public virtual void TestRecreateAndRefresh()
{
// DirTaxoWriter lost the INDEX_EPOCH property if it was opened in
// CREATE_OR_APPEND (or commit(userData) called twice), which could lead to
// DirTaxoReader succeeding to refresh().
Directory dir = NewDirectory();
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE_OR_APPEND, NO_OP_CACHE);
TouchTaxo(taxoWriter, new FacetLabel("a"));
var taxoReader = new DirectoryTaxonomyReader(dir);
TouchTaxo(taxoWriter, new FacetLabel("b"));
var newtr = TaxonomyReader.OpenIfChanged(taxoReader);
taxoReader.Dispose();
taxoReader = newtr;
Assert.AreEqual(1, Convert.ToInt32(taxoReader.CommitUserData[DirectoryTaxonomyWriter.INDEX_EPOCH]));
// now recreate the taxonomy, and check that the epoch is preserved after opening DirTW again.
taxoWriter.Dispose();
taxoWriter = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE, NO_OP_CACHE);
TouchTaxo(taxoWriter, new FacetLabel("c"));
taxoWriter.Dispose();
taxoWriter = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE_OR_APPEND, NO_OP_CACHE);
TouchTaxo(taxoWriter, new FacetLabel("d"));
taxoWriter.Dispose();
newtr = TaxonomyReader.OpenIfChanged(taxoReader);
taxoReader.Dispose();
taxoReader = newtr;
Assert.AreEqual(2, Convert.ToInt32(taxoReader.CommitUserData[DirectoryTaxonomyWriter.INDEX_EPOCH]));
taxoReader.Dispose();
dir.Dispose();
}
public virtual void TestConcurrency()
{
// tests that addTaxonomy and addCategory work in parallel
int numCategories = AtLeast(10000);
// build an input taxonomy index
Directory src = NewDirectory();
var tw = new DirectoryTaxonomyWriter(src);
for (int i = 0; i < numCategories; i++)
{
tw.AddCategory(new FacetLabel("a", Convert.ToString(i)));
}
tw.Dispose();
// now add the taxonomy to an empty taxonomy, while adding the categories
// again, in parallel -- in the end, no duplicate categories should exist.
Directory dest = NewDirectory();
var destTw = new DirectoryTaxonomyWriter(dest);
var t = new ThreadAnonymousInnerClassHelper2(this, numCategories, destTw);
t.Start();
IOrdinalMap map = new MemoryOrdinalMap();
destTw.AddTaxonomy(src, map);
t.Join();
destTw.Dispose();
// now validate
var dtr = new DirectoryTaxonomyReader(dest);
// +2 to account for the root category + "a"
Assert.AreEqual(numCategories + 2, dtr.Count);
var categories = new JCG.HashSet <FacetLabel>();
for (int i = 1; i < dtr.Count; i++)
{
FacetLabel cat = dtr.GetPath(i);
Assert.True(categories.Add(cat), "category " + cat + " already existed");
}
dtr.Dispose();
IOUtils.Dispose(src, dest);
}
public virtual void TestCloseAfterIncRef()
{
Directory dir = NewDirectory();
var ltw = new DirectoryTaxonomyWriter(dir);
ltw.AddCategory(new FacetLabel("a"));
ltw.Dispose();
DirectoryTaxonomyReader ltr = new DirectoryTaxonomyReader(dir);
ltr.IncRef();
ltr.Dispose();
// should not fail as we IncRef() before close
var tmpSie = ltr.Size;
ltr.DecRef();
dir.Dispose();
}
public virtual void TestOpenIfChangedReuse()
{
// test the reuse of data from the old DTR instance
foreach (bool nrt in new bool[] { false, true })
{
Directory dir = NewDirectory();
DirectoryTaxonomyWriter writer = new DirectoryTaxonomyWriter(dir);
FacetLabel cp_a = new FacetLabel("a");
writer.AddCategory(cp_a);
if (!nrt)
{
writer.Commit();
}
DirectoryTaxonomyReader r1 = nrt ? new DirectoryTaxonomyReader(writer) : new DirectoryTaxonomyReader(dir);
// fill r1's caches
Assert.AreEqual(1, r1.GetOrdinal(cp_a));
Assert.AreEqual(cp_a, r1.GetPath(1));
FacetLabel cp_b = new FacetLabel("b");
writer.AddCategory(cp_b);
if (!nrt)
{
writer.Commit();
}
DirectoryTaxonomyReader r2 = TaxonomyReader.OpenIfChanged(r1);
Assert.IsNotNull(r2);
// add r2's categories to the caches
Assert.AreEqual(2, r2.GetOrdinal(cp_b));
Assert.AreEqual(cp_b, r2.GetPath(2));
// check that r1 doesn't see cp_b
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, r1.GetOrdinal(cp_b));
Assert.IsNull(r1.GetPath(2));
r1.Dispose();
r2.Dispose();
writer.Dispose();
dir.Dispose();
}
}
public virtual void TestOpenIfChangedReuseAfterRecreate()
{
// tests that if the taxonomy is recreated, no data is reused from the previous taxonomy
Directory dir = NewDirectory();
DirectoryTaxonomyWriter writer = new DirectoryTaxonomyWriter(dir);
FacetLabel cp_a = new FacetLabel("a");
writer.AddCategory(cp_a);
writer.Dispose();
DirectoryTaxonomyReader r1 = new DirectoryTaxonomyReader(dir);
// fill r1's caches
Assert.AreEqual(1, r1.GetOrdinal(cp_a));
Assert.AreEqual(cp_a, r1.GetPath(1));
// now recreate, add a different category
writer = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE);
FacetLabel cp_b = new FacetLabel("b");
writer.AddCategory(cp_b);
writer.Dispose();
DirectoryTaxonomyReader r2 = TaxonomyReader.OpenIfChanged(r1);
Assert.IsNotNull(r2);
// fill r2's caches
Assert.AreEqual(1, r2.GetOrdinal(cp_b));
Assert.AreEqual(cp_b, r2.GetPath(1));
// check that r1 doesn't see cp_b
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, r1.GetOrdinal(cp_b));
Assert.AreEqual(cp_a, r1.GetPath(1));
// check that r2 doesn't see cp_a
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, r2.GetOrdinal(cp_a));
Assert.AreEqual(cp_b, r2.GetPath(1));
r2.Dispose();
r1.Dispose();
dir.Dispose();
}
public virtual void TestCloseAfterIncRef()
{
Directory dir = NewDirectory();
var ltw = new DirectoryTaxonomyWriter(dir);
ltw.AddCategory(new FacetLabel("a"));
ltw.Dispose();
DirectoryTaxonomyReader ltr = new DirectoryTaxonomyReader(dir);
ltr.IncRef();
ltr.Dispose();
// should not fail as we IncRef() before close
var _ = ltr.Count;
ltr.DecRef();
dir.Dispose();
}
public virtual void TestAlreadyClosed()
{
Directory dir = NewDirectory();
var ltw = new DirectoryTaxonomyWriter(dir);
ltw.AddCategory(new FacetLabel("a"));
ltw.Dispose();
var ltr = new DirectoryTaxonomyReader(dir);
ltr.Dispose();
try
{
var tmpSize = ltr.Size;
Fail("An AlreadyClosedException should have been thrown here");
}
catch (AlreadyClosedException)
{
// good!
}
dir.Dispose();
}
public virtual void TestBackwardsCompatibility()
{
// tests that if the taxonomy index doesn't have the INDEX_EPOCH
// property (supports pre-3.6 indexes), all still works.
Directory dir = NewDirectory();
// create an empty index first, so that DirTaxoWriter initializes indexEpoch to 1.
new IndexWriter(dir, new IndexWriterConfig(TEST_VERSION_CURRENT, null)).Dispose();
var taxoWriter = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE_OR_APPEND, NO_OP_CACHE);
taxoWriter.Dispose();
var taxoReader = new DirectoryTaxonomyReader(dir);
Assert.AreEqual(1, Convert.ToInt32(taxoReader.CommitUserData[DirectoryTaxonomyWriter.INDEX_EPOCH], CultureInfo.InvariantCulture));
Assert.IsNull(TaxonomyReader.OpenIfChanged(taxoReader));
taxoReader.Dispose();
dir.Dispose();
}
public virtual void TestAlreadyClosed()
{
Directory dir = NewDirectory();
var ltw = new DirectoryTaxonomyWriter(dir);
ltw.AddCategory(new FacetLabel("a"));
ltw.Dispose();
var ltr = new DirectoryTaxonomyReader(dir);
ltr.Dispose();
try
{
var _ = ltr.Count;
fail("An ObjectDisposedException should have been thrown here");
}
catch (ObjectDisposedException)
{
// good!
}
dir.Dispose();
}
public virtual void TestGetChildren()
{
Directory dir = NewDirectory();
var taxoWriter = new DirectoryTaxonomyWriter(dir);
int numCategories = AtLeast(10);
int numA = 0, numB = 0;
Random random = Random;
// add the two categories for which we'll also add children (so asserts are simpler)
taxoWriter.AddCategory(new FacetLabel("a"));
taxoWriter.AddCategory(new FacetLabel("b"));
for (int i = 0; i < numCategories; i++)
{
if (random.NextBoolean())
{
taxoWriter.AddCategory(new FacetLabel("a", Convert.ToString(i, CultureInfo.InvariantCulture)));
++numA;
}
else
{
taxoWriter.AddCategory(new FacetLabel("b", Convert.ToString(i, CultureInfo.InvariantCulture)));
++numB;
}
}
// add category with no children
taxoWriter.AddCategory(new FacetLabel("c"));
taxoWriter.Dispose();
var taxoReader = new DirectoryTaxonomyReader(dir);
// non existing category
TaxonomyReader.ChildrenEnumerator it = taxoReader.GetChildren(taxoReader.GetOrdinal(new FacetLabel("invalid")));
Assert.AreEqual(false, it.MoveNext());
// a category with no children
it = taxoReader.GetChildren(taxoReader.GetOrdinal(new FacetLabel("c")));
Assert.AreEqual(false, it.MoveNext());
// arbitrary negative ordinal
it = taxoReader.GetChildren(-2);
Assert.AreEqual(false, it.MoveNext());
// root's children
var roots = new JCG.HashSet <string> {
"a", "b", "c"
};
it = taxoReader.GetChildren(TaxonomyReader.ROOT_ORDINAL);
while (roots.Count > 0)
{
it.MoveNext();
FacetLabel root = taxoReader.GetPath(it.Current);
Assert.AreEqual(1, root.Length);
Assert.IsTrue(roots.Remove(root.Components[0]));
}
Assert.AreEqual(false, it.MoveNext());
for (int i = 0; i < 2; i++)
{
FacetLabel cp = i == 0 ? new FacetLabel("a") : new FacetLabel("b");
int ordinal = taxoReader.GetOrdinal(cp);
it = taxoReader.GetChildren(ordinal);
int numChildren = 0;
int child;
while (it.MoveNext())
{
child = it.Current;
FacetLabel path = taxoReader.GetPath(child);
Assert.AreEqual(2, path.Length);
Assert.AreEqual(path.Components[0], i == 0 ? "a" : "b");
++numChildren;
}
int expected = i == 0 ? numA : numB;
Assert.AreEqual(expected, numChildren, "invalid num children");
}
taxoReader.Dispose();
dir.Dispose();
}
public virtual void TestGetChildren()
{
Directory dir = NewDirectory();
var taxoWriter = new DirectoryTaxonomyWriter(dir);
int numCategories = AtLeast(10);
int numA = 0, numB = 0;
Random random = Random();
// add the two categories for which we'll also add children (so asserts are simpler)
taxoWriter.AddCategory(new FacetLabel("a"));
taxoWriter.AddCategory(new FacetLabel("b"));
for (int i = 0; i < numCategories; i++)
{
if (random.NextBoolean())
{
taxoWriter.AddCategory(new FacetLabel("a", Convert.ToString(i)));
++numA;
}
else
{
taxoWriter.AddCategory(new FacetLabel("b", Convert.ToString(i)));
++numB;
}
}
// add category with no children
taxoWriter.AddCategory(new FacetLabel("c"));
taxoWriter.Dispose();
var taxoReader = new DirectoryTaxonomyReader(dir);
// non existing category
TaxonomyReader.ChildrenIterator it = taxoReader.GetChildren(taxoReader.GetOrdinal(new FacetLabel("invalid")));
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, it.Next());
// a category with no children
it = taxoReader.GetChildren(taxoReader.GetOrdinal(new FacetLabel("c")));
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, it.Next());
// arbitrary negative ordinal
it = taxoReader.GetChildren(-2);
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, it.Next());
// root's children
var roots = new HashSet<string>(Arrays.AsList("a", "b", "c"));
it = taxoReader.GetChildren(TaxonomyReader.ROOT_ORDINAL);
while (roots.Count > 0)
{
FacetLabel root = taxoReader.GetPath(it.Next());
Assert.AreEqual(1, root.Length);
Assert.True(roots.Remove(root.Components[0]));
}
Assert.AreEqual(TaxonomyReader.INVALID_ORDINAL, it.Next());
for (int i = 0; i < 2; i++)
{
FacetLabel cp = i == 0 ? new FacetLabel("a") : new FacetLabel("b");
int ordinal = taxoReader.GetOrdinal(cp);
it = taxoReader.GetChildren(ordinal);
int numChildren = 0;
int child;
while ((child = it.Next()) != TaxonomyReader.INVALID_ORDINAL)
{
FacetLabel path = taxoReader.GetPath(child);
Assert.AreEqual(2, path.Length);
Assert.AreEqual(path.Components[0], i == 0 ? "a" : "b");
++numChildren;
}
int expected = i == 0 ? numA : numB;
Assert.AreEqual(expected, numChildren, "invalid num children");
}
taxoReader.Dispose();
dir.Dispose();
}
public virtual void TestOpenIfChangedAndRefCount()
{
Directory dir = new RAMDirectory(); // no need for random directories here
var taxoWriter = new DirectoryTaxonomyWriter(dir);
taxoWriter.AddCategory(new FacetLabel("a"));
taxoWriter.Commit();
var taxoReader = new DirectoryTaxonomyReader(dir);
Assert.AreEqual(1, taxoReader.RefCount, "wrong refCount");
taxoReader.IncRef();
Assert.AreEqual(2, taxoReader.RefCount, "wrong refCount");
taxoWriter.AddCategory(new FacetLabel("a", "b"));
taxoWriter.Commit();
var newtr = TaxonomyReader.OpenIfChanged(taxoReader);
Assert.NotNull(newtr);
taxoReader.Dispose();
taxoReader = newtr;
Assert.AreEqual(1, taxoReader.RefCount, "wrong refCount");
taxoWriter.Dispose();
taxoReader.Dispose();
dir.Dispose();
}
public virtual void TestOpenIfChangedManySegments()
{
// test openIfChanged() when the taxonomy contains many segments
Directory dir = NewDirectory();
DirectoryTaxonomyWriter writer = new DirectoryTaxonomyWriterAnonymousInnerClassHelper(this, dir);
var reader = new DirectoryTaxonomyReader(writer);
int numRounds = Random().Next(10) + 10;
int numCategories = 1; // one for root
for (int i = 0; i < numRounds; i++)
{
int numCats = Random().Next(4) + 1;
for (int j = 0; j < numCats; j++)
{
writer.AddCategory(new FacetLabel(Convert.ToString(i), Convert.ToString(j)));
}
numCategories += numCats + 1; // one for round-parent
var newtr = TaxonomyReader.OpenIfChanged(reader);
Assert.NotNull(newtr);
reader.Dispose();
reader = newtr;
// assert categories
Assert.AreEqual(numCategories, reader.Size);
int roundOrdinal = reader.GetOrdinal(new FacetLabel(Convert.ToString(i)));
int[] parents = reader.ParallelTaxonomyArrays.Parents();
Assert.AreEqual(0, parents[roundOrdinal]); // round's parent is root
for (int j = 0; j < numCats; j++)
{
int ord = reader.GetOrdinal(new FacetLabel(Convert.ToString(i), Convert.ToString(j)));
Assert.AreEqual(roundOrdinal, parents[ord]); // round's parent is root
}
}
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
public virtual void TestConcurrency()
{
// tests that addTaxonomy and addCategory work in parallel
int numCategories = AtLeast(10000);
// build an input taxonomy index
Directory src = NewDirectory();
var tw = new DirectoryTaxonomyWriter(src);
for (int i = 0; i < numCategories; i++)
{
tw.AddCategory(new FacetLabel("a", Convert.ToString(i)));
}
tw.Dispose();
// now add the taxonomy to an empty taxonomy, while adding the categories
// again, in parallel -- in the end, no duplicate categories should exist.
Directory dest = NewDirectory();
var destTw = new DirectoryTaxonomyWriter(dest);
ThreadClass t = new ThreadAnonymousInnerClassHelper2(this, numCategories, destTw);
t.Start();
OrdinalMap map = new MemoryOrdinalMap();
destTw.AddTaxonomy(src, map);
t.Join();
destTw.Dispose();
// now validate
var dtr = new DirectoryTaxonomyReader(dest);
// +2 to account for the root category + "a"
Assert.AreEqual(numCategories + 2, dtr.Size);
var categories = new HashSet<FacetLabel>();
for (int i = 1; i < dtr.Size; i++)
{
FacetLabel cat = dtr.GetPath(i);
Assert.True(categories.Add(cat), "category " + cat + " already existed");
}
dtr.Dispose();
IOUtils.Close(src, dest);
}
public virtual void TestOpenIfChangedMergedSegment()
{
// test openIfChanged() when all index segments were merged - used to be
// a bug in ParentArray, caught by testOpenIfChangedManySegments - only
// this test is not random
Directory dir = NewDirectory();
// hold onto IW to forceMerge
// note how we don't close it, since DTW will close it.
IndexWriter iw = new IndexWriter(dir, new IndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(new LogByteSizeMergePolicy()));
var writer = new DirectoryTaxonomyWriterAnonymousInnerClassHelper2(this, dir, iw) as DirectoryTaxonomyWriter;
var reader = new DirectoryTaxonomyReader(writer);
Assert.AreEqual(1, reader.Size);
Assert.AreEqual(1, reader.ParallelTaxonomyArrays.Parents().Length);
// add category and call forceMerge -- this should flush IW and merge segments down to 1
// in ParentArray.initFromReader, this used to fail assuming there are no parents.
writer.AddCategory(new FacetLabel("1"));
iw.ForceMerge(1);
// now calling openIfChanged should trip on the bug
var newtr = TaxonomyReader.OpenIfChanged(reader);
Assert.NotNull(newtr);
reader.Dispose();
reader = newtr;
Assert.AreEqual(2, reader.Size);
Assert.AreEqual(2, reader.ParallelTaxonomyArrays.Parents().Length);
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
public virtual void TestRecreateAndRefresh()
{
// DirTaxoWriter lost the INDEX_EPOCH property if it was opened in
// CREATE_OR_APPEND (or commit(userData) called twice), which could lead to
// DirTaxoReader succeeding to refresh().
Directory dir = NewDirectory();
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE_OR_APPEND, NO_OP_CACHE);
TouchTaxo(taxoWriter, new FacetLabel("a"));
var taxoReader = new DirectoryTaxonomyReader(dir);
TouchTaxo(taxoWriter, new FacetLabel("b"));
var newtr = TaxonomyReader.OpenIfChanged(taxoReader);
taxoReader.Dispose();
taxoReader = newtr;
Assert.AreEqual(1, Convert.ToInt32(taxoReader.CommitUserData[DirectoryTaxonomyWriter.INDEX_EPOCH]));
// now recreate the taxonomy, and check that the epoch is preserved after opening DirTW again.
taxoWriter.Dispose();
taxoWriter = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE, NO_OP_CACHE);
TouchTaxo(taxoWriter, new FacetLabel("c"));
taxoWriter.Dispose();
taxoWriter = new DirectoryTaxonomyWriter(dir, OpenMode.CREATE_OR_APPEND, NO_OP_CACHE);
TouchTaxo(taxoWriter, new FacetLabel("d"));
taxoWriter.Dispose();
newtr = TaxonomyReader.OpenIfChanged(taxoReader);
taxoReader.Dispose();
taxoReader = newtr;
Assert.AreEqual(2, Convert.ToInt32(taxoReader.CommitUserData[DirectoryTaxonomyWriter.INDEX_EPOCH]));
taxoReader.Dispose();
dir.Dispose();
}
public virtual void TestOpenIfChangedNoChangesButSegmentMerges()
{
// test openIfChanged() when the taxonomy hasn't really changed, but segments
// were merged. The NRT reader will be reopened, and ParentArray used to assert
// that the new reader contains more ordinals than were given from the old
// TaxReader version
Directory dir = NewDirectory();
// hold onto IW to forceMerge
// note how we don't close it, since DTW will close it.
var iw = new IndexWriter(dir, new IndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetMergePolicy(new LogByteSizeMergePolicy()));
DirectoryTaxonomyWriter writer = new DirectoryTaxonomyWriterAnonymousInnerClassHelper3(this, dir, iw);
// add a category so that the following DTR open will cause a flush and
// a new segment will be created
writer.AddCategory(new FacetLabel("a"));
var reader = new DirectoryTaxonomyReader(writer);
Assert.AreEqual(2, reader.Size);
Assert.AreEqual(2, reader.ParallelTaxonomyArrays.Parents().Length);
// merge all the segments so that NRT reader thinks there's a change
iw.ForceMerge(1);
// now calling openIfChanged should trip on the wrong assert in ParetArray's ctor
var newtr = TaxonomyReader.OpenIfChanged(reader);
Assert.NotNull(newtr);
reader.Dispose();
reader = newtr;
Assert.AreEqual(2, reader.Size);
Assert.AreEqual(2, reader.ParallelTaxonomyArrays.Parents().Length);
reader.Dispose();
writer.Dispose();
dir.Dispose();
}
private void doTestReadRecreatedTaxonomy(Random random, bool closeReader)
{
Directory dir = null;
TaxonomyWriter tw = null;
TaxonomyReader tr = null;
// prepare a few categories
int n = 10;
FacetLabel[] cp = new FacetLabel[n];
for (int i = 0; i < n; i++)
{
cp[i] = new FacetLabel("a", Convert.ToString(i));
}
try
{
dir = NewDirectory();
tw = new DirectoryTaxonomyWriter(dir);
tw.AddCategory(new FacetLabel("a"));
tw.Dispose();
tr = new DirectoryTaxonomyReader(dir);
int baseNumCategories = tr.Size;
for (int i = 0; i < n; i++)
{
int k = random.Next(n);
tw = new DirectoryTaxonomyWriter(dir, IndexWriterConfig.OpenMode_e.CREATE);
for (int j = 0; j <= k; j++)
{
tw.AddCategory(cp[j]);
}
tw.Dispose();
if (closeReader)
{
tr.Dispose(true);
tr = new DirectoryTaxonomyReader(dir);
}
else
{
var newtr = TaxonomyReader.OpenIfChanged(tr);
Assert.NotNull(newtr);
tr.Dispose(true);
tr = newtr;
}
Assert.AreEqual(baseNumCategories + 1 + k, tr.Size, "Wrong #categories in taxonomy (i=" + i + ", k=" + k + ")");
}
}
finally
{
IOUtils.Close(tr as DirectoryTaxonomyReader, tw, dir);
}
}
