public virtual void TestPhrasePrefix()
{
RAMDirectory indexStore = new RAMDirectory();
IndexWriter writer = new IndexWriter(indexStore, new SimpleAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
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);
writer.Optimize();
writer.Close();
IndexSearcher searcher = new IndexSearcher(indexStore);
// 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"));
System.Collections.ArrayList termsWithPrefix = new System.Collections.ArrayList();
IndexReader ir = IndexReader.Open(indexStore);
// this TermEnum gives "piccadilly", "pie" and "pizza".
System.String prefix = "pi";
TermEnum te = ir.Terms(new Term("body", prefix));
do
{
if (te.Term().Text().StartsWith(prefix))
{
termsWithPrefix.Add(te.Term());
}
}
while (te.Next());
query1.Add((Term[]) termsWithPrefix.ToArray(typeof(Term)));
Assert.AreEqual("body:\"blueberry (piccadilly pie pizza)\"", query1.ToString());
query2.Add((Term[]) termsWithPrefix.ToArray(typeof(Term)));
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 = ir.Terms(new Term("body", prefix));
do
{
if (te.Term().Text().StartsWith(prefix))
{
termsWithPrefix.Add(te.Term());
}
}
while (te.Next());
query3.Add((Term[]) termsWithPrefix.ToArray(typeof(Term)));
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.SetSlop(1);
result = searcher.Search(query3, null, 1000).scoreDocs;
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
}
searcher.Close();
indexStore.Close();
}
public virtual void TestPhrasePrefix()
{
Directory indexStore = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), indexStore, Similarity, TimeZone);
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);
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()
{
RAMDirectory indexStore = new RAMDirectory();
IndexWriter writer = new IndexWriter(indexStore, new SimpleAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
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);
writer.Optimize();
writer.Close();
IndexSearcher searcher = new IndexSearcher(indexStore, true);
// 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"));
System.Collections.ArrayList termsWithPrefix = new System.Collections.ArrayList();
IndexReader ir = IndexReader.Open(indexStore, true);
// this TermEnum gives "piccadilly", "pie" and "pizza".
System.String prefix = "pi";
TermEnum te = ir.Terms(new Term("body", prefix));
do
{
if (te.Term.Text.StartsWith(prefix))
{
termsWithPrefix.Add(te.Term);
}
}while (te.Next());
query1.Add((Term[])termsWithPrefix.ToArray(typeof(Term)));
Assert.AreEqual("body:\"blueberry (piccadilly pie pizza)\"", query1.ToString());
query2.Add((Term[])termsWithPrefix.ToArray(typeof(Term)));
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 = ir.Terms(new Term("body", prefix));
do
{
if (te.Term.Text.StartsWith(prefix))
{
termsWithPrefix.Add(te.Term);
}
}while (te.Next());
query3.Add((Term[])termsWithPrefix.ToArray(typeof(Term)));
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;
Assert.AreEqual(3, result.Length); // blueberry pizza, bluebird pizza, bluebird foobar pizza
MultiPhraseQuery query4 = new MultiPhraseQuery();
// okay, all terms must belong to the same field
Assert.Throws <ArgumentException>(() =>
{
query4.Add(new Term("field1", "foo"));
query4.Add(new Term("field2", "foobar"));
});
searcher.Close();
indexStore.Close();
}
