public virtual void TestAddSameDocTwice()
{
// LUCENE-5367: this was a problem with the previous code, making sure it
// works with the new code.
Directory indexDir = NewDirectory(), taxoDir = NewDirectory();
IndexWriter indexWriter = new IndexWriter(indexDir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())));
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
FacetsConfig facetsConfig = new FacetsConfig();
Document doc = new Document();
doc.Add(new FacetField("a", "b"));
doc = facetsConfig.Build(taxoWriter, doc);
// these two addDocument() used to fail
indexWriter.AddDocument(doc);
indexWriter.AddDocument(doc);
IOUtils.Close(indexWriter, taxoWriter);
DirectoryReader indexReader = DirectoryReader.Open(indexDir);
DirectoryTaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoDir);
IndexSearcher searcher = NewSearcher(indexReader);
FacetsCollector fc = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), fc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, facetsConfig, fc);
FacetResult res = facets.GetTopChildren(10, "a");
Assert.AreEqual(1, res.LabelValues.Length);
Assert.AreEqual(2, res.LabelValues[0].value);
IOUtils.Close(indexReader, taxoReader);
IOUtils.Close(indexDir, taxoDir);
}
public static void BeforeClassCountingFacetsAggregatorTest()
{
indexDir = NewDirectory();
taxoDir = NewDirectory();
// create an index which has:
// 1. Segment with no categories, but matching results
// 2. Segment w/ categories, but no results
// 3. Segment w/ categories and results
// 4. Segment w/ categories, but only some results
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
//conf.MergePolicy = NoMergePolicy.INSTANCE; // prevent merges, so we can control the index segments
IndexWriter indexWriter = new IndexWriter(indexDir, conf);
TaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
allExpectedCounts = newCounts();
termExpectedCounts = newCounts();
// segment w/ no categories
IndexDocsNoFacets(indexWriter);
// segment w/ categories, no content
IndexDocsWithFacetsNoTerms(indexWriter, taxoWriter, allExpectedCounts);
// segment w/ categories and content
IndexDocsWithFacetsAndTerms(indexWriter, taxoWriter, allExpectedCounts);
// segment w/ categories and some content
IndexDocsWithFacetsAndSomeTerms(indexWriter, taxoWriter, allExpectedCounts);
IOUtils.Close(indexWriter, taxoWriter);
}
public ThreadAnonymousInnerClassHelper(TestDirectoryTaxonomyWriter outerInstance, int range, AtomicInteger numCats, ConcurrentDictionary <string, string> values, Lucene.Net.Facet.Taxonomy.Directory.DirectoryTaxonomyWriter tw)
{
this.outerInstance = outerInstance;
this.range = range;
this.numCats = numCats;
this.values = values;
this.tw = tw;
}
public ThreadAnonymousInnerClassHelper(TestConcurrentFacetedIndexing outerInstance, AtomicInt32 numDocs, ConcurrentDictionary <string, string> values, IndexWriter iw, Lucene.Net.Facet.Taxonomy.Directory.DirectoryTaxonomyWriter tw, FacetsConfig config)
{
this.outerInstance = outerInstance;
this.numDocs = numDocs;
this.values = values;
this.iw = iw;
this.tw = tw;
this.config = config;
}
public virtual void TestWithThreads()
{
// LUCENE-5303: OrdinalsCache used the ThreadLocal BinaryDV instead of reader.getCoreCacheKey().
Store.Directory indexDir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
IndexWriter writer = new IndexWriter(indexDir, conf);
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
FacetsConfig config = new FacetsConfig();
Document doc = new Document();
doc.Add(new FacetField("A", "1"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new FacetField("A", "2"));
writer.AddDocument(config.Build(taxoWriter, doc));
var reader = DirectoryReader.Open(writer, true);
CachedOrdinalsReader ordsReader = new CachedOrdinalsReader(new DocValuesOrdinalsReader(FacetsConfig.DEFAULT_INDEX_FIELD_NAME));
ThreadClass[] threads = new ThreadClass[3];
for (int i = 0; i < threads.Length; i++)
{
threads[i] = new ThreadAnonymousInnerClassHelper(this, "CachedOrdsThread-" + i, reader, ordsReader);
}
long ramBytesUsed = 0;
foreach (ThreadClass t in threads)
{
t.Start();
t.Join();
if (ramBytesUsed == 0)
{
ramBytesUsed = ordsReader.RamBytesUsed();
}
else
{
Assert.AreEqual(ramBytesUsed, ordsReader.RamBytesUsed());
}
}
IOUtils.Close(writer, taxoWriter, reader, indexDir, taxoDir);
}
public static void BeforeClass()
{
dir = NewDirectory();
taxoDir = NewDirectory();
// preparations - index, taxonomy, content
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
// Cannot mix ints & floats in the same indexed field:
config = new FacetsConfig();
config.SetIndexFieldName("int", "$facets.int");
config.SetMultiValued("int", true);
config.SetIndexFieldName("float", "$facets.float");
config.SetMultiValued("float", true);
var writer = new RandomIndexWriter(Random(), dir);
// index documents, 50% have only 'b' and all have 'a'
for (int i = 0; i < 110; i++)
{
Document doc = new Document();
// every 11th document is added empty, this used to cause the association
// aggregators to go into an infinite loop
if (i % 11 != 0)
{
doc.Add(new IntAssociationFacetField(2, "int", "a"));
doc.Add(new FloatAssociationFacetField(0.5f, "float", "a"));
if (i % 2 == 0) // 50
{
doc.Add(new IntAssociationFacetField(3, "int", "b"));
doc.Add(new FloatAssociationFacetField(0.2f, "float", "b"));
}
}
writer.AddDocument(config.Build(taxoWriter, doc));
}
taxoWriter.Dispose();
reader = writer.Reader;
writer.Dispose();
taxoReader = new DirectoryTaxonomyReader(taxoDir);
}
private void AssertConsistentYoungestChild(FacetLabel abPath, int abOrd, int abYoungChildBase1, int abYoungChildBase2, int retry, int numCategories)
{
var indexDir = new SlowRAMDirectory(-1, null); // no slowness for intialization
var tw = new DirectoryTaxonomyWriter(indexDir);
tw.AddCategory(new FacetLabel("a", "0"));
tw.AddCategory(abPath);
tw.Commit();
var tr = new DirectoryTaxonomyReader(indexDir);
for (int i = 0; i < numCategories; i++)
{
var cp = new FacetLabel("a", "b", Convert.ToString(i));
tw.AddCategory(cp);
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.GetOrdinal(cp), "Ordinal of " + cp + " must be invalid until Taxonomy Reader was refreshed");
}
tw.Dispose();
var stop = new AtomicBoolean(false);
Exception[] error = new Exception[] { null };
int[] retrieval = new int[] { 0 };
var thread = new ThreadAnonymousInnerClassHelper(this, abPath, abOrd, abYoungChildBase1, abYoungChildBase2, retry, tr, stop, error, retrieval);
thread.Start();
indexDir.SleepMillis = 1; // some delay for refresh
var newTaxoReader = TaxonomyReader.OpenIfChanged(tr);
if (newTaxoReader != null)
{
newTaxoReader.Dispose();
}
stop.Set(true);
thread.Join();
Assert.Null(error[0], "Unexpcted exception at retry " + retry + " retrieval " + retrieval[0] + ": \n" + stackTraceStr(error[0]));
tr.Dispose();
}
public ThreadAnonymousInnerClassHelper(TestDirectoryTaxonomyWriter outerInstance, int range, AtomicInteger numCats, ConcurrentDictionary<string, string> values, Lucene.Net.Facet.Taxonomy.Directory.DirectoryTaxonomyWriter tw)
{
this.outerInstance = outerInstance;
this.range = range;
this.numCats = numCats;
this.values = values;
this.tw = tw;
}
public virtual void TestWrongIndexFieldName()
{
Store.Directory dir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
// Writes facet ords to a separate directory from the
// main index:
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode_e.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetIndexFieldName("a", "$facets2");
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
Document doc = new Document();
doc.Add(new FacetField("a", "foo1"));
writer.AddDocument(config.Build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.Reader);
// NRT open
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector c = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), c);
// Uses default $facets field:
Facets facets;
if (Random().NextBoolean())
{
facets = new FastTaxonomyFacetCounts(taxoReader, config, c);
}
else
{
OrdinalsReader ordsReader = new DocValuesOrdinalsReader();
if (Random().NextBoolean())
{
ordsReader = new CachedOrdinalsReader(ordsReader);
}
facets = new TaxonomyFacetCounts(ordsReader, taxoReader, config, c);
}
// Ask for top 10 labels for any dims that have counts:
IList<FacetResult> results = facets.GetAllDims(10);
Assert.True(results.Count == 0);
try
{
facets.GetSpecificValue("a");
Fail("should have hit exc");
}
catch (System.ArgumentException)
{
// expected
}
try
{
facets.GetTopChildren(10, "a");
Fail("should have hit exc");
}
catch (System.ArgumentException)
{
// expected
}
IOUtils.Close(writer, taxoWriter, searcher.IndexReader, taxoReader, taxoDir, dir);
}
public virtual void TestWriterTwice()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
// run fillTaxonomy again - this will try to add the same categories
// again, and check that we see the same ordinal paths again, not
// different ones.
FillTaxonomy(tw);
// Let's check the number of categories again, to see that no
// extraneous categories were created:
Assert.AreEqual(ExpectedCategories.Length, tw.Size);
tw.Dispose();
indexDir.Dispose();
}
public virtual void TestWriterTwice3()
{
var indexDir = NewDirectory();
// First, create and fill the taxonomy
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
tw.Dispose();
// Now, open the same taxonomy and add the same categories again.
// After a few categories, the LuceneTaxonomyWriter implementation
// will stop looking for each category on disk, and rather read them
// all into memory and close it's reader. The bug was that it closed
// the reader, but forgot that it did (because it didn't set the reader
// reference to null).
tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
// Add one new category, just to make commit() do something:
tw.AddCategory(new FacetLabel("hi"));
// Do a commit(). Here was a bug - if tw had a reader open, it should
// be reopened after the commit. However, in our case the reader should
// not be open (as explained above) but because it was not set to null,
// we forgot that, tried to reopen it, and got an AlreadyClosedException.
tw.Commit();
Assert.AreEqual(ExpectedCategories.Length + 1, tw.Size);
tw.Dispose();
indexDir.Dispose();
}
public virtual void TestWriterCheckPaths2()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
CheckPaths(tw);
FillTaxonomy(tw);
CheckPaths(tw);
tw.Dispose();
tw = new DirectoryTaxonomyWriter(indexDir);
CheckPaths(tw);
FillTaxonomy(tw);
CheckPaths(tw);
tw.Dispose();
indexDir.Dispose();
}
public virtual void TestWriterParent2()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
tw.Commit();
var tr = new DirectoryTaxonomyReader(indexDir);
CheckWriterParent(tr, tw);
tw.Dispose();
tr.Dispose();
indexDir.Dispose();
}
public virtual void TestTaxonomyReaderRefreshRaces()
{
// compute base child arrays - after first chunk, and after the other
var indexDirBase = NewDirectory();
var twBase = new DirectoryTaxonomyWriter(indexDirBase);
twBase.AddCategory(new FacetLabel("a", "0"));
FacetLabel abPath = new FacetLabel("a", "b");
twBase.AddCategory(abPath);
twBase.Commit();
var trBase = new DirectoryTaxonomyReader(indexDirBase);
ParallelTaxonomyArrays ca1 = trBase.ParallelTaxonomyArrays;
int abOrd = trBase.GetOrdinal(abPath);
int abYoungChildBase1 = ca1.Children()[abOrd];
int numCategories = AtLeast(800);
for (int i = 0; i < numCategories; i++)
{
twBase.AddCategory(new FacetLabel("a", "b", Convert.ToString(i)));
}
twBase.Dispose();
var newTaxoReader = TaxonomyReader.OpenIfChanged(trBase);
Assert.NotNull(newTaxoReader);
trBase.Dispose();
trBase = newTaxoReader;
ParallelTaxonomyArrays ca2 = trBase.ParallelTaxonomyArrays;
int abYoungChildBase2 = ca2.Children()[abOrd];
int numRetries = AtLeast(50);
for (int retry = 0; retry < numRetries; retry++)
{
AssertConsistentYoungestChild(abPath, abOrd, abYoungChildBase1, abYoungChildBase2, retry, numCategories);
}
trBase.Dispose();
indexDirBase.Dispose();
}
public virtual void TestSeparateReaderAndWriter()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
tw.Commit();
var tr = new DirectoryTaxonomyReader(indexDir);
Assert.AreEqual(1, tr.Size); // the empty taxonomy has size 1 (the root)
tw.AddCategory(new FacetLabel("Author"));
Assert.AreEqual(1, tr.Size); // still root only...
Assert.Null(TaxonomyReader.OpenIfChanged(tr)); // this is not enough, because tw.Commit() hasn't been done yet
Assert.AreEqual(1, tr.Size); // still root only...
tw.Commit();
Assert.AreEqual(1, tr.Size); // still root only...
var newTaxoReader = TaxonomyReader.OpenIfChanged(tr);
Assert.NotNull(newTaxoReader);
tr.Dispose();
tr = newTaxoReader;
int author = 1;
try
{
Assert.AreEqual(TaxonomyReader.ROOT_ORDINAL, tr.ParallelTaxonomyArrays.Parents()[author]);
// ok
}
catch (System.IndexOutOfRangeException)
{
Fail("After category addition, commit() and refresh(), getParent for " + author + " should NOT throw exception");
}
Assert.AreEqual(2, tr.Size); // finally, see there are two categories
// now, add another category, and verify that after commit and refresh
// the parent of this category is correct (this requires the reader
// to correctly update its prefetched parent vector), and that the
// old information also wasn't ruined:
tw.AddCategory(new FacetLabel("Author", "Richard Dawkins"));
int dawkins = 2;
tw.Commit();
newTaxoReader = TaxonomyReader.OpenIfChanged(tr);
Assert.NotNull(newTaxoReader);
tr.Dispose();
tr = newTaxoReader;
int[] parents = tr.ParallelTaxonomyArrays.Parents();
Assert.AreEqual(author, parents[dawkins]);
Assert.AreEqual(TaxonomyReader.ROOT_ORDINAL, parents[author]);
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, parents[TaxonomyReader.ROOT_ORDINAL]);
Assert.AreEqual(3, tr.Size);
tw.Dispose();
tr.Dispose();
indexDir.Dispose();
}
public virtual void TestBasic()
{
Store.Directory dir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
// Writes facet ords to a separate directory from the
// main index:
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode_e.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetHierarchical("Publish Date", true);
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
Document doc = new Document();
doc.Add(new FacetField("Author", "Bob"));
doc.Add(new FacetField("Publish Date", "2010", "10", "15"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new FacetField("Author", "Lisa"));
doc.Add(new FacetField("Publish Date", "2010", "10", "20"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new FacetField("Author", "Lisa"));
doc.Add(new FacetField("Publish Date", "2012", "1", "1"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new FacetField("Author", "Susan"));
doc.Add(new FacetField("Publish Date", "2012", "1", "7"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new FacetField("Author", "Frank"));
doc.Add(new FacetField("Publish Date", "1999", "5", "5"));
writer.AddDocument(config.Build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.Reader);
// NRT open
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
// Aggregate the facet counts:
FacetsCollector c = new FacetsCollector();
// MatchAllDocsQuery is for "browsing" (counts facets
// for all non-deleted docs in the index); normally
// you'd use a "normal" query, and use MultiCollector to
// wrap collecting the "normal" hits and also facets:
searcher.Search(new MatchAllDocsQuery(), c);
Facets facets = new FastTaxonomyFacetCounts(taxoReader, config, c);
// Retrieve & verify results:
Assert.AreEqual("dim=Publish Date path=[] value=5 childCount=3\n 2010 (2)\n 2012 (2)\n 1999 (1)\n", facets.GetTopChildren(10, "Publish Date").ToString());
Assert.AreEqual("dim=Author path=[] value=5 childCount=4\n Lisa (2)\n Bob (1)\n Susan (1)\n Frank (1)\n", facets.GetTopChildren(10, "Author").ToString());
// Now user drills down on Publish Date/2010:
DrillDownQuery q2 = new DrillDownQuery(config);
q2.Add("Publish Date", "2010");
c = new FacetsCollector();
searcher.Search(q2, c);
facets = new FastTaxonomyFacetCounts(taxoReader, config, c);
Assert.AreEqual("dim=Author path=[] value=2 childCount=2\n Bob (1)\n Lisa (1)\n", facets.GetTopChildren(10, "Author").ToString());
Assert.AreEqual(1, facets.GetSpecificValue("Author", "Lisa"));
Assert.Null(facets.GetTopChildren(10, "Non exitent dim"));
// Smoke test PrintTaxonomyStats:
ByteArrayOutputStream bos = new ByteArrayOutputStream();
PrintTaxonomyStats.PrintStats(taxoReader, Console.Out, true);
string result = bos.ToString();
Assert.True(result.IndexOf("/Author: 4 immediate children; 5 total categories", StringComparison.Ordinal) != -1);
Assert.True(result.IndexOf("/Publish Date: 3 immediate children; 12 total categories", StringComparison.Ordinal) != -1);
// Make sure at least a few nodes of the tree came out:
Assert.True(result.IndexOf(" /1999", StringComparison.Ordinal) != -1);
Assert.True(result.IndexOf(" /2012", StringComparison.Ordinal) != -1);
Assert.True(result.IndexOf(" /20", StringComparison.Ordinal) != -1);
IOUtils.Close(writer, taxoWriter, searcher.IndexReader, taxoReader, taxoDir, dir);
}
public virtual void TestDefault()
{
Directory indexDir = NewDirectory();
Directory taxoDir = NewDirectory();
// create and open an index writer
var iw = new RandomIndexWriter(Random(), indexDir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random(), MockTokenizer.WHITESPACE, false)));
// create and open a taxonomy writer
var tw = new DirectoryTaxonomyWriter(taxoDir, OpenMode.CREATE);
var config = Config;
seedIndex(tw, iw, config);
IndexReader ir = iw.Reader;
tw.Commit();
// prepare index reader and taxonomy.
var tr = new DirectoryTaxonomyReader(taxoDir);
// prepare searcher to search against
IndexSearcher searcher = NewSearcher(ir);
FacetsCollector sfc = PerformSearch(tr, ir, searcher);
// Obtain facets results and hand-test them
AssertCorrectResults(GetTaxonomyFacetCounts(tr, config, sfc));
assertOrdinalsExist("$facets", ir);
IOUtils.Close(tr, ir, iw, tw, indexDir, taxoDir);
}
public virtual void TestCustom()
{
Directory indexDir = NewDirectory();
Directory taxoDir = NewDirectory();
// create and open an index writer
RandomIndexWriter iw = new RandomIndexWriter(Random(), indexDir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random(), MockTokenizer.WHITESPACE, false)));
// create and open a taxonomy writer
var tw = new DirectoryTaxonomyWriter(taxoDir, OpenMode.CREATE);
FacetsConfig config = Config;
config.SetIndexFieldName("Author", "$author");
seedIndex(tw, iw, config);
IndexReader ir = iw.Reader;
tw.Commit();
// prepare index reader and taxonomy.
var tr = new DirectoryTaxonomyReader(taxoDir);
// prepare searcher to search against
IndexSearcher searcher = NewSearcher(ir);
FacetsCollector sfc = PerformSearch(tr, ir, searcher);
IDictionary<string, Facets> facetsMap = new Dictionary<string, Facets>();
facetsMap["Author"] = GetTaxonomyFacetCounts(tr, config, sfc, "$author");
Facets facets = new MultiFacets(facetsMap, GetTaxonomyFacetCounts(tr, config, sfc));
// Obtain facets results and hand-test them
AssertCorrectResults(facets);
assertOrdinalsExist("$facets", ir);
assertOrdinalsExist("$author", ir);
IOUtils.Close(tr, ir, iw, tw, indexDir, taxoDir);
}
public virtual void TestSeparateReaderAndWriter2()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
tw.Commit();
var tr = new DirectoryTaxonomyReader(indexDir);
// Test getOrdinal():
FacetLabel author = new FacetLabel("Author");
Assert.AreEqual(1, tr.Size); // the empty taxonomy has size 1 (the root)
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.GetOrdinal(author));
tw.AddCategory(author);
// before commit and refresh, no change:
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.GetOrdinal(author));
Assert.AreEqual(1, tr.Size); // still root only...
Assert.Null(TaxonomyReader.OpenIfChanged(tr)); // this is not enough, because tw.Commit() hasn't been done yet
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.GetOrdinal(author));
Assert.AreEqual(1, tr.Size); // still root only...
tw.Commit();
// still not enough before refresh:
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.GetOrdinal(author));
Assert.AreEqual(1, tr.Size); // still root only...
var newTaxoReader = TaxonomyReader.OpenIfChanged(tr);
Assert.NotNull(newTaxoReader);
tr.Dispose();
tr = newTaxoReader;
Assert.AreEqual(1, tr.GetOrdinal(author));
Assert.AreEqual(2, tr.Size);
tw.Dispose();
tr.Dispose();
indexDir.Dispose();
}
public virtual void TestChildrenArrays()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
tw.Dispose();
var tr = new DirectoryTaxonomyReader(indexDir);
ParallelTaxonomyArrays ca = tr.ParallelTaxonomyArrays;
int[] youngestChildArray = ca.Children();
Assert.AreEqual(tr.Size, youngestChildArray.Length);
int[] olderSiblingArray = ca.Siblings();
Assert.AreEqual(tr.Size, olderSiblingArray.Length);
for (int i = 0; i < ExpectedCategories.Length; i++)
{
// find expected children by looking at all expectedCategories
// for children
List<int?> expectedChildren = new List<int?>();
for (int j = ExpectedCategories.Length - 1; j >= 0; j--)
{
if (ExpectedCategories[j].Length != ExpectedCategories[i].Length + 1)
{
continue; // not longer by 1, so can't be a child
}
bool ischild = true;
for (int k = 0; k < ExpectedCategories[i].Length; k++)
{
if (!ExpectedCategories[j][k].Equals(ExpectedCategories[i][k]))
{
ischild = false;
break;
}
}
if (ischild)
{
expectedChildren.Add(j);
}
}
// check that children and expectedChildren are the same, with the
// correct reverse (youngest to oldest) order:
if (expectedChildren.Count == 0)
{
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, youngestChildArray[i]);
}
else
{
int child = youngestChildArray[i];
Assert.AreEqual((int)expectedChildren[0], child);
for (int j = 1; j < expectedChildren.Count; j++)
{
child = olderSiblingArray[child];
Assert.AreEqual((int)expectedChildren[j], child);
// if child is INVALID_ORDINAL we should stop, but
// AssertEquals would fail in this case anyway.
}
// When we're done comparing, olderSiblingArray should now point
// to INVALID_ORDINAL, saying there are no more children. If it
// doesn't, we found too many children...
Assert.AreEqual(-1, olderSiblingArray[child]);
}
}
tr.Dispose();
indexDir.Dispose();
}
public virtual void TestWriterCheckPaths()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomyCheckPaths(tw);
// Also check TaxonomyWriter.getSize() - see that the taxonomy's size
// is what we expect it to be.
Assert.AreEqual(ExpectedCategories.Length, tw.Size);
tw.Dispose();
indexDir.Dispose();
}
public virtual void TestChildrenArraysGrowth()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
tw.AddCategory(new FacetLabel("hi", "there"));
tw.Commit();
var tr = new DirectoryTaxonomyReader(indexDir);
ParallelTaxonomyArrays ca = tr.ParallelTaxonomyArrays;
Assert.AreEqual(3, tr.Size);
Assert.AreEqual(3, ca.Siblings().Length);
Assert.AreEqual(3, ca.Children().Length);
Assert.True(Arrays.Equals(new int[] { 1, 2, -1 }, ca.Children()));
Assert.True(Arrays.Equals(new int[] { -1, -1, -1 }, ca.Siblings()));
tw.AddCategory(new FacetLabel("hi", "ho"));
tw.AddCategory(new FacetLabel("hello"));
tw.Commit();
// Before refresh, nothing changed..
ParallelTaxonomyArrays newca = tr.ParallelTaxonomyArrays;
Assert.AreSame(newca, ca); // we got exactly the same object
Assert.AreEqual(3, tr.Size);
Assert.AreEqual(3, ca.Siblings().Length);
Assert.AreEqual(3, ca.Children().Length);
// After the refresh, things change:
var newtr = TaxonomyReader.OpenIfChanged(tr);
Assert.NotNull(newtr);
tr.Dispose();
tr = newtr;
ca = tr.ParallelTaxonomyArrays;
Assert.AreEqual(5, tr.Size);
Assert.AreEqual(5, ca.Siblings().Length);
Assert.AreEqual(5, ca.Children().Length);
Assert.True(Arrays.Equals(new int[] { 4, 3, -1, -1, -1 }, ca.Children()));
Assert.True(Arrays.Equals(new int[] { -1, -1, -1, 2, 1 }, ca.Siblings()));
tw.Dispose();
tr.Dispose();
indexDir.Dispose();
}
public virtual void TestWriterLock()
{
// native fslock impl gets angry if we use it, so use RAMDirectory explicitly.
var indexDir = new RAMDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
tw.AddCategory(new FacetLabel("hi", "there"));
tw.Commit();
// we deliberately not close the write now, and keep it open and
// locked.
// Verify that the writer worked:
var tr = new DirectoryTaxonomyReader(indexDir);
Assert.AreEqual(2, tr.GetOrdinal(new FacetLabel("hi", "there")));
// Try to open a second writer, with the first one locking the directory.
// We expect to get a LockObtainFailedException.
try
{
Assert.Null(new DirectoryTaxonomyWriter(indexDir));
Fail("should have failed to write in locked directory");
}
catch (LockObtainFailedException)
{
// this is what we expect to happen.
}
// Remove the lock, and now the open should succeed, and we can
// write to the new writer.
DirectoryTaxonomyWriter.Unlock(indexDir);
var tw2 = new DirectoryTaxonomyWriter(indexDir);
tw2.AddCategory(new FacetLabel("hey"));
tw2.Dispose();
// See that the writer indeed wrote:
var newtr = TaxonomyReader.OpenIfChanged(tr);
Assert.NotNull(newtr);
tr.Dispose();
tr = newtr;
Assert.AreEqual(3, tr.GetOrdinal(new FacetLabel("hey")));
tr.Dispose();
tw.Dispose();
indexDir.Dispose();
}
public virtual void TestChildrenArraysInvariants()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
tw.Dispose();
var tr = new DirectoryTaxonomyReader(indexDir);
ParallelTaxonomyArrays ca = tr.ParallelTaxonomyArrays;
int[] children = ca.Children();
Assert.AreEqual(tr.Size, children.Length);
int[] olderSiblingArray = ca.Siblings();
Assert.AreEqual(tr.Size, olderSiblingArray.Length);
// test that the "youngest child" of every category is indeed a child:
int[] parents = tr.ParallelTaxonomyArrays.Parents();
for (int i = 0; i < tr.Size; i++)
{
int youngestChild = children[i];
if (youngestChild != TaxonomyReader.INVALID_ORDINAL)
{
Assert.AreEqual(i, parents[youngestChild]);
}
}
// test that the "older sibling" of every category is indeed older (lower)
// (it can also be INVALID_ORDINAL, which is lower than any ordinal)
for (int i = 0; i < tr.Size; i++)
{
Assert.True(olderSiblingArray[i] < i, "olderSiblingArray[" + i + "] should be <" + i);
}
// test that the "older sibling" of every category is indeed a sibling
// (they share the same parent)
for (int i = 0; i < tr.Size; i++)
{
int sibling = olderSiblingArray[i];
if (sibling == TaxonomyReader.INVALID_ORDINAL)
{
continue;
}
Assert.AreEqual(parents[i], parents[sibling]);
}
// And now for slightly more complex (and less "invariant-like"...)
// tests:
// test that the "youngest child" is indeed the youngest (so we don't
// miss the first children in the chain)
for (int i = 0; i < tr.Size; i++)
{
// Find the really youngest child:
int j;
for (j = tr.Size - 1; j > i; j--)
{
if (parents[j] == i)
{
break; // found youngest child
}
}
if (j == i) // no child found
{
j = TaxonomyReader.INVALID_ORDINAL;
}
Assert.AreEqual(j, children[i]);
}
// test that the "older sibling" is indeed the least oldest one - and
// not a too old one or -1 (so we didn't miss some children in the
// middle or the end of the chain).
for (int i = 0; i < tr.Size; i++)
{
// Find the youngest older sibling:
int j;
for (j = i - 1; j >= 0; j--)
{
if (parents[j] == parents[i])
{
break; // found youngest older sibling
}
}
if (j < 0) // no sibling found
{
j = TaxonomyReader.INVALID_ORDINAL;
}
Assert.AreEqual(j, olderSiblingArray[i]);
}
tr.Dispose();
indexDir.Dispose();
}
public virtual void TestWriterSimpler()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
Assert.AreEqual(1, tw.Size); // the root only
// Test that adding a new top-level category works
Assert.AreEqual(1, tw.AddCategory(new FacetLabel("a")));
Assert.AreEqual(2, tw.Size);
// Test that adding the same category again is noticed, and the
// same ordinal (and not a new one) is returned.
Assert.AreEqual(1, tw.AddCategory(new FacetLabel("a")));
Assert.AreEqual(2, tw.Size);
// Test that adding another top-level category returns a new ordinal,
// not the same one
Assert.AreEqual(2, tw.AddCategory(new FacetLabel("b")));
Assert.AreEqual(3, tw.Size);
// Test that adding a category inside one of the above adds just one
// new ordinal:
Assert.AreEqual(3, tw.AddCategory(new FacetLabel("a", "c")));
Assert.AreEqual(4, tw.Size);
// Test that adding the same second-level category doesn't do anything:
Assert.AreEqual(3, tw.AddCategory(new FacetLabel("a", "c")));
Assert.AreEqual(4, tw.Size);
// Test that adding a second-level category with two new components
// indeed adds two categories
Assert.AreEqual(5, tw.AddCategory(new FacetLabel("d", "e")));
Assert.AreEqual(6, tw.Size);
// Verify that the parents were added above in the order we expected
Assert.AreEqual(4, tw.AddCategory(new FacetLabel("d")));
// Similar, but inside a category that already exists:
Assert.AreEqual(7, tw.AddCategory(new FacetLabel("b", "d", "e")));
Assert.AreEqual(8, tw.Size);
// And now inside two levels of categories that already exist:
Assert.AreEqual(8, tw.AddCategory(new FacetLabel("b", "d", "f")));
Assert.AreEqual(9, tw.Size);
tw.Dispose();
indexDir.Dispose();
}
public virtual void TestNrt()
{
var dir = NewDirectory();
var writer = new DirectoryTaxonomyWriter(dir);
var reader = new DirectoryTaxonomyReader(writer);
FacetLabel cp = new FacetLabel("a");
writer.AddCategory(cp);
var newReader = TaxonomyReader.OpenIfChanged(reader);
Assert.NotNull(newReader, "expected a new instance");
Assert.AreEqual(2, newReader.Size);
Assert.AreNotSame(TaxonomyReader.INVALID_ORDINAL, newReader.GetOrdinal(cp));
reader.Dispose();
reader = newReader;
writer.Dispose();
reader.Dispose();
dir.Dispose();
}
public virtual void TestWriterTwice2()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
tw.Dispose();
tw = new DirectoryTaxonomyWriter(indexDir);
// run fillTaxonomy again - this will try to add the same categories
// again, and check that we see the same ordinals again, not different
// ones, and that the number of categories hasn't grown by the new
// additions
FillTaxonomy(tw);
Assert.AreEqual(ExpectedCategories.Length, tw.Size);
tw.Dispose();
indexDir.Dispose();
}
public virtual void TestReaderBasic()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
tw.Dispose();
var tr = new DirectoryTaxonomyReader(indexDir);
// test TaxonomyReader.getSize():
Assert.AreEqual(ExpectedCategories.Length, tr.Size);
// test round trips of ordinal => category => ordinal
for (int i = 0; i < tr.Size; i++)
{
Assert.AreEqual(i, tr.GetOrdinal(tr.GetPath(i)));
}
// test TaxonomyReader.getCategory():
for (int i = 1; i < tr.Size; i++)
{
FacetLabel expectedCategory = new FacetLabel(ExpectedCategories[i]);
FacetLabel category = tr.GetPath(i);
if (!expectedCategory.Equals(category))
{
Fail("For ordinal " + i + " expected category " + Showcat(expectedCategory) + ", but got " + Showcat(category));
}
}
// (also test invalid ordinals:)
Assert.Null(tr.GetPath(-1));
Assert.Null(tr.GetPath(tr.Size));
Assert.Null(tr.GetPath(TaxonomyReader.INVALID_ORDINAL));
// test TaxonomyReader.GetOrdinal():
for (int i = 1; i < ExpectedCategories.Length; i++)
{
int expectedOrdinal = i;
int ordinal = tr.GetOrdinal(new FacetLabel(ExpectedCategories[i]));
if (expectedOrdinal != ordinal)
{
Fail("For category " + Showcat(ExpectedCategories[i]) + " expected ordinal " + expectedOrdinal + ", but got " + ordinal);
}
}
// (also test invalid categories:)
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.GetOrdinal(new FacetLabel("non-existant")));
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.GetOrdinal(new FacetLabel("Author", "Jules Verne")));
tr.Dispose();
indexDir.Dispose();
}
public virtual void TestRandomSampling()
{
Directory dir = NewDirectory();
Directory taxoDir = NewDirectory();
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
FacetsConfig config = new FacetsConfig();
int numDocs = AtLeast(10000);
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
doc.Add(new StringField("EvenOdd", (i % 2 == 0) ? "even" : "odd", Store.NO));
doc.Add(new FacetField("iMod10", Convert.ToString(i % 10)));
writer.AddDocument(config.Build(taxoWriter, doc));
}
Random random = Random();
// NRT open
IndexSearcher searcher = NewSearcher(writer.Reader);
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
IOUtils.Close(writer, taxoWriter);
// Test empty results
RandomSamplingFacetsCollector collectRandomZeroResults = new RandomSamplingFacetsCollector(numDocs / 10, random.NextLong());
// There should be no divisions by zero
searcher.Search(new TermQuery(new Term("EvenOdd", "NeverMatches")), collectRandomZeroResults);
// There should be no divisions by zero and no null result
Assert.NotNull(collectRandomZeroResults.GetMatchingDocs);
// There should be no results at all
foreach (MatchingDocs doc in collectRandomZeroResults.GetMatchingDocs)
{
Assert.AreEqual(0, doc.totalHits);
}
// Now start searching and retrieve results.
// Use a query to select half of the documents.
TermQuery query = new TermQuery(new Term("EvenOdd", "even"));
// there will be 5 facet values (0, 2, 4, 6 and 8), as only the even (i %
// 10) are hits.
// there is a REAL small chance that one of the 5 values will be missed when
// sampling.
// but is that 0.8 (chance not to take a value) ^ 2000 * 5 (any can be
// missing) ~ 10^-193
// so that is probably not going to happen.
int maxNumChildren = 5;
RandomSamplingFacetsCollector random100Percent = new RandomSamplingFacetsCollector(numDocs, random.NextLong()); // no sampling
RandomSamplingFacetsCollector random10Percent = new RandomSamplingFacetsCollector(numDocs / 10, random.NextLong()); // 10 % of total docs, 20% of the hits
FacetsCollector fc = new FacetsCollector();
searcher.Search(query, MultiCollector.Wrap(fc, random100Percent, random10Percent));
FastTaxonomyFacetCounts random10FacetCounts = new FastTaxonomyFacetCounts(taxoReader, config, random10Percent);
FastTaxonomyFacetCounts random100FacetCounts = new FastTaxonomyFacetCounts(taxoReader, config, random100Percent);
FastTaxonomyFacetCounts exactFacetCounts = new FastTaxonomyFacetCounts(taxoReader, config, fc);
FacetResult random10Result = random10Percent.AmortizeFacetCounts(random10FacetCounts.GetTopChildren(10, "iMod10"), config, searcher);
FacetResult random100Result = random100FacetCounts.GetTopChildren(10, "iMod10");
FacetResult exactResult = exactFacetCounts.GetTopChildren(10, "iMod10");
Assert.AreEqual(random100Result, exactResult);
// we should have five children, but there is a small chance we have less.
// (see above).
Assert.True(random10Result.ChildCount <= maxNumChildren);
// there should be one child at least.
Assert.True(random10Result.ChildCount >= 1);
// now calculate some statistics to determine if the sampled result is 'ok'.
// because random sampling is used, the results will vary each time.
int sum = 0;
foreach (LabelAndValue lav in random10Result.LabelValues)
{
sum += (int)lav.value;
}
float mu = (float)sum / (float)maxNumChildren;
float variance = 0;
foreach (LabelAndValue lav in random10Result.LabelValues)
{
variance += (float)Math.Pow((mu - (int)lav.value), 2);
}
variance = variance / maxNumChildren;
float sigma = (float)Math.Sqrt(variance);
// we query only half the documents and have 5 categories. The average
// number of docs in a category will thus be the total divided by 5*2
float targetMu = numDocs / (5.0f * 2.0f);
// the average should be in the range and the standard deviation should not
// be too great
Assert.True(sigma < 200);
Assert.True(targetMu - 3 * sigma < mu && mu < targetMu + 3 * sigma);
IOUtils.Close(searcher.IndexReader, taxoReader, dir, taxoDir);
}
public virtual void TestReaderParent()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
FillTaxonomy(tw);
tw.Dispose();
var tr = new DirectoryTaxonomyReader(indexDir);
// check that the parent of the root ordinal is the invalid ordinal:
int[] parents = tr.ParallelTaxonomyArrays.Parents();
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, parents[0]);
// check parent of non-root ordinals:
for (int ordinal = 1; ordinal < tr.Size; ordinal++)
{
FacetLabel me = tr.GetPath(ordinal);
int parentOrdinal = parents[ordinal];
FacetLabel parent = tr.GetPath(parentOrdinal);
if (parent == null)
{
Fail("Parent of " + ordinal + " is " + parentOrdinal + ", but this is not a valid category.");
}
// verify that the parent is indeed my parent, according to the strings
if (!me.Subpath(me.Length - 1).Equals(parent))
{
Fail("Got parent " + parentOrdinal + " for ordinal " + ordinal + " but categories are " + Showcat(parent) + " and " + Showcat(me) + " respectively.");
}
}
tr.Dispose();
indexDir.Dispose();
}
public ThreadAnonymousInnerClassHelper2(TestAddTaxonomy outerInstance, int numCategories, Lucene.Net.Facet.Taxonomy.Directory.DirectoryTaxonomyWriter destTW)
{
this.outerInstance = outerInstance;
this.numCategories = numCategories;
this.destTW = destTW;
}
public virtual void TestRootOnly()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
// right after opening the index, it should already contain the
// root, so have size 1:
Assert.AreEqual(1, tw.Size);
tw.Dispose();
var tr = new DirectoryTaxonomyReader(indexDir);
Assert.AreEqual(1, tr.Size);
Assert.AreEqual(0, tr.GetPath(0).Length);
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.ParallelTaxonomyArrays.Parents()[0]);
Assert.AreEqual(0, tr.GetOrdinal(new FacetLabel()));
tr.Dispose(true);
indexDir.Dispose();
}
public static void BeforeClassDrillDownQueryTest()
{
dir = NewDirectory();
Random r = Random();
RandomIndexWriter writer = new RandomIndexWriter(r, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(r, MockTokenizer.KEYWORD, false)));
taxoDir = NewDirectory();
TaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
config = new FacetsConfig();
// Randomize the per-dim config:
config.SetHierarchical("a", Random().NextBoolean());
config.SetMultiValued("a", Random().NextBoolean());
if (Random().NextBoolean())
{
config.SetIndexFieldName("a", "$a");
}
config.SetRequireDimCount("a", true);
config.SetHierarchical("b", Random().NextBoolean());
config.SetMultiValued("b", Random().NextBoolean());
if (Random().NextBoolean())
{
config.SetIndexFieldName("b", "$b");
}
config.SetRequireDimCount("b", true);
for (int i = 0; i < 100; i++)
{
Document doc = new Document();
if (i % 2 == 0) // 50
{
doc.Add(new TextField("content", "foo", Field.Store.NO));
}
if (i % 3 == 0) // 33
{
doc.Add(new TextField("content", "bar", Field.Store.NO));
}
if (i % 4 == 0) // 25
{
if (r.NextBoolean())
{
doc.Add(new FacetField("a", "1"));
}
else
{
doc.Add(new FacetField("a", "2"));
}
}
if (i % 5 == 0) // 20
{
doc.Add(new FacetField("b", "1"));
}
writer.AddDocument(config.Build(taxoWriter, doc));
}
taxoWriter.Dispose();
reader = writer.Reader;
writer.Dispose();
taxo = new DirectoryTaxonomyReader(taxoDir);
}
public virtual void TestRootOnly2()
{
var indexDir = NewDirectory();
var tw = new DirectoryTaxonomyWriter(indexDir);
tw.Commit();
var tr = new DirectoryTaxonomyReader(indexDir);
Assert.AreEqual(1, tr.Size);
Assert.AreEqual(0, tr.GetPath(0).Length);
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, tr.ParallelTaxonomyArrays.Parents()[0]);
Assert.AreEqual(0, tr.GetOrdinal(new FacetLabel()));
tw.Dispose();
tr.Dispose(true);
indexDir.Dispose();
}
