Search in sources :

Example 6 with LabelAndValue

use of org.apache.lucene.facet.LabelAndValue in project lucene-solr by apache.

the class TestSortedSetDocValuesFacets method testRandom.

public void testRandom() throws Exception {
    String[] tokens = getRandomTokens(10);
    Directory indexDir = newDirectory();
    Directory taxoDir = newDirectory();
    RandomIndexWriter w = new RandomIndexWriter(random(), indexDir);
    FacetsConfig config = new FacetsConfig();
    int numDocs = atLeast(1000);
    int numDims = TestUtil.nextInt(random(), 1, 7);
    List<TestDoc> testDocs = getRandomDocs(tokens, numDocs, numDims);
    for (TestDoc testDoc : 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 SortedSetDocValuesFacetField("dim" + j, testDoc.dims[j]));
            }
        }
        w.addDocument(config.build(doc));
    }
    // NRT open
    IndexSearcher searcher = newSearcher(w.getReader());
    // Per-top-reader state:
    SortedSetDocValuesReaderState state = new DefaultSortedSetDocValuesReaderState(searcher.getIndexReader());
    ExecutorService exec = randomExecutorServiceOrNull();
    int iters = atLeast(100);
    for (int iter = 0; iter < iters; iter++) {
        String searchToken = tokens[random().nextInt(tokens.length)];
        if (VERBOSE) {
            System.out.println("\nTEST: iter content=" + searchToken);
        }
        FacetsCollector fc = new FacetsCollector();
        FacetsCollector.search(searcher, new TermQuery(new Term("content", searchToken)), 10, fc);
        Facets facets;
        if (exec != null) {
            facets = new ConcurrentSortedSetDocValuesFacetCounts(state, fc, exec);
        } else {
            facets = new SortedSetDocValuesFacetCounts(state, fc);
        }
        // Slow, yet hopefully bug-free, faceting:
        @SuppressWarnings({ "rawtypes", "unchecked" }) Map<String, Integer>[] expectedCounts = new HashMap[numDims];
        for (int i = 0; i < numDims; i++) {
            expectedCounts[i] = new HashMap<>();
        }
        for (TestDoc doc : testDocs) {
            if (doc.content.equals(searchToken)) {
                for (int j = 0; j < numDims; j++) {
                    if (doc.dims[j] != null) {
                        Integer v = expectedCounts[j].get(doc.dims[j]);
                        if (v == null) {
                            expectedCounts[j].put(doc.dims[j], 1);
                        } else {
                            expectedCounts[j].put(doc.dims[j], v.intValue() + 1);
                        }
                    }
                }
            }
        }
        List<FacetResult> expected = new ArrayList<>();
        for (int i = 0; i < numDims; i++) {
            List<LabelAndValue> labelValues = new ArrayList<>();
            int totCount = 0;
            for (Map.Entry<String, Integer> ent : expectedCounts[i].entrySet()) {
                labelValues.add(new LabelAndValue(ent.getKey(), ent.getValue()));
                totCount += ent.getValue();
            }
            sortLabelValues(labelValues);
            if (totCount > 0) {
                expected.add(new FacetResult("dim" + i, new String[0], totCount, labelValues.toArray(new LabelAndValue[labelValues.size()]), labelValues.size()));
            }
        }
        // Sort by highest value, tie break by value:
        sortFacetResults(expected);
        List<FacetResult> actual = facets.getAllDims(10);
        // Messy: fixup ties
        //sortTies(actual);
        assertEquals(expected, actual);
    }
    if (exec != null) {
        exec.shutdownNow();
    }
    w.close();
    IOUtils.close(searcher.getIndexReader(), indexDir, taxoDir);
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) Facets(org.apache.lucene.facet.Facets) HashMap(java.util.HashMap) ArrayList(java.util.ArrayList) Document(org.apache.lucene.document.Document) LabelAndValue(org.apache.lucene.facet.LabelAndValue) Directory(org.apache.lucene.store.Directory) TermQuery(org.apache.lucene.search.TermQuery) FacetsConfig(org.apache.lucene.facet.FacetsConfig) Term(org.apache.lucene.index.Term) FacetsCollector(org.apache.lucene.facet.FacetsCollector) ExecutorService(java.util.concurrent.ExecutorService) FacetResult(org.apache.lucene.facet.FacetResult) HashMap(java.util.HashMap) Map(java.util.Map) RandomIndexWriter(org.apache.lucene.index.RandomIndexWriter)

Example 7 with LabelAndValue

use of org.apache.lucene.facet.LabelAndValue in project lucene-solr by apache.

the class TestTaxonomyFacetCounts method testRandom.

public void testRandom() throws Exception {
    String[] tokens = getRandomTokens(10);
    Directory indexDir = newDirectory();
    Directory taxoDir = newDirectory();
    RandomIndexWriter w = new RandomIndexWriter(random(), indexDir);
    DirectoryTaxonomyWriter tw = new DirectoryTaxonomyWriter(taxoDir);
    FacetsConfig config = new FacetsConfig();
    int numDocs = atLeast(1000);
    int numDims = TestUtil.nextInt(random(), 1, 7);
    List<TestDoc> testDocs = getRandomDocs(tokens, numDocs, numDims);
    for (TestDoc testDoc : 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
    TaxonomyReader tr = new DirectoryTaxonomyReader(tw);
    int iters = atLeast(100);
    for (int iter = 0; iter < iters; iter++) {
        String searchToken = tokens[random().nextInt(tokens.length)];
        if (VERBOSE) {
            System.out.println("\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:
        @SuppressWarnings({ "rawtypes", "unchecked" }) Map<String, Integer>[] expectedCounts = new HashMap[numDims];
        for (int i = 0; i < numDims; i++) {
            expectedCounts[i] = new HashMap<>();
        }
        for (TestDoc doc : testDocs) {
            if (doc.content.equals(searchToken)) {
                for (int j = 0; j < numDims; j++) {
                    if (doc.dims[j] != null) {
                        Integer v = expectedCounts[j].get(doc.dims[j]);
                        if (v == null) {
                            expectedCounts[j].put(doc.dims[j], 1);
                        } else {
                            expectedCounts[j].put(doc.dims[j], v.intValue() + 1);
                        }
                    }
                }
            }
        }
        List<FacetResult> expected = new ArrayList<>();
        for (int i = 0; i < numDims; i++) {
            List<LabelAndValue> labelValues = new ArrayList<>();
            int totCount = 0;
            for (Map.Entry<String, Integer> ent : expectedCounts[i].entrySet()) {
                labelValues.add(new LabelAndValue(ent.getKey(), ent.getValue()));
                totCount += ent.getValue();
            }
            sortLabelValues(labelValues);
            if (totCount > 0) {
                expected.add(new FacetResult("dim" + i, new String[0], totCount, labelValues.toArray(new LabelAndValue[labelValues.size()]), labelValues.size()));
            }
        }
        // Sort by highest value, tie break by value:
        sortFacetResults(expected);
        List<FacetResult> actual = facets.getAllDims(10);
        // Messy: fixup ties
        sortTies(actual);
        assertEquals(expected, actual);
    }
    w.close();
    IOUtils.close(tw, searcher.getIndexReader(), tr, indexDir, taxoDir);
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) Facets(org.apache.lucene.facet.Facets) HashMap(java.util.HashMap) ArrayList(java.util.ArrayList) FacetField(org.apache.lucene.facet.FacetField) Document(org.apache.lucene.document.Document) LabelAndValue(org.apache.lucene.facet.LabelAndValue) DirectoryTaxonomyWriter(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyWriter) Directory(org.apache.lucene.store.Directory) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) TermQuery(org.apache.lucene.search.TermQuery) FacetsConfig(org.apache.lucene.facet.FacetsConfig) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) Term(org.apache.lucene.index.Term) FacetsCollector(org.apache.lucene.facet.FacetsCollector) FacetResult(org.apache.lucene.facet.FacetResult) HashMap(java.util.HashMap) Map(java.util.Map) RandomIndexWriter(org.apache.lucene.index.RandomIndexWriter)

Example 8 with LabelAndValue

use of org.apache.lucene.facet.LabelAndValue in project lucene-solr by apache.

the class TestTaxonomyFacetSumValueSource method testRandom.

public void testRandom() throws Exception {
    String[] tokens = getRandomTokens(10);
    Directory indexDir = newDirectory();
    Directory taxoDir = newDirectory();
    RandomIndexWriter w = new RandomIndexWriter(random(), indexDir);
    DirectoryTaxonomyWriter tw = new DirectoryTaxonomyWriter(taxoDir);
    FacetsConfig config = new FacetsConfig();
    int numDocs = atLeast(1000);
    int numDims = TestUtil.nextInt(random(), 1, 7);
    List<TestDoc> testDocs = getRandomDocs(tokens, numDocs, numDims);
    for (TestDoc testDoc : 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.getReader());
    // NRT open
    TaxonomyReader tr = new DirectoryTaxonomyReader(tw);
    int iters = atLeast(100);
    for (int iter = 0; iter < iters; iter++) {
        String searchToken = tokens[random().nextInt(tokens.length)];
        if (VERBOSE) {
            System.out.println("\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, DoubleValuesSource.fromFloatField("value"));
        // Slow, yet hopefully bug-free, faceting:
        @SuppressWarnings({ "rawtypes", "unchecked" }) Map<String, Float>[] expectedValues = new HashMap[numDims];
        for (int i = 0; i < numDims; i++) {
            expectedValues[i] = new HashMap<>();
        }
        for (TestDoc doc : testDocs) {
            if (doc.content.equals(searchToken)) {
                for (int j = 0; j < numDims; j++) {
                    if (doc.dims[j] != null) {
                        Float v = expectedValues[j].get(doc.dims[j]);
                        if (v == null) {
                            expectedValues[j].put(doc.dims[j], doc.value);
                        } else {
                            expectedValues[j].put(doc.dims[j], v + doc.value);
                        }
                    }
                }
            }
        }
        List<FacetResult> expected = new ArrayList<>();
        for (int i = 0; i < numDims; i++) {
            List<LabelAndValue> labelValues = new ArrayList<>();
            double totValue = 0;
            for (Map.Entry<String, Float> ent : expectedValues[i].entrySet()) {
                labelValues.add(new LabelAndValue(ent.getKey(), ent.getValue()));
                totValue += ent.getValue();
            }
            sortLabelValues(labelValues);
            if (totValue > 0) {
                expected.add(new FacetResult("dim" + i, new String[0], totValue, labelValues.toArray(new LabelAndValue[labelValues.size()]), labelValues.size()));
            }
        }
        // Sort by highest value, tie break by value:
        sortFacetResults(expected);
        List<FacetResult> actual = facets.getAllDims(10);
        // Messy: fixup ties
        sortTies(actual);
        if (VERBOSE) {
            System.out.println("expected=\n" + expected.toString());
            System.out.println("actual=\n" + actual.toString());
        }
        assertFloatValuesEquals(expected, actual);
    }
    w.close();
    IOUtils.close(tw, searcher.getIndexReader(), tr, indexDir, taxoDir);
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) Facets(org.apache.lucene.facet.Facets) HashMap(java.util.HashMap) ArrayList(java.util.ArrayList) FloatDocValuesField(org.apache.lucene.document.FloatDocValuesField) FacetField(org.apache.lucene.facet.FacetField) Document(org.apache.lucene.document.Document) LabelAndValue(org.apache.lucene.facet.LabelAndValue) DirectoryTaxonomyWriter(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyWriter) Directory(org.apache.lucene.store.Directory) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) TermQuery(org.apache.lucene.search.TermQuery) FacetsConfig(org.apache.lucene.facet.FacetsConfig) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) Term(org.apache.lucene.index.Term) FacetsCollector(org.apache.lucene.facet.FacetsCollector) FacetResult(org.apache.lucene.facet.FacetResult) HashMap(java.util.HashMap) Map(java.util.Map) RandomIndexWriter(org.apache.lucene.index.RandomIndexWriter)

Example 9 with LabelAndValue

use of org.apache.lucene.facet.LabelAndValue in project lucene-solr by apache.

the class TestTaxonomyFacetCounts2 method testAllCounts.

@Test
public void testAllCounts() throws Exception {
    DirectoryReader indexReader = DirectoryReader.open(indexDir);
    TaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoDir);
    IndexSearcher searcher = newSearcher(indexReader);
    FacetsCollector sfc = new FacetsCollector();
    searcher.search(new MatchAllDocsQuery(), sfc);
    Facets facets = getTaxonomyFacetCounts(taxoReader, getConfig(), sfc);
    FacetResult result = facets.getTopChildren(NUM_CHILDREN_CP_A, CP_A);
    assertEquals(-1, result.value.intValue());
    int prevValue = Integer.MAX_VALUE;
    for (LabelAndValue labelValue : result.labelValues) {
        assertEquals(allExpectedCounts.get(CP_A + "/" + labelValue.label), labelValue.value);
        assertTrue("wrong sort order of sub results: labelValue.value=" + labelValue.value + " prevValue=" + prevValue, labelValue.value.intValue() <= prevValue);
        prevValue = labelValue.value.intValue();
    }
    result = facets.getTopChildren(NUM_CHILDREN_CP_B, CP_B);
    assertEquals(allExpectedCounts.get(CP_B), result.value);
    prevValue = Integer.MAX_VALUE;
    for (LabelAndValue labelValue : result.labelValues) {
        assertEquals(allExpectedCounts.get(CP_B + "/" + labelValue.label), labelValue.value);
        assertTrue("wrong sort order of sub results: labelValue.value=" + labelValue.value + " prevValue=" + prevValue, labelValue.value.intValue() <= prevValue);
        prevValue = labelValue.value.intValue();
    }
    IOUtils.close(indexReader, taxoReader);
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) Facets(org.apache.lucene.facet.Facets) DirectoryReader(org.apache.lucene.index.DirectoryReader) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) FacetResult(org.apache.lucene.facet.FacetResult) MatchAllDocsQuery(org.apache.lucene.search.MatchAllDocsQuery) LabelAndValue(org.apache.lucene.facet.LabelAndValue) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) FacetsCollector(org.apache.lucene.facet.FacetsCollector) Test(org.junit.Test)

Example 10 with LabelAndValue

use of org.apache.lucene.facet.LabelAndValue in project lucene-solr by apache.

the class TestTaxonomyFacetCounts2 method testBigNumResults.

@Test
public void testBigNumResults() throws Exception {
    DirectoryReader indexReader = DirectoryReader.open(indexDir);
    TaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoDir);
    IndexSearcher searcher = newSearcher(indexReader);
    FacetsCollector sfc = new FacetsCollector();
    searcher.search(new MatchAllDocsQuery(), sfc);
    Facets facets = getTaxonomyFacetCounts(taxoReader, getConfig(), sfc);
    FacetResult result = facets.getTopChildren(Integer.MAX_VALUE, CP_A);
    assertEquals(-1, result.value.intValue());
    for (LabelAndValue labelValue : result.labelValues) {
        assertEquals(allExpectedCounts.get(CP_A + "/" + labelValue.label), labelValue.value);
    }
    result = facets.getTopChildren(Integer.MAX_VALUE, CP_B);
    assertEquals(allExpectedCounts.get(CP_B), result.value);
    for (LabelAndValue labelValue : result.labelValues) {
        assertEquals(allExpectedCounts.get(CP_B + "/" + labelValue.label), labelValue.value);
    }
    IOUtils.close(indexReader, taxoReader);
}
Also used : IndexSearcher(org.apache.lucene.search.IndexSearcher) Facets(org.apache.lucene.facet.Facets) DirectoryReader(org.apache.lucene.index.DirectoryReader) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) FacetResult(org.apache.lucene.facet.FacetResult) MatchAllDocsQuery(org.apache.lucene.search.MatchAllDocsQuery) LabelAndValue(org.apache.lucene.facet.LabelAndValue) DirectoryTaxonomyReader(org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader) FacetsCollector(org.apache.lucene.facet.FacetsCollector) Test(org.junit.Test)

Aggregations

LabelAndValue (org.apache.lucene.facet.LabelAndValue)29 FacetResult (org.apache.lucene.facet.FacetResult)28 Facets (org.apache.lucene.facet.Facets)23 FacetsCollector (org.apache.lucene.facet.FacetsCollector)22 IndexSearcher (org.apache.lucene.search.IndexSearcher)12 IOException (java.io.IOException)11 ArrayList (java.util.ArrayList)11 DefaultSortedSetDocValuesReaderState (org.apache.lucene.facet.sortedset.DefaultSortedSetDocValuesReaderState)11 SortedSetDocValuesFacetCounts (org.apache.lucene.facet.sortedset.SortedSetDocValuesFacetCounts)11 SortedSetDocValuesReaderState (org.apache.lucene.facet.sortedset.SortedSetDocValuesReaderState)11 DirectoryTaxonomyReader (org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader)9 Document (org.apache.lucene.document.Document)7 FacetsConfig (org.apache.lucene.facet.FacetsConfig)7 RandomIndexWriter (org.apache.lucene.index.RandomIndexWriter)7 MatchAllDocsQuery (org.apache.lucene.search.MatchAllDocsQuery)7 Directory (org.apache.lucene.store.Directory)7 DirectoryReader (org.apache.lucene.index.DirectoryReader)6 TermQuery (org.apache.lucene.search.TermQuery)5 HashMap (java.util.HashMap)4 Map (java.util.Map)4