/// <summary>
/// Aggreggates float facet values from the provided
/// <see cref="ValueSource"/>, and pulls ordinals from the
/// provided <see cref="OrdinalsReader"/>.
/// </summary>
public TaxonomyFacetSumValueSource(OrdinalsReader ordinalsReader, TaxonomyReader taxoReader,
FacetsConfig config, FacetsCollector fc, ValueSource valueSource)
: base(ordinalsReader.IndexFieldName, taxoReader, config)
{
this.ordinalsReader = ordinalsReader;
SumValues(fc.GetMatchingDocs(), fc.KeepScores, valueSource);
}
public virtual void TestBasic()
{
AssumeTrue("Test requires SortedSetDV support", DefaultCodecSupportsSortedSet());
Directory dir = NewDirectory();
FacetsConfig config = new FacetsConfig();
config.SetMultiValued("a", true);
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, Similarity, TimeZone);
Document doc = new Document();
doc.Add(new SortedSetDocValuesFacetField("a", "foo"));
doc.Add(new SortedSetDocValuesFacetField("a", "bar"));
doc.Add(new SortedSetDocValuesFacetField("a", "zoo"));
doc.Add(new SortedSetDocValuesFacetField("b", "baz"));
writer.AddDocument(config.Build(doc));
if (Random().NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new SortedSetDocValuesFacetField("a", "foo"));
writer.AddDocument(config.Build(doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.Reader);
// Per-top-reader state:
SortedSetDocValuesReaderState state = new DefaultSortedSetDocValuesReaderState(searcher.IndexReader);
FacetsCollector c = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), c);
SortedSetDocValuesFacetCounts facets = new SortedSetDocValuesFacetCounts(state, c);
Assert.AreEqual("dim=a path=[] value=4 childCount=3\n foo (2)\n bar (1)\n zoo (1)\n", facets.GetTopChildren(10, "a").ToString());
Assert.AreEqual("dim=b path=[] value=1 childCount=1\n baz (1)\n", facets.GetTopChildren(10, "b").ToString());
// DrillDown:
DrillDownQuery q = new DrillDownQuery(config);
q.Add("a", "foo");
q.Add("b", "baz");
TopDocs hits = searcher.Search(q, 1);
Assert.AreEqual(1, hits.TotalHits);
IOUtils.Close(writer, searcher.IndexReader, dir);
}
public virtual void TestChildCount()
{
// LUCENE-4885: FacetResult.numValidDescendants was not set properly by FacetsAccumulator
var indexDir = NewDirectory();
var taxoDir = NewDirectory();
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
IndexWriter iw = new IndexWriter(indexDir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())));
FacetsConfig config = new FacetsConfig();
for (int i = 0; i < 10; i++)
{
Document doc = new Document();
doc.Add(new FacetField("a", Convert.ToString(i)));
iw.AddDocument(config.Build(taxoWriter, doc));
}
DirectoryReader r = DirectoryReader.Open(iw, true);
DirectoryTaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector sfc = new FacetsCollector();
NewSearcher(r).Search(new MatchAllDocsQuery(), sfc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, config, sfc);
Assert.AreEqual(10, facets.GetTopChildren(2, "a").ChildCount);
IOUtils.Close(taxoWriter, iw, taxoReader, taxoDir, r, indexDir);
}
public virtual void TestMissingValues()
{
AssumeTrue("codec does not support docsWithField", DefaultCodecSupportsDocsWithField());
Directory d = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), d);
Document doc = new Document();
NumericDocValuesField field = new NumericDocValuesField("field", 0L);
doc.Add(field);
for (long l = 0; l < 100; l++)
{
if (l % 5 == 0)
{
// Every 5th doc is missing the value:
w.AddDocument(new Document());
continue;
}
field.LongValue = l;
w.AddDocument(doc);
}
IndexReader r = w.Reader;
FacetsCollector fc = new FacetsCollector();
IndexSearcher s = NewSearcher(r);
s.Search(new MatchAllDocsQuery(), fc);
Facets facets = new LongRangeFacetCounts("field", fc, new LongRange("less than 10", 0L, true, 10L, false), new LongRange("less than or equal to 10", 0L, true, 10L, true), new LongRange("over 90", 90L, false, 100L, false), new LongRange("90 or above", 90L, true, 100L, false), new LongRange("over 1000", 1000L, false, long.MaxValue, false));
Assert.AreEqual("dim=field path=[] value=16 childCount=5\n less than 10 (8)\n less than or equal to 10 (8)\n over 90 (8)\n 90 or above (8)\n over 1000 (0)\n", facets.GetTopChildren(10, "field").ToString());
IOUtils.Close(w, r, d);
}
public virtual void TestRandomFloats()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), dir);
int numDocs = AtLeast(1000);
float[] values = new float[numDocs];
float minValue = float.PositiveInfinity;
float maxValue = float.NegativeInfinity;
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
float v = Random().NextFloat();
values[i] = v;
doc.Add(new FloatDocValuesField("field", v));
doc.Add(new FloatField("field", v, Field.Store.NO));
w.AddDocument(doc);
minValue = Math.Min(minValue, v);
maxValue = Math.Max(maxValue, v);
}
IndexReader r = w.Reader;
IndexSearcher s = NewSearcher(r);
FacetsConfig config = new FacetsConfig();
int numIters = AtLeast(10);
for (int iter = 0; iter < numIters; iter++)
{
if (VERBOSE)
{
Console.WriteLine("TEST: iter=" + iter);
}
int numRange = TestUtil.NextInt(Random(), 1, 5);
DoubleRange[] ranges = new DoubleRange[numRange];
int[] expectedCounts = new int[numRange];
float minAcceptedValue = float.PositiveInfinity;
float maxAcceptedValue = float.NegativeInfinity;
if (VERBOSE)
{
Console.WriteLine("TEST: " + numRange + " ranges");
}
for (int rangeID = 0; rangeID < numRange; rangeID++)
{
double min;
if (rangeID > 0 && Random().Next(10) == 7)
{
// Use an existing boundary:
DoubleRange prevRange = ranges[Random().Next(rangeID)];
if (Random().NextBoolean())
{
min = prevRange.Min;
}
else
{
min = prevRange.Max;
}
}
else
{
min = Random().NextDouble();
}
double max;
if (rangeID > 0 && Random().Next(10) == 7)
{
// Use an existing boundary:
DoubleRange prevRange = ranges[Random().Next(rangeID)];
if (Random().NextBoolean())
{
max = prevRange.Min;
}
else
{
max = prevRange.Max;
}
}
else
{
max = Random().NextDouble();
}
if (min > max)
{
double x = min;
min = max;
max = x;
}
// Must truncate to float precision so that the
// drill-down counts (which use NRQ.newFloatRange)
// are correct:
min = (float)min;
max = (float)max;
bool minIncl;
bool maxIncl;
if (min == max)
{
minIncl = true;
maxIncl = true;
}
else
{
minIncl = Random().NextBoolean();
maxIncl = Random().NextBoolean();
}
ranges[rangeID] = new DoubleRange("r" + rangeID, min, minIncl, max, maxIncl);
if (VERBOSE)
{
Console.WriteLine("TEST: range " + rangeID + ": " + ranges[rangeID]);
}
// Do "slow but hopefully correct" computation of
// expected count:
for (int i = 0; i < numDocs; i++)
{
bool accept = true;
if (minIncl)
{
accept &= values[i] >= min;
}
else
{
accept &= values[i] > min;
}
if (maxIncl)
{
accept &= values[i] <= max;
}
else
{
accept &= values[i] < max;
}
if (VERBOSE)
{
Console.WriteLine("TEST: check doc=" + i + " val=" + values[i] + " accept=" + accept);
}
if (accept)
{
expectedCounts[rangeID]++;
minAcceptedValue = Math.Min(minAcceptedValue, values[i]);
maxAcceptedValue = Math.Max(maxAcceptedValue, values[i]);
}
}
}
FacetsCollector sfc = new FacetsCollector();
s.Search(new MatchAllDocsQuery(), sfc);
Filter fastMatchFilter;
if (Random().NextBoolean())
{
if (Random().NextBoolean())
{
fastMatchFilter = NumericRangeFilter.NewFloatRange("field", minValue, maxValue, true, true);
}
else
{
fastMatchFilter = NumericRangeFilter.NewFloatRange("field", minAcceptedValue, maxAcceptedValue, true, true);
}
}
else
{
fastMatchFilter = null;
}
ValueSource vs = new FloatFieldSource("field");
Facets facets = new DoubleRangeFacetCounts("field", vs, sfc, fastMatchFilter, ranges);
FacetResult result = facets.GetTopChildren(10, "field");
Assert.AreEqual(numRange, result.LabelValues.Length);
for (int rangeID = 0; rangeID < numRange; rangeID++)
{
if (VERBOSE)
{
Console.WriteLine("TEST: verify range " + rangeID + " expectedCount=" + expectedCounts[rangeID]);
}
LabelAndValue subNode = result.LabelValues[rangeID];
Assert.AreEqual("r" + rangeID, subNode.label);
Assert.AreEqual(expectedCounts[rangeID], (int)subNode.value);
DoubleRange range = ranges[rangeID];
// Test drill-down:
DrillDownQuery ddq = new DrillDownQuery(config);
if (Random().NextBoolean())
{
if (Random().NextBoolean())
{
ddq.Add("field", NumericRangeFilter.NewFloatRange("field", (float)range.Min, (float)range.Max, range.MinInclusive, range.MaxInclusive));
}
else
{
ddq.Add("field", NumericRangeQuery.NewFloatRange("field", (float)range.Min, (float)range.Max, range.MinInclusive, range.MaxInclusive));
}
}
else
{
ddq.Add("field", range.GetFilter(fastMatchFilter, vs));
}
Assert.AreEqual(expectedCounts[rangeID], s.Search(ddq, 10).TotalHits);
}
}
IOUtils.Close(w, r, dir);
}
public virtual void TestAllCounts()
{
DirectoryReader indexReader = DirectoryReader.Open(indexDir);
var taxoReader = new DirectoryTaxonomyReader(taxoDir);
IndexSearcher searcher = NewSearcher(indexReader);
FacetsCollector sfc = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), sfc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, Config, sfc);
FacetResult result = facets.GetTopChildren(NUM_CHILDREN_CP_A, CP_A);
Assert.AreEqual(-1, (int)result.Value);
int prevValue = int.MaxValue;
foreach (LabelAndValue labelValue in result.LabelValues)
{
Assert.AreEqual(allExpectedCounts[CP_A + "/" + labelValue.label], labelValue.value);
Assert.True((int)labelValue.value <= prevValue, "wrong sort order of sub results: labelValue.value=" + labelValue.value + " prevValue=" + prevValue);
prevValue = (int)labelValue.value;
}
result = facets.GetTopChildren(NUM_CHILDREN_CP_B, CP_B);
Assert.AreEqual(allExpectedCounts[CP_B], result.Value);
prevValue = int.MaxValue;
foreach (LabelAndValue labelValue in result.LabelValues)
{
Assert.AreEqual(allExpectedCounts[CP_B + "/" + labelValue.label], labelValue.value);
Assert.True((int)labelValue.value <= prevValue, "wrong sort order of sub results: labelValue.value=" + labelValue.value + " prevValue=" + prevValue);
prevValue = (int)labelValue.value;
}
IOUtils.Close(indexReader, taxoReader);
}
/// <summary>
/// Create <see cref="TaxonomyFacetSumIntAssociations"/> against
/// the default index field.
/// </summary>
public TaxonomyFacetSumIntAssociations(TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector fc)
: this(FacetsConfig.DEFAULT_INDEX_FIELD_NAME, taxoReader, config, fc)
{
}
public virtual void TestRandom()
{
string[] tokens = GetRandomTokens(10);
Store.Directory indexDir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), indexDir);
var tw = new DirectoryTaxonomyWriter(taxoDir);
FacetsConfig config = new FacetsConfig();
int numDocs = AtLeast(1000);
int numDims = TestUtil.NextInt(Random(), 1, 7);
IList<TestDoc> testDocs = GetRandomDocs(tokens, numDocs, numDims);
foreach (TestDoc testDoc in testDocs)
{
Document doc = new Document();
doc.Add(NewStringField("content", testDoc.content, Field.Store.NO));
for (int j = 0; j < numDims; j++)
{
if (testDoc.dims[j] != null)
{
doc.Add(new FacetField("dim" + j, testDoc.dims[j]));
}
}
w.AddDocument(config.Build(tw, doc));
}
// NRT open
IndexSearcher searcher = NewSearcher(w.Reader);
// NRT open
var tr = new DirectoryTaxonomyReader(tw);
int iters = AtLeast(100);
for (int iter = 0; iter < iters; iter++)
{
string searchToken = tokens[Random().Next(tokens.Length)];
if (VERBOSE)
{
Console.WriteLine("\nTEST: iter content=" + searchToken);
}
FacetsCollector fc = new FacetsCollector();
FacetsCollector.Search(searcher, new TermQuery(new Term("content", searchToken)), 10, fc);
Facets facets = GetTaxonomyFacetCounts(tr, config, fc);
// Slow, yet hopefully bug-free, faceting:
var expectedCounts = new List<Dictionary<string, int?>>();
for (int i = 0; i < numDims; i++)
{
expectedCounts[i] = new Dictionary<string, int?>();
}
foreach (TestDoc doc in testDocs)
{
if (doc.content.Equals(searchToken))
{
for (int j = 0; j < numDims; j++)
{
if (doc.dims[j] != null)
{
int? v = expectedCounts[j][doc.dims[j]];
if (v == null)
{
expectedCounts[j][doc.dims[j]] = 1;
}
else
{
expectedCounts[j][doc.dims[j]] = (int)v + 1;
}
}
}
}
}
IList<FacetResult> expected = new List<FacetResult>();
for (int i = 0; i < numDims; i++)
{
IList<LabelAndValue> labelValues = new List<LabelAndValue>();
int totCount = 0;
foreach (KeyValuePair<string, int?> ent in expectedCounts[i])
{
labelValues.Add(new LabelAndValue(ent.Key, ent.Value.Value));
totCount += ent.Value.Value;
}
SortLabelValues(labelValues);
if (totCount > 0)
{
expected.Add(new FacetResult("dim" + i, new string[0], totCount, labelValues.ToArray(), labelValues.Count));
}
}
// Sort by highest value, tie break by value:
SortFacetResults(expected);
IList<FacetResult> actual = facets.GetAllDims(10);
// Messy: fixup ties
SortTies(actual);
Assert.AreEqual(expected, actual);
}
IOUtils.Close(w, tw, searcher.IndexReader, tr, indexDir, taxoDir);
}
public virtual void TestSparseFacets()
{
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);
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
FacetsConfig config = new FacetsConfig();
Document doc = new Document();
doc.Add(new FacetField("a", "foo1"));
writer.AddDocument(config.Build(taxoWriter, doc));
if (Random().NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new FacetField("a", "foo2"));
doc.Add(new FacetField("b", "bar1"));
writer.AddDocument(config.Build(taxoWriter, doc));
if (Random().NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new FacetField("a", "foo3"));
doc.Add(new FacetField("b", "bar2"));
doc.Add(new FacetField("c", "baz1"));
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);
Facets facets = GetTaxonomyFacetCounts(taxoReader, new FacetsConfig(), c);
// Ask for top 10 labels for any dims that have counts:
IList<FacetResult> results = facets.GetAllDims(10);
Assert.AreEqual(3, results.Count);
Assert.AreEqual("dim=a path=[] value=3 childCount=3\n foo1 (1)\n foo2 (1)\n foo3 (1)\n", results[0].ToString());
Assert.AreEqual("dim=b path=[] value=2 childCount=2\n bar1 (1)\n bar2 (1)\n", results[1].ToString());
Assert.AreEqual("dim=c path=[] value=1 childCount=1\n baz1 (1)\n", results[2].ToString());
IOUtils.Close(writer, taxoWriter, searcher.IndexReader, taxoReader, taxoDir, dir);
}
protected override Facets BuildFacetsResult(FacetsCollector drillDowns, FacetsCollector[] drillSideways, string[] drillSidewaysDims)
{
Debug.Assert(drillSideways.Length == 1);
return(new DoubleRangeFacetCounts("field", vs, drillSideways[0], fastMatchFilter, ranges));
}
public virtual void TestRandomDoubles()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
int numDocs = AtLeast(1000);
double[] values = new double[numDocs];
double minValue = double.PositiveInfinity;
double maxValue = double.NegativeInfinity;
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
double v = Random.NextDouble();
values[i] = v;
doc.Add(new DoubleDocValuesField("field", v));
doc.Add(new DoubleField("field", v, Field.Store.NO));
w.AddDocument(doc);
minValue = Math.Min(minValue, v);
maxValue = Math.Max(maxValue, v);
}
IndexReader r = w.GetReader();
IndexSearcher s = NewSearcher(r);
FacetsConfig config = new FacetsConfig();
int numIters = AtLeast(10);
for (int iter = 0; iter < numIters; iter++)
{
if (VERBOSE)
{
Console.WriteLine("TEST: iter=" + iter);
}
int numRange = TestUtil.NextInt32(Random, 1, 5);
DoubleRange[] ranges = new DoubleRange[numRange];
int[] expectedCounts = new int[numRange];
double minAcceptedValue = double.PositiveInfinity;
double maxAcceptedValue = double.NegativeInfinity;
for (int rangeID = 0; rangeID < numRange; rangeID++)
{
double min;
if (rangeID > 0 && Random.Next(10) == 7)
{
// Use an existing boundary:
DoubleRange prevRange = ranges[Random.Next(rangeID)];
if (Random.NextBoolean())
{
min = prevRange.Min;
}
else
{
min = prevRange.Max;
}
}
else
{
min = Random.NextDouble();
}
double max;
if (rangeID > 0 && Random.Next(10) == 7)
{
// Use an existing boundary:
DoubleRange prevRange = ranges[Random.Next(rangeID)];
if (Random.NextBoolean())
{
max = prevRange.Min;
}
else
{
max = prevRange.Max;
}
}
else
{
max = Random.NextDouble();
}
if (min > max)
{
double x = min;
min = max;
max = x;
}
bool minIncl;
bool maxIncl;
if (min == max)
{
minIncl = true;
maxIncl = true;
}
else
{
minIncl = Random.NextBoolean();
maxIncl = Random.NextBoolean();
}
ranges[rangeID] = new DoubleRange("r" + rangeID, min, minIncl, max, maxIncl);
// Do "slow but hopefully correct" computation of
// expected count:
for (int i = 0; i < numDocs; i++)
{
bool accept = true;
if (minIncl)
{
accept &= values[i] >= min;
}
else
{
accept &= values[i] > min;
}
if (maxIncl)
{
accept &= values[i] <= max;
}
else
{
accept &= values[i] < max;
}
if (accept)
{
expectedCounts[rangeID]++;
minAcceptedValue = Math.Min(minAcceptedValue, values[i]);
maxAcceptedValue = Math.Max(maxAcceptedValue, values[i]);
}
}
}
FacetsCollector sfc = new FacetsCollector();
s.Search(new MatchAllDocsQuery(), sfc);
Filter fastMatchFilter;
if (Random.NextBoolean())
{
if (Random.NextBoolean())
{
fastMatchFilter = NumericRangeFilter.NewDoubleRange("field", minValue, maxValue, true, true);
}
else
{
fastMatchFilter = NumericRangeFilter.NewDoubleRange("field", minAcceptedValue, maxAcceptedValue, true, true);
}
}
else
{
fastMatchFilter = null;
}
ValueSource vs = new DoubleFieldSource("field");
Facets facets = new DoubleRangeFacetCounts("field", vs, sfc, fastMatchFilter, ranges);
FacetResult result = facets.GetTopChildren(10, "field");
Assert.AreEqual(numRange, result.LabelValues.Length);
for (int rangeID = 0; rangeID < numRange; rangeID++)
{
if (VERBOSE)
{
Console.WriteLine(" range " + rangeID + " expectedCount=" + expectedCounts[rangeID]);
}
LabelAndValue subNode = result.LabelValues[rangeID];
Assert.AreEqual("r" + rangeID, subNode.Label);
Assert.AreEqual(expectedCounts[rangeID], (int)subNode.Value);
DoubleRange range = ranges[rangeID];
// Test drill-down:
DrillDownQuery ddq = new DrillDownQuery(config);
if (Random.NextBoolean())
{
if (Random.NextBoolean())
{
ddq.Add("field", NumericRangeFilter.NewDoubleRange("field", range.Min, range.Max, range.MinInclusive, range.MaxInclusive));
}
else
{
ddq.Add("field", NumericRangeQuery.NewDoubleRange("field", range.Min, range.Max, range.MinInclusive, range.MaxInclusive));
}
}
else
{
ddq.Add("field", range.GetFilter(fastMatchFilter, vs));
}
Assert.AreEqual(expectedCounts[rangeID], s.Search(ddq, 10).TotalHits);
}
}
IOUtils.Dispose(w, r, dir);
}
public virtual void TestRandom()
{
string[] tokens = GetRandomTokens(10);
Store.Directory indexDir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, indexDir);
var tw = new DirectoryTaxonomyWriter(taxoDir);
FacetsConfig config = new FacetsConfig();
int numDocs = AtLeast(1000);
int numDims = TestUtil.NextInt32(Random, 1, 7);
IList <TestDoc> testDocs = GetRandomDocs(tokens, numDocs, numDims);
foreach (TestDoc testDoc in testDocs)
{
Document doc = new Document();
doc.Add(NewStringField("content", testDoc.content, Field.Store.NO));
for (int j = 0; j < numDims; j++)
{
if (testDoc.dims[j] != null)
{
doc.Add(new FacetField("dim" + j, testDoc.dims[j]));
}
}
w.AddDocument(config.Build(tw, doc));
}
// NRT open
IndexSearcher searcher = NewSearcher(w.GetReader());
// NRT open
var tr = new DirectoryTaxonomyReader(tw);
int iters = AtLeast(100);
for (int iter = 0; iter < iters; iter++)
{
string searchToken = tokens[Random.Next(tokens.Length)];
if (VERBOSE)
{
Console.WriteLine("\nTEST: iter content=" + searchToken);
}
FacetsCollector fc = new FacetsCollector();
FacetsCollector.Search(searcher, new TermQuery(new Term("content", searchToken)), 10, fc);
Facets facets = GetTaxonomyFacetCounts(tr, config, fc);
// Slow, yet hopefully bug-free, faceting:
var expectedCounts = new List <Dictionary <string, int?> >();
for (int i = 0; i < numDims; i++)
{
expectedCounts.Add(new Dictionary <string, int?>());
}
foreach (TestDoc doc in testDocs)
{
if (doc.content.Equals(searchToken, StringComparison.Ordinal))
{
for (int j = 0; j < numDims; j++)
{
if (doc.dims[j] != null)
{
if (!expectedCounts[j].TryGetValue(doc.dims[j], out int?v) || v == null)
{
expectedCounts[j][doc.dims[j]] = 1;
}
else
{
expectedCounts[j][doc.dims[j]] = (int)v + 1;
}
}
}
}
}
List <FacetResult> expected = new List <FacetResult>();
for (int i = 0; i < numDims; i++)
{
List <LabelAndValue> labelValues = new List <LabelAndValue>();
int totCount = 0;
foreach (KeyValuePair <string, int?> ent in expectedCounts[i])
{
labelValues.Add(new LabelAndValue(ent.Key, ent.Value.Value));
totCount += ent.Value.Value;
}
SortLabelValues(labelValues);
if (totCount > 0)
{
expected.Add(new FacetResult("dim" + i, new string[0], totCount, labelValues.ToArray(), labelValues.Count));
}
}
// Sort by highest value, tie break by value:
SortFacetResults(expected);
IList <FacetResult> actual = facets.GetAllDims(10);
// Messy: fixup ties
SortTies(actual);
Assert.AreEqual(expected, actual);
}
IOUtils.Dispose(w, tw, searcher.IndexReader, tr, indexDir, taxoDir);
}
public virtual void TestCustomDoublesValueSource()
{
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
Document doc = new Document();
writer.AddDocument(doc);
writer.AddDocument(doc);
writer.AddDocument(doc);
// Test wants 3 docs in one segment:
writer.ForceMerge(1);
var vs = new ValueSourceAnonymousInnerClassHelper(this, doc);
FacetsConfig config = new FacetsConfig();
FacetsCollector fc = new FacetsCollector();
IndexReader r = writer.GetReader();
IndexSearcher s = NewSearcher(r);
s.Search(new MatchAllDocsQuery(), fc);
DoubleRange[] ranges = new DoubleRange[] { new DoubleRange("< 1", 0.0, true, 1.0, false), new DoubleRange("< 2", 0.0, true, 2.0, false), new DoubleRange("< 5", 0.0, true, 5.0, false), new DoubleRange("< 10", 0.0, true, 10.0, false), new DoubleRange("< 20", 0.0, true, 20.0, false), new DoubleRange("< 50", 0.0, true, 50.0, false) };
Filter fastMatchFilter;
AtomicBoolean filterWasUsed = new AtomicBoolean();
if (Random.NextBoolean())
{
// Sort of silly:
fastMatchFilter = new CachingWrapperFilterAnonymousInnerClassHelper(this, new QueryWrapperFilter(new MatchAllDocsQuery()), filterWasUsed);
}
else
{
fastMatchFilter = null;
}
if (VERBOSE)
{
Console.WriteLine("TEST: fastMatchFilter=" + fastMatchFilter);
}
Facets facets = new DoubleRangeFacetCounts("field", vs, fc, fastMatchFilter, ranges);
Assert.AreEqual("dim=field path=[] value=3 childCount=6\n < 1 (0)\n < 2 (1)\n < 5 (3)\n < 10 (3)\n < 20 (3)\n < 50 (3)\n", facets.GetTopChildren(10, "field").ToString());
Assert.True(fastMatchFilter == null || filterWasUsed);
DrillDownQuery ddq = new DrillDownQuery(config);
ddq.Add("field", ranges[1].GetFilter(fastMatchFilter, vs));
// Test simple drill-down:
Assert.AreEqual(1, s.Search(ddq, 10).TotalHits);
// Test drill-sideways after drill-down
DrillSideways ds = new DrillSidewaysAnonymousInnerClassHelper2(this, s, config, (TaxonomyReader)null, vs, ranges, fastMatchFilter);
DrillSidewaysResult dsr = ds.Search(ddq, 10);
Assert.AreEqual(1, dsr.Hits.TotalHits);
Assert.AreEqual("dim=field path=[] value=3 childCount=6\n < 1 (0)\n < 2 (1)\n < 5 (3)\n < 10 (3)\n < 20 (3)\n < 50 (3)\n", dsr.Facets.GetTopChildren(10, "field").ToString());
IOUtils.Dispose(r, writer, dir);
}
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, OpenMode.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetHierarchical("Publish Date", true);
RandomIndexWriter writer = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
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.GetReader());
// 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:
string result;
using (ByteArrayOutputStream bos = new ByteArrayOutputStream())
{
using (StreamWriter w = new StreamWriter(bos, Encoding.UTF8, 2048, true)
{
AutoFlush = true
})
{
PrintTaxonomyStats.PrintStats(taxoReader, w, true);
}
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.Dispose(writer, taxoWriter, searcher.IndexReader, taxoReader, taxoDir, dir);
}
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, OpenMode.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetIndexFieldName("a", "$facets2");
RandomIndexWriter writer = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
Document doc = new Document();
doc.Add(new FacetField("a", "foo1"));
writer.AddDocument(config.Build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.GetReader());
// 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.Dispose(writer, taxoWriter, searcher.IndexReader, taxoReader, taxoDir, dir);
}
public virtual void TestSparseFacets()
{
Store.Directory dir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
// Writes facet ords to a separate directory from the
// main index:
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir, OpenMode.CREATE);
RandomIndexWriter writer = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
FacetsConfig config = new FacetsConfig();
Document doc = new Document();
doc.Add(new FacetField("a", "foo1"));
writer.AddDocument(config.Build(taxoWriter, doc));
if (Random.NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new FacetField("a", "foo2"));
doc.Add(new FacetField("b", "bar1"));
writer.AddDocument(config.Build(taxoWriter, doc));
if (Random.NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new FacetField("a", "foo3"));
doc.Add(new FacetField("b", "bar2"));
doc.Add(new FacetField("c", "baz1"));
writer.AddDocument(config.Build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.GetReader());
// NRT open
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector c = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), c);
Facets facets = GetTaxonomyFacetCounts(taxoReader, new FacetsConfig(), c);
// Ask for top 10 labels for any dims that have counts:
IList <FacetResult> results = facets.GetAllDims(10);
Assert.AreEqual(3, results.Count);
Assert.AreEqual("dim=a path=[] value=3 childCount=3\n foo1 (1)\n foo2 (1)\n foo3 (1)\n", results[0].ToString());
Assert.AreEqual("dim=b path=[] value=2 childCount=2\n bar1 (1)\n bar2 (1)\n", results[1].ToString());
Assert.AreEqual("dim=c path=[] value=1 childCount=1\n baz1 (1)\n", results[2].ToString());
IOUtils.Dispose(writer, taxoWriter, searcher.IndexReader, taxoReader, taxoDir, dir);
}
public virtual void TestGetFacetResultsTwice()
{
// LUCENE-4893: counts were multiplied as many times as getFacetResults was called.
Store.Directory indexDir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
IndexWriter iw = new IndexWriter(indexDir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())));
FacetsConfig config = new FacetsConfig();
Document doc = new Document();
doc.Add(new FacetField("a", "1"));
doc.Add(new FacetField("b", "1"));
iw.AddDocument(config.Build(taxoWriter, doc));
DirectoryReader r = DirectoryReader.Open(iw, true);
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector sfc = new FacetsCollector();
NewSearcher(r).Search(new MatchAllDocsQuery(), sfc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, config, sfc);
IList<FacetResult> res1 = facets.GetAllDims(10);
IList<FacetResult> res2 = facets.GetAllDims(10);
Assert.AreEqual(res1, res2, "calling getFacetResults twice should return the .equals()=true result");
IOUtils.Close(taxoWriter, iw, taxoReader, taxoDir, r, indexDir);
}
/// <summary>
/// Create <see cref="TaxonomyFacetCounts"/>, which also
/// counts all facet labels. Use this for a non-default
/// <see cref="OrdinalsReader"/>; otherwise use <see cref="FastTaxonomyFacetCounts"/>.
/// </summary>
public TaxonomyFacetCounts(OrdinalsReader ordinalsReader, TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector fc)
: base(ordinalsReader.IndexFieldName, taxoReader, config)
{
this.ordinalsReader = ordinalsReader;
Count(fc.GetMatchingDocs());
}
public virtual void TestManyFacetsInOneDocument()
{
AssumeTrue("default Codec doesn't support huge BinaryDocValues", TestUtil.FieldSupportsHugeBinaryDocValues(FacetsConfig.DEFAULT_INDEX_FIELD_NAME));
Store.Directory dir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, iwc);
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode_e.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetMultiValued("dim", true);
int numLabels = TestUtil.NextInt(Random(), 40000, 100000);
Document doc = new Document();
doc.Add(NewTextField("field", "text", Field.Store.NO));
for (int i = 0; i < numLabels; i++)
{
doc.Add(new FacetField("dim", "" + i));
}
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 = GetTaxonomyFacetCounts(taxoReader, config, c);
FacetResult result = facets.GetTopChildren(int.MaxValue, "dim");
Assert.AreEqual(numLabels, result.LabelValues.Length);
var allLabels = new HashSet<string>();
foreach (LabelAndValue labelValue in result.LabelValues)
{
allLabels.Add(labelValue.label);
Assert.AreEqual(1, (int)labelValue.value);
}
Assert.AreEqual(numLabels, allLabels.Count);
IOUtils.Close(searcher.IndexReader, taxoWriter, writer, taxoReader, dir, taxoDir);
}
protected override Facets BuildFacetsResult(FacetsCollector drillDowns, FacetsCollector[] drillSideways, string[] drillSidewaysDims)
{
Debug.Assert(drillSideways.Length == 1);
return(new DoubleRangeFacetCounts("field", vs, drillSideways[0], ONE_KM, TWO_KM, FIVE_KM, TEN_KM));
}
public virtual void TestSegmentsWithoutCategoriesOrResults()
{
// tests the accumulator when there are segments with no results
var indexDir = NewDirectory();
var taxoDir = NewDirectory();
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random()));
//iwc.MergePolicy = NoMergePolicy.INSTANCE; // prevent merges
IndexWriter indexWriter = new IndexWriter(indexDir, iwc);
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
FacetsConfig config = new FacetsConfig();
indexTwoDocs(taxoWriter, indexWriter, config, false); // 1st segment, no content, with categories
indexTwoDocs(taxoWriter, indexWriter, null, true); // 2nd segment, with content, no categories
indexTwoDocs(taxoWriter, indexWriter, config, true); // 3rd segment ok
indexTwoDocs(taxoWriter, indexWriter, null, false); // 4th segment, no content, or categories
indexTwoDocs(taxoWriter, indexWriter, null, true); // 5th segment, with content, no categories
indexTwoDocs(taxoWriter, indexWriter, config, true); // 6th segment, with content, with categories
indexTwoDocs(taxoWriter, indexWriter, null, true); // 7th segment, with content, no categories
IOUtils.Close(indexWriter, taxoWriter);
DirectoryReader indexReader = DirectoryReader.Open(indexDir);
var taxoReader = new DirectoryTaxonomyReader(taxoDir);
IndexSearcher indexSearcher = NewSearcher(indexReader);
// search for "f:a", only segments 1 and 3 should match results
Query q = new TermQuery(new Term("f", "a"));
FacetsCollector sfc = new FacetsCollector();
indexSearcher.Search(q, sfc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, config, sfc);
FacetResult result = facets.GetTopChildren(10, "A");
Assert.AreEqual(2, result.LabelValues.Length, "wrong number of children");
foreach (LabelAndValue labelValue in result.LabelValues)
{
Assert.AreEqual(2, (int)labelValue.value, "wrong weight for child " + labelValue.label);
}
IOUtils.Close(indexReader, taxoReader, indexDir, taxoDir);
}
/// <summary>
/// Search facet content in the existing index.
/// </summary>
/// <param name="text">The text to search for.</param>
/// <param name="indexFields">The array of index fields to search in.</param>
/// <param name="facetPaths">The array of facet paths to perform a drill down search on.</param>
/// <param name="numberToReturn">The maximum number of documents to return.</param>
/// <returns>The facet document.</returns>
/// <remarks>Use wildcard chars ('*', '?', '\'), logical ('AND', 'OR'), Quoted exact phrase ("search this").</remarks>
public Nequeo.Search.Engine.FacetDocument SearchFacetDocument(string text, FacetData.IndexField[] indexFields, FacetPath[] facetPaths, int numberToReturn = Int32.MaxValue)
{
Nequeo.Search.Engine.FacetDocument documents = new FacetDocument();
documents.TotalHits = 0;
try
{
// If text exists.
if (!String.IsNullOrEmpty(text))
{
// Load the searcher.
Lucene.Net.Search.IndexSearcher searcher = new Lucene.Net.Search.IndexSearcher(_reader);
string searchFieldName = "facetcontent";
Query query = null;
DrillDownQuery queryFacet = null;
TopDocs results = null;
// Build the facet configuration information.
FacetsConfig config = new FacetsConfig();
// Add the config.
foreach (FacetData.IndexField item in indexFields)
{
config.SetIndexFieldName(item.DimensionName, item.IndexFieldName);
}
// Get bytes
char[] textArray = text.ToCharArray();
// Search logical.
if (text.Contains("AND") || text.Contains("OR"))
{
// Create the query.
query = CreateLogicalQuery(text, searchFieldName);
}
else if (textArray[0].Equals('"') && textArray[textArray.Length - 1].Equals('"'))
{
// Create the query.
query = CreateQuotedQuery(new string(textArray, 1, textArray.Length - 2), searchFieldName);
}
else
{
// Create the query.
query = CreateBoolenQuery(text, BooleanClause.Occur.SHOULD, searchFieldName);
}
// Create the facet query.
queryFacet = new DrillDownQuery(config, query);
foreach (FacetPath facetPath in facetPaths)
{
// Add the path.
queryFacet.Add(facetPath.DimensionName, facetPath.Path);
}
// The collector.
FacetsCollector collector = new FacetsCollector();
// Search.
if (queryFacet != null)
{
results = FacetsCollector.Search(searcher, queryFacet, numberToReturn, collector);
}
// Get the total number of results that was asked for.
int totalResult = ((results.ScoreDocs != null && results.ScoreDocs.Length > 0) ? results.ScoreDocs.Length : 0);
// If result found.
if (results != null && results.TotalHits > 0)
{
List <TextDataResult> textDataResults = new List <TextDataResult>();
List <FileDocumentResult> fileDocResults = new List <FileDocumentResult>();
List <FacetPathResult> facetPathResults = new List <FacetPathResult>();
IDictionary <string, Facets> facetsMap = new Dictionary <string, Facets>();
// Add the facet count.
foreach (FacetData.IndexField item in indexFields)
{
// Add the facet for each index field.
facetsMap[item.DimensionName] = GetTaxonomyFacetCounts(_facetReader, config, collector, item.IndexFieldName);
}
// Create the multi facet list.
foreach (FacetPath facetPath in facetPaths)
{
try
{
// Add the facets.
Facets facets = facetsMap.First(u => u.Key.ToLower().Contains(facetPath.DimensionName.ToLower())).Value;
float number = facets.GetSpecificValue(facetPath.DimensionName, facetPath.Path);
// Add the path.
facetPathResults.Add(new FacetPathResult(facetPath.DimensionName, number, facetPath.Path));
}
catch { }
}
// For each document found.
for (int i = 0; i < totalResult; i++)
{
FileDocumentResult fileDocument = null;
TextDataResult textData = null;
int docID = results.ScoreDocs[i].Doc;
Lucene.Net.Documents.Document doc = searcher.Doc(docID);
try
{
// Get the data for each field.
IndexableField[] textNameFields = doc.GetFields("textname");
// If this field exists then text data.
if (textNameFields.Length > 0)
{
// Assign the data to the text document.
textData = new TextDataResult();
textData.Name = textNameFields.Length > 0 ? textNameFields[0].StringValue : null;
textData.Score = results.ScoreDocs[i].Score;
textData.Doc = docID;
// Do not know if the text was stored.
IndexableField[] textValueFields = doc.GetFields("textcomplete");
textData.Text = textValueFields.Length > 0 ? textValueFields[0].StringValue : null;
}
}
catch { }
// If text data exists then add.
if (textData != null)
{
textDataResults.Add(textData);
}
try
{
// Get the data for each field.
IndexableField[] pathNameFields = doc.GetFields("path");
IndexableField[] modifiedNameFields = doc.GetFields("modified");
// If this field exists then file document.
if (pathNameFields.Length > 0)
{
// Assign the data to the path document.
fileDocument = new FileDocumentResult();
fileDocument.Path = pathNameFields.Length > 0 ? pathNameFields[0].StringValue : null;
fileDocument.Modified = modifiedNameFields.Length > 0 ? modifiedNameFields[0].StringValue : null;
fileDocument.Score = results.ScoreDocs[i].Score;
fileDocument.Doc = docID;
}
}
catch { }
// If file data exists then add.
if (fileDocument != null)
{
fileDocResults.Add(fileDocument);
}
}
// Assign the facet document values.
documents.MaxScore = results.MaxScore;
documents.TotalHits = results.TotalHits;
documents.FacetPathResults = facetPathResults.ToArray();
documents.TextDataResults = textDataResults.ToArray();
documents.FileDocumentResults = fileDocResults.ToArray();
}
}
// Return the documents.
return(documents);
}
catch (Exception)
{
throw;
}
}
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);
}
/// <summary>
/// Get the taxonomy facet caounts
/// </summary>
/// <param name="taxoReader">The taxonomy reader.</param>
/// <param name="config">The facet configuration.</param>
/// <param name="collector">The result collector.</param>
/// <param name="indexFieldName">The index field name.</param>
/// <returns>The facets.</returns>
private Facets GetTaxonomyFacetCounts(TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector collector, string indexFieldName)
{
Facets facets = new FastTaxonomyFacetCounts(indexFieldName, taxoReader, config, collector);
return(facets);
}
public virtual void TestSparseFacets()
{
AssumeTrue("Test requires SortedSetDV support", DefaultCodecSupportsSortedSet);
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(
#if FEATURE_INSTANCE_TESTDATA_INITIALIZATION
this,
#endif
Random, dir);
FacetsConfig config = new FacetsConfig();
Document doc = new Document();
doc.Add(new SortedSetDocValuesFacetField("a", "foo1"));
writer.AddDocument(config.Build(doc));
if (Random.NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new SortedSetDocValuesFacetField("a", "foo2"));
doc.Add(new SortedSetDocValuesFacetField("b", "bar1"));
writer.AddDocument(config.Build(doc));
if (Random.NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new SortedSetDocValuesFacetField("a", "foo3"));
doc.Add(new SortedSetDocValuesFacetField("b", "bar2"));
doc.Add(new SortedSetDocValuesFacetField("c", "baz1"));
writer.AddDocument(config.Build(doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.GetReader());
writer.Dispose();
// Per-top-reader state:
SortedSetDocValuesReaderState state = new DefaultSortedSetDocValuesReaderState(searcher.IndexReader);
FacetsCollector c = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), c);
SortedSetDocValuesFacetCounts facets = new SortedSetDocValuesFacetCounts(state, c);
// Ask for top 10 labels for any dims that have counts:
IList <FacetResult> results = facets.GetAllDims(10);
Assert.AreEqual(3, results.Count);
Assert.AreEqual("dim=a path=[] value=3 childCount=3\n foo1 (1)\n foo2 (1)\n foo3 (1)\n", results[0].ToString());
Assert.AreEqual("dim=b path=[] value=2 childCount=2\n bar1 (1)\n bar2 (1)\n", results[1].ToString());
Assert.AreEqual("dim=c path=[] value=1 childCount=1\n baz1 (1)\n", results[2].ToString());
searcher.IndexReader.Dispose();
dir.Dispose();
}
public virtual void TestSparseFacets()
{
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);
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
FacetsConfig config = new FacetsConfig();
Document doc = new Document();
doc.Add(new IntField("num", 10, Field.Store.NO));
doc.Add(new FacetField("a", "foo1"));
writer.AddDocument(config.Build(taxoWriter, doc));
if (Random().NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new IntField("num", 20, Field.Store.NO));
doc.Add(new FacetField("a", "foo2"));
doc.Add(new FacetField("b", "bar1"));
writer.AddDocument(config.Build(taxoWriter, doc));
if (Random().NextBoolean())
{
writer.Commit();
}
doc = new Document();
doc.Add(new IntField("num", 30, Field.Store.NO));
doc.Add(new FacetField("a", "foo3"));
doc.Add(new FacetField("b", "bar2"));
doc.Add(new FacetField("c", "baz1"));
writer.AddDocument(config.Build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.Reader);
writer.Dispose();
// NRT open
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
taxoWriter.Dispose();
FacetsCollector c = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), c);
TaxonomyFacetSumValueSource facets = new TaxonomyFacetSumValueSource(taxoReader, new FacetsConfig(), c, new IntFieldSource("num"));
// Ask for top 10 labels for any dims that have counts:
IList <FacetResult> results = facets.GetAllDims(10);
Assert.AreEqual(3, results.Count);
Assert.AreEqual("dim=a path=[] value=60.0 childCount=3\n foo3 (30.0)\n foo2 (20.0)\n foo1 (10.0)\n", results[0].ToString());
Assert.AreEqual("dim=b path=[] value=50.0 childCount=2\n bar2 (30.0)\n bar1 (20.0)\n", results[1].ToString());
Assert.AreEqual("dim=c path=[] value=30.0 childCount=1\n baz1 (30.0)\n", results[2].ToString());
IOUtils.Close(searcher.IndexReader, taxoReader, dir, taxoDir);
}
/// <summary>
/// Create {@code TaxonomyFacetSumIntAssociations} against
/// the default index field.
/// </summary>
public TaxonomyFacetSumIntAssociations(TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector fc)
: this(FacetsConfig.DEFAULT_INDEX_FIELD_NAME, taxoReader, config, fc)
{
}
public virtual void TestWrongIndexFieldName()
{
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.SetIndexFieldName("a", "$facets2");
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
Document doc = new Document();
doc.Add(new IntField("num", 10, Field.Store.NO));
doc.Add(new FacetField("a", "foo1"));
writer.AddDocument(config.Build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.Reader);
writer.Dispose();
// NRT open
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
taxoWriter.Dispose();
FacetsCollector c = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), c);
TaxonomyFacetSumValueSource facets = new TaxonomyFacetSumValueSource(taxoReader, config, c, new IntFieldSource("num"));
// 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(searcher.IndexReader, taxoReader, dir, taxoDir);
}
public virtual void TestOverlappedEndStart()
{
Directory d = NewDirectory();
var w = new RandomIndexWriter(Random(), d);
Document doc = new Document();
NumericDocValuesField field = new NumericDocValuesField("field", 0L);
doc.Add(field);
for (long l = 0; l < 100; l++)
{
field.LongValue = l;
w.AddDocument(doc);
}
field.LongValue = long.MaxValue;
w.AddDocument(doc);
IndexReader r = w.Reader;
w.Dispose();
FacetsCollector fc = new FacetsCollector();
IndexSearcher s = NewSearcher(r);
s.Search(new MatchAllDocsQuery(), fc);
Facets facets = new LongRangeFacetCounts("field", fc, new LongRange("0-10", 0L, true, 10L, true), new LongRange("10-20", 10L, true, 20L, true), new LongRange("20-30", 20L, true, 30L, true), new LongRange("30-40", 30L, true, 40L, true));
FacetResult result = facets.GetTopChildren(10, "field");
Assert.AreEqual("dim=field path=[] value=41 childCount=4\n 0-10 (11)\n 10-20 (11)\n 20-30 (11)\n 30-40 (11)\n", result.ToString());
r.Dispose();
d.Dispose();
}
public virtual void TestRandom()
{
string[] tokens = GetRandomTokens(10);
Store.Directory indexDir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), indexDir);
var tw = new DirectoryTaxonomyWriter(taxoDir);
FacetsConfig config = new FacetsConfig();
int numDocs = AtLeast(1000);
int numDims = TestUtil.NextInt(Random(), 1, 7);
IList <TestDoc> testDocs = GetRandomDocs(tokens, numDocs, numDims);
foreach (TestDoc testDoc in testDocs)
{
Document doc = new Document();
doc.Add(NewStringField("content", testDoc.content, Field.Store.NO));
testDoc.value = Random().NextFloat();
doc.Add(new FloatDocValuesField("value", testDoc.value));
for (int j = 0; j < numDims; j++)
{
if (testDoc.dims[j] != null)
{
doc.Add(new FacetField("dim" + j, testDoc.dims[j]));
}
}
w.AddDocument(config.Build(tw, doc));
}
// NRT open
IndexSearcher searcher = NewSearcher(w.Reader);
// NRT open
var tr = new DirectoryTaxonomyReader(tw);
ValueSource values = new FloatFieldSource("value");
int iters = AtLeast(100);
for (int iter = 0; iter < iters; iter++)
{
string searchToken = tokens[Random().Next(tokens.Length)];
if (VERBOSE)
{
Console.WriteLine("\nTEST: iter content=" + searchToken);
}
FacetsCollector fc = new FacetsCollector();
FacetsCollector.Search(searcher, new TermQuery(new Term("content", searchToken)), 10, fc);
Facets facets = new TaxonomyFacetSumValueSource(tr, config, fc, values);
// Slow, yet hopefully bug-free, faceting:
var expectedValues = new List <Dictionary <string, float?> >();
for (int i = 0; i < numDims; i++)
{
expectedValues[i] = new Dictionary <string, float?>();
}
foreach (TestDoc doc in testDocs)
{
if (doc.content.Equals(searchToken))
{
for (int j = 0; j < numDims; j++)
{
if (doc.dims[j] != null)
{
float?v = expectedValues[j][doc.dims[j]];
if (v == null)
{
expectedValues[j][doc.dims[j]] = doc.value;
}
else
{
expectedValues[j][doc.dims[j]] = (float)v + doc.value;
}
}
}
}
}
IList <FacetResult> expected = new List <FacetResult>();
for (int i = 0; i < numDims; i++)
{
IList <LabelAndValue> labelValues = new List <LabelAndValue>();
float totValue = 0;
foreach (KeyValuePair <string, float?> ent in expectedValues[i])
{
labelValues.Add(new LabelAndValue(ent.Key, ent.Value.Value));
totValue += ent.Value.Value;
}
SortLabelValues(labelValues);
if (totValue > 0)
{
expected.Add(new FacetResult("dim" + i, new string[0], totValue, labelValues.ToArray(), labelValues.Count));
}
}
// Sort by highest value, tie break by value:
SortFacetResults(expected);
IList <FacetResult> actual = facets.GetAllDims(10);
// Messy: fixup ties
SortTies(actual);
if (VERBOSE)
{
Console.WriteLine("expected=\n" + expected.ToString());
Console.WriteLine("actual=\n" + actual.ToString());
}
AssertFloatValuesEquals(expected, actual);
}
IOUtils.Close(w, tw, searcher.IndexReader, tr, indexDir, taxoDir);
}
public virtual void TestRandomLongs()
{
Directory dir = NewDirectory();
var w = new RandomIndexWriter(Random(), dir);
int numDocs = AtLeast(1000);
if (VERBOSE)
{
Console.WriteLine("TEST: numDocs=" + numDocs);
}
long[] values = new long[numDocs];
long minValue = long.MaxValue;
long maxValue = long.MinValue;
for (int i = 0; i < numDocs; i++)
{
Document doc = new Document();
long v = Random().NextLong();
values[i] = v;
doc.Add(new NumericDocValuesField("field", v));
doc.Add(new LongField("field", v, Field.Store.NO));
w.AddDocument(doc);
minValue = Math.Min(minValue, v);
maxValue = Math.Max(maxValue, v);
}
IndexReader r = w.Reader;
IndexSearcher s = NewSearcher(r);
FacetsConfig config = new FacetsConfig();
int numIters = AtLeast(10);
for (int iter = 0; iter < numIters; iter++)
{
if (VERBOSE)
{
Console.WriteLine("TEST: iter=" + iter);
}
int numRange = TestUtil.NextInt(Random(), 1, 100);
LongRange[] ranges = new LongRange[numRange];
int[] expectedCounts = new int[numRange];
long minAcceptedValue = long.MaxValue;
long maxAcceptedValue = long.MinValue;
for (int rangeID = 0; rangeID < numRange; rangeID++)
{
long min;
if (rangeID > 0 && Random().Next(10) == 7)
{
// Use an existing boundary:
LongRange prevRange = ranges[Random().Next(rangeID)];
if (Random().NextBoolean())
{
min = prevRange.min;
}
else
{
min = prevRange.max;
}
}
else
{
min = Random().NextLong();
}
long max;
if (rangeID > 0 && Random().Next(10) == 7)
{
// Use an existing boundary:
LongRange prevRange = ranges[Random().Next(rangeID)];
if (Random().NextBoolean())
{
max = prevRange.min;
}
else
{
max = prevRange.max;
}
}
else
{
max = Random().NextLong();
}
if (min > max)
{
long x = min;
min = max;
max = x;
}
bool minIncl;
bool maxIncl;
if (min == max)
{
minIncl = true;
maxIncl = true;
}
else
{
minIncl = Random().NextBoolean();
maxIncl = Random().NextBoolean();
}
ranges[rangeID] = new LongRange("r" + rangeID, min, minIncl, max, maxIncl);
if (VERBOSE)
{
Console.WriteLine(" range " + rangeID + ": " + ranges[rangeID]);
}
// Do "slow but hopefully correct" computation of
// expected count:
for (int i = 0; i < numDocs; i++)
{
bool accept = true;
if (minIncl)
{
accept &= values[i] >= min;
}
else
{
accept &= values[i] > min;
}
if (maxIncl)
{
accept &= values[i] <= max;
}
else
{
accept &= values[i] < max;
}
if (accept)
{
expectedCounts[rangeID]++;
minAcceptedValue = Math.Min(minAcceptedValue, values[i]);
maxAcceptedValue = Math.Max(maxAcceptedValue, values[i]);
}
}
}
FacetsCollector sfc = new FacetsCollector();
s.Search(new MatchAllDocsQuery(), sfc);
Filter fastMatchFilter;
if (Random().NextBoolean())
{
if (Random().NextBoolean())
{
fastMatchFilter = NumericRangeFilter.NewLongRange("field", minValue, maxValue, true, true);
}
else
{
fastMatchFilter = NumericRangeFilter.NewLongRange("field", minAcceptedValue, maxAcceptedValue, true, true);
}
}
else
{
fastMatchFilter = null;
}
ValueSource vs = new LongFieldSource("field");
Facets facets = new LongRangeFacetCounts("field", vs, sfc, fastMatchFilter, ranges);
FacetResult result = facets.GetTopChildren(10, "field");
Assert.AreEqual(numRange, result.LabelValues.Length);
for (int rangeID = 0; rangeID < numRange; rangeID++)
{
if (VERBOSE)
{
Console.WriteLine(" range " + rangeID + " expectedCount=" + expectedCounts[rangeID]);
}
LabelAndValue subNode = result.LabelValues[rangeID];
Assert.AreEqual("r" + rangeID, subNode.label);
Assert.AreEqual(expectedCounts[rangeID], (int)subNode.value);
LongRange range = ranges[rangeID];
// Test drill-down:
DrillDownQuery ddq = new DrillDownQuery(config);
if (Random().NextBoolean())
{
if (Random().NextBoolean())
{
ddq.Add("field", NumericRangeFilter.NewLongRange("field", range.min, range.max, range.minInclusive, range.maxInclusive));
}
else
{
ddq.Add("field", NumericRangeQuery.NewLongRange("field", range.min, range.max, range.minInclusive, range.maxInclusive));
}
}
else
{
ddq.Add("field", range.GetFilter(fastMatchFilter, vs));
}
Assert.AreEqual(expectedCounts[rangeID], s.Search(ddq, 10).TotalHits);
}
}
IOUtils.Close(w, r, dir);
}
public virtual void TestBasic()
{
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);
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
FacetsConfig config = new FacetsConfig();
// Reused across documents, to add the necessary facet
// fields:
Document doc = new Document();
doc.Add(new IntField("num", 10, Field.Store.NO));
doc.Add(new FacetField("Author", "Bob"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new IntField("num", 20, Field.Store.NO));
doc.Add(new FacetField("Author", "Lisa"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new IntField("num", 30, Field.Store.NO));
doc.Add(new FacetField("Author", "Lisa"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new IntField("num", 40, Field.Store.NO));
doc.Add(new FacetField("Author", "Susan"));
writer.AddDocument(config.Build(taxoWriter, doc));
doc = new Document();
doc.Add(new IntField("num", 45, Field.Store.NO));
doc.Add(new FacetField("Author", "Frank"));
writer.AddDocument(config.Build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = NewSearcher(writer.Reader);
writer.Dispose();
// NRT open
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
taxoWriter.Dispose();
// 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 one of the
// Facets.search utility methods:
searcher.Search(new MatchAllDocsQuery(), c);
TaxonomyFacetSumValueSource facets = new TaxonomyFacetSumValueSource(taxoReader, new FacetsConfig(), c, new IntFieldSource("num"));
// Retrieve & verify results:
Assert.AreEqual("dim=Author path=[] value=145.0 childCount=4\n Lisa (50.0)\n Frank (45.0)\n Susan (40.0)\n Bob (10.0)\n", facets.GetTopChildren(10, "Author").ToString());
taxoReader.Dispose();
searcher.IndexReader.Dispose();
dir.Dispose();
taxoDir.Dispose();
}
public virtual void TestCountRoot()
{
// LUCENE-4882: FacetsAccumulator threw NPE if a FacetRequest was defined on CP.EMPTY
Store.Directory indexDir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
IndexWriter iw = new IndexWriter(indexDir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())));
FacetsConfig config = new FacetsConfig();
for (int i = AtLeast(30); i > 0; --i)
{
Document doc = new Document();
doc.Add(new FacetField("a", "1"));
doc.Add(new FacetField("b", "1"));
iw.AddDocument(config.Build(taxoWriter, doc));
}
DirectoryReader r = DirectoryReader.Open(iw, true);
DirectoryTaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector sfc = new FacetsCollector();
NewSearcher(r).Search(new MatchAllDocsQuery(), sfc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, config, sfc);
foreach (FacetResult result in facets.GetAllDims(10))
{
Assert.AreEqual(r.NumDocs, (int)result.Value);
}
IOUtils.Close(taxoWriter, iw, taxoReader, taxoDir, r, indexDir);
}
/// <summary>
/// Create <see cref="FastTaxonomyFacetCounts"/>, which also
/// counts all facet labels.
/// </summary>
public FastTaxonomyFacetCounts(TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector fc)
: this(FacetsConfig.DEFAULT_INDEX_FIELD_NAME, taxoReader, config, fc)
{
}
public virtual void TestLabelWithDelimiter()
{
Store.Directory dir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode_e.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetMultiValued("dim", true);
Document doc = new Document();
doc.Add(NewTextField("field", "text", Field.Store.NO));
doc.Add(new FacetField("dim", "test\u001Fone"));
doc.Add(new FacetField("dim", "test\u001Etwo"));
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);
Facets facets = GetTaxonomyFacetCounts(taxoReader, config, c);
Assert.AreEqual(1, facets.GetSpecificValue("dim", "test\u001Fone"));
Assert.AreEqual(1, facets.GetSpecificValue("dim", "test\u001Etwo"));
FacetResult result = facets.GetTopChildren(10, "dim");
Assert.AreEqual("dim=dim path=[] value=-1 childCount=2\n test\u001Fone (1)\n test\u001Etwo (1)\n", result.ToString());
IOUtils.Close(writer, taxoWriter, searcher.IndexReader, taxoReader, dir, taxoDir);
}
/// <summary>
/// Create <see cref="FastTaxonomyFacetCounts"/>, using the
/// specified <paramref name="indexFieldName"/> for ordinals. Use
/// this if you had set <see cref="FacetsConfig.SetIndexFieldName"/>
/// to change the index
/// field name for certain dimensions.
/// </summary>
public FastTaxonomyFacetCounts(string indexFieldName, TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector fc)
: base(indexFieldName, taxoReader, config)
{
Count(fc.GetMatchingDocs());
}
public virtual void TestMultiValuedHierarchy()
{
Store.Directory dir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode_e.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetHierarchical("a", true);
config.SetMultiValued("a", true);
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
Document doc = new Document();
doc.Add(NewTextField("field", "text", Field.Store.NO));
doc.Add(new FacetField("a", "path", "x"));
doc.Add(new FacetField("a", "path", "y"));
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 = GetTaxonomyFacetCounts(taxoReader, config, c);
try
{
facets.GetSpecificValue("a");
Fail("didn't hit expected exception");
}
catch (System.ArgumentException)
{
// expected
}
FacetResult result = facets.GetTopChildren(10, "a");
Assert.AreEqual(1, result.LabelValues.Length);
Assert.AreEqual(1, (int)result.LabelValues[0].value);
IOUtils.Close(writer, taxoWriter, searcher.IndexReader, taxoReader, dir, taxoDir);
}
/// <summary>
/// Create <see cref="RangeFacetCounts"/>, using the provided
/// <see cref="ValueSource"/>, and using the provided Filter as
/// a fastmatch: only documents passing the filter are
/// checked for the matching ranges. The filter must be
/// random access (implement <see cref="DocIdSet.Bits"/>).
/// </summary>
public DoubleRangeFacetCounts(string field, ValueSource valueSource, FacetsCollector hits, Filter fastMatchFilter, params DoubleRange[] ranges)
: base(field, ranges, fastMatchFilter)
{
Count(valueSource, hits.GetMatchingDocs());
}
public virtual void TestRequireDimCount()
{
Store.Directory dir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode_e.CREATE);
FacetsConfig config = new FacetsConfig();
config.SetRequireDimCount("dim", true);
config.SetMultiValued("dim2", true);
config.SetRequireDimCount("dim2", true);
config.SetMultiValued("dim3", true);
config.SetHierarchical("dim3", true);
config.SetRequireDimCount("dim3", true);
Document doc = new Document();
doc.Add(NewTextField("field", "text", Field.Store.NO));
doc.Add(new FacetField("dim", "a"));
doc.Add(new FacetField("dim2", "a"));
doc.Add(new FacetField("dim2", "b"));
doc.Add(new FacetField("dim3", "a", "b"));
doc.Add(new FacetField("dim3", "a", "c"));
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);
Facets facets = GetTaxonomyFacetCounts(taxoReader, config, c);
Assert.AreEqual(1, facets.GetTopChildren(10, "dim").Value);
Assert.AreEqual(1, facets.GetTopChildren(10, "dim2").Value);
Assert.AreEqual(1, facets.GetTopChildren(10, "dim3").Value);
try
{
Assert.AreEqual(1, facets.GetSpecificValue("dim"));
Fail("didn't hit expected exception");
}
catch (System.ArgumentException)
{
// expected
}
Assert.AreEqual(1, facets.GetSpecificValue("dim2"));
Assert.AreEqual(1, facets.GetSpecificValue("dim3"));
IOUtils.Close(writer, taxoWriter, searcher.IndexReader, taxoReader, dir, taxoDir);
}
/// <summary>
/// Create <see cref="RangeFacetCounts"/>, using
/// <see cref="DoubleFieldSource"/> from the specified field.
/// </summary>
public DoubleRangeFacetCounts(string field, FacetsCollector hits, params DoubleRange[] ranges)
: this(field, new DoubleFieldSource(field), hits, ranges)
{
}
public virtual void TestSeparateIndexedFields()
{
Store.Directory indexDir = NewDirectory();
Store.Directory taxoDir = NewDirectory();
var taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
IndexWriter iw = new IndexWriter(indexDir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())));
FacetsConfig config = new FacetsConfig();
config.SetIndexFieldName("b", "$b");
for (int i = AtLeast(30); i > 0; --i)
{
Document doc = new Document();
doc.Add(new StringField("f", "v", Field.Store.NO));
doc.Add(new FacetField("a", "1"));
doc.Add(new FacetField("b", "1"));
iw.AddDocument(config.Build(taxoWriter, doc));
}
DirectoryReader r = DirectoryReader.Open(iw, true);
var taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector sfc = new FacetsCollector();
NewSearcher(r).Search(new MatchAllDocsQuery(), sfc);
Facets facets1 = GetTaxonomyFacetCounts(taxoReader, config, sfc);
Facets facets2 = GetTaxonomyFacetCounts(taxoReader, config, sfc, "$b");
Assert.AreEqual(r.MaxDoc, (int)facets1.GetTopChildren(10, "a").Value);
Assert.AreEqual(r.MaxDoc, (int)facets2.GetTopChildren(10, "b").Value);
IOUtils.Close(taxoWriter, iw, taxoReader, taxoDir, r, indexDir);
}
/// <summary>
/// Create <see cref="RangeFacetCounts"/>, using the provided
/// <see cref="ValueSource"/>.
/// </summary>
public DoubleRangeFacetCounts(string field, ValueSource valueSource, FacetsCollector hits, params DoubleRange[] ranges)
: this(field, valueSource, hits, null, ranges)
{
}
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);
}
/// <summary>
/// Create <see cref="LongRangeFacetCounts"/>, using
/// <see cref="LongFieldSource"/> from the specified field.
/// </summary>
public LongRangeFacetCounts(string field, FacetsCollector hits, params LongRange[] ranges)
: this(field, new LongFieldSource(field), hits, ranges)
{
}
public virtual void TestDifferentNumResults()
{
// test the collector w/ FacetRequests and different numResults
DirectoryReader indexReader = DirectoryReader.Open(indexDir);
var taxoReader = new DirectoryTaxonomyReader(taxoDir);
IndexSearcher searcher = NewSearcher(indexReader);
FacetsCollector sfc = new FacetsCollector();
TermQuery q = new TermQuery(A);
searcher.Search(q, sfc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, Config, sfc);
FacetResult result = facets.GetTopChildren(NUM_CHILDREN_CP_A, CP_A);
Assert.AreEqual(-1, (int)result.Value);
foreach (LabelAndValue labelValue in result.LabelValues)
{
Assert.AreEqual(termExpectedCounts[CP_A + "/" + labelValue.label], labelValue.value);
}
result = facets.GetTopChildren(NUM_CHILDREN_CP_B, CP_B);
Assert.AreEqual(termExpectedCounts[CP_B], result.Value);
foreach (LabelAndValue labelValue in result.LabelValues)
{
Assert.AreEqual(termExpectedCounts[CP_B + "/" + labelValue.label], labelValue.value);
}
IOUtils.Close(indexReader, taxoReader);
}
/// <summary>
/// Create <see cref="RangeFacetCounts"/>, using the provided
/// <see cref="ValueSource"/>, and using the provided Filter as
/// a fastmatch: only documents passing the filter are
/// checked for the matching ranges. The filter must be
/// random access (implement <see cref="DocIdSet.GetBits"/>).
/// </summary>
public LongRangeFacetCounts(string field, ValueSource valueSource,
FacetsCollector hits, Filter fastMatchFilter, params LongRange[] ranges)
: base(field, ranges, fastMatchFilter)
{
Count(valueSource, hits.GetMatchingDocs());
}
public virtual void TestNoParents()
{
DirectoryReader indexReader = DirectoryReader.Open(indexDir);
var taxoReader = new DirectoryTaxonomyReader(taxoDir);
IndexSearcher searcher = NewSearcher(indexReader);
var sfc = new FacetsCollector();
searcher.Search(new MatchAllDocsQuery(), sfc);
Facets facets = GetTaxonomyFacetCounts(taxoReader, Config, sfc);
FacetResult result = facets.GetTopChildren(NUM_CHILDREN_CP_C, CP_C);
Assert.AreEqual(allExpectedCounts[CP_C], result.Value);
foreach (LabelAndValue labelValue in result.LabelValues)
{
Assert.AreEqual(allExpectedCounts[CP_C + "/" + labelValue.label], labelValue.value);
}
result = facets.GetTopChildren(NUM_CHILDREN_CP_D, CP_D);
Assert.AreEqual(allExpectedCounts[CP_C], result.Value);
foreach (LabelAndValue labelValue in result.LabelValues)
{
Assert.AreEqual(allExpectedCounts[CP_D + "/" + labelValue.label], labelValue.value);
}
IOUtils.Close(indexReader, taxoReader);
}
/// <summary>
/// Create {@code TaxonomyFacetSumIntAssociations} against
/// the specified index field.
/// </summary>
public TaxonomyFacetSumIntAssociations(string indexFieldName, TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector fc)
: base(indexFieldName, taxoReader, config)
{
SumValues(fc.GetMatchingDocs);
}
/// <summary>
/// Create <see cref="RangeFacetCounts"/>, using the provided
/// <see cref="ValueSource"/>.
/// </summary>
public LongRangeFacetCounts(string field, ValueSource valueSource, FacetsCollector hits, params LongRange[] ranges)
: this(field, valueSource, hits, null, ranges)
{
}
public virtual void TestCustomDoublesValueSource()
{
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir);
Document doc = new Document();
writer.AddDocument(doc);
writer.AddDocument(doc);
writer.AddDocument(doc);
// Test wants 3 docs in one segment:
writer.ForceMerge(1);
var vs = new ValueSourceAnonymousInnerClassHelper(this, doc);
FacetsConfig config = new FacetsConfig();
FacetsCollector fc = new FacetsCollector();
IndexReader r = writer.Reader;
IndexSearcher s = NewSearcher(r);
s.Search(new MatchAllDocsQuery(), fc);
DoubleRange[] ranges = new DoubleRange[] { new DoubleRange("< 1", 0.0, true, 1.0, false), new DoubleRange("< 2", 0.0, true, 2.0, false), new DoubleRange("< 5", 0.0, true, 5.0, false), new DoubleRange("< 10", 0.0, true, 10.0, false), new DoubleRange("< 20", 0.0, true, 20.0, false), new DoubleRange("< 50", 0.0, true, 50.0, false) };
Filter fastMatchFilter;
AtomicBoolean filterWasUsed = new AtomicBoolean();
if (Random().NextBoolean())
{
// Sort of silly:
fastMatchFilter = new CachingWrapperFilterAnonymousInnerClassHelper(this, new QueryWrapperFilter(new MatchAllDocsQuery()), filterWasUsed);
}
else
{
fastMatchFilter = null;
}
if (VERBOSE)
{
Console.WriteLine("TEST: fastMatchFilter=" + fastMatchFilter);
}
Facets facets = new DoubleRangeFacetCounts("field", vs, fc, fastMatchFilter, ranges);
Assert.AreEqual("dim=field path=[] value=3 childCount=6\n < 1 (0)\n < 2 (1)\n < 5 (3)\n < 10 (3)\n < 20 (3)\n < 50 (3)\n", facets.GetTopChildren(10, "field").ToString());
Assert.True(fastMatchFilter == null || filterWasUsed.Get());
DrillDownQuery ddq = new DrillDownQuery(config);
ddq.Add("field", ranges[1].GetFilter(fastMatchFilter, vs));
// Test simple drill-down:
Assert.AreEqual(1, s.Search(ddq, 10).TotalHits);
// Test drill-sideways after drill-down
DrillSideways ds = new DrillSidewaysAnonymousInnerClassHelper2(this, s, config, (TaxonomyReader)null, vs, ranges, fastMatchFilter);
DrillSidewaysResult dsr = ds.Search(ddq, 10);
Assert.AreEqual(1, dsr.Hits.TotalHits);
Assert.AreEqual("dim=field path=[] value=3 childCount=6\n < 1 (0)\n < 2 (1)\n < 5 (3)\n < 10 (3)\n < 20 (3)\n < 50 (3)\n", dsr.Facets.GetTopChildren(10, "field").ToString());
IOUtils.Close(r, writer, dir);
}
/// <summary>
/// Searches the Lucene index for Documents that match the specified search criteria.
/// </summary>
/// <param name="criteria">The search criteria.</param>
/// <returns></returns>
/// <exception cref="ArgumentNullException"></exception>
/// <exception cref="System.ArgumentNullException"></exception>
public SearchResult<Guid> Search(SearchCriteria criteria)
{
if (criteria == null)
throw new ArgumentNullException(nameof(criteria));
criteria.Query = String.IsNullOrWhiteSpace(criteria.Query) ? ALL_DOCS_QUERY : criteria.Query;
criteria.TopN = criteria.TopN > 0 ? criteria.TopN : SearchCriteria.DEFAULT_TOP_N;
criteria.ItemsPerPage = criteria.ItemsPerPage > 0 ? criteria.ItemsPerPage : SearchCriteria.DEFAULT_ITEMS_PER_PAGE;
criteria.PageNumber = criteria.PageNumber > 0 ? criteria.PageNumber : 1;
criteria.Validate();
var result = new SearchResult<Guid>(criteria);
var queryParser = new LuceneQueryParser(Schema.StandardField.FULL_TEXT, _compositeAnalyzer, Schema);
var query = queryParser.Parse(criteria.Query);
var instance = _searcherTaxonomyManager.Acquire() as SearcherTaxonomyManagerSearcherAndTaxonomy;
if (instance != null)
{
var searcher = instance.Searcher;
var taxonomyReader = instance.TaxonomyReader;
try
{
var sort = GetSortCriteria(criteria.SortByField);
var selectedFacets = criteria.SelectCategories.ToFacetFields();
var topDocs = (TopDocs)null;
var categories = (IEnumerable<Category>)null;
if (selectedFacets.Count() == 0)
{
// We are not going to do a drill-down on specific facets.
// Instead we will just take the top N facets from the matching Documents.
var facetsCollector = new FacetsCollector();
// Get the matching Documents
topDocs = FacetsCollector.Search(searcher, query, criteria.TopN, sort, facetsCollector);
// Get the Facet counts from the matching Documents
var facetCounts = new FastTaxonomyFacetCounts(taxonomyReader, _facetBuilder.FacetsConfig, facetsCollector);
categories = facetCounts.GetCategories(criteria.TopNCategories);
}
else
{
// Perform a drill-sideways query
var drillDownQuery = new DrillDownQuery(_facetBuilder.FacetsConfig, query);
foreach (var facetField in selectedFacets)
drillDownQuery.Add(facetField.Dim, facetField.Path);
var drillSideways = new DrillSideways(searcher, _facetBuilder.FacetsConfig, taxonomyReader);
var drillSidewaysResult = drillSideways.Search(drillDownQuery, null, null, criteria.TopN, sort, false, false);
// Get the matching documents
topDocs = drillSidewaysResult.Hits;
// Get the Facet counts from the matching Documents
categories = drillSidewaysResult.Facets.GetCategories(criteria.TopNCategories, selectedFacets);
}
// TODO: Don't pass TopDocs; pass an IEnumerable<Guid>
result.PopulateWith(topDocs, categories, id => searcher.Doc(id));
}
finally
{
_searcherTaxonomyManager.Release(instance);
searcher = null;
taxonomyReader = null;
}
}
return result;
}
public virtual void TestLongMinMax()
{
Directory d = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random(), d);
Document doc = new Document();
NumericDocValuesField field = new NumericDocValuesField("field", 0L);
doc.Add(field);
field.LongValue = long.MinValue;
w.AddDocument(doc);
field.LongValue = 0;
w.AddDocument(doc);
field.LongValue = long.MaxValue;
w.AddDocument(doc);
IndexReader r = w.Reader;
w.Dispose();
FacetsCollector fc = new FacetsCollector();
IndexSearcher s = NewSearcher(r);
s.Search(new MatchAllDocsQuery(), fc);
Facets facets = new LongRangeFacetCounts("field", fc, new LongRange("min", long.MinValue, true, long.MinValue, true), new LongRange("max", long.MaxValue, true, long.MaxValue, true), new LongRange("all0", long.MinValue, true, long.MaxValue, true), new LongRange("all1", long.MinValue, false, long.MaxValue, true), new LongRange("all2", long.MinValue, true, long.MaxValue, false), new LongRange("all3", long.MinValue, false, long.MaxValue, false));
FacetResult result = facets.GetTopChildren(10, "field");
Assert.AreEqual("dim=field path=[] value=3 childCount=6\n min (1)\n max (1)\n all0 (3)\n all1 (2)\n all2 (2)\n all3 (1)\n", result.ToString());
r.Dispose();
d.Dispose();
}
/// <summary>
/// Create <see cref="TaxonomyFacetSumIntAssociations"/> against
/// the specified index field.
/// </summary>
public TaxonomyFacetSumIntAssociations(string indexFieldName, TaxonomyReader taxoReader, FacetsConfig config, FacetsCollector fc)
: base(indexFieldName, taxoReader, config)
{
SumValues(fc.GetMatchingDocs());
}
/// <summary>
/// Aggreggates float facet values from the provided
/// <see cref="ValueSource"/>, pulling ordinals using <see cref="DocValuesOrdinalsReader"/>
/// against the default indexed
/// facet field <see cref="FacetsConfig.DEFAULT_INDEX_FIELD_NAME"/>.
/// </summary>
public TaxonomyFacetSumValueSource(TaxonomyReader taxoReader, FacetsConfig config,
FacetsCollector fc, ValueSource valueSource)
: this(new DocValuesOrdinalsReader(FacetsConfig.DEFAULT_INDEX_FIELD_NAME),
taxoReader, config, fc, valueSource)
{
}
