TestSortpublic class TestSort extends TestCase implements Serializable| Unit tests for sorting code.
Created: Feb 17, 2004 4:55:10 PM |
| Fields Summary |
|---|
| private Searcher | full | | private Searcher | searchX | | private Searcher | searchY | | private Query | queryX | | private Query | queryY | | private Query | queryA | | private Query | queryE | | private Query | queryF | | private Query | queryG | | private Sort | sort | | private String[] | data |
| Constructors Summary |
|---|
public TestSort(String name)
super (name);
|
| Methods Summary |
|---|
| private void | assertMatches(org.apache.lucene.search.Searcher searcher, org.apache.lucene.search.Query query, org.apache.lucene.search.Sort sort, java.lang.String expectedResult)
Hits result = searcher.search (query, sort);
StringBuffer buff = new StringBuffer(10);
int n = result.length();
for (int i=0; i<n; ++i) {
Document doc = result.doc(i);
String[] v = doc.getValues("tracer");
for (int j=0; j<v.length; ++j) {
buff.append (v[j]);
}
}
assertEquals (expectedResult, buff.toString());
| | private void | assertMatchesPattern(org.apache.lucene.search.Searcher searcher, org.apache.lucene.search.Query query, org.apache.lucene.search.Sort sort, java.lang.String pattern)
Hits result = searcher.search (query, sort);
StringBuffer buff = new StringBuffer(10);
int n = result.length();
for (int i=0; i<n; ++i) {
Document doc = result.doc(i);
String[] v = doc.getValues("tracer");
for (int j=0; j<v.length; ++j) {
buff.append (v[j]);
}
}
// System.out.println ("matching \""+buff+"\" against pattern \""+pattern+"\"");
assertTrue (Pattern.compile(pattern).matcher(buff.toString()).matches());
| | private void | assertSameValues(java.util.HashMap m1, java.util.HashMap m2)
int n = m1.size();
int m = m2.size();
assertEquals (n, m);
Iterator iter = m1.keySet().iterator();
while (iter.hasNext()) {
Object key = iter.next();
Object o1 = m1.get(key);
Object o2 = m2.get(key);
if (o1 instanceof Float) {
assertEquals(((Float)o1).floatValue(), ((Float)o2).floatValue(), 1e-6);
} else {
assertEquals (m1.get(key), m2.get(key));
}
}
| | private org.apache.lucene.search.Searcher | getEmptyIndex()
return getIndex (false, false);
| | private org.apache.lucene.search.Searcher | getFullIndex()
return getIndex (true, true);
| | private org.apache.lucene.search.Searcher | getIndex(boolean even, boolean odd)
// create an index of all the documents, or just the x, or just the y documents
RAMDirectory indexStore = new RAMDirectory ();
IndexWriter writer = new IndexWriter (indexStore, new SimpleAnalyzer(), true);
for (int i=0; i<data.length; ++i) {
if (((i%2)==0 && even) || ((i%2)==1 && odd)) {
Document doc = new Document();
doc.add (new Field ("tracer", data[i][0], Field.Store.YES, Field.Index.NO));
doc.add (new Field ("contents", data[i][1], Field.Store.NO, Field.Index.TOKENIZED));
if (data[i][2] != null) doc.add (new Field ("int", data[i][2], Field.Store.NO, Field.Index.UN_TOKENIZED));
if (data[i][3] != null) doc.add (new Field ("float", data[i][3], Field.Store.NO, Field.Index.UN_TOKENIZED));
if (data[i][4] != null) doc.add (new Field ("string", data[i][4], Field.Store.NO, Field.Index.UN_TOKENIZED));
if (data[i][5] != null) doc.add (new Field ("custom", data[i][5], Field.Store.NO, Field.Index.UN_TOKENIZED));
if (data[i][6] != null) doc.add (new Field ("i18n", data[i][6], Field.Store.NO, Field.Index.UN_TOKENIZED));
doc.setBoost(2); // produce some scores above 1.0
writer.addDocument (doc);
}
}
writer.optimize ();
writer.close ();
return new IndexSearcher (indexStore);
| | private org.apache.lucene.search.Searchable | getRemote()
try {
return lookupRemote ();
} catch (Throwable e) {
startServer ();
return lookupRemote ();
}
| | private java.util.HashMap | getScores(org.apache.lucene.search.Hits hits)
HashMap scoreMap = new HashMap();
int n = hits.length();
for (int i=0; i<n; ++i) {
Document doc = hits.doc(i);
String[] v = doc.getValues("tracer");
assertEquals (v.length, 1);
scoreMap.put (v[0], new Float(hits.score(i)));
}
return scoreMap;
| | private org.apache.lucene.search.Searcher | getXIndex()
return getIndex (true, false);
| | private org.apache.lucene.search.Searcher | getYIndex()
return getIndex (false, true);
| | private org.apache.lucene.search.Searchable | lookupRemote()
return (Searchable) Naming.lookup ("//localhost/SortedSearchable");
| | public static void | main(java.lang.String[] argv)
if (argv == null || argv.length < 1)
TestRunner.run (suite());
else if ("server".equals (argv[0])) {
TestSort test = new TestSort (null);
try {
test.startServer();
Thread.sleep (500000);
} catch (Exception e) {
System.out.println (e);
e.printStackTrace();
}
}
| | private void | runMultiSorts(org.apache.lucene.search.Searcher multi)
sort.setSort (SortField.FIELD_DOC);
assertMatchesPattern (multi, queryA, sort, "[AB]{2}[CD]{2}[EF]{2}[GH]{2}[IJ]{2}");
sort.setSort (new SortField ("int", SortField.INT));
assertMatchesPattern (multi, queryA, sort, "IDHFGJ[ABE]{3}C");
sort.setSort (new SortField[] {new SortField ("int", SortField.INT), SortField.FIELD_DOC});
assertMatchesPattern (multi, queryA, sort, "IDHFGJ[AB]{2}EC");
sort.setSort ("int");
assertMatchesPattern (multi, queryA, sort, "IDHFGJ[AB]{2}EC");
sort.setSort (new SortField[] {new SortField ("float", SortField.FLOAT), SortField.FIELD_DOC});
assertMatchesPattern (multi, queryA, sort, "GDHJ[CI]{2}EFAB");
sort.setSort ("float");
assertMatchesPattern (multi, queryA, sort, "GDHJ[CI]{2}EFAB");
sort.setSort ("string");
assertMatches (multi, queryA, sort, "DJAIHGFEBC");
sort.setSort ("int", true);
assertMatchesPattern (multi, queryA, sort, "C[AB]{2}EJGFHDI");
sort.setSort ("float", true);
assertMatchesPattern (multi, queryA, sort, "BAFE[IC]{2}JHDG");
sort.setSort ("string", true);
assertMatches (multi, queryA, sort, "CBEFGHIAJD");
sort.setSort (new SortField[] { new SortField ("string", Locale.US) });
assertMatches (multi, queryA, sort, "DJAIHGFEBC");
sort.setSort (new SortField[] { new SortField ("string", Locale.US, true) });
assertMatches (multi, queryA, sort, "CBEFGHIAJD");
sort.setSort (new String[] {"int","float"});
assertMatches (multi, queryA, sort, "IDHFGJEABC");
sort.setSort (new String[] {"float","string"});
assertMatches (multi, queryA, sort, "GDHJICEFAB");
sort.setSort ("int");
assertMatches (multi, queryF, sort, "IZJ");
sort.setSort ("int", true);
assertMatches (multi, queryF, sort, "JZI");
sort.setSort ("float");
assertMatches (multi, queryF, sort, "ZJI");
sort.setSort ("string");
assertMatches (multi, queryF, sort, "ZJI");
sort.setSort ("string", true);
assertMatches (multi, queryF, sort, "IJZ");
| | public void | setUp()
full = getFullIndex();
searchX = getXIndex();
searchY = getYIndex();
queryX = new TermQuery (new Term ("contents", "x"));
queryY = new TermQuery (new Term ("contents", "y"));
queryA = new TermQuery (new Term ("contents", "a"));
queryE = new TermQuery (new Term ("contents", "e"));
queryF = new TermQuery (new Term ("contents", "f"));
queryG = new TermQuery (new Term ("contents", "g"));
sort = new Sort();
| | private void | startServer()
// construct an index
Searcher local = getFullIndex();
// local.search (queryA, new Sort());
// publish it
Registry reg = LocateRegistry.createRegistry (1099);
RemoteSearchable impl = new RemoteSearchable (local);
Naming.rebind ("//localhost/SortedSearchable", impl);
| | public static junit.framework.Test | suite()
return new TestSuite (TestSort.class);
| | public void | testAutoSort()
sort.setSort("int");
assertMatches (full, queryX, sort, "IGAEC");
assertMatches (full, queryY, sort, "DHFJB");
sort.setSort("float");
assertMatches (full, queryX, sort, "GCIEA");
assertMatches (full, queryY, sort, "DHJFB");
sort.setSort("string");
assertMatches (full, queryX, sort, "AIGEC");
assertMatches (full, queryY, sort, "DJHFB");
| | public void | testBuiltInSorts()
sort = new Sort();
assertMatches (full, queryX, sort, "ACEGI");
assertMatches (full, queryY, sort, "BDFHJ");
sort.setSort(SortField.FIELD_DOC);
assertMatches (full, queryX, sort, "ACEGI");
assertMatches (full, queryY, sort, "BDFHJ");
| | public void | testCustomSorts()
sort.setSort (new SortField ("custom", SampleComparable.getComparatorSource()));
assertMatches (full, queryX, sort, "CAIEG");
sort.setSort (new SortField ("custom", SampleComparable.getComparatorSource(), true));
assertMatches (full, queryY, sort, "HJDBF");
SortComparator custom = SampleComparable.getComparator();
sort.setSort (new SortField ("custom", custom));
assertMatches (full, queryX, sort, "CAIEG");
sort.setSort (new SortField ("custom", custom, true));
assertMatches (full, queryY, sort, "HJDBF");
| | public void | testEmptyFieldSort()
sort.setSort ("string");
assertMatches (full, queryF, sort, "ZJI");
sort.setSort ("string", true);
assertMatches (full, queryF, sort, "IJZ");
sort.setSort ("int");
assertMatches (full, queryF, sort, "IZJ");
sort.setSort ("int", true);
assertMatches (full, queryF, sort, "JZI");
sort.setSort ("float");
assertMatches (full, queryF, sort, "ZJI");
// using a nonexisting field as first sort key shouldn't make a difference:
sort.setSort (new SortField[] { new SortField ("nosuchfield", SortField.STRING),
new SortField ("float") });
assertMatches (full, queryF, sort, "ZJI");
sort.setSort ("float", true);
assertMatches (full, queryF, sort, "IJZ");
// When a field is null for both documents, the next SortField should be used.
// Works for
sort.setSort (new SortField[] { new SortField ("int"),
new SortField ("string", SortField.STRING),
new SortField ("float") });
assertMatches (full, queryG, sort, "ZWXY");
// Reverse the last criterium to make sure the test didn't pass by chance
sort.setSort (new SortField[] { new SortField ("int"),
new SortField ("string", SortField.STRING),
new SortField ("float", true) });
assertMatches (full, queryG, sort, "ZYXW");
// Do the same for a MultiSearcher
Searcher multiSearcher=new MultiSearcher (new Searchable[] { full });
sort.setSort (new SortField[] { new SortField ("int"),
new SortField ("string", SortField.STRING),
new SortField ("float") });
assertMatches (multiSearcher, queryG, sort, "ZWXY");
sort.setSort (new SortField[] { new SortField ("int"),
new SortField ("string", SortField.STRING),
new SortField ("float", true) });
assertMatches (multiSearcher, queryG, sort, "ZYXW");
// Don't close the multiSearcher. it would close the full searcher too!
// Do the same for a ParallelMultiSearcher
Searcher parallelSearcher=new ParallelMultiSearcher (new Searchable[] { full });
sort.setSort (new SortField[] { new SortField ("int"),
new SortField ("string", SortField.STRING),
new SortField ("float") });
assertMatches (parallelSearcher, queryG, sort, "ZWXY");
sort.setSort (new SortField[] { new SortField ("int"),
new SortField ("string", SortField.STRING),
new SortField ("float", true) });
assertMatches (parallelSearcher, queryG, sort, "ZYXW");
// Don't close the parallelSearcher. it would close the full searcher too!
| | public void | testEmptyIndex()
Searcher empty = getEmptyIndex();
sort = new Sort();
assertMatches (empty, queryX, sort, "");
sort.setSort(SortField.FIELD_DOC);
assertMatches (empty, queryX, sort, "");
sort.setSort (new SortField[] { new SortField ("int", SortField.INT), SortField.FIELD_DOC });
assertMatches (empty, queryX, sort, "");
sort.setSort (new SortField[] { new SortField ("string", SortField.STRING, true), SortField.FIELD_DOC });
assertMatches (empty, queryX, sort, "");
sort.setSort (new SortField[] { new SortField ("float", SortField.FLOAT), new SortField ("string", SortField.STRING) });
assertMatches (empty, queryX, sort, "");
| | public void | testInternationalMultiSearcherSort()
Searcher multiSearcher = new MultiSearcher (new Searchable[] { full });
sort.setSort (new SortField ("i18n", new Locale("sv", "se")));
assertMatches (multiSearcher, queryY, sort, "BJDFH");
sort.setSort (new SortField ("i18n", Locale.US));
assertMatches (multiSearcher, queryY, sort, "BFJDH");
sort.setSort (new SortField ("i18n", new Locale("da", "dk")));
assertMatches (multiSearcher, queryY, sort, "BJDHF");
| | public void | testInternationalSort()
sort.setSort (new SortField ("i18n", Locale.US));
assertMatches (full, queryY, sort, "BFJDH");
sort.setSort (new SortField ("i18n", new Locale("sv", "se")));
assertMatches (full, queryY, sort, "BJDFH");
sort.setSort (new SortField ("i18n", new Locale("da", "dk")));
assertMatches (full, queryY, sort, "BJDHF");
sort.setSort (new SortField ("i18n", Locale.US));
assertMatches (full, queryX, sort, "ECAGI");
sort.setSort (new SortField ("i18n", Locale.FRANCE));
assertMatches (full, queryX, sort, "EACGI");
| | public void | testLocaleSort()
sort.setSort (new SortField[] { new SortField ("string", Locale.US) });
assertMatches (full, queryX, sort, "AIGEC");
assertMatches (full, queryY, sort, "DJHFB");
sort.setSort (new SortField[] { new SortField ("string", Locale.US, true) });
assertMatches (full, queryX, sort, "CEGIA");
assertMatches (full, queryY, sort, "BFHJD");
| | public void | testMultiSort()
MultiSearcher searcher = new MultiSearcher (new Searchable[] { searchX, searchY });
runMultiSorts (searcher);
| | public void | testNormalizedScores()
// capture relevancy scores
HashMap scoresX = getScores (full.search (queryX));
HashMap scoresY = getScores (full.search (queryY));
HashMap scoresA = getScores (full.search (queryA));
// we'll test searching locally, remote and multi
MultiSearcher remote = new MultiSearcher (new Searchable[] { getRemote() });
MultiSearcher multi = new MultiSearcher (new Searchable[] { searchX, searchY });
// change sorting and make sure relevancy stays the same
sort = new Sort();
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
sort.setSort(SortField.FIELD_DOC);
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
sort.setSort ("int");
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
sort.setSort ("float");
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
sort.setSort ("string");
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
sort.setSort (new String[] {"int","float"});
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
sort.setSort (new SortField[] { new SortField ("int", true), new SortField (null, SortField.DOC, true) });
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
sort.setSort (new String[] {"float","string"});
assertSameValues (scoresX, getScores(full.search(queryX,sort)));
assertSameValues (scoresX, getScores(remote.search(queryX,sort)));
assertSameValues (scoresX, getScores(multi.search(queryX,sort)));
assertSameValues (scoresY, getScores(full.search(queryY,sort)));
assertSameValues (scoresY, getScores(remote.search(queryY,sort)));
assertSameValues (scoresY, getScores(multi.search(queryY,sort)));
assertSameValues (scoresA, getScores(full.search(queryA,sort)));
assertSameValues (scoresA, getScores(remote.search(queryA,sort)));
assertSameValues (scoresA, getScores(multi.search(queryA,sort)));
| | public void | testParallelMultiSort()
Searcher searcher = new ParallelMultiSearcher (new Searchable[] { searchX, searchY });
runMultiSorts (searcher);
| | public void | testRemoteCustomSort()
Searchable searcher = getRemote();
MultiSearcher multi = new MultiSearcher (new Searchable[] { searcher });
sort.setSort (new SortField ("custom", SampleComparable.getComparatorSource()));
assertMatches (multi, queryX, sort, "CAIEG");
sort.setSort (new SortField ("custom", SampleComparable.getComparatorSource(), true));
assertMatches (multi, queryY, sort, "HJDBF");
SortComparator custom = SampleComparable.getComparator();
sort.setSort (new SortField ("custom", custom));
assertMatches (multi, queryX, sort, "CAIEG");
sort.setSort (new SortField ("custom", custom, true));
assertMatches (multi, queryY, sort, "HJDBF");
| | public void | testRemoteSort()
Searchable searcher = getRemote();
MultiSearcher multi = new MultiSearcher (new Searchable[] { searcher });
runMultiSorts (multi);
| | public void | testReverseSort()
sort.setSort (new SortField[] { new SortField (null, SortField.SCORE, true), SortField.FIELD_DOC });
assertMatches (full, queryX, sort, "IEGCA");
assertMatches (full, queryY, sort, "JFHDB");
sort.setSort (new SortField (null, SortField.DOC, true));
assertMatches (full, queryX, sort, "IGECA");
assertMatches (full, queryY, sort, "JHFDB");
sort.setSort ("int", true);
assertMatches (full, queryX, sort, "CAEGI");
assertMatches (full, queryY, sort, "BJFHD");
sort.setSort ("float", true);
assertMatches (full, queryX, sort, "AECIG");
assertMatches (full, queryY, sort, "BFJHD");
sort.setSort ("string", true);
assertMatches (full, queryX, sort, "CEGIA");
assertMatches (full, queryY, sort, "BFHJD");
| | public void | testSortCombos()
sort.setSort (new String[] {"int","float"});
assertMatches (full, queryX, sort, "IGEAC");
sort.setSort (new SortField[] { new SortField ("int", true), new SortField (null, SortField.DOC, true) });
assertMatches (full, queryX, sort, "CEAGI");
sort.setSort (new String[] {"float","string"});
assertMatches (full, queryX, sort, "GICEA");
| | public void | testTopDocsScores()
// There was previously a bug in FieldSortedHitQueue.maxscore when only a single
// doc was added. That is what the following tests for.
Sort sort = new Sort();
int nDocs=10;
// try to pick a query that will result in an unnormalized
// score greater than 1 to test for correct normalization
final TopDocs docs1 = full.search(queryE,null,nDocs,sort);
// a filter that only allows through the first hit
Filter filt = new Filter() {
public BitSet bits(IndexReader reader) throws IOException {
BitSet bs = new BitSet(reader.maxDoc());
bs.set(docs1.scoreDocs[0].doc);
return bs;
}
};
TopDocs docs2 = full.search(queryE, filt, nDocs, sort);
assertEquals(docs1.scoreDocs[0].score, docs2.scoreDocs[0].score, 1e-6);
| | public void | testTypedSort()
sort.setSort (new SortField[] { new SortField ("int", SortField.INT), SortField.FIELD_DOC });
assertMatches (full, queryX, sort, "IGAEC");
assertMatches (full, queryY, sort, "DHFJB");
sort.setSort (new SortField[] { new SortField ("float", SortField.FLOAT), SortField.FIELD_DOC });
assertMatches (full, queryX, sort, "GCIEA");
assertMatches (full, queryY, sort, "DHJFB");
sort.setSort (new SortField[] { new SortField ("string", SortField.STRING), SortField.FIELD_DOC });
assertMatches (full, queryX, sort, "AIGEC");
assertMatches (full, queryY, sort, "DJHFB");
|
|
