public void TestMultiValuedSortByScore()
{
Directory dir = NewDirectory();
IndexWriter writer = new IndexWriter(dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)));
Document doc = new Document();
FieldType type = new FieldType(TextField.TYPE_STORED);
type.StoreTermVectorOffsets = (true);
type.StoreTermVectorPositions = (true);
type.StoreTermVectors = (true);
type.Freeze();
doc.Add(new Field("field", "zero if naught", type)); // The first two fields contain the best match
doc.Add(new Field("field", "hero of legend", type)); // but total a lower score (3) than the bottom
doc.Add(new Field("field", "naught of hero", type)); // two fields (4)
doc.Add(new Field("field", "naught of hero", type));
writer.AddDocument(doc);
FastVectorHighlighter highlighter = new FastVectorHighlighter();
ScoreOrderFragmentsBuilder fragmentsBuilder = new ScoreOrderFragmentsBuilder();
fragmentsBuilder.IsDiscreteMultiValueHighlighting = (true);
IndexReader reader = DirectoryReader.Open(writer, true);
String[] preTags = new String[] { "<b>" };
String[] postTags = new String[] { "</b>" };
IEncoder encoder = new DefaultEncoder();
int docId = 0;
BooleanQuery query = new BooleanQuery();
query.Add(clause("field", "hero"), Occur.SHOULD);
query.Add(clause("field", "of"), Occur.SHOULD);
query.Add(clause("field", "legend"), Occur.SHOULD);
FieldQuery fieldQuery = highlighter.GetFieldQuery(query, reader);
foreach (IFragListBuilder fragListBuilder in new IFragListBuilder[] {
new SimpleFragListBuilder(), new WeightedFragListBuilder()
})
{
String[] bestFragments = highlighter.GetBestFragments(fieldQuery, reader, docId, "field", 20, 1,
fragListBuilder, fragmentsBuilder, preTags, postTags, encoder);
assertEquals("<b>hero</b> <b>of</b> <b>legend</b>", bestFragments[0]);
bestFragments = highlighter.GetBestFragments(fieldQuery, reader, docId, "field", 28, 1,
fragListBuilder, fragmentsBuilder, preTags, postTags, encoder);
assertEquals("<b>hero</b> <b>of</b> <b>legend</b>", bestFragments[0]);
bestFragments = highlighter.GetBestFragments(fieldQuery, reader, docId, "field", 30000, 1,
fragListBuilder, fragmentsBuilder, preTags, postTags, encoder);
assertEquals("<b>hero</b> <b>of</b> <b>legend</b>", bestFragments[0]);
}
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
public void Test3Frags()
{
BooleanQuery query = new BooleanQuery();
query.Add(new TermQuery(new Term(F, "a")), Occur.SHOULD);
query.Add(new TermQuery(new Term(F, "c")), Occur.SHOULD);
FieldFragList ffl = Ffl(query, "a b b b b b b b b b b b a b a b b b b b c a a b b");
ScoreOrderFragmentsBuilder sofb = new ScoreOrderFragmentsBuilder();
String[] f = sofb.CreateFragments(reader, 0, F, ffl, 3);
assertEquals(3, f.Length);
// check score order
assertEquals("<b>c</b> <b>a</b> <b>a</b> b b", f[0]);
assertEquals("b b <b>a</b> b <b>a</b> b b b b b c", f[1]);
assertEquals("<b>a</b> b b b b b b b b b b", f[2]);
}
private void matchedFieldsTestCase(bool useMatchedFields, bool fieldMatch, String fieldValue, String expected, params Query[] queryClauses)
{
Document doc = new Document();
FieldType stored = new FieldType(TextField.TYPE_STORED);
stored.StoreTermVectorOffsets = (true);
stored.StoreTermVectorPositions = (true);
stored.StoreTermVectors = (true);
stored.Freeze();
FieldType matched = new FieldType(TextField.TYPE_NOT_STORED);
matched.StoreTermVectorOffsets = (true);
matched.StoreTermVectorPositions = (true);
matched.StoreTermVectors = (true);
matched.Freeze();
doc.Add(new Field("field", fieldValue, stored)); // Whitespace tokenized with English stop words
doc.Add(new Field("field_exact", fieldValue, matched)); // Whitespace tokenized without stop words
doc.Add(new Field("field_super_exact", fieldValue, matched)); // Whitespace tokenized without toLower
doc.Add(new Field("field_characters", fieldValue, matched)); // Each letter is a token
doc.Add(new Field("field_tripples", fieldValue, matched)); // Every three letters is a token
doc.Add(new Field("field_sliced", fieldValue.Substring(0, // Sliced at 10 chars then analyzed just like field
Math.Min(fieldValue.Length - 1, 10) - 0), matched));
doc.Add(new Field("field_der_red", new CannedTokenStream( // Hacky field containing "der" and "red" at pos = 0
token("der", 1, 0, 3),
token("red", 0, 0, 3)
), matched));
Analyzer analyzer = new AnalyzerWrapperAnonymousHelper();
Directory dir = NewDirectory();
IndexWriter writer = new IndexWriter(dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, analyzer));
writer.AddDocument(doc);
FastVectorHighlighter highlighter = new FastVectorHighlighter();
IFragListBuilder fragListBuilder = new SimpleFragListBuilder();
IFragmentsBuilder fragmentsBuilder = new ScoreOrderFragmentsBuilder();
IndexReader reader = DirectoryReader.Open(writer, true);
String[] preTags = new String[] { "<b>" };
String[] postTags = new String[] { "</b>" };
IEncoder encoder = new DefaultEncoder();
int docId = 0;
BooleanQuery query = new BooleanQuery();
foreach (Query clause in queryClauses)
{
query.Add(clause, Occur.MUST);
}
FieldQuery fieldQuery = new FieldQuery(query, reader, true, fieldMatch);
String[] bestFragments;
if (useMatchedFields)
{
ISet <String> matchedFields = new HashSet <String>();
matchedFields.Add("field");
matchedFields.Add("field_exact");
matchedFields.Add("field_super_exact");
matchedFields.Add("field_characters");
matchedFields.Add("field_tripples");
matchedFields.Add("field_sliced");
matchedFields.Add("field_der_red");
bestFragments = highlighter.GetBestFragments(fieldQuery, reader, docId, "field", matchedFields, 25, 1,
fragListBuilder, fragmentsBuilder, preTags, postTags, encoder);
}
else
{
bestFragments = highlighter.GetBestFragments(fieldQuery, reader, docId, "field", 25, 1,
fragListBuilder, fragmentsBuilder, preTags, postTags, encoder);
}
assertEquals(expected, bestFragments[0]);
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
