private void VerifyDocFreq() { IndexReader reader = DirectoryReader.Open(Dir); TermsEnum termEnum = MultiFields.GetTerms(reader, "content").Iterator(null); // create enumeration of all terms // go to the first term (aaa) termEnum.Next(); // assert that term is 'aaa' Assert.AreEqual("aaa", termEnum.Term().Utf8ToString()); Assert.AreEqual(200, termEnum.DocFreq()); // go to the second term (bbb) termEnum.Next(); // assert that term is 'bbb' Assert.AreEqual("bbb", termEnum.Term().Utf8ToString()); Assert.AreEqual(100, termEnum.DocFreq()); // create enumeration of terms after term 'aaa', // including 'aaa' termEnum.SeekCeil(new BytesRef("aaa")); // assert that term is 'aaa' Assert.AreEqual("aaa", termEnum.Term().Utf8ToString()); Assert.AreEqual(200, termEnum.DocFreq()); // go to term 'bbb' termEnum.Next(); // assert that term is 'bbb' Assert.AreEqual("bbb", termEnum.Term().Utf8ToString()); Assert.AreEqual(100, termEnum.DocFreq()); reader.Dispose(); }
public override sealed int DocFreq(Term term) { Fields fields = Fields; if (fields == null) { return(0); } Terms terms = fields.Terms(term.Field()); if (terms == null) { return(0); } TermsEnum termsEnum = terms.Iterator(null); if (termsEnum.SeekExact(term.Bytes())) { return(termsEnum.DocFreq()); } else { return(0); } }
private void AssertSumDocFreq(IndexReader ir) { // compute sumDocFreq across all fields Fields fields = MultiFields.GetFields(ir); foreach (string f in fields) { Terms terms = fields.Terms(f); long sumDocFreq = terms.SumDocFreq; if (sumDocFreq == -1) { if (VERBOSE) { Console.WriteLine("skipping field: " + f + ", codec does not support sumDocFreq"); } continue; } long computedSumDocFreq = 0; TermsEnum termsEnum = terms.Iterator(null); while (termsEnum.Next() != null) { computedSumDocFreq += termsEnum.DocFreq(); } Assert.AreEqual(computedSumDocFreq, sumDocFreq); } }
/// <summary> /// Creates a <seealso cref="TermContext"/> from a top-level <seealso cref="IndexReaderContext"/> and the /// given <seealso cref="Term"/>. this method will lookup the given term in all context's leaf readers /// and register each of the readers containing the term in the returned <seealso cref="TermContext"/> /// using the leaf reader's ordinal. /// <p> /// Note: the given context must be a top-level context. /// </summary> public static TermContext Build(IndexReaderContext context, Term term) { Debug.Assert(context != null && context.IsTopLevel); string field = term.Field(); BytesRef bytes = term.Bytes(); TermContext perReaderTermState = new TermContext(context); //if (DEBUG) System.out.println("prts.build term=" + term); foreach (AtomicReaderContext ctx in context.Leaves) { //if (DEBUG) System.out.println(" r=" + leaves[i].reader); Fields fields = ctx.AtomicReader.Fields; if (fields != null) { Terms terms = fields.Terms(field); if (terms != null) { TermsEnum termsEnum = terms.Iterator(null); if (termsEnum.SeekExact(bytes)) { TermState termState = termsEnum.TermState(); //if (DEBUG) System.out.println(" found"); perReaderTermState.Register(termState, ctx.Ord, termsEnum.DocFreq(), termsEnum.TotalTermFreq()); } } } } return(perReaderTermState); }
private void PrintSegment(StreamWriter @out, SegmentCommitInfo si) { SegmentReader reader = new SegmentReader(si, DirectoryReader.DEFAULT_TERMS_INDEX_DIVISOR, NewIOContext(Random())); for (int i = 0; i < reader.NumDocs; i++) { @out.WriteLine(reader.Document(i)); } Fields fields = reader.Fields; foreach (string field in fields) { Terms terms = fields.Terms(field); Assert.IsNotNull(terms); TermsEnum tis = terms.Iterator(null); while (tis.Next() != null) { @out.Write(" term=" + field + ":" + tis.Term()); @out.WriteLine(" DF=" + tis.DocFreq()); DocsAndPositionsEnum positions = tis.DocsAndPositions(reader.LiveDocs, null); while (positions.NextDoc() != DocIdSetIterator.NO_MORE_DOCS) { @out.Write(" doc=" + positions.DocID()); @out.Write(" TF=" + positions.Freq()); @out.Write(" pos="); @out.Write(positions.NextPosition()); for (int j = 1; j < positions.Freq(); j++) { @out.Write("," + positions.NextPosition()); } @out.WriteLine(""); } } } reader.Dispose(); }
/// <summary> /// checks the terms enum sequentially /// if deep is false, it does a 'shallow' test that doesnt go down to the docsenums /// </summary> public void AssertTermsEnumEquals(string info, IndexReader leftReader, TermsEnum leftTermsEnum, TermsEnum rightTermsEnum, bool deep) { BytesRef term; Bits randomBits = new RandomBits(leftReader.MaxDoc, Random().NextDouble(), Random()); DocsAndPositionsEnum leftPositions = null; DocsAndPositionsEnum rightPositions = null; DocsEnum leftDocs = null; DocsEnum rightDocs = null; while ((term = leftTermsEnum.Next()) != null) { Assert.AreEqual(term, rightTermsEnum.Next(), info); AssertTermStatsEquals(info, leftTermsEnum, rightTermsEnum); if (deep) { AssertDocsAndPositionsEnumEquals(info, leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions)); AssertDocsAndPositionsEnumEquals(info, leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions)); AssertPositionsSkippingEquals(info, leftReader, leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions)); AssertPositionsSkippingEquals(info, leftReader, leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions)); // with freqs: AssertDocsEnumEquals(info, leftDocs = leftTermsEnum.Docs(null, leftDocs), rightDocs = rightTermsEnum.Docs(null, rightDocs), true); AssertDocsEnumEquals(info, leftDocs = leftTermsEnum.Docs(randomBits, leftDocs), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs), true); // w/o freqs: AssertDocsEnumEquals(info, leftDocs = leftTermsEnum.Docs(null, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(null, rightDocs, DocsEnum.FLAG_NONE), false); AssertDocsEnumEquals(info, leftDocs = leftTermsEnum.Docs(randomBits, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs, DocsEnum.FLAG_NONE), false); // with freqs: AssertDocsSkippingEquals(info, leftReader, leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(null, leftDocs), rightDocs = rightTermsEnum.Docs(null, rightDocs), true); AssertDocsSkippingEquals(info, leftReader, leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(randomBits, leftDocs), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs), true); // w/o freqs: AssertDocsSkippingEquals(info, leftReader, leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(null, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(null, rightDocs, DocsEnum.FLAG_NONE), false); AssertDocsSkippingEquals(info, leftReader, leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(randomBits, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs, DocsEnum.FLAG_NONE), false); } } Assert.IsNull(rightTermsEnum.Next(), info); }
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); } }
public override int DocFreq() { return(TermsEnum.DocFreq()); }
/// <summary> /// Call this only once (if you subclass!) </summary> protected internal virtual void Uninvert(AtomicReader reader, Bits liveDocs, BytesRef termPrefix) { FieldInfo info = reader.FieldInfos.FieldInfo(Field); if (info != null && info.HasDocValues()) { throw new InvalidOperationException("Type mismatch: " + Field + " was indexed as " + info.DocValuesType); } //System.out.println("DTO uninvert field=" + field + " prefix=" + termPrefix); long startTime = DateTime.Now.Millisecond; Prefix = termPrefix == null ? null : BytesRef.DeepCopyOf(termPrefix); int maxDoc = reader.MaxDoc; int[] index = new int[maxDoc]; // immediate term numbers, or the index into the byte[] representing the last number int[] lastTerm = new int[maxDoc]; // last term we saw for this document sbyte[][] bytes = new sbyte[maxDoc][]; // list of term numbers for the doc (delta encoded vInts) Fields fields = reader.Fields; if (fields == null) { // No terms return; } Terms terms = fields.Terms(Field); if (terms == null) { // No terms return; } TermsEnum te = terms.Iterator(null); BytesRef seekStart = termPrefix != null ? termPrefix : new BytesRef(); //System.out.println("seekStart=" + seekStart.utf8ToString()); if (te.SeekCeil(seekStart) == TermsEnum.SeekStatus.END) { // No terms match return; } // If we need our "term index wrapper", these will be // init'd below: IList <BytesRef> indexedTerms = null; PagedBytes indexedTermsBytes = null; bool testedOrd = false; // we need a minimum of 9 bytes, but round up to 12 since the space would // be wasted with most allocators anyway. sbyte[] tempArr = new sbyte[12]; // // enumerate all terms, and build an intermediate form of the un-inverted field. // // During this intermediate form, every document has a (potential) byte[] // and the int[maxDoc()] array either contains the termNumber list directly // or the *end* offset of the termNumber list in it's byte array (for faster // appending and faster creation of the final form). // // idea... if things are too large while building, we could do a range of docs // at a time (but it would be a fair amount slower to build) // could also do ranges in parallel to take advantage of multiple CPUs // OPTIONAL: remap the largest df terms to the lowest 128 (single byte) // values. this requires going over the field first to find the most // frequent terms ahead of time. int termNum = 0; DocsEnum = null; // Loop begins with te positioned to first term (we call // seek above): for (; ;) { BytesRef t = te.Term(); if (t == null || (termPrefix != null && !StringHelper.StartsWith(t, termPrefix))) { break; } //System.out.println("visit term=" + t.utf8ToString() + " " + t + " termNum=" + termNum); if (!testedOrd) { try { OrdBase = (int)te.Ord(); //System.out.println("got ordBase=" + ordBase); } catch (System.NotSupportedException uoe) { // Reader cannot provide ord support, so we wrap // our own support by creating our own terms index: indexedTerms = new List <BytesRef>(); indexedTermsBytes = new PagedBytes(15); //System.out.println("NO ORDS"); } testedOrd = true; } VisitTerm(te, termNum); if (indexedTerms != null && (termNum & IndexIntervalMask) == 0) { // Index this term SizeOfIndexedStrings += t.Length; BytesRef indexedTerm = new BytesRef(); indexedTermsBytes.Copy(t, indexedTerm); // TODO: really should 1) strip off useless suffix, // and 2) use FST not array/PagedBytes indexedTerms.Add(indexedTerm); } int df = te.DocFreq(); if (df <= MaxTermDocFreq) { DocsEnum = te.Docs(liveDocs, DocsEnum, DocsEnum.FLAG_NONE); // dF, but takes deletions into account int actualDF = 0; for (; ;) { int doc = DocsEnum.NextDoc(); if (doc == DocIdSetIterator.NO_MORE_DOCS) { break; } //System.out.println(" chunk=" + chunk + " docs"); actualDF++; TermInstances++; //System.out.println(" docID=" + doc); // add TNUM_OFFSET to the term number to make room for special reserved values: // 0 (end term) and 1 (index into byte array follows) int delta = termNum - lastTerm[doc] + TNUM_OFFSET; lastTerm[doc] = termNum; int val = index[doc]; if ((val & 0xff) == 1) { // index into byte array (actually the end of // the doc-specific byte[] when building) int pos = (int)((uint)val >> 8); int ilen = VIntSize(delta); sbyte[] arr = bytes[doc]; int newend = pos + ilen; if (newend > arr.Length) { // We avoid a doubling strategy to lower memory usage. // this faceting method isn't for docs with many terms. // In hotspot, objects have 2 words of overhead, then fields, rounded up to a 64-bit boundary. // TODO: figure out what array lengths we can round up to w/o actually using more memory // (how much space does a byte[] take up? Is data preceded by a 32 bit length only? // It should be safe to round up to the nearest 32 bits in any case. int newLen = (newend + 3) & unchecked ((int)0xfffffffc); // 4 byte alignment sbyte[] newarr = new sbyte[newLen]; Array.Copy(arr, 0, newarr, 0, pos); arr = newarr; bytes[doc] = newarr; } pos = WriteInt(delta, arr, pos); index[doc] = (pos << 8) | 1; // update pointer to end index in byte[] } else { // OK, this int has data in it... find the end (a zero starting byte - not // part of another number, hence not following a byte with the high bit set). int ipos; if (val == 0) { ipos = 0; } else if ((val & 0x0000ff80) == 0) { ipos = 1; } else if ((val & 0x00ff8000) == 0) { ipos = 2; } else if ((val & 0xff800000) == 0) { ipos = 3; } else { ipos = 4; } //System.out.println(" ipos=" + ipos); int endPos = WriteInt(delta, tempArr, ipos); //System.out.println(" endpos=" + endPos); if (endPos <= 4) { //System.out.println(" fits!"); // value will fit in the integer... move bytes back for (int j = ipos; j < endPos; j++) { val |= (tempArr[j] & 0xff) << (j << 3); } index[doc] = val; } else { // value won't fit... move integer into byte[] for (int j = 0; j < ipos; j++) { tempArr[j] = (sbyte)val; val = (int)((uint)val >> 8); } // point at the end index in the byte[] index[doc] = (endPos << 8) | 1; bytes[doc] = tempArr; tempArr = new sbyte[12]; } } } SetActualDocFreq(termNum, actualDF); } termNum++; if (te.Next() == null) { break; } } NumTermsInField = termNum; long midPoint = DateTime.Now.Millisecond; if (TermInstances == 0) { // we didn't invert anything // lower memory consumption. Tnums = null; } else { this.Index = index; // // transform intermediate form into the final form, building a single byte[] // at a time, and releasing the intermediate byte[]s as we go to avoid // increasing the memory footprint. // for (int pass = 0; pass < 256; pass++) { sbyte[] target = Tnums[pass]; int pos = 0; // end in target; if (target != null) { pos = target.Length; } else { target = new sbyte[4096]; } // loop over documents, 0x00ppxxxx, 0x01ppxxxx, 0x02ppxxxx // where pp is the pass (which array we are building), and xx is all values. // each pass shares the same byte[] for termNumber lists. for (int docbase = pass << 16; docbase < maxDoc; docbase += (1 << 24)) { int lim = Math.Min(docbase + (1 << 16), maxDoc); for (int doc = docbase; doc < lim; doc++) { //System.out.println(" pass="******" process docID=" + doc); int val = index[doc]; if ((val & 0xff) == 1) { int len = (int)((uint)val >> 8); //System.out.println(" ptr pos=" + pos); index[doc] = (pos << 8) | 1; // change index to point to start of array if ((pos & 0xff000000) != 0) { // we only have 24 bits for the array index throw new InvalidOperationException("Too many values for UnInvertedField faceting on field " + Field); } sbyte[] arr = bytes[doc]; /* * for(byte b : arr) { * //System.out.println(" b=" + Integer.toHexString((int) b)); * } */ bytes[doc] = null; // IMPORTANT: allow GC to avoid OOM if (target.Length <= pos + len) { int newlen = target.Length; /// <summary> ///* we don't have to worry about the array getting too large /// since the "pos" param will overflow first (only 24 bits available) /// if ((newlen<<1) <= 0) { /// // overflow... /// newlen = Integer.MAX_VALUE; /// if (newlen <= pos + len) { /// throw new SolrException(400,"Too many terms to uninvert field!"); /// } /// } else { /// while (newlen <= pos + len) newlen<<=1; // doubling strategy /// } /// *** /// </summary> while (newlen <= pos + len) // doubling strategy { newlen <<= 1; } sbyte[] newtarget = new sbyte[newlen]; Array.Copy(target, 0, newtarget, 0, pos); target = newtarget; } Array.Copy(arr, 0, target, pos, len); pos += len + 1; // skip single byte at end and leave it 0 for terminator } } } // shrink array if (pos < target.Length) { sbyte[] newtarget = new sbyte[pos]; Array.Copy(target, 0, newtarget, 0, pos); target = newtarget; } Tnums[pass] = target; if ((pass << 16) > maxDoc) { break; } } } if (indexedTerms != null) { IndexedTermsArray = indexedTerms.ToArray(); } long endTime = DateTime.Now.Millisecond; Total_time = (int)(endTime - startTime); Phase1_time = (int)(midPoint - startTime); }
/// <summary> /// checks term-level statistics /// </summary> public virtual void AssertTermStats(TermsEnum leftTermsEnum, TermsEnum rightTermsEnum) { Assert.AreEqual(leftTermsEnum.DocFreq(), rightTermsEnum.DocFreq()); if (leftTermsEnum.TotalTermFreq() != -1 && rightTermsEnum.TotalTermFreq() != -1) { Assert.AreEqual(leftTermsEnum.TotalTermFreq(), rightTermsEnum.TotalTermFreq()); } }
public virtual void TestIntersectRandom() { Directory dir = NewDirectory(); RandomIndexWriter w = new RandomIndexWriter(Random(), dir, Similarity, TimeZone); int numTerms = AtLeast(300); //final int numTerms = 50; HashSet <string> terms = new HashSet <string>(); ICollection <string> pendingTerms = new List <string>(); IDictionary <BytesRef, int?> termToID = new Dictionary <BytesRef, int?>(); int id = 0; while (terms.Count != numTerms) { string s = RandomString; if (!terms.Contains(s)) { terms.Add(s); pendingTerms.Add(s); if (Random().Next(20) == 7) { AddDoc(w, pendingTerms, termToID, id++); } } } AddDoc(w, pendingTerms, termToID, id++); BytesRef[] termsArray = new BytesRef[terms.Count]; HashSet <BytesRef> termsSet = new HashSet <BytesRef>(); { int upto = 0; foreach (string s in terms) { BytesRef b = new BytesRef(s); termsArray[upto++] = b; termsSet.Add(b); } Array.Sort(termsArray); } if (VERBOSE) { Console.WriteLine("\nTEST: indexed terms (unicode order):"); foreach (BytesRef t in termsArray) { Console.WriteLine(" " + t.Utf8ToString() + " -> id:" + termToID[t]); } } IndexReader r = w.Reader; w.Dispose(); // NOTE: intentional insanity!! FieldCache.Ints docIDToID = FieldCache.DEFAULT.GetInts(SlowCompositeReaderWrapper.Wrap(r), "id", false); for (int iter = 0; iter < 10 * RANDOM_MULTIPLIER; iter++) { // TODO: can we also test infinite As here...? // From the random terms, pick some ratio and compile an // automaton: HashSet <string> acceptTerms = new HashSet <string>(); SortedSet <BytesRef> sortedAcceptTerms = new SortedSet <BytesRef>(); double keepPct = Random().NextDouble(); Automaton a; if (iter == 0) { if (VERBOSE) { Console.WriteLine("\nTEST: empty automaton"); } a = BasicAutomata.MakeEmpty(); } else { if (VERBOSE) { Console.WriteLine("\nTEST: keepPct=" + keepPct); } foreach (string s in terms) { string s2; if (Random().NextDouble() <= keepPct) { s2 = s; } else { s2 = RandomString; } acceptTerms.Add(s2); sortedAcceptTerms.Add(new BytesRef(s2)); } a = BasicAutomata.MakeStringUnion(sortedAcceptTerms); } if (Random().NextBoolean()) { if (VERBOSE) { Console.WriteLine("TEST: reduce the automaton"); } a.Reduce(); } CompiledAutomaton c = new CompiledAutomaton(a, true, false); BytesRef[] acceptTermsArray = new BytesRef[acceptTerms.Count]; HashSet <BytesRef> acceptTermsSet = new HashSet <BytesRef>(); int upto = 0; foreach (string s in acceptTerms) { BytesRef b = new BytesRef(s); acceptTermsArray[upto++] = b; acceptTermsSet.Add(b); Assert.IsTrue(Accepts(c, b)); } Array.Sort(acceptTermsArray); if (VERBOSE) { Console.WriteLine("\nTEST: accept terms (unicode order):"); foreach (BytesRef t in acceptTermsArray) { Console.WriteLine(" " + t.Utf8ToString() + (termsSet.Contains(t) ? " (exists)" : "")); } Console.WriteLine(a.ToDot()); } for (int iter2 = 0; iter2 < 100; iter2++) { BytesRef startTerm = acceptTermsArray.Length == 0 || Random().NextBoolean() ? null : acceptTermsArray[Random().Next(acceptTermsArray.Length)]; if (VERBOSE) { Console.WriteLine("\nTEST: iter2=" + iter2 + " startTerm=" + (startTerm == null ? "<null>" : startTerm.Utf8ToString())); if (startTerm != null) { int state = c.RunAutomaton.InitialState; for (int idx = 0; idx < startTerm.Length; idx++) { int label = startTerm.Bytes[startTerm.Offset + idx] & 0xff; Console.WriteLine(" state=" + state + " label=" + label); state = c.RunAutomaton.Step(state, label); Assert.IsTrue(state != -1); } Console.WriteLine(" state=" + state); } } TermsEnum te = MultiFields.GetTerms(r, "f").Intersect(c, startTerm); int loc; if (startTerm == null) { loc = 0; } else { loc = Array.BinarySearch(termsArray, BytesRef.DeepCopyOf(startTerm)); if (loc < 0) { loc = -(loc + 1); } else { // startTerm exists in index loc++; } } while (loc < termsArray.Length && !acceptTermsSet.Contains(termsArray[loc])) { loc++; } DocsEnum docsEnum = null; while (loc < termsArray.Length) { BytesRef expected = termsArray[loc]; BytesRef actual = te.Next(); if (VERBOSE) { Console.WriteLine("TEST: next() expected=" + expected.Utf8ToString() + " actual=" + (actual == null ? "null" : actual.Utf8ToString())); } Assert.AreEqual(expected, actual); Assert.AreEqual(1, te.DocFreq()); docsEnum = TestUtil.Docs(Random(), te, null, docsEnum, DocsEnum.FLAG_NONE); int docID = docsEnum.NextDoc(); Assert.IsTrue(docID != DocIdSetIterator.NO_MORE_DOCS); Assert.AreEqual(docIDToID.Get(docID), (int)termToID[expected]); do { loc++; } while (loc < termsArray.Length && !acceptTermsSet.Contains(termsArray[loc])); } Assert.IsNull(te.Next()); } } r.Dispose(); dir.Dispose(); }
public virtual void Test() { IList <string> postingsList = new List <string>(); int numTerms = AtLeast(300); int maxTermsPerDoc = TestUtil.NextInt(Random(), 10, 20); bool isSimpleText = "SimpleText".Equals(TestUtil.GetPostingsFormat("field")); IndexWriterConfig iwc = NewIndexWriterConfig(Random(), TEST_VERSION_CURRENT, new MockAnalyzer(Random())); if ((isSimpleText || iwc.MergePolicy is MockRandomMergePolicy) && (TEST_NIGHTLY || RANDOM_MULTIPLIER > 1)) { // Otherwise test can take way too long (> 2 hours) numTerms /= 2; } if (VERBOSE) { Console.WriteLine("maxTermsPerDoc=" + maxTermsPerDoc); Console.WriteLine("numTerms=" + numTerms); } for (int i = 0; i < numTerms; i++) { string term = Convert.ToString(i); for (int j = 0; j < i; j++) { postingsList.Add(term); } } postingsList = CollectionsHelper.Shuffle(postingsList); ConcurrentQueue <string> postings = new ConcurrentQueue <string>(postingsList); Directory dir = NewFSDirectory(CreateTempDir("bagofpostings")); RandomIndexWriter iw = new RandomIndexWriter(Random(), dir, iwc); int threadCount = TestUtil.NextInt(Random(), 1, 5); if (VERBOSE) { Console.WriteLine("config: " + iw.w.Config); Console.WriteLine("threadCount=" + threadCount); } ThreadClass[] threads = new ThreadClass[threadCount]; CountdownEvent startingGun = new CountdownEvent(1); for (int threadID = 0; threadID < threadCount; threadID++) { threads[threadID] = new ThreadAnonymousInnerClassHelper(this, maxTermsPerDoc, postings, iw, startingGun); threads[threadID].Start(); } startingGun.Signal(); foreach (ThreadClass t in threads) { t.Join(); } iw.ForceMerge(1); DirectoryReader ir = iw.Reader; Assert.AreEqual(1, ir.Leaves.Count); AtomicReader air = (AtomicReader)ir.Leaves[0].Reader; Terms terms = air.Terms("field"); // numTerms-1 because there cannot be a term 0 with 0 postings: Assert.AreEqual(numTerms - 1, air.Fields.UniqueTermCount); if (iwc.Codec is Lucene3xCodec == false) { Assert.AreEqual(numTerms - 1, terms.Size()); } TermsEnum termsEnum = terms.Iterator(null); BytesRef term_; while ((term_ = termsEnum.Next()) != null) { int value = Convert.ToInt32(term_.Utf8ToString()); Assert.AreEqual(value, termsEnum.DocFreq()); // don't really need to check more than this, as CheckIndex // will verify that docFreq == actual number of documents seen // from a docsAndPositionsEnum. } ir.Dispose(); iw.Dispose(); dir.Dispose(); }
private void VerifyEnum(ThreadState threadState, string field, BytesRef term, TermsEnum termsEnum, FieldInfo.IndexOptions maxTestOptions, FieldInfo.IndexOptions maxIndexOptions, ISet<Option> options, bool alwaysTestMax) // Maximum options (docs/freqs/positions/offsets) to test: { if (VERBOSE) { Console.WriteLine(" verifyEnum: options=" + options + " maxTestOptions=" + maxTestOptions); } // Make sure TermsEnum really is positioned on the // expected term: Assert.AreEqual(term, termsEnum.Term()); // 50% of the time time pass liveDocs: bool useLiveDocs = options.Contains(Option.LIVE_DOCS) && Random().NextBoolean(); Bits liveDocs; if (useLiveDocs) { liveDocs = GlobalLiveDocs; if (VERBOSE) { Console.WriteLine(" use liveDocs"); } } else { liveDocs = null; if (VERBOSE) { Console.WriteLine(" no liveDocs"); } } FieldInfo fieldInfo = CurrentFieldInfos.FieldInfo(field); // NOTE: can be empty list if we are using liveDocs: SeedPostings expected = GetSeedPostings(term.Utf8ToString(), Fields[field][term], useLiveDocs, maxIndexOptions); Assert.AreEqual(expected.DocFreq, termsEnum.DocFreq()); bool allowFreqs = fieldInfo.FieldIndexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS && maxTestOptions.CompareTo(FieldInfo.IndexOptions.DOCS_AND_FREQS) >= 0; bool doCheckFreqs = allowFreqs && (alwaysTestMax || Random().Next(3) <= 2); bool allowPositions = fieldInfo.FieldIndexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS && maxTestOptions.CompareTo(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0; bool doCheckPositions = allowPositions && (alwaysTestMax || Random().Next(3) <= 2); bool allowOffsets = fieldInfo.FieldIndexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS && maxTestOptions.CompareTo(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) >= 0; bool doCheckOffsets = allowOffsets && (alwaysTestMax || Random().Next(3) <= 2); bool doCheckPayloads = options.Contains(Option.PAYLOADS) && allowPositions && fieldInfo.HasPayloads() && (alwaysTestMax || Random().Next(3) <= 2); DocsEnum prevDocsEnum = null; DocsEnum docsEnum; DocsAndPositionsEnum docsAndPositionsEnum; if (!doCheckPositions) { if (allowPositions && Random().Next(10) == 7) { // 10% of the time, even though we will not check positions, pull a DocsAndPositions enum if (options.Contains(Option.REUSE_ENUMS) && Random().Next(10) < 9) { prevDocsEnum = threadState.ReuseDocsAndPositionsEnum; } int flags = 0; if (alwaysTestMax || Random().NextBoolean()) { flags |= DocsAndPositionsEnum.FLAG_OFFSETS; } if (alwaysTestMax || Random().NextBoolean()) { flags |= DocsAndPositionsEnum.FLAG_PAYLOADS; } if (VERBOSE) { Console.WriteLine(" get DocsAndPositionsEnum (but we won't check positions) flags=" + flags); } threadState.ReuseDocsAndPositionsEnum = termsEnum.DocsAndPositions(liveDocs, (DocsAndPositionsEnum)prevDocsEnum, flags); docsEnum = threadState.ReuseDocsAndPositionsEnum; docsAndPositionsEnum = threadState.ReuseDocsAndPositionsEnum; } else { if (VERBOSE) { Console.WriteLine(" get DocsEnum"); } if (options.Contains(Option.REUSE_ENUMS) && Random().Next(10) < 9) { prevDocsEnum = threadState.ReuseDocsEnum; } threadState.ReuseDocsEnum = termsEnum.Docs(liveDocs, prevDocsEnum, doCheckFreqs ? DocsEnum.FLAG_FREQS : DocsEnum.FLAG_NONE); docsEnum = threadState.ReuseDocsEnum; docsAndPositionsEnum = null; } } else { if (options.Contains(Option.REUSE_ENUMS) && Random().Next(10) < 9) { prevDocsEnum = threadState.ReuseDocsAndPositionsEnum; } int flags = 0; if (alwaysTestMax || doCheckOffsets || Random().Next(3) == 1) { flags |= DocsAndPositionsEnum.FLAG_OFFSETS; } if (alwaysTestMax || doCheckPayloads || Random().Next(3) == 1) { flags |= DocsAndPositionsEnum.FLAG_PAYLOADS; } if (VERBOSE) { Console.WriteLine(" get DocsAndPositionsEnum flags=" + flags); } threadState.ReuseDocsAndPositionsEnum = termsEnum.DocsAndPositions(liveDocs, (DocsAndPositionsEnum)prevDocsEnum, flags); docsEnum = threadState.ReuseDocsAndPositionsEnum; docsAndPositionsEnum = threadState.ReuseDocsAndPositionsEnum; } Assert.IsNotNull(docsEnum, "null DocsEnum"); int initialDocID = docsEnum.DocID(); Assert.AreEqual(-1, initialDocID, "inital docID should be -1" + docsEnum); if (VERBOSE) { if (prevDocsEnum == null) { Console.WriteLine(" got enum=" + docsEnum); } else if (prevDocsEnum == docsEnum) { Console.WriteLine(" got reuse enum=" + docsEnum); } else { Console.WriteLine(" got enum=" + docsEnum + " (reuse of " + prevDocsEnum + " failed)"); } } // 10% of the time don't consume all docs: int stopAt; if (!alwaysTestMax && options.Contains(Option.PARTIAL_DOC_CONSUME) && expected.DocFreq > 1 && Random().Next(10) == 7) { stopAt = Random().Next(expected.DocFreq - 1); if (VERBOSE) { Console.WriteLine(" will not consume all docs (" + stopAt + " vs " + expected.DocFreq + ")"); } } else { stopAt = expected.DocFreq; if (VERBOSE) { Console.WriteLine(" consume all docs"); } } double skipChance = alwaysTestMax ? 0.5 : Random().NextDouble(); int numSkips = expected.DocFreq < 3 ? 1 : TestUtil.NextInt(Random(), 1, Math.Min(20, expected.DocFreq / 3)); int skipInc = expected.DocFreq / numSkips; int skipDocInc = MaxDoc / numSkips; // Sometimes do 100% skipping: bool doAllSkipping = options.Contains(Option.SKIPPING) && Random().Next(7) == 1; double freqAskChance = alwaysTestMax ? 1.0 : Random().NextDouble(); double payloadCheckChance = alwaysTestMax ? 1.0 : Random().NextDouble(); double offsetCheckChance = alwaysTestMax ? 1.0 : Random().NextDouble(); if (VERBOSE) { if (options.Contains(Option.SKIPPING)) { Console.WriteLine(" skipChance=" + skipChance + " numSkips=" + numSkips); } else { Console.WriteLine(" no skipping"); } if (doCheckFreqs) { Console.WriteLine(" freqAskChance=" + freqAskChance); } if (doCheckPayloads) { Console.WriteLine(" payloadCheckChance=" + payloadCheckChance); } if (doCheckOffsets) { Console.WriteLine(" offsetCheckChance=" + offsetCheckChance); } } while (expected.Upto <= stopAt) { if (expected.Upto == stopAt) { if (stopAt == expected.DocFreq) { Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.NextDoc(), "DocsEnum should have ended but didn't"); // Common bug is to forget to set this.Doc=NO_MORE_DOCS in the enum!: Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.DocID(), "DocsEnum should have ended but didn't"); } break; } if (options.Contains(Option.SKIPPING) && (doAllSkipping || Random().NextDouble() <= skipChance)) { int targetDocID = -1; if (expected.Upto < stopAt && Random().NextBoolean()) { // Pick target we know exists: int skipCount = TestUtil.NextInt(Random(), 1, skipInc); for (int skip = 0; skip < skipCount; skip++) { if (expected.NextDoc() == DocsEnum.NO_MORE_DOCS) { break; } } } else { // Pick random target (might not exist): int skipDocIDs = TestUtil.NextInt(Random(), 1, skipDocInc); if (skipDocIDs > 0) { targetDocID = expected.DocID() + skipDocIDs; expected.Advance(targetDocID); } } if (expected.Upto >= stopAt) { int target = Random().NextBoolean() ? MaxDoc : DocsEnum.NO_MORE_DOCS; if (VERBOSE) { Console.WriteLine(" now advance to end (target=" + target + ")"); } Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.Advance(target), "DocsEnum should have ended but didn't"); break; } else { if (VERBOSE) { if (targetDocID != -1) { Console.WriteLine(" now advance to random target=" + targetDocID + " (" + expected.Upto + " of " + stopAt + ") current=" + docsEnum.DocID()); } else { Console.WriteLine(" now advance to known-exists target=" + expected.DocID() + " (" + expected.Upto + " of " + stopAt + ") current=" + docsEnum.DocID()); } } int docID = docsEnum.Advance(targetDocID != -1 ? targetDocID : expected.DocID()); Assert.AreEqual(expected.DocID(), docID, "docID is wrong"); } } else { expected.NextDoc(); if (VERBOSE) { Console.WriteLine(" now nextDoc to " + expected.DocID() + " (" + expected.Upto + " of " + stopAt + ")"); } int docID = docsEnum.NextDoc(); Assert.AreEqual(expected.DocID(), docID, "docID is wrong"); if (docID == DocsEnum.NO_MORE_DOCS) { break; } } if (doCheckFreqs && Random().NextDouble() <= freqAskChance) { if (VERBOSE) { Console.WriteLine(" now freq()=" + expected.Freq()); } int freq = docsEnum.Freq(); Assert.AreEqual(expected.Freq(), freq, "freq is wrong"); } if (doCheckPositions) { int freq = docsEnum.Freq(); int numPosToConsume; if (!alwaysTestMax && options.Contains(Option.PARTIAL_POS_CONSUME) && Random().Next(5) == 1) { numPosToConsume = Random().Next(freq); } else { numPosToConsume = freq; } for (int i = 0; i < numPosToConsume; i++) { int pos = expected.NextPosition(); if (VERBOSE) { Console.WriteLine(" now nextPosition to " + pos); } Assert.AreEqual(pos, docsAndPositionsEnum.NextPosition(), "position is wrong"); if (doCheckPayloads) { BytesRef expectedPayload = expected.Payload; if (Random().NextDouble() <= payloadCheckChance) { if (VERBOSE) { Console.WriteLine(" now check expectedPayload length=" + (expectedPayload == null ? 0 : expectedPayload.Length)); } if (expectedPayload == null || expectedPayload.Length == 0) { Assert.IsNull(docsAndPositionsEnum.Payload, "should not have payload"); } else { BytesRef payload = docsAndPositionsEnum.Payload; Assert.IsNotNull(payload, "should have payload but doesn't"); Assert.AreEqual(expectedPayload.Length, payload.Length, "payload length is wrong"); for (int byteUpto = 0; byteUpto < expectedPayload.Length; byteUpto++) { Assert.AreEqual(expectedPayload.Bytes[expectedPayload.Offset + byteUpto], payload.Bytes[payload.Offset + byteUpto], "payload bytes are wrong"); } // make a deep copy payload = BytesRef.DeepCopyOf(payload); Assert.AreEqual(payload, docsAndPositionsEnum.Payload, "2nd call to getPayload returns something different!"); } } else { if (VERBOSE) { Console.WriteLine(" skip check payload length=" + (expectedPayload == null ? 0 : expectedPayload.Length)); } } } if (doCheckOffsets) { if (Random().NextDouble() <= offsetCheckChance) { if (VERBOSE) { Console.WriteLine(" now check offsets: startOff=" + expected.StartOffset() + " endOffset=" + expected.EndOffset()); } Assert.AreEqual(expected.StartOffset(), docsAndPositionsEnum.StartOffset(), "startOffset is wrong"); Assert.AreEqual(expected.EndOffset(), docsAndPositionsEnum.EndOffset(), "endOffset is wrong"); } else { if (VERBOSE) { Console.WriteLine(" skip check offsets"); } } } else if (fieldInfo.FieldIndexOptions < FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) { if (VERBOSE) { Console.WriteLine(" now check offsets are -1"); } Assert.AreEqual(-1, docsAndPositionsEnum.StartOffset(), "startOffset isn't -1"); Assert.AreEqual(-1, docsAndPositionsEnum.EndOffset(), "endOffset isn't -1"); } } } } }
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.Terms(idField); if (terms1 == null) { Assert.IsTrue(MultiFields.GetFields(r2) == null || MultiFields.GetFields(r2).Terms(idField) == null); return; } TermsEnum termsEnum = terms1.Iterator(null); Bits liveDocs1 = MultiFields.GetLiveDocs(r1); Bits 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): Bits liveDocs = MultiFields.GetLiveDocs(r1); DocsEnum docs = null; while (termsEnum.Next() != null) { docs = TestUtil.Docs(Random(), termsEnum, liveDocs, docs, DocsEnum.FLAG_NONE); while (docs.NextDoc() != DocIdSetIterator.NO_MORE_DOCS) { Assert.Fail("r1 is not empty but r2 is"); } } return; } Terms terms2 = fields.Terms(idField); TermsEnum termsEnum2 = terms2.Iterator(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, DocsEnum.FLAG_NONE); if (termsEnum2.SeekExact(term)) { termDocs2 = TestUtil.Docs(Random(), termsEnum2, liveDocs2, termDocs2, DocsEnum.FLAG_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.Terms(field); Assert.IsNotNull(terms3); TermsEnum termsEnum3 = terms3.Iterator(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, DocsEnum.FLAG_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.Terms(field); Assert.IsNotNull(terms3); TermsEnum termsEnum3 = terms3.Iterator(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, DocsEnum.FLAG_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.Terms(field1); if (terms == null) { continue; } termsEnum1 = terms.Iterator(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, DocsEnum.FLAG_FREQS); while (docs1.NextDoc() != DocIdSetIterator.NO_MORE_DOCS) { int d = docs1.DocID(); int f = docs1.Freq(); info1[len1] = (((long)d) << 32) | 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.Terms(field2); if (terms == null) { continue; } termsEnum2 = terms.Iterator(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, DocsEnum.FLAG_FREQS); while (docs2.NextDoc() != DocIdSetIterator.NO_MORE_DOCS) { int d = r2r1[docs2.DocID()]; int f = docs2.Freq(); info2[len2] = (((long)d) << 32) | 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()); } } }
public override int DocFreq() { return(@in.DocFreq()); }
/// <summary> /// checks term-level statistics /// </summary> public void AssertTermStatsEquals(string info, TermsEnum leftTermsEnum, TermsEnum rightTermsEnum) { Assert.AreEqual(leftTermsEnum.DocFreq(), rightTermsEnum.DocFreq(), info); if (leftTermsEnum.TotalTermFreq() != -1 && rightTermsEnum.TotalTermFreq() != -1) { Assert.AreEqual(leftTermsEnum.TotalTermFreq(), rightTermsEnum.TotalTermFreq(), info); } }
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.Size()); 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.Comparator); TermsEnum termsEnum = terms.Iterator(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.Payload)) { foundPayload = true; break; } } Assert.IsTrue(foundPayload); } } try { docsAndPositionsEnum.NextPosition(); Assert.Fail(); } catch (Exception e) { // 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> /// checks the terms enum sequentially /// if deep is false, it does a 'shallow' test that doesnt go down to the docsenums /// </summary> public virtual void AssertTermsEnum(TermsEnum leftTermsEnum, TermsEnum rightTermsEnum, bool deep) { BytesRef term; Bits randomBits = new RandomBits(MAXDOC, Random().NextDouble(), Random()); DocsAndPositionsEnum leftPositions = null; DocsAndPositionsEnum rightPositions = null; DocsEnum leftDocs = null; DocsEnum rightDocs = null; while ((term = leftTermsEnum.Next()) != null) { Assert.AreEqual(term, rightTermsEnum.Next()); AssertTermStats(leftTermsEnum, rightTermsEnum); if (deep) { // with payloads + off AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions)); AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions)); // with payloads only AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions, DocsAndPositionsEnum.FLAG_PAYLOADS), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions, DocsAndPositionsEnum.FLAG_PAYLOADS)); AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions, DocsAndPositionsEnum.FLAG_PAYLOADS), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions, DocsAndPositionsEnum.FLAG_PAYLOADS)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions, DocsAndPositionsEnum.FLAG_PAYLOADS), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions, DocsAndPositionsEnum.FLAG_PAYLOADS)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions, DocsAndPositionsEnum.FLAG_PAYLOADS), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions, DocsAndPositionsEnum.FLAG_PAYLOADS)); // with offsets only AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions, DocsAndPositionsEnum.FLAG_OFFSETS), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions, DocsAndPositionsEnum.FLAG_OFFSETS)); AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions, DocsAndPositionsEnum.FLAG_OFFSETS), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions, DocsAndPositionsEnum.FLAG_OFFSETS)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions, DocsAndPositionsEnum.FLAG_OFFSETS), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions, DocsAndPositionsEnum.FLAG_OFFSETS)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions, DocsAndPositionsEnum.FLAG_OFFSETS), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions, DocsAndPositionsEnum.FLAG_OFFSETS)); // with positions only AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions, DocsEnum.FLAG_NONE), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions, DocsEnum.FLAG_NONE)); AssertDocsAndPositionsEnum(leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions, DocsEnum.FLAG_NONE), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions, DocsEnum.FLAG_NONE)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(null, leftPositions, DocsEnum.FLAG_NONE), rightPositions = rightTermsEnum.DocsAndPositions(null, rightPositions, DocsEnum.FLAG_NONE)); AssertPositionsSkipping(leftTermsEnum.DocFreq(), leftPositions = leftTermsEnum.DocsAndPositions(randomBits, leftPositions, DocsEnum.FLAG_NONE), rightPositions = rightTermsEnum.DocsAndPositions(randomBits, rightPositions, DocsEnum.FLAG_NONE)); // with freqs: AssertDocsEnum(leftDocs = leftTermsEnum.Docs(null, leftDocs), rightDocs = rightTermsEnum.Docs(null, rightDocs)); AssertDocsEnum(leftDocs = leftTermsEnum.Docs(randomBits, leftDocs), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs)); // w/o freqs: AssertDocsEnum(leftDocs = leftTermsEnum.Docs(null, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(null, rightDocs, DocsEnum.FLAG_NONE)); AssertDocsEnum(leftDocs = leftTermsEnum.Docs(randomBits, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs, DocsEnum.FLAG_NONE)); // with freqs: AssertDocsSkipping(leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(null, leftDocs), rightDocs = rightTermsEnum.Docs(null, rightDocs)); AssertDocsSkipping(leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(randomBits, leftDocs), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs)); // w/o freqs: AssertDocsSkipping(leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(null, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(null, rightDocs, DocsEnum.FLAG_NONE)); AssertDocsSkipping(leftTermsEnum.DocFreq(), leftDocs = leftTermsEnum.Docs(randomBits, leftDocs, DocsEnum.FLAG_NONE), rightDocs = rightTermsEnum.Docs(randomBits, rightDocs, DocsEnum.FLAG_NONE)); } } Assert.IsNull(rightTermsEnum.Next()); }