TestQueryParserpublic class TestQueryParser extends TestCase
| Fields Summary |
|---|
| public static Analyzer | qpAnalyzer | | private int | originalMaxClauses |
| Methods Summary |
|---|
| private static void | addDateDoc(java.lang.String content, int year, int month, int day, int hour, int minute, int second, org.apache.lucene.index.IndexWriter iw)
Document d = new Document();
d.add(new Field("f", content, Field.Store.YES, Field.Index.TOKENIZED));
Calendar cal = Calendar.getInstance();
cal.set(year, month-1, day, hour, minute, second);
d.add(new Field("date", DateField.dateToString(cal.getTime()), Field.Store.YES, Field.Index.UN_TOKENIZED));
iw.addDocument(d);
| | public void | assertEscapedQueryEquals(java.lang.String query, org.apache.lucene.analysis.Analyzer a, java.lang.String result)
String escapedQuery = QueryParser.escape(query);
if (!escapedQuery.equals(result)) {
fail("Query /" + query + "/ yielded /" + escapedQuery
+ "/, expecting /" + result + "/");
}
| | private void | assertHits(int expected, java.lang.String query, org.apache.lucene.search.IndexSearcher is)
QueryParser qp = new QueryParser("date", new WhitespaceAnalyzer());
qp.setLocale(Locale.ENGLISH);
Query q = qp.parse(query);
Hits hits = is.search(q);
assertEquals(expected, hits.length());
| | public void | assertQueryEquals(java.lang.String query, org.apache.lucene.analysis.Analyzer a, java.lang.String result)
Query q = getQuery(query, a);
String s = q.toString("field");
if (!s.equals(result)) {
fail("Query /" + query + "/ yielded /" + s
+ "/, expecting /" + result + "/");
}
| | public void | assertQueryEqualsDOA(java.lang.String query, org.apache.lucene.analysis.Analyzer a, java.lang.String result)
Query q = getQueryDOA(query, a);
String s = q.toString("field");
if (!s.equals(result)) {
fail("Query /" + query + "/ yielded /" + s
+ "/, expecting /" + result + "/");
}
| | public void | assertWildcardQueryEquals(java.lang.String query, boolean lowercase, java.lang.String result)
QueryParser qp = getParser(null);
qp.setLowercaseExpandedTerms(lowercase);
Query q = qp.parse(query);
String s = q.toString("field");
if (!s.equals(result)) {
fail("WildcardQuery /" + query + "/ yielded /" + s
+ "/, expecting /" + result + "/");
}
| | public void | assertWildcardQueryEquals(java.lang.String query, java.lang.String result)
QueryParser qp = getParser(null);
Query q = qp.parse(query);
String s = q.toString("field");
if (!s.equals(result)) {
fail("WildcardQuery /" + query + "/ yielded /" + s + "/, expecting /"
+ result + "/");
}
| | private java.lang.String | getDate(java.lang.String s)
DateFormat df = DateFormat.getDateInstance(DateFormat.SHORT);
return DateField.dateToString(df.parse(s));
| | private java.lang.String | getLocalizedDate(int year, int month, int day, boolean extendLastDate)
DateFormat df = DateFormat.getDateInstance(DateFormat.SHORT);
Calendar calendar = Calendar.getInstance();
calendar.set(year, month, day);
if (extendLastDate) {
calendar.set(Calendar.HOUR_OF_DAY, 23);
calendar.set(Calendar.MINUTE, 59);
calendar.set(Calendar.SECOND, 59);
calendar.set(Calendar.MILLISECOND, 999);
}
return df.format(calendar.getTime());
| | public org.apache.lucene.queryParser.QueryParser | getParser(org.apache.lucene.analysis.Analyzer a)
if (a == null)
a = new SimpleAnalyzer();
QueryParser qp = new QueryParser("field", a);
qp.setDefaultOperator(QueryParser.OR_OPERATOR);
return qp;
| | public org.apache.lucene.search.Query | getQuery(java.lang.String query, org.apache.lucene.analysis.Analyzer a)
return getParser(a).parse(query);
| | public org.apache.lucene.search.Query | getQueryDOA(java.lang.String query, org.apache.lucene.analysis.Analyzer a)
if (a == null)
a = new SimpleAnalyzer();
QueryParser qp = new QueryParser("field", a);
qp.setDefaultOperator(QueryParser.AND_OPERATOR);
return qp.parse(query);
| | public void | setUp()
originalMaxClauses = BooleanQuery.getMaxClauseCount();
| | public void | tearDown()
BooleanQuery.setMaxClauseCount(originalMaxClauses);
| | public void | testBooleanQuery()
BooleanQuery.setMaxClauseCount(2);
try {
QueryParser qp = new QueryParser("field", new WhitespaceAnalyzer());
qp.parse("one two three");
fail("ParseException expected due to too many boolean clauses");
} catch (ParseException expected) {
// too many boolean clauses, so ParseException is expected
}
| | public void | testBoost()
StandardAnalyzer oneStopAnalyzer = new StandardAnalyzer(new String[]{"on"});
QueryParser qp = new QueryParser("field", oneStopAnalyzer);
Query q = qp.parse("on^1.0");
assertNotNull(q);
q = qp.parse("\"hello\"^2.0");
assertNotNull(q);
assertEquals(q.getBoost(), (float) 2.0, (float) 0.5);
q = qp.parse("hello^2.0");
assertNotNull(q);
assertEquals(q.getBoost(), (float) 2.0, (float) 0.5);
q = qp.parse("\"on\"^1.0");
assertNotNull(q);
QueryParser qp2 = new QueryParser("field", new StandardAnalyzer());
q = qp2.parse("the^3");
// "the" is a stop word so the result is an empty query:
assertNotNull(q);
assertEquals("", q.toString());
assertEquals(1.0f, q.getBoost(), 0.01f);
| | public void | testCustomQueryParserFuzzy()
try {
new QPTestParser("contents", new WhitespaceAnalyzer()).parse("xunit~");
fail("Fuzzy queries should not be allowed");
} catch (ParseException expected) {
// expected exception
}
| | public void | testCustomQueryParserWildcard()
try {
new QPTestParser("contents", new WhitespaceAnalyzer()).parse("a?t");
fail("Wildcard queries should not be allowed");
} catch (ParseException expected) {
// expected exception
}
| | public void | testDateRange()
String startDate = getLocalizedDate(2002, 1, 1, false);
String endDate = getLocalizedDate(2002, 1, 4, false);
Calendar endDateExpected = Calendar.getInstance();
endDateExpected.set(2002, 1, 4, 23, 59, 59);
endDateExpected.set(Calendar.MILLISECOND, 999);
assertQueryEquals("[ " + startDate + " TO " + endDate + "]", null,
"[" + getDate(startDate) + " TO " + DateField.dateToString(endDateExpected.getTime()) + "]");
assertQueryEquals("{ " + startDate + " " + endDate + " }", null,
"{" + getDate(startDate) + " TO " + getDate(endDate) + "}");
| | public void | testEscaped()
Analyzer a = new WhitespaceAnalyzer();
/*assertQueryEquals("\\[brackets", a, "\\[brackets");
assertQueryEquals("\\[brackets", null, "brackets");
assertQueryEquals("\\\\", a, "\\\\");
assertQueryEquals("\\+blah", a, "\\+blah");
assertQueryEquals("\\(blah", a, "\\(blah");
assertQueryEquals("\\-blah", a, "\\-blah");
assertQueryEquals("\\!blah", a, "\\!blah");
assertQueryEquals("\\{blah", a, "\\{blah");
assertQueryEquals("\\}blah", a, "\\}blah");
assertQueryEquals("\\:blah", a, "\\:blah");
assertQueryEquals("\\^blah", a, "\\^blah");
assertQueryEquals("\\[blah", a, "\\[blah");
assertQueryEquals("\\]blah", a, "\\]blah");
assertQueryEquals("\\\"blah", a, "\\\"blah");
assertQueryEquals("\\(blah", a, "\\(blah");
assertQueryEquals("\\)blah", a, "\\)blah");
assertQueryEquals("\\~blah", a, "\\~blah");
assertQueryEquals("\\*blah", a, "\\*blah");
assertQueryEquals("\\?blah", a, "\\?blah");
//assertQueryEquals("foo \\&\\& bar", a, "foo \\&\\& bar");
//assertQueryEquals("foo \\|| bar", a, "foo \\|| bar");
//assertQueryEquals("foo \\AND bar", a, "foo \\AND bar");*/
assertQueryEquals("a\\-b:c", a, "a-b:c");
assertQueryEquals("a\\+b:c", a, "a+b:c");
assertQueryEquals("a\\:b:c", a, "a:b:c");
assertQueryEquals("a\\\\b:c", a, "a\\b:c");
assertQueryEquals("a:b\\-c", a, "a:b-c");
assertQueryEquals("a:b\\+c", a, "a:b+c");
assertQueryEquals("a:b\\:c", a, "a:b:c");
assertQueryEquals("a:b\\\\c", a, "a:b\\c");
assertQueryEquals("a:b\\-c*", a, "a:b-c*");
assertQueryEquals("a:b\\+c*", a, "a:b+c*");
assertQueryEquals("a:b\\:c*", a, "a:b:c*");
assertQueryEquals("a:b\\\\c*", a, "a:b\\c*");
assertQueryEquals("a:b\\-?c", a, "a:b-?c");
assertQueryEquals("a:b\\+?c", a, "a:b+?c");
assertQueryEquals("a:b\\:?c", a, "a:b:?c");
assertQueryEquals("a:b\\\\?c", a, "a:b\\?c");
assertQueryEquals("a:b\\-c~", a, "a:b-c~0.5");
assertQueryEquals("a:b\\+c~", a, "a:b+c~0.5");
assertQueryEquals("a:b\\:c~", a, "a:b:c~0.5");
assertQueryEquals("a:b\\\\c~", a, "a:b\\c~0.5");
assertQueryEquals("[ a\\- TO a\\+ ]", null, "[a- TO a+]");
assertQueryEquals("[ a\\: TO a\\~ ]", null, "[a: TO a~]");
assertQueryEquals("[ a\\\\ TO a\\* ]", null, "[a\\ TO a*]");
| | public void | testException()
try {
assertQueryEquals("\"some phrase", null, "abc");
fail("ParseException expected, not thrown");
} catch (ParseException expected) {
}
| | public void | testLocalDateFormat()
RAMDirectory ramDir = new RAMDirectory();
IndexWriter iw = new IndexWriter(ramDir, new WhitespaceAnalyzer(), true);
addDateDoc("a", 2005, 12, 2, 10, 15, 33, iw);
addDateDoc("b", 2005, 12, 4, 22, 15, 00, iw);
iw.close();
IndexSearcher is = new IndexSearcher(ramDir);
assertHits(1, "[12/1/2005 TO 12/3/2005]", is);
assertHits(2, "[12/1/2005 TO 12/4/2005]", is);
assertHits(1, "[12/3/2005 TO 12/4/2005]", is);
assertHits(1, "{12/1/2005 TO 12/3/2005}", is);
assertHits(1, "{12/1/2005 TO 12/4/2005}", is);
assertHits(0, "{12/3/2005 TO 12/4/2005}", is);
is.close();
| | public void | testNumber()
// The numbers go away because SimpleAnalzyer ignores them
assertQueryEquals("3", null, "");
assertQueryEquals("term 1.0 1 2", null, "term");
assertQueryEquals("term term1 term2", null, "term term term");
Analyzer a = new StandardAnalyzer();
assertQueryEquals("3", a, "3");
assertQueryEquals("term 1.0 1 2", a, "term 1.0 1 2");
assertQueryEquals("term term1 term2", a, "term term1 term2");
| | public void | testPrecedence()This test differs from TestPrecedenceQueryParser
QueryParser qp = new QueryParser("field", new WhitespaceAnalyzer());
Query query1 = qp.parse("A AND B OR C AND D");
Query query2 = qp.parse("+A +B +C +D");
assertEquals(query1, query2);
| | public void | testPunct()
Analyzer a = new WhitespaceAnalyzer();
assertQueryEquals("a&b", a, "a&b");
assertQueryEquals("a&&b", a, "a&&b");
assertQueryEquals(".NET", a, ".NET");
| | public void | testQPA()
assertQueryEquals("term term term", qpAnalyzer, "term term term");
assertQueryEquals("term +stop term", qpAnalyzer, "term term");
assertQueryEquals("term -stop term", qpAnalyzer, "term term");
assertQueryEquals("drop AND stop AND roll", qpAnalyzer, "+drop +roll");
assertQueryEquals("term phrase term", qpAnalyzer,
"term \"phrase1 phrase2\" term");
assertQueryEquals("term AND NOT phrase term", qpAnalyzer,
"+term -\"phrase1 phrase2\" term");
assertQueryEquals("stop", qpAnalyzer, "");
assertTrue(getQuery("term term term", qpAnalyzer) instanceof BooleanQuery);
assertTrue(getQuery("term +stop", qpAnalyzer) instanceof TermQuery);
| | public void | testQueryStringEscaping()
Analyzer a = new WhitespaceAnalyzer();
assertEscapedQueryEquals("a-b:c", a, "a\\-b\\:c");
assertEscapedQueryEquals("a+b:c", a, "a\\+b\\:c");
assertEscapedQueryEquals("a:b:c", a, "a\\:b\\:c");
assertEscapedQueryEquals("a\\b:c", a, "a\\\\b\\:c");
assertEscapedQueryEquals("a:b-c", a, "a\\:b\\-c");
assertEscapedQueryEquals("a:b+c", a, "a\\:b\\+c");
assertEscapedQueryEquals("a:b:c", a, "a\\:b\\:c");
assertEscapedQueryEquals("a:b\\c", a, "a\\:b\\\\c");
assertEscapedQueryEquals("a:b-c*", a, "a\\:b\\-c\\*");
assertEscapedQueryEquals("a:b+c*", a, "a\\:b\\+c\\*");
assertEscapedQueryEquals("a:b:c*", a, "a\\:b\\:c\\*");
assertEscapedQueryEquals("a:b\\\\c*", a, "a\\:b\\\\\\\\c\\*");
assertEscapedQueryEquals("a:b-?c", a, "a\\:b\\-\\?c");
assertEscapedQueryEquals("a:b+?c", a, "a\\:b\\+\\?c");
assertEscapedQueryEquals("a:b:?c", a, "a\\:b\\:\\?c");
assertEscapedQueryEquals("a:b?c", a, "a\\:b\\?c");
assertEscapedQueryEquals("a:b-c~", a, "a\\:b\\-c\\~");
assertEscapedQueryEquals("a:b+c~", a, "a\\:b\\+c\\~");
assertEscapedQueryEquals("a:b:c~", a, "a\\:b\\:c\\~");
assertEscapedQueryEquals("a:b\\c~", a, "a\\:b\\\\c\\~");
assertEscapedQueryEquals("[ a - TO a+ ]", null, "\\[ a \\- TO a\\+ \\]");
assertEscapedQueryEquals("[ a : TO a~ ]", null, "\\[ a \\: TO a\\~ \\]");
assertEscapedQueryEquals("[ a\\ TO a* ]", null, "\\[ a\\\\ TO a\\* \\]");
| | public void | testRange()
assertQueryEquals("[ a TO z]", null, "[a TO z]");
assertTrue(getQuery("[ a TO z]", null) instanceof RangeQuery);
assertQueryEquals("[ a TO z ]", null, "[a TO z]");
assertQueryEquals("{ a TO z}", null, "{a TO z}");
assertQueryEquals("{ a TO z }", null, "{a TO z}");
assertQueryEquals("{ a TO z }^2.0", null, "{a TO z}^2.0");
assertQueryEquals("[ a TO z] OR bar", null, "[a TO z] bar");
assertQueryEquals("[ a TO z] AND bar", null, "+[a TO z] +bar");
assertQueryEquals("( bar blar { a TO z}) ", null, "bar blar {a TO z}");
assertQueryEquals("gack ( bar blar { a TO z}) ", null, "gack (bar blar {a TO z})");
| | public void | testSimple()
assertQueryEquals("term term term", null, "term term term");
assertQueryEquals("türm term term", null, "türm term term");
assertQueryEquals("ümlaut", null, "ümlaut");
assertQueryEquals("a AND b", null, "+a +b");
assertQueryEquals("(a AND b)", null, "+a +b");
assertQueryEquals("c OR (a AND b)", null, "c (+a +b)");
assertQueryEquals("a AND NOT b", null, "+a -b");
assertQueryEquals("a AND -b", null, "+a -b");
assertQueryEquals("a AND !b", null, "+a -b");
assertQueryEquals("a && b", null, "+a +b");
assertQueryEquals("a && ! b", null, "+a -b");
assertQueryEquals("a OR b", null, "a b");
assertQueryEquals("a || b", null, "a b");
assertQueryEquals("a OR !b", null, "a -b");
assertQueryEquals("a OR ! b", null, "a -b");
assertQueryEquals("a OR -b", null, "a -b");
assertQueryEquals("+term -term term", null, "+term -term term");
assertQueryEquals("foo:term AND field:anotherTerm", null,
"+foo:term +anotherterm");
assertQueryEquals("term AND \"phrase phrase\"", null,
"+term +\"phrase phrase\"");
assertQueryEquals("\"hello there\"", null, "\"hello there\"");
assertTrue(getQuery("a AND b", null) instanceof BooleanQuery);
assertTrue(getQuery("hello", null) instanceof TermQuery);
assertTrue(getQuery("\"hello there\"", null) instanceof PhraseQuery);
assertQueryEquals("germ term^2.0", null, "germ term^2.0");
assertQueryEquals("(term)^2.0", null, "term^2.0");
assertQueryEquals("(germ term)^2.0", null, "(germ term)^2.0");
assertQueryEquals("term^2.0", null, "term^2.0");
assertQueryEquals("term^2", null, "term^2.0");
assertQueryEquals("\"germ term\"^2.0", null, "\"germ term\"^2.0");
assertQueryEquals("\"term germ\"^2", null, "\"term germ\"^2.0");
assertQueryEquals("(foo OR bar) AND (baz OR boo)", null,
"+(foo bar) +(baz boo)");
assertQueryEquals("((a OR b) AND NOT c) OR d", null,
"(+(a b) -c) d");
assertQueryEquals("+(apple \"steve jobs\") -(foo bar baz)", null,
"+(apple \"steve jobs\") -(foo bar baz)");
assertQueryEquals("+title:(dog OR cat) -author:\"bob dole\"", null,
"+(title:dog title:cat) -author:\"bob dole\"");
QueryParser qp = new QueryParser("field", new StandardAnalyzer());
// make sure OR is the default:
assertEquals(QueryParser.OR_OPERATOR, qp.getDefaultOperator());
qp.setDefaultOperator(QueryParser.AND_OPERATOR);
assertEquals(QueryParser.AND_OPERATOR, qp.getDefaultOperator());
qp.setDefaultOperator(QueryParser.OR_OPERATOR);
assertEquals(QueryParser.OR_OPERATOR, qp.getDefaultOperator());
| | public void | testSimpleDAO()
assertQueryEqualsDOA("term term term", null, "+term +term +term");
assertQueryEqualsDOA("term +term term", null, "+term +term +term");
assertQueryEqualsDOA("term term +term", null, "+term +term +term");
assertQueryEqualsDOA("term +term +term", null, "+term +term +term");
assertQueryEqualsDOA("-term term term", null, "-term +term +term");
| | public void | testSlop()
assertQueryEquals("\"term germ\"~2", null, "\"term germ\"~2");
assertQueryEquals("\"term germ\"~2 flork", null, "\"term germ\"~2 flork");
assertQueryEquals("\"term\"~2", null, "term");
assertQueryEquals("\" \"~2 germ", null, "germ");
assertQueryEquals("\"term germ\"~2^2", null, "\"term germ\"~2^2.0");
| | public void | testTabNewlineCarriageReturn()
assertQueryEqualsDOA("+weltbank +worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("+weltbank\n+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \n+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \n +worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("+weltbank\r+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \r+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \r +worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("+weltbank\r\n+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \r\n+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \r\n +worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \r \n +worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("+weltbank\t+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \t+worlbank", null,
"+weltbank +worlbank");
assertQueryEqualsDOA("weltbank \t +worlbank", null,
"+weltbank +worlbank");
| | public void | testWildcard()
assertQueryEquals("term*", null, "term*");
assertQueryEquals("term*^2", null, "term*^2.0");
assertQueryEquals("term~", null, "term~0.5");
assertQueryEquals("term~0.7", null, "term~0.7");
assertQueryEquals("term~^2", null, "term~0.5^2.0");
assertQueryEquals("term^2~", null, "term~0.5^2.0");
assertQueryEquals("term*germ", null, "term*germ");
assertQueryEquals("term*germ^3", null, "term*germ^3.0");
assertTrue(getQuery("term*", null) instanceof PrefixQuery);
assertTrue(getQuery("term*^2", null) instanceof PrefixQuery);
assertTrue(getQuery("term~", null) instanceof FuzzyQuery);
assertTrue(getQuery("term~0.7", null) instanceof FuzzyQuery);
FuzzyQuery fq = (FuzzyQuery)getQuery("term~0.7", null);
assertEquals(0.7f, fq.getMinSimilarity(), 0.1f);
assertEquals(FuzzyQuery.defaultPrefixLength, fq.getPrefixLength());
fq = (FuzzyQuery)getQuery("term~", null);
assertEquals(0.5f, fq.getMinSimilarity(), 0.1f);
assertEquals(FuzzyQuery.defaultPrefixLength, fq.getPrefixLength());
try {
getQuery("term~1.1", null); // value > 1, throws exception
fail();
} catch(ParseException pe) {
// expected exception
}
assertTrue(getQuery("term*germ", null) instanceof WildcardQuery);
/* Tests to see that wild card terms are (or are not) properly
* lower-cased with propery parser configuration
*/
// First prefix queries:
// by default, convert to lowercase:
assertWildcardQueryEquals("Term*", true, "term*");
// explicitly set lowercase:
assertWildcardQueryEquals("term*", true, "term*");
assertWildcardQueryEquals("Term*", true, "term*");
assertWildcardQueryEquals("TERM*", true, "term*");
// explicitly disable lowercase conversion:
assertWildcardQueryEquals("term*", false, "term*");
assertWildcardQueryEquals("Term*", false, "Term*");
assertWildcardQueryEquals("TERM*", false, "TERM*");
// Then 'full' wildcard queries:
// by default, convert to lowercase:
assertWildcardQueryEquals("Te?m", "te?m");
// explicitly set lowercase:
assertWildcardQueryEquals("te?m", true, "te?m");
assertWildcardQueryEquals("Te?m", true, "te?m");
assertWildcardQueryEquals("TE?M", true, "te?m");
assertWildcardQueryEquals("Te?m*gerM", true, "te?m*germ");
// explicitly disable lowercase conversion:
assertWildcardQueryEquals("te?m", false, "te?m");
assertWildcardQueryEquals("Te?m", false, "Te?m");
assertWildcardQueryEquals("TE?M", false, "TE?M");
assertWildcardQueryEquals("Te?m*gerM", false, "Te?m*gerM");
// Fuzzy queries:
assertWildcardQueryEquals("Term~", "term~0.5");
assertWildcardQueryEquals("Term~", true, "term~0.5");
assertWildcardQueryEquals("Term~", false, "Term~0.5");
// Range queries:
assertWildcardQueryEquals("[A TO C]", "[a TO c]");
assertWildcardQueryEquals("[A TO C]", true, "[a TO c]");
assertWildcardQueryEquals("[A TO C]", false, "[A TO C]");
|
|
