public void OverlappingSpans_Part4()
{
var lm = MakeSpan(0, cc, dd);
var rm = MakeSpan(0, dd, cc);
// Rewriting with SpanNotQueries works!
Console.WriteLine("((cc dd) - (dd cc)) ** (dd cc)");
var not = new SpanNotQuery(lm, rm);
var q4 = MakeSpan(16, not, rm);
TestIndex.AssertQuery(q4, 3);
// This is now implemented int the parser
TestIndex.AssertQuery(Content("cc dd ** dd cc"), 3);
}
public void OverlappingSpans_Part6()
{
// But the reverse ways of excluding spans works
// (dd ee * ee) => (dd ee) * (ee -(dd ee))
var span = MakeSpan(0, dd, ee);
var not = new SpanNotQuery(ee, span);
var q6 = MakeSpan(1, span, not);
TestIndex.AssertQuery(q6, 3);
// The parser now implements this strategy
var q = Content("dd ee * ee");
TestIndex.AssertQuery(Content("dd ee * ee"), 3);
}
public override SpanQuery GetSpanQuery(XmlElement e)
{
XmlElement includeElem = DOMUtils.GetChildByTagOrFail(e, "Include");
includeElem = DOMUtils.GetFirstChildOrFail(includeElem);
XmlElement excludeElem = DOMUtils.GetChildByTagOrFail(e, "Exclude");
excludeElem = DOMUtils.GetFirstChildOrFail(excludeElem);
SpanQuery include = factory.GetSpanQuery(includeElem);
SpanQuery exclude = factory.GetSpanQuery(excludeElem);
SpanNotQuery snq = new SpanNotQuery(include, exclude);
snq.Boost = DOMUtils.GetAttribute(e, "boost", 1.0f);
return(snq);
}
public void TestSpanNot()
{
Directory dir = NewDirectory();
// use simpleanalyzer for more natural tokenization (else "test." is a token)
Analyzer analyzer = new MockAnalyzer(Random, MockTokenizer.SIMPLE, true);
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, analyzer);
iwc.SetMergePolicy(NewLogMergePolicy());
RandomIndexWriter iw = new RandomIndexWriter(Random, dir, iwc);
FieldType offsetsType = new FieldType(TextField.TYPE_STORED);
offsetsType.IndexOptions = (IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS);
Field body = new Field("body", "", offsetsType);
Document doc = new Document();
doc.Add(body);
body.SetStringValue("This is a test.");
iw.AddDocument(doc);
body.SetStringValue("Test a one sentence document.");
iw.AddDocument(doc);
IndexReader ir = iw.GetReader();
iw.Dispose();
IndexSearcher searcher = NewSearcher(ir);
ICUPostingsHighlighter highlighter = new PostingsHighlighterAnalyzerHelper(analyzer);
SpanQuery include = new SpanMultiTermQueryWrapper <WildcardQuery>(new WildcardQuery(new Term("body", "te*")));
SpanQuery exclude = new SpanTermQuery(new Term("body", "bogus"));
Query query = new SpanNotQuery(include, exclude);
TopDocs topDocs = searcher.Search(query, null, 10, Sort.INDEXORDER);
assertEquals(2, topDocs.TotalHits);
String[] snippets = highlighter.Highlight("body", query, searcher, topDocs);
assertEquals(2, snippets.Length);
assertEquals("This is a <b>test</b>.", snippets[0]);
assertEquals("<b>Test</b> a one sentence document.", snippets[1]);
ir.Dispose();
dir.Dispose();
}
public override Query Rewrite(IndexReader reader)
{
// ArrayList spanClauses = new ArrayList();
if (contents is TermQuery)
{
return(contents);
}
// Build a sequence of Span clauses arranged in a SpanNear - child
// clauses can be complex
// Booleans e.g. nots and ors etc
int numNegatives = 0;
if (!(contents is BooleanQuery))
{
throw new ArgumentException("Unknown query type \""
+ contents.GetType().Name
+ "\" found in phrase query string \"" + phrasedQueryStringContents
+ "\"");
}
BooleanQuery bq = (BooleanQuery)contents;
BooleanClause[] bclauses = bq.GetClauses();
SpanQuery[] allSpanClauses = new SpanQuery[bclauses.Length];
// For all clauses e.g. one* two~
for (int i = 0; i < bclauses.Length; i++)
{
// HashSet bclauseterms=new HashSet();
Query qc = bclauses[i].Query;
// Rewrite this clause e.g one* becomes (one OR onerous)
qc = qc.Rewrite(reader);
if (bclauses[i].Occur.Equals(Occur.MUST_NOT))
{
numNegatives++;
}
if (qc is BooleanQuery booleanQuery)
{
IList <SpanQuery> sc = new JCG.List <SpanQuery>();
AddComplexPhraseClause(sc, booleanQuery);
if (sc.Count > 0)
{
allSpanClauses[i] = sc[0];
}
else
{
// Insert fake term e.g. phrase query was for "Fred Smithe*" and
// there were no "Smithe*" terms - need to
// prevent match on just "Fred".
allSpanClauses[i] = new SpanTermQuery(new Term(field,
"Dummy clause because no terms found - must match nothing"));
}
}
else
{
if (qc is TermQuery tq)
{
allSpanClauses[i] = new SpanTermQuery(tq.Term);
}
else
{
throw new ArgumentException("Unknown query type \""
+ qc.GetType().Name
+ "\" found in phrase query string \""
+ phrasedQueryStringContents + "\"");
}
}
}
if (numNegatives == 0)
{
// The simple case - no negative elements in phrase
return(new SpanNearQuery(allSpanClauses, slopFactor, inOrder));
}
// Complex case - we have mixed positives and negatives in the
// sequence.
// Need to return a SpanNotQuery
JCG.List <SpanQuery> positiveClauses = new JCG.List <SpanQuery>();
for (int j = 0; j < allSpanClauses.Length; j++)
{
if (!bclauses[j].Occur.Equals(Occur.MUST_NOT))
{
positiveClauses.Add(allSpanClauses[j]);
}
}
SpanQuery[] includeClauses = positiveClauses
.ToArray();
SpanQuery include; // LUCENENET: IDE0059: Remove unnecessary value assignment
if (includeClauses.Length == 1)
{
include = includeClauses[0]; // only one positive clause
}
else
{
// need to increase slop factor based on gaps introduced by
// negatives
include = new SpanNearQuery(includeClauses, slopFactor + numNegatives,
inOrder);
}
// Use sequence of positive and negative values as the exclude.
SpanNearQuery exclude = new SpanNearQuery(allSpanClauses, slopFactor,
inOrder);
SpanNotQuery snot = new SpanNotQuery(include, exclude);
return(snot);
}
public virtual Query VisitSpanNotQuery(SpanNotQuery spanNotq)
{
throw new NotImplementedException();
}
public virtual Query VisitSpanNotQuery(SpanNotQuery spanNotq)
{
throw new SnNotSupportedException();
}
