOffsets() && memTerms.HasOffsets();
while (iwTermsIter.Next() != null)
{
assertNotNull(memTermsIter.Next());
assertEquals(iwTermsIter.Term(), memTermsIter.Term());
DocsAndPositionsEnum iwDocsAndPos = iwTermsIter.DocsAndPositions(null, null);
DocsAndPositionsEnum memDocsAndPos = memTermsIter.DocsAndPositions(null, null);
while (iwDocsAndPos.NextDoc() != DocsAndPositionsEnum.NO_MORE_DOCS)
{
assertEquals(iwDocsAndPos.DocID(), memDocsAndPos.NextDoc());
assertEquals(iwDocsAndPos.Freq(), memDocsAndPos.Freq());
for (int i = 0; i < iwDocsAndPos.Freq(); i++)
{
assertEquals("term: " + iwTermsIter.Term().Utf8ToString(), iwDocsAndPos.NextPosition(), memDocsAndPos.NextPosition());
if (offsets)
{
assertEquals(iwDocsAndPos.StartOffset(), memDocsAndPos.StartOffset());
assertEquals(iwDocsAndPos.EndOffset(), memDocsAndPos.EndOffset());
}
}
}
}
}
else
{
while (iwTermsIter.Next() != null)
{
assertEquals(iwTermsIter.Term(), memTermsIter.Term());
DocsEnum iwDocsAndPos = iwTermsIter.Docs(null, null);
DocsEnum memDocsAndPos = memTermsIter.Docs(null, null);
while (iwDocsAndPos.NextDoc() != DocsAndPositionsEnum.NO_MORE_DOCS)
{
assertEquals(iwDocsAndPos.DocID(), memDocsAndPos.NextDoc());
assertEquals(iwDocsAndPos.Freq(), memDocsAndPos.Freq());
}
}
}
}
}
}
// we need to guarantee that if several threads call this concurrently, only
// one executes it, and after it returns, the cache is updated and is either
// complete or not.
private void PerhapsFillCache()
{
lock (this)
{
if (cacheMisses.Get() < cacheMissesUntilFill)
{
return;
}
if (!shouldFillCache)
{
// we already filled the cache once, there's no need to re-fill it
return;
}
shouldFillCache = false;
InitReaderManager();
bool aborted = false;
DirectoryReader reader = readerManager.Acquire();
try
{
TermsEnum termsEnum = null;
DocsEnum docsEnum = null;
foreach (AtomicReaderContext ctx in reader.Leaves)
{
Terms terms = ctx.AtomicReader.Terms(Consts.FULL);
if (terms != null) // cannot really happen, but be on the safe side
{
termsEnum = terms.Iterator(termsEnum);
while (termsEnum.Next() != null)
{
if (!cache.Full)
{
BytesRef t = termsEnum.Term();
// Since we guarantee uniqueness of categories, each term has exactly
// one document. Also, since we do not allow removing categories (and
// hence documents), there are no deletions in the index. Therefore, it
// is sufficient to call next(), and then doc(), exactly once with no
// 'validation' checks.
FacetLabel cp = new FacetLabel(FacetsConfig.StringToPath(t.Utf8ToString()));
docsEnum = termsEnum.Docs(null, docsEnum, DocsEnum.FLAG_NONE);
bool res = cache.Put(cp, docsEnum.NextDoc() + ctx.DocBase);
Debug.Assert(!res, "entries should not have been evicted from the cache");
}
else
{
// the cache is full and the next put() will evict entries from it, therefore abort the iteration.
aborted = true;
break;
}
}
}
if (aborted)
{
break;
}
}
}
finally
{
readerManager.Release(reader);
}
cacheIsComplete = !aborted;
if (cacheIsComplete)
{
lock (this)
{
// everything is in the cache, so no need to keep readerManager open.
// this block is executed in a sync block so that it works well with
// initReaderManager called in parallel.
readerManager.Dispose();
readerManager = null;
initializedReaderManager = false;
}
}
}
}
/// <summary>
/// Takes the categories from the given taxonomy directory, and adds the
/// missing ones to this taxonomy. Additionally, it fills the given
/// <seealso cref="OrdinalMap"/> with a mapping from the original ordinal to the new
/// ordinal.
/// </summary>
public virtual void AddTaxonomy(Directory taxoDir, OrdinalMap map)
{
EnsureOpen();
DirectoryReader r = DirectoryReader.Open(taxoDir);
try
{
int size = r.NumDocs;
OrdinalMap ordinalMap = map;
ordinalMap.Size = size;
int @base = 0;
TermsEnum te = null;
DocsEnum docs = null;
foreach (AtomicReaderContext ctx in r.Leaves)
{
AtomicReader ar = ctx.AtomicReader;
Terms terms = ar.Terms(Consts.FULL);
te = terms.Iterator(te);
while (te.Next() != null)
{
FacetLabel cp = new FacetLabel(FacetsConfig.StringToPath(te.Term().Utf8ToString()));
int ordinal = AddCategory(cp);
docs = te.Docs(null, docs, DocsEnum.FLAG_NONE);
ordinalMap.AddMapping(docs.NextDoc() + @base, ordinal);
}
@base += ar.MaxDoc; // no deletions, so we're ok
}
ordinalMap.AddDone();
}
finally
{
r.Dispose();
}
}
public virtual void TestPhrasePrefix()
{
Directory indexStore = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), indexStore);
Add("blueberry pie", writer);
Add("blueberry strudel", writer);
Add("blueberry pizza", writer);
Add("blueberry chewing gum", writer);
Add("bluebird pizza", writer);
Add("bluebird foobar pizza", writer);
Add("piccadilly circus", writer);
IndexReader reader = writer.Reader;
IndexSearcher searcher = NewSearcher(reader);
// search for "blueberry pi*":
MultiPhraseQuery query1 = new MultiPhraseQuery();
// search for "strawberry pi*":
MultiPhraseQuery query2 = new MultiPhraseQuery();
query1.Add(new Term("body", "blueberry"));
query2.Add(new Term("body", "strawberry"));
LinkedList <Term> termsWithPrefix = new LinkedList <Term>();
// this TermEnum gives "piccadilly", "pie" and "pizza".
string prefix = "pi";
TermsEnum te = MultiFields.GetFields(reader).Terms("body").Iterator(null);
te.SeekCeil(new BytesRef(prefix));
do
{
string s = te.Term().Utf8ToString();
if (s.StartsWith(prefix))
{
termsWithPrefix.AddLast(new Term("body", s));
}
else
{
break;
}
} while (te.Next() != null);
query1.Add(termsWithPrefix.ToArray(/*new Term[0]*/));
Assert.AreEqual("body:\"blueberry (piccadilly pie pizza)\"", query1.ToString());
query2.Add(termsWithPrefix.ToArray(/*new Term[0]*/));
Assert.AreEqual("body:\"strawberry (piccadilly pie pizza)\"", query2.ToString());
ScoreDoc[] result;
result = searcher.Search(query1, null, 1000).ScoreDocs;
Assert.AreEqual(2, result.Length);
result = searcher.Search(query2, null, 1000).ScoreDocs;
Assert.AreEqual(0, result.Length);
// search for "blue* pizza":
MultiPhraseQuery query3 = new MultiPhraseQuery();
termsWithPrefix.Clear();
prefix = "blue";
te.SeekCeil(new BytesRef(prefix));
do
{
if (te.Term().Utf8ToString().StartsWith(prefix))
{
termsWithPrefix.AddLast(new Term("body", te.Term().Utf8ToString()));
}
} while (te.Next() != null);
query3.Add(termsWithPrefix.ToArray(/*new Term[0]*/));
query3.Add(new Term("body", "pizza"));
result = searcher.Search(query3, null, 1000).ScoreDocs;
Assert.AreEqual(2, result.Length); // blueberry pizza, bluebird pizza
Assert.AreEqual("body:\"(blueberry bluebird) pizza\"", query3.ToString());
// test slop:
query3.Slop = 1;
result = searcher.Search(query3, null, 1000).ScoreDocs;
// just make sure no exc:
searcher.Explain(query3, 0);
Assert.AreEqual(3, result.Length); // blueberry pizza, bluebird pizza, bluebird
// foobar pizza
MultiPhraseQuery query4 = new MultiPhraseQuery();
try
{
query4.Add(new Term("field1", "foo"));
query4.Add(new Term("field2", "foobar"));
Assert.Fail();
}
catch (System.ArgumentException e)
{
// okay, all terms must belong to the same field
}
writer.Dispose();
reader.Dispose();
indexStore.Dispose();
}
public virtual void TestPhrasePrefix()
{
Directory indexStore = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), indexStore);
Document doc1 = new Document();
Document doc2 = new Document();
Document doc3 = new Document();
Document doc4 = new Document();
Document doc5 = new Document();
doc1.Add(NewTextField("body", "blueberry pie", Field.Store.YES));
doc2.Add(NewTextField("body", "blueberry strudel", Field.Store.YES));
doc3.Add(NewTextField("body", "blueberry pizza", Field.Store.YES));
doc4.Add(NewTextField("body", "blueberry chewing gum", Field.Store.YES));
doc5.Add(NewTextField("body", "piccadilly circus", Field.Store.YES));
writer.AddDocument(doc1);
writer.AddDocument(doc2);
writer.AddDocument(doc3);
writer.AddDocument(doc4);
writer.AddDocument(doc5);
IndexReader reader = writer.Reader;
writer.Dispose();
IndexSearcher searcher = NewSearcher(reader);
// PhrasePrefixQuery query1 = new PhrasePrefixQuery();
MultiPhraseQuery query1 = new MultiPhraseQuery();
// PhrasePrefixQuery query2 = new PhrasePrefixQuery();
MultiPhraseQuery query2 = new MultiPhraseQuery();
query1.Add(new Term("body", "blueberry"));
query2.Add(new Term("body", "strawberry"));
LinkedList <Term> termsWithPrefix = new LinkedList <Term>();
// this TermEnum gives "piccadilly", "pie" and "pizza".
string prefix = "pi";
TermsEnum te = MultiFields.GetFields(reader).Terms("body").Iterator(null);
te.SeekCeil(new BytesRef(prefix));
do
{
string s = te.Term().Utf8ToString();
if (s.StartsWith(prefix))
{
termsWithPrefix.AddLast(new Term("body", s));
}
else
{
break;
}
} while (te.Next() != null);
query1.Add(termsWithPrefix.ToArray(/*new Term[0]*/));
query2.Add(termsWithPrefix.ToArray(/*new Term[0]*/));
ScoreDoc[] result;
result = searcher.Search(query1, null, 1000).ScoreDocs;
Assert.AreEqual(2, result.Length);
result = searcher.Search(query2, null, 1000).ScoreDocs;
Assert.AreEqual(0, result.Length);
reader.Dispose();
indexStore.Dispose();
}
public virtual void Test()
{
IFieldCache cache = FieldCache.DEFAULT;
FieldCache.Doubles doubles = cache.GetDoubles(Reader, "theDouble", Random().NextBoolean());
Assert.AreSame(doubles, cache.GetDoubles(Reader, "theDouble", Random().NextBoolean()), "Second request to cache return same array");
Assert.AreSame(doubles, cache.GetDoubles(Reader, "theDouble", FieldCache.DEFAULT_DOUBLE_PARSER, Random().NextBoolean()), "Second request with explicit parser return same array");
for (int i = 0; i < NUM_DOCS; i++)
{
Assert.IsTrue(doubles.Get(i) == (double.MaxValue - i), doubles.Get(i) + " does not equal: " + (double.MaxValue - i));
}
FieldCache.Longs longs = cache.GetLongs(Reader, "theLong", Random().NextBoolean());
Assert.AreSame(longs, cache.GetLongs(Reader, "theLong", Random().NextBoolean()), "Second request to cache return same array");
Assert.AreSame(longs, cache.GetLongs(Reader, "theLong", FieldCache.DEFAULT_LONG_PARSER, Random().NextBoolean()), "Second request with explicit parser return same array");
for (int i = 0; i < NUM_DOCS; i++)
{
Assert.IsTrue(longs.Get(i) == (long.MaxValue - i), longs.Get(i) + " does not equal: " + (long.MaxValue - i) + " i=" + i);
}
FieldCache.Bytes bytes = cache.GetBytes(Reader, "theByte", Random().NextBoolean());
Assert.AreSame(bytes, cache.GetBytes(Reader, "theByte", Random().NextBoolean()), "Second request to cache return same array");
Assert.AreSame(bytes, cache.GetBytes(Reader, "theByte", FieldCache.DEFAULT_BYTE_PARSER, Random().NextBoolean()), "Second request with explicit parser return same array");
for (int i = 0; i < NUM_DOCS; i++)
{
Assert.IsTrue(bytes.Get(i) == (sbyte)(sbyte.MaxValue - i), bytes.Get(i) + " does not equal: " + (sbyte.MaxValue - i));
}
FieldCache.Shorts shorts = cache.GetShorts(Reader, "theShort", Random().NextBoolean());
Assert.AreSame(shorts, cache.GetShorts(Reader, "theShort", Random().NextBoolean()), "Second request to cache return same array");
Assert.AreSame(shorts, cache.GetShorts(Reader, "theShort", FieldCache.DEFAULT_SHORT_PARSER, Random().NextBoolean()), "Second request with explicit parser return same array");
for (int i = 0; i < NUM_DOCS; i++)
{
Assert.IsTrue(shorts.Get(i) == (short)(short.MaxValue - i), shorts.Get(i) + " does not equal: " + (short.MaxValue - i));
}
FieldCache.Ints ints = cache.GetInts(Reader, "theInt", Random().NextBoolean());
Assert.AreSame(ints, cache.GetInts(Reader, "theInt", Random().NextBoolean()), "Second request to cache return same array");
Assert.AreSame(ints, cache.GetInts(Reader, "theInt", FieldCache.DEFAULT_INT_PARSER, Random().NextBoolean()), "Second request with explicit parser return same array");
for (int i = 0; i < NUM_DOCS; i++)
{
Assert.IsTrue(ints.Get(i) == (int.MaxValue - i), ints.Get(i) + " does not equal: " + (int.MaxValue - i));
}
FieldCache.Floats floats = cache.GetFloats(Reader, "theFloat", Random().NextBoolean());
Assert.AreSame(floats, cache.GetFloats(Reader, "theFloat", Random().NextBoolean()), "Second request to cache return same array");
Assert.AreSame(floats, cache.GetFloats(Reader, "theFloat", FieldCache.DEFAULT_FLOAT_PARSER, Random().NextBoolean()), "Second request with explicit parser return same array");
for (int i = 0; i < NUM_DOCS; i++)
{
Assert.IsTrue(floats.Get(i) == (float.MaxValue - i), floats.Get(i) + " does not equal: " + (float.MaxValue - i));
}
Bits docsWithField = cache.GetDocsWithField(Reader, "theLong");
Assert.AreSame(docsWithField, cache.GetDocsWithField(Reader, "theLong"), "Second request to cache return same array");
Assert.IsTrue(docsWithField is Bits_MatchAllBits, "docsWithField(theLong) must be class Bits.MatchAllBits");
Assert.IsTrue(docsWithField.Length() == NUM_DOCS, "docsWithField(theLong) Size: " + docsWithField.Length() + " is not: " + NUM_DOCS);
for (int i = 0; i < docsWithField.Length(); i++)
{
Assert.IsTrue(docsWithField.Get(i));
}
docsWithField = cache.GetDocsWithField(Reader, "sparse");
Assert.AreSame(docsWithField, cache.GetDocsWithField(Reader, "sparse"), "Second request to cache return same array");
Assert.IsFalse(docsWithField is Bits_MatchAllBits, "docsWithField(sparse) must not be class Bits.MatchAllBits");
Assert.IsTrue(docsWithField.Length() == NUM_DOCS, "docsWithField(sparse) Size: " + docsWithField.Length() + " is not: " + NUM_DOCS);
for (int i = 0; i < docsWithField.Length(); i++)
{
Assert.AreEqual(i % 2 == 0, docsWithField.Get(i));
}
// getTermsIndex
SortedDocValues termsIndex = cache.GetTermsIndex(Reader, "theRandomUnicodeString");
Assert.AreSame(termsIndex, cache.GetTermsIndex(Reader, "theRandomUnicodeString"), "Second request to cache return same array");
BytesRef br = new BytesRef();
for (int i = 0; i < NUM_DOCS; i++)
{
BytesRef term;
int ord = termsIndex.GetOrd(i);
if (ord == -1)
{
term = null;
}
else
{
termsIndex.LookupOrd(ord, br);
term = br;
}
string s = term == null ? null : term.Utf8ToString();
Assert.IsTrue(UnicodeStrings[i] == null || UnicodeStrings[i].Equals(s), "for doc " + i + ": " + s + " does not equal: " + UnicodeStrings[i]);
}
int nTerms = termsIndex.ValueCount;
TermsEnum tenum = termsIndex.TermsEnum();
BytesRef val = new BytesRef();
for (int i = 0; i < nTerms; i++)
{
BytesRef val1 = tenum.Next();
termsIndex.LookupOrd(i, val);
// System.out.println("i="+i);
Assert.AreEqual(val, val1);
}
// seek the enum around (note this isn't a great test here)
int num = AtLeast(100);
for (int i = 0; i < num; i++)
{
int k = Random().Next(nTerms);
termsIndex.LookupOrd(k, val);
Assert.AreEqual(TermsEnum.SeekStatus.FOUND, tenum.SeekCeil(val));
Assert.AreEqual(val, tenum.Term());
}
for (int i = 0; i < nTerms; i++)
{
termsIndex.LookupOrd(i, val);
Assert.AreEqual(TermsEnum.SeekStatus.FOUND, tenum.SeekCeil(val));
Assert.AreEqual(val, tenum.Term());
}
// test bad field
termsIndex = cache.GetTermsIndex(Reader, "bogusfield");
// getTerms
BinaryDocValues terms = cache.GetTerms(Reader, "theRandomUnicodeString", true);
Assert.AreSame(terms, cache.GetTerms(Reader, "theRandomUnicodeString", true), "Second request to cache return same array");
Bits bits = cache.GetDocsWithField(Reader, "theRandomUnicodeString");
for (int i = 0; i < NUM_DOCS; i++)
{
terms.Get(i, br);
BytesRef term;
if (!bits.Get(i))
{
term = null;
}
else
{
term = br;
}
string s = term == null ? null : term.Utf8ToString();
Assert.IsTrue(UnicodeStrings[i] == null || UnicodeStrings[i].Equals(s), "for doc " + i + ": " + s + " does not equal: " + UnicodeStrings[i]);
}
// test bad field
terms = cache.GetTerms(Reader, "bogusfield", false);
// getDocTermOrds
SortedSetDocValues termOrds = cache.GetDocTermOrds(Reader, "theRandomUnicodeMultiValuedField");
int numEntries = cache.CacheEntries.Length;
// ask for it again, and check that we didnt create any additional entries:
termOrds = cache.GetDocTermOrds(Reader, "theRandomUnicodeMultiValuedField");
Assert.AreEqual(numEntries, cache.CacheEntries.Length);
for (int i = 0; i < NUM_DOCS; i++)
{
termOrds.Document = i;
// this will remove identical terms. A DocTermOrds doesn't return duplicate ords for a docId
IList <BytesRef> values = new List <BytesRef>(new /*Linked*/ HashSet <BytesRef>(Arrays.AsList(MultiValued[i])));
foreach (BytesRef v in values)
{
if (v == null)
{
// why does this test use null values... instead of an empty list: confusing
break;
}
long ord = termOrds.NextOrd();
Debug.Assert(ord != SortedSetDocValues.NO_MORE_ORDS);
BytesRef scratch = new BytesRef();
termOrds.LookupOrd(ord, scratch);
Assert.AreEqual(v, scratch);
}
Assert.AreEqual(SortedSetDocValues.NO_MORE_ORDS, termOrds.NextOrd());
}
// test bad field
termOrds = cache.GetDocTermOrds(Reader, "bogusfield");
Assert.IsTrue(termOrds.ValueCount == 0);
FieldCache.DEFAULT.PurgeByCacheKey(Reader.CoreCacheKey);
}
private void AssertTermsSeeking(Terms leftTerms, Terms rightTerms)
{
TermsEnum leftEnum = null;
TermsEnum rightEnum = null;
// just an upper bound
int numTests = AtLeast(20);
Random random = Random();
// collect this number of terms from the left side
HashSet <BytesRef> tests = new HashSet <BytesRef>();
int numPasses = 0;
while (numPasses < 10 && tests.Count < numTests)
{
leftEnum = leftTerms.Iterator(leftEnum);
BytesRef term = null;
while ((term = leftEnum.Next()) != null)
{
int code = random.Next(10);
if (code == 0)
{
// the term
tests.Add(BytesRef.DeepCopyOf(term));
}
else if (code == 1)
{
// truncated subsequence of term
term = BytesRef.DeepCopyOf(term);
if (term.Length > 0)
{
// truncate it
term.Length = random.Next(term.Length);
}
}
else if (code == 2)
{
// term, but ensure a non-zero offset
sbyte[] newbytes = new sbyte[term.Length + 5];
Array.Copy(term.Bytes, term.Offset, newbytes, 5, term.Length);
tests.Add(new BytesRef(newbytes, 5, term.Length));
}
}
numPasses++;
}
List <BytesRef> shuffledTests = new List <BytesRef>(tests);
shuffledTests = (List <BytesRef>)CollectionsHelper.Shuffle(shuffledTests);
foreach (BytesRef b in shuffledTests)
{
leftEnum = leftTerms.Iterator(leftEnum);
rightEnum = rightTerms.Iterator(rightEnum);
Assert.AreEqual(leftEnum.SeekExact(b), rightEnum.SeekExact(b));
Assert.AreEqual(leftEnum.SeekExact(b), rightEnum.SeekExact(b));
SeekStatus leftStatus;
SeekStatus rightStatus;
leftStatus = leftEnum.SeekCeil(b);
rightStatus = rightEnum.SeekCeil(b);
Assert.AreEqual(leftStatus, rightStatus);
if (leftStatus != SeekStatus.END)
{
Assert.AreEqual(leftEnum.Term(), rightEnum.Term());
}
leftStatus = leftEnum.SeekCeil(b);
rightStatus = rightEnum.SeekCeil(b);
Assert.AreEqual(leftStatus, rightStatus);
if (leftStatus != SeekStatus.END)
{
Assert.AreEqual(leftEnum.Term(), rightEnum.Term());
}
}
}
public virtual void TestSimple()
{
int numNodes = TestUtil.NextInt(Random(), 1, 10);
double runTimeSec = AtLeast(3);
int minDocsToMakeTerms = TestUtil.NextInt(Random(), 5, 20);
int maxSearcherAgeSeconds = TestUtil.NextInt(Random(), 1, 3);
if (VERBOSE)
{
Console.WriteLine("TEST: numNodes=" + numNodes + " runTimeSec=" + runTimeSec + " maxSearcherAgeSeconds=" + maxSearcherAgeSeconds);
}
Start(numNodes, runTimeSec, maxSearcherAgeSeconds);
List <PreviousSearchState> priorSearches = new List <PreviousSearchState>();
List <BytesRef> terms = null;
while (TimeHelper.NanoTime() < EndTimeNanos)
{
bool doFollowon = priorSearches.Count > 0 && Random().Next(7) == 1;
// Pick a random node; we will run the query on this node:
int myNodeID = Random().Next(numNodes);
NodeState.ShardIndexSearcher localShardSearcher;
PreviousSearchState prevSearchState;
if (doFollowon)
{
// Pretend user issued a followon query:
prevSearchState = priorSearches[Random().Next(priorSearches.Count)];
if (VERBOSE)
{
Console.WriteLine("\nTEST: follow-on query age=" + ((TimeHelper.NanoTime() - prevSearchState.SearchTimeNanos) / 1000000000.0));
}
try
{
localShardSearcher = Nodes[myNodeID].Acquire(prevSearchState.Versions);
}
catch (SearcherExpiredException see)
{
// Expected, sometimes; in a "real" app we would
// either forward this error to the user ("too
// much time has passed; please re-run your
// search") or sneakily just switch to newest
// searcher w/o telling them...
if (VERBOSE)
{
Console.WriteLine(" searcher expired during local shard searcher init: " + see);
}
priorSearches.Remove(prevSearchState);
continue;
}
}
else
{
if (VERBOSE)
{
Console.WriteLine("\nTEST: fresh query");
}
// Do fresh query:
localShardSearcher = Nodes[myNodeID].Acquire();
prevSearchState = null;
}
IndexReader[] subs = new IndexReader[numNodes];
PreviousSearchState searchState = null;
try
{
// Mock: now make a single reader (MultiReader) from all node
// searchers. In a real shard env you can't do this... we
// do it to confirm results from the shard searcher
// are correct:
int docCount = 0;
try
{
for (int nodeID = 0; nodeID < numNodes; nodeID++)
{
long subVersion = localShardSearcher.NodeVersions[nodeID];
IndexSearcher sub = Nodes[nodeID].Searchers.Acquire(subVersion);
if (sub == null)
{
nodeID--;
while (nodeID >= 0)
{
subs[nodeID].DecRef();
subs[nodeID] = null;
nodeID--;
}
throw new SearcherExpiredException("nodeID=" + nodeID + " version=" + subVersion);
}
subs[nodeID] = sub.IndexReader;
docCount += subs[nodeID].MaxDoc;
}
}
catch (SearcherExpiredException see)
{
// Expected
if (VERBOSE)
{
Console.WriteLine(" searcher expired during mock reader init: " + see);
}
continue;
}
IndexReader mockReader = new MultiReader(subs);
IndexSearcher mockSearcher = new IndexSearcher(mockReader);
Query query;
Sort sort;
if (prevSearchState != null)
{
query = prevSearchState.Query;
sort = prevSearchState.Sort;
}
else
{
if (terms == null && docCount > minDocsToMakeTerms)
{
// TODO: try to "focus" on high freq terms sometimes too
// TODO: maybe also periodically reset the terms...?
TermsEnum termsEnum = MultiFields.GetTerms(mockReader, "body").Iterator(null);
terms = new List <BytesRef>();
while (termsEnum.Next() != null)
{
terms.Add(BytesRef.DeepCopyOf(termsEnum.Term()));
}
if (VERBOSE)
{
Console.WriteLine("TEST: init terms: " + terms.Count + " terms");
}
if (terms.Count == 0)
{
terms = null;
}
}
if (VERBOSE)
{
Console.WriteLine(" maxDoc=" + mockReader.MaxDoc);
}
if (terms != null)
{
if (Random().NextBoolean())
{
query = new TermQuery(new Term("body", terms[Random().Next(terms.Count)]));
}
else
{
string t = terms[Random().Next(terms.Count)].Utf8ToString();
string prefix;
if (t.Length <= 1)
{
prefix = t;
}
else
{
prefix = t.Substring(0, TestUtil.NextInt(Random(), 1, 2));
}
query = new PrefixQuery(new Term("body", prefix));
}
if (Random().NextBoolean())
{
sort = null;
}
else
{
// TODO: sort by more than 1 field
int what = Random().Next(3);
if (what == 0)
{
sort = new Sort(SortField.FIELD_SCORE);
}
else if (what == 1)
{
// TODO: this sort doesn't merge
// correctly... it's tricky because you
// could have > 2.1B docs across all shards:
//sort = new Sort(SortField.FIELD_DOC);
sort = null;
}
else if (what == 2)
{
sort = new Sort(new SortField[] { new SortField("docid", SortField.Type_e.INT, Random().NextBoolean()) });
}
else
{
sort = new Sort(new SortField[] { new SortField("title", SortField.Type_e.STRING, Random().NextBoolean()) });
}
}
}
else
{
query = null;
sort = null;
}
}
if (query != null)
{
try
{
searchState = AssertSame(mockSearcher, localShardSearcher, query, sort, prevSearchState);
}
catch (SearcherExpiredException see)
{
// Expected; in a "real" app we would
// either forward this error to the user ("too
// much time has passed; please re-run your
// search") or sneakily just switch to newest
// searcher w/o telling them...
if (VERBOSE)
{
Console.WriteLine(" searcher expired during search: " + see);
Console.Out.Write(see.StackTrace);
}
// We can't do this in general: on a very slow
// computer it's possible the local searcher
// expires before we can finish our search:
// assert prevSearchState != null;
if (prevSearchState != null)
{
priorSearches.Remove(prevSearchState);
}
}
}
}
finally
{
Nodes[myNodeID].Release(localShardSearcher);
foreach (IndexReader sub in subs)
{
if (sub != null)
{
sub.DecRef();
}
}
}
if (searchState != null && searchState.SearchAfterLocal != null && Random().Next(5) == 3)
{
priorSearches.Add(searchState);
if (priorSearches.Count > 200)
{
priorSearches = (List <PreviousSearchState>)CollectionsHelper.Shuffle(priorSearches);
priorSearches.SubList(100, priorSearches.Count).Clear();
}
}
}
Finish();
}
public virtual void TestMixedVectrosVectors()
{
RandomIndexWriter writer = new RandomIndexWriter(Random(), Directory, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random(), MockTokenizer.SIMPLE, true)).SetOpenMode(OpenMode.CREATE));
Document doc = new Document();
FieldType ft2 = new FieldType(TextField.TYPE_STORED);
ft2.StoreTermVectors = true;
FieldType ft3 = new FieldType(TextField.TYPE_STORED);
ft3.StoreTermVectors = true;
ft3.StoreTermVectorPositions = true;
FieldType ft4 = new FieldType(TextField.TYPE_STORED);
ft4.StoreTermVectors = true;
ft4.StoreTermVectorOffsets = true;
FieldType ft5 = new FieldType(TextField.TYPE_STORED);
ft5.StoreTermVectors = true;
ft5.StoreTermVectorOffsets = true;
ft5.StoreTermVectorPositions = true;
doc.Add(NewTextField("field", "one", Field.Store.YES));
doc.Add(NewField("field", "one", ft2));
doc.Add(NewField("field", "one", ft3));
doc.Add(NewField("field", "one", ft4));
doc.Add(NewField("field", "one", ft5));
writer.AddDocument(doc);
IndexReader reader = writer.Reader;
writer.Dispose();
IndexSearcher searcher = NewSearcher(reader);
Query query = new TermQuery(new Term("field", "one"));
ScoreDoc[] hits = searcher.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(1, hits.Length);
Fields vectors = searcher.IndexReader.GetTermVectors(hits[0].Doc);
Assert.IsNotNull(vectors);
Assert.AreEqual(1, vectors.Size);
Terms vector = vectors.Terms("field");
Assert.IsNotNull(vector);
Assert.AreEqual(1, vector.Size());
TermsEnum termsEnum = vector.Iterator(null);
Assert.IsNotNull(termsEnum.Next());
Assert.AreEqual("one", termsEnum.Term().Utf8ToString());
Assert.AreEqual(5, termsEnum.TotalTermFreq());
DocsAndPositionsEnum dpEnum = termsEnum.DocsAndPositions(null, null);
Assert.IsNotNull(dpEnum);
Assert.IsTrue(dpEnum.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
Assert.AreEqual(5, dpEnum.Freq());
for (int i = 0; i < 5; i++)
{
Assert.AreEqual(i, dpEnum.NextPosition());
}
dpEnum = termsEnum.DocsAndPositions(null, dpEnum);
Assert.IsNotNull(dpEnum);
Assert.IsTrue(dpEnum.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
Assert.AreEqual(5, dpEnum.Freq());
for (int i = 0; i < 5; i++)
{
dpEnum.NextPosition();
Assert.AreEqual(4 * i, dpEnum.StartOffset());
Assert.AreEqual(4 * i + 3, dpEnum.EndOffset());
}
reader.Dispose();
}
/// <summary>
/// Safe (but, slowish) default method to write every
/// vector field in the document.
/// </summary>
protected internal void AddAllDocVectors(Fields vectors, MergeState mergeState)
{
if (vectors == null)
{
StartDocument(0);
FinishDocument();
return;
}
int numFields = vectors.Size;
if (numFields == -1)
{
// count manually! TODO: Maybe enforce that Fields.size() returns something valid?
numFields = 0;
//for (IEnumerator<string> it = vectors.Iterator(); it.hasNext();)
foreach (string it in vectors)
{
numFields++;
}
}
StartDocument(numFields);
string lastFieldName = null;
TermsEnum termsEnum = null;
DocsAndPositionsEnum docsAndPositionsEnum = null;
int fieldCount = 0;
foreach (string fieldName in vectors)
{
fieldCount++;
FieldInfo fieldInfo = mergeState.FieldInfos.FieldInfo(fieldName);
Debug.Assert(lastFieldName == null || fieldName.CompareTo(lastFieldName) > 0, "lastFieldName=" + lastFieldName + " fieldName=" + fieldName);
lastFieldName = fieldName;
Terms terms = vectors.Terms(fieldName);
if (terms == null)
{
// FieldsEnum shouldn't lie...
continue;
}
bool hasPositions = terms.HasPositions();
bool hasOffsets = terms.HasOffsets();
bool hasPayloads = terms.HasPayloads();
Debug.Assert(!hasPayloads || hasPositions);
int numTerms = (int)terms.Size();
if (numTerms == -1)
{
// count manually. It is stupid, but needed, as Terms.size() is not a mandatory statistics function
numTerms = 0;
termsEnum = terms.Iterator(termsEnum);
while (termsEnum.Next() != null)
{
numTerms++;
}
}
StartField(fieldInfo, numTerms, hasPositions, hasOffsets, hasPayloads);
termsEnum = terms.Iterator(termsEnum);
int termCount = 0;
while (termsEnum.Next() != null)
{
termCount++;
int freq = (int)termsEnum.TotalTermFreq();
StartTerm(termsEnum.Term(), freq);
if (hasPositions || hasOffsets)
{
docsAndPositionsEnum = termsEnum.DocsAndPositions(null, docsAndPositionsEnum);
Debug.Assert(docsAndPositionsEnum != null);
int docID = docsAndPositionsEnum.NextDoc();
Debug.Assert(docID != DocIdSetIterator.NO_MORE_DOCS);
Debug.Assert(docsAndPositionsEnum.Freq() == freq);
for (int posUpto = 0; posUpto < freq; posUpto++)
{
int pos = docsAndPositionsEnum.NextPosition();
int startOffset = docsAndPositionsEnum.StartOffset();
int endOffset = docsAndPositionsEnum.EndOffset();
BytesRef payload = docsAndPositionsEnum.Payload;
Debug.Assert(!hasPositions || pos >= 0);
AddPosition(pos, startOffset, endOffset, payload);
}
}
FinishTerm();
}
Debug.Assert(termCount == numTerms);
FinishField();
}
Debug.Assert(fieldCount == numFields);
FinishDocument();
}
/// <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();
}
}
/// <exception cref="System.IO.IOException"></exception>
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)
{
tenum != null.SeekExact(startFacetOrd == -1 ? 0 : startFacetOrd);
mergeTerm = tenum.Term();
}
}