public abstract Idf ( int docFreq, int numDocs ) : float | ||
docFreq | int | the number of documents which contain the term /// |
numDocs | int | the total number of documents in the collection /// |
return | float |
public PhraseWeight(PhraseQuery enclosingInstance, Searcher searcher) { InitBlock(enclosingInstance); this.similarity = Enclosing_Instance.GetSimilarity(searcher); idf = similarity.Idf(Enclosing_Instance.terms, searcher); }
public MultiPhraseWeight(MultiPhraseQuery enclosingInstance, Searcher searcher) { InitBlock(enclosingInstance); this.similarity = Enclosing_Instance.GetSimilarity(searcher); // compute idf int maxDoc = searcher.MaxDoc; foreach (Term[] terms in enclosingInstance.termArrays) { foreach (Term term in terms) { idf += similarity.Idf(searcher.DocFreq(term), maxDoc); } } }
public PhraseWeight(PhraseQuery enclosingInstance, Searcher searcher) { InitBlock(enclosingInstance); this.similarity = Enclosing_Instance.GetSimilarity(searcher); idf = similarity.Idf(Enclosing_Instance.terms, searcher); }
public virtual void TestKnownSetOfDocuments() { System.String[] termArray = new System.String[] { "eating", "chocolate", "in", "a", "computer", "lab", "grows", "old", "colored", "with", "an" }; System.String test1 = "eating chocolate in a computer lab"; //6 terms System.String test2 = "computer in a computer lab"; //5 terms System.String test3 = "a chocolate lab grows old"; //5 terms System.String test4 = "eating chocolate with a chocolate lab in an old chocolate colored computer lab"; //13 terms System.Collections.IDictionary test4Map = new System.Collections.Hashtable(); test4Map["chocolate"] = 3; test4Map["lab"] = 2; test4Map["eating"] = 1; test4Map["computer"] = 1; test4Map["with"] = 1; test4Map["a"] = 1; test4Map["colored"] = 1; test4Map["in"] = 1; test4Map["an"] = 1; test4Map["computer"] = 1; test4Map["old"] = 1; Document testDoc1 = new Document(); SetupDoc(testDoc1, test1); Document testDoc2 = new Document(); SetupDoc(testDoc2, test2); Document testDoc3 = new Document(); SetupDoc(testDoc3, test3); Document testDoc4 = new Document(); SetupDoc(testDoc4, test4); Directory dir = new RAMDirectory(); try { IndexWriter writer = new IndexWriter(dir, new SimpleAnalyzer(), true); Assert.IsTrue(writer != null); writer.AddDocument(testDoc1); writer.AddDocument(testDoc2); writer.AddDocument(testDoc3); writer.AddDocument(testDoc4); writer.Close(); IndexSearcher knownSearcher = new IndexSearcher(dir); TermEnum termEnum = knownSearcher.reader.Terms(); TermDocs termDocs = knownSearcher.reader.TermDocs(); //System.out.println("Terms: " + termEnum.size() + " Orig Len: " + termArray.length); Similarity sim = knownSearcher.GetSimilarity(); while (termEnum.Next() == true) { Term term = termEnum.Term(); //System.out.println("Term: " + term); termDocs.Seek(term); while (termDocs.Next()) { int docId = termDocs.Doc(); int freq = termDocs.Freq(); //System.out.println("Doc Id: " + docId + " freq " + freq); TermFreqVector vector = knownSearcher.reader.GetTermFreqVector(docId, "Field"); float tf = sim.Tf(freq); float idf = sim.Idf(term, knownSearcher); //float qNorm = sim.queryNorm() //This is fine since we don't have stop words float lNorm = sim.LengthNorm("Field", vector.GetTerms().Length); //float coord = sim.coord() //System.out.println("TF: " + tf + " IDF: " + idf + " LenNorm: " + lNorm); Assert.IsTrue(vector != null); System.String[] vTerms = vector.GetTerms(); int[] freqs = vector.GetTermFrequencies(); for (int i = 0; i < vTerms.Length; i++) { if (term.Text().Equals(vTerms[i]) == true) { Assert.IsTrue(freqs[i] == freq); } } } //System.out.println("--------"); } Query query = new TermQuery(new Term("Field", "chocolate")); Hits hits = knownSearcher.Search(query); //doc 3 should be the first hit b/c it is the shortest match Assert.IsTrue(hits.Length() == 3); float score = hits.Score(0); /*System.out.println("Hit 0: " + hits.id(0) + " Score: " + hits.score(0) + " String: " + hits.doc(0).toString()); * System.out.println("Explain: " + knownSearcher.explain(query, hits.id(0))); * System.out.println("Hit 1: " + hits.id(1) + " Score: " + hits.score(1) + " String: " + hits.doc(1).toString()); * System.out.println("Explain: " + knownSearcher.explain(query, hits.id(1))); * System.out.println("Hit 2: " + hits.id(2) + " Score: " + hits.score(2) + " String: " + hits.doc(2).toString()); * System.out.println("Explain: " + knownSearcher.explain(query, hits.id(2)));*/ Assert.IsTrue(testDoc3.ToString().Equals(hits.Doc(0).ToString())); Assert.IsTrue(testDoc4.ToString().Equals(hits.Doc(1).ToString())); Assert.IsTrue(testDoc1.ToString().Equals(hits.Doc(2).ToString())); TermFreqVector vector2 = knownSearcher.reader.GetTermFreqVector(hits.Id(1), "Field"); Assert.IsTrue(vector2 != null); //System.out.println("Vector: " + vector); System.String[] terms = vector2.GetTerms(); int[] freqs2 = vector2.GetTermFrequencies(); Assert.IsTrue(terms != null && terms.Length == 10); for (int i = 0; i < terms.Length; i++) { System.String term = terms[i]; //System.out.println("Term: " + term); int freq = freqs2[i]; Assert.IsTrue(test4.IndexOf(term) != -1); System.Int32 freqInt = (System.Int32)test4Map[term]; System.Object tmpFreqInt = test4Map[term]; Assert.IsTrue(tmpFreqInt != null); Assert.IsTrue(freqInt == freq); } knownSearcher.Close(); } catch (System.IO.IOException e) { System.Console.Error.WriteLine(e.StackTrace); Assert.IsTrue(false); } }
private void AddTerms(IndexReader reader, FieldVals f) { if (f.queryString == null) { return; } TokenStream ts = analyzer.TokenStream(f.fieldName, new System.IO.StringReader(f.queryString)); TermAttribute termAtt = (TermAttribute)ts.AddAttribute(typeof(TermAttribute)); int corpusNumDocs = reader.NumDocs(); Term internSavingTemplateTerm = new Term(f.fieldName); //optimization to avoid constructing new Term() objects Hashtable processedTerms = new Hashtable(); while (ts.IncrementToken()) { String term = termAtt.Term(); if (!processedTerms.Contains(term)) { processedTerms.Add(term, term); ScoreTermQueue variantsQ = new ScoreTermQueue(MAX_VARIANTS_PER_TERM); //maxNum variants considered for any one term float minScore = 0; Term startTerm = internSavingTemplateTerm.CreateTerm(term); FuzzyTermEnum fe = new FuzzyTermEnum(reader, startTerm, f.minSimilarity, f.prefixLength); TermEnum origEnum = reader.Terms(startTerm); int df = 0; if (startTerm.Equals(origEnum.Term())) { df = origEnum.DocFreq(); //store the df so all variants use same idf } int numVariants = 0; int totalVariantDocFreqs = 0; do { Term possibleMatch = fe.Term(); if (possibleMatch != null) { numVariants++; totalVariantDocFreqs += fe.DocFreq(); float score = fe.Difference(); if (variantsQ.Size() < MAX_VARIANTS_PER_TERM || score > minScore) { ScoreTerm st = new ScoreTerm(possibleMatch, score, startTerm); variantsQ.Insert(st); minScore = ((ScoreTerm)variantsQ.Top()).score; // maintain minScore } } }while (fe.Next()); if (numVariants > 0) { int avgDf = totalVariantDocFreqs / numVariants; if (df == 0) //no direct match we can use as df for all variants { df = avgDf; //use avg df of all variants } // take the top variants (scored by edit distance) and reset the score // to include an IDF factor then add to the global queue for ranking // overall top query terms int size = variantsQ.Size(); for (int i = 0; i < size; i++) { ScoreTerm st = (ScoreTerm)variantsQ.Pop(); st.score = (st.score * st.score) * sim.Idf(df, corpusNumDocs); q.Insert(st); } } } } }
public TermWeight(TermQuery enclosingInstance, Searcher searcher) { InitBlock(enclosingInstance); this.similarity = Enclosing_Instance.GetSimilarity(searcher); idf = similarity.Idf(Enclosing_Instance.term, searcher); // compute idf }
public MultiPhraseWeight(MultiPhraseQuery enclosingInstance, Searcher searcher) { InitBlock(enclosingInstance); this.similarity = Enclosing_Instance.GetSimilarity(searcher); // compute idf int maxDoc = searcher.MaxDoc; foreach (Term[] terms in enclosingInstance.termArrays) { foreach (Term term in terms) { idf += similarity.Idf(searcher.DocFreq(term), maxDoc); } } }
public virtual void TestKnownSetOfDocuments() { System.String test1 = "eating chocolate in a computer lab"; //6 terms System.String test2 = "computer in a computer lab"; //5 terms System.String test3 = "a chocolate lab grows old"; //5 terms System.String test4 = "eating chocolate with a chocolate lab in an old chocolate colored computer lab"; //13 terms System.Collections.IDictionary test4Map = new System.Collections.Hashtable(); test4Map["chocolate"] = 3; test4Map["lab"] = 2; test4Map["eating"] = 1; test4Map["computer"] = 1; test4Map["with"] = 1; test4Map["a"] = 1; test4Map["colored"] = 1; test4Map["in"] = 1; test4Map["an"] = 1; test4Map["computer"] = 1; test4Map["old"] = 1; Document testDoc1 = new Document(); SetupDoc(testDoc1, test1); Document testDoc2 = new Document(); SetupDoc(testDoc2, test2); Document testDoc3 = new Document(); SetupDoc(testDoc3, test3); Document testDoc4 = new Document(); SetupDoc(testDoc4, test4); Directory dir = new MockRAMDirectory(); try { IndexWriter writer = new IndexWriter(dir, new SimpleAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED); Assert.IsTrue(writer != null); writer.AddDocument(testDoc1); writer.AddDocument(testDoc2); writer.AddDocument(testDoc3); writer.AddDocument(testDoc4); writer.Close(); IndexSearcher knownSearcher = new IndexSearcher(dir); TermEnum termEnum = knownSearcher.reader_ForNUnit.Terms(); TermDocs termDocs = knownSearcher.reader_ForNUnit.TermDocs(); //System.out.println("Terms: " + termEnum.size() + " Orig Len: " + termArray.length); Similarity sim = knownSearcher.GetSimilarity(); while (termEnum.Next() == true) { Term term = termEnum.Term(); //System.out.println("Term: " + term); termDocs.Seek(term); while (termDocs.Next()) { int docId = termDocs.Doc(); int freq = termDocs.Freq(); //System.out.println("Doc Id: " + docId + " freq " + freq); TermFreqVector vector = knownSearcher.reader_ForNUnit.GetTermFreqVector(docId, "field"); float tf = sim.Tf(freq); float idf = sim.Idf(term, knownSearcher); //float qNorm = sim.queryNorm() //This is fine since we don't have stop words float lNorm = sim.LengthNorm("field", vector.GetTerms().Length); //float coord = sim.coord() //System.out.println("TF: " + tf + " IDF: " + idf + " LenNorm: " + lNorm); Assert.IsTrue(vector != null); System.String[] vTerms = vector.GetTerms(); int[] freqs = vector.GetTermFrequencies(); for (int i = 0; i < vTerms.Length; i++) { if (term.Text().Equals(vTerms[i])) { Assert.IsTrue(freqs[i] == freq); } } } //System.out.println("--------"); } Query query = new TermQuery(new Term("field", "chocolate")); ScoreDoc[] hits = knownSearcher.Search(query, null, 1000).scoreDocs; //doc 3 should be the first hit b/c it is the shortest match Assert.IsTrue(hits.Length == 3); float score = hits[0].score; /*System.out.println("Hit 0: " + hits.id(0) + " Score: " + hits.score(0) + " String: " + hits.doc(0).toString()); * System.out.println("Explain: " + knownSearcher.explain(query, hits.id(0))); * System.out.println("Hit 1: " + hits.id(1) + " Score: " + hits.score(1) + " String: " + hits.doc(1).toString()); * System.out.println("Explain: " + knownSearcher.explain(query, hits.id(1))); * System.out.println("Hit 2: " + hits.id(2) + " Score: " + hits.score(2) + " String: " + hits.doc(2).toString()); * System.out.println("Explain: " + knownSearcher.explain(query, hits.id(2)));*/ Assert.IsTrue(hits[0].doc == 2); Assert.IsTrue(hits[1].doc == 3); Assert.IsTrue(hits[2].doc == 0); TermFreqVector vector2 = knownSearcher.reader_ForNUnit.GetTermFreqVector(hits[1].doc, "field"); Assert.IsTrue(vector2 != null); //System.out.println("Vector: " + vector); System.String[] terms = vector2.GetTerms(); int[] freqs2 = vector2.GetTermFrequencies(); Assert.IsTrue(terms != null && terms.Length == 10); for (int i = 0; i < terms.Length; i++) { System.String term = terms[i]; //System.out.println("Term: " + term); int freq = freqs2[i]; Assert.IsTrue(test4.IndexOf(term) != -1); System.Int32 freqInt = -1; try { freqInt = (System.Int32)test4Map[term]; } catch (Exception) { Assert.IsTrue(false); } Assert.IsTrue(freqInt == freq); } SortedTermVectorMapper mapper = new SortedTermVectorMapper(new TermVectorEntryFreqSortedComparator()); knownSearcher.reader_ForNUnit.GetTermFreqVector(hits[1].doc, mapper); System.Collections.Generic.SortedDictionary <object, object> vectorEntrySet = mapper.GetTermVectorEntrySet(); Assert.IsTrue(vectorEntrySet.Count == 10, "mapper.getTermVectorEntrySet() Size: " + vectorEntrySet.Count + " is not: " + 10); TermVectorEntry last = null; foreach (TermVectorEntry tve in vectorEntrySet.Keys) { if (tve != null && last != null) { Assert.IsTrue(last.GetFrequency() >= tve.GetFrequency(), "terms are not properly sorted"); System.Int32 expectedFreq = (System.Int32)test4Map[tve.GetTerm()]; //we expect double the expectedFreq, since there are two fields with the exact same text and we are collapsing all fields Assert.IsTrue(tve.GetFrequency() == 2 * expectedFreq, "Frequency is not correct:"); } last = tve; } FieldSortedTermVectorMapper fieldMapper = new FieldSortedTermVectorMapper(new TermVectorEntryFreqSortedComparator()); knownSearcher.reader_ForNUnit.GetTermFreqVector(hits[1].doc, fieldMapper); System.Collections.IDictionary map = fieldMapper.GetFieldToTerms(); Assert.IsTrue(map.Count == 2, "map Size: " + map.Count + " is not: " + 2); vectorEntrySet = (System.Collections.Generic.SortedDictionary <Object, Object>)map["field"]; Assert.IsTrue(vectorEntrySet != null, "vectorEntrySet is null and it shouldn't be"); Assert.IsTrue(vectorEntrySet.Count == 10, "vectorEntrySet Size: " + vectorEntrySet.Count + " is not: " + 10); knownSearcher.Close(); } catch (System.IO.IOException e) { System.Console.Error.WriteLine(e.StackTrace); Assert.IsTrue(false); } }
public TermWeight(TermQuery enclosingInstance, Searcher searcher) { InitBlock(enclosingInstance); this.similarity = Enclosing_Instance.GetSimilarity(searcher); idf = similarity.Idf(Enclosing_Instance.term, searcher); // compute idf }
public override float Idf(int docFreq, int numDocs) { return(delegee.Idf(docFreq, numDocs)); }