use of org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader in project lucene-solr by apache.
the class TestTaxonomyCombined method testChildrenArraysInvariants.
/**
* Similar to testChildrenArrays, except rather than look at
* expected results, we test for several "invariants" that the results
* should uphold, e.g., that a child of a category indeed has this category
* as its parent. This sort of test can more easily be extended to larger
* example taxonomies, because we do not need to build the expected list
* of categories like we did in the above test.
*/
@Test
public void testChildrenArraysInvariants() throws Exception {
Directory indexDir = newDirectory();
TaxonomyWriter tw = new DirectoryTaxonomyWriter(indexDir);
fillTaxonomy(tw);
tw.close();
TaxonomyReader tr = new DirectoryTaxonomyReader(indexDir);
ParallelTaxonomyArrays ca = tr.getParallelTaxonomyArrays();
int[] children = ca.children();
assertEquals(tr.getSize(), children.length);
int[] olderSiblingArray = ca.siblings();
assertEquals(tr.getSize(), olderSiblingArray.length);
// test that the "youngest child" of every category is indeed a child:
int[] parents = tr.getParallelTaxonomyArrays().parents();
for (int i = 0; i < tr.getSize(); i++) {
int youngestChild = children[i];
if (youngestChild != TaxonomyReader.INVALID_ORDINAL) {
assertEquals(i, parents[youngestChild]);
}
}
// (it can also be INVALID_ORDINAL, which is lower than any ordinal)
for (int i = 0; i < tr.getSize(); i++) {
assertTrue("olderSiblingArray[" + i + "] should be <" + i, olderSiblingArray[i] < i);
}
// (they share the same parent)
for (int i = 0; i < tr.getSize(); i++) {
int sibling = olderSiblingArray[i];
if (sibling == TaxonomyReader.INVALID_ORDINAL) {
continue;
}
assertEquals(parents[i], parents[sibling]);
}
// miss the first children in the chain)
for (int i = 0; i < tr.getSize(); i++) {
// Find the really youngest child:
int j;
for (j = tr.getSize() - 1; j > i; j--) {
if (parents[j] == i) {
// found youngest child
break;
}
}
if (j == i) {
// no child found
j = TaxonomyReader.INVALID_ORDINAL;
}
assertEquals(j, children[i]);
}
// middle or the end of the chain).
for (int i = 0; i < tr.getSize(); i++) {
// Find the youngest older sibling:
int j;
for (j = i - 1; j >= 0; j--) {
if (parents[j] == parents[i]) {
// found youngest older sibling
break;
}
}
if (j < 0) {
// no sibling found
j = TaxonomyReader.INVALID_ORDINAL;
}
assertEquals(j, olderSiblingArray[i]);
}
tr.close();
indexDir.close();
}
use of org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader in project lucene-solr by apache.
the class TestTaxonomyCombined method testChildrenArraysGrowth.
/**
* Test how getChildrenArrays() deals with the taxonomy's growth:
*/
@Test
public void testChildrenArraysGrowth() throws Exception {
Directory indexDir = newDirectory();
TaxonomyWriter tw = new DirectoryTaxonomyWriter(indexDir);
tw.addCategory(new FacetLabel("hi", "there"));
tw.commit();
TaxonomyReader tr = new DirectoryTaxonomyReader(indexDir);
ParallelTaxonomyArrays ca = tr.getParallelTaxonomyArrays();
assertEquals(3, tr.getSize());
assertEquals(3, ca.siblings().length);
assertEquals(3, ca.children().length);
assertTrue(Arrays.equals(new int[] { 1, 2, -1 }, ca.children()));
assertTrue(Arrays.equals(new int[] { -1, -1, -1 }, ca.siblings()));
tw.addCategory(new FacetLabel("hi", "ho"));
tw.addCategory(new FacetLabel("hello"));
tw.commit();
// Before refresh, nothing changed..
ParallelTaxonomyArrays newca = tr.getParallelTaxonomyArrays();
// we got exactly the same object
assertSame(newca, ca);
assertEquals(3, tr.getSize());
assertEquals(3, ca.siblings().length);
assertEquals(3, ca.children().length);
// After the refresh, things change:
TaxonomyReader newtr = TaxonomyReader.openIfChanged(tr);
assertNotNull(newtr);
tr.close();
tr = newtr;
ca = tr.getParallelTaxonomyArrays();
assertEquals(5, tr.getSize());
assertEquals(5, ca.siblings().length);
assertEquals(5, ca.children().length);
assertTrue(Arrays.equals(new int[] { 4, 3, -1, -1, -1 }, ca.children()));
assertTrue(Arrays.equals(new int[] { -1, -1, -1, 2, 1 }, ca.siblings()));
tw.close();
tr.close();
indexDir.close();
}
use of org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader in project lucene-solr by apache.
the class TestTaxonomyFacetCounts method testLabelWithDelimiter.
public void testLabelWithDelimiter() throws Exception {
Directory dir = newDirectory();
Directory taxoDir = newDirectory();
RandomIndexWriter writer = new RandomIndexWriter(random(), dir);
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode.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", "testone"));
doc.add(new FacetField("dim", "testtwo"));
writer.addDocument(config.build(taxoWriter, doc));
// NRT open
IndexSearcher searcher = newSearcher(writer.getReader());
// NRT open
TaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoWriter);
Facets facets = getAllFacets(FacetsConfig.DEFAULT_INDEX_FIELD_NAME, searcher, taxoReader, config);
assertEquals(1, facets.getSpecificValue("dim", "testone"));
assertEquals(1, facets.getSpecificValue("dim", "testtwo"));
FacetResult result = facets.getTopChildren(10, "dim");
assertEquals("dim=dim path=[] value=-1 childCount=2\n testone (1)\n testtwo (1)\n", result.toString());
writer.close();
IOUtils.close(taxoWriter, searcher.getIndexReader(), taxoReader, dir, taxoDir);
}
use of org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader in project lucene-solr by apache.
the class TestTaxonomyFacetCounts method testSparseFacets.
// LUCENE-5333
public void testSparseFacets() throws Exception {
Directory dir = newDirectory();
Directory taxoDir = newDirectory();
// Writes facet ords to a separate directory from the
// main index:
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir, IndexWriterConfig.OpenMode.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.getReader());
// NRT open
TaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoWriter);
Facets facets = getAllFacets(FacetsConfig.DEFAULT_INDEX_FIELD_NAME, searcher, taxoReader, config);
// Ask for top 10 labels for any dims that have counts:
List<FacetResult> results = facets.getAllDims(10);
assertEquals(3, results.size());
assertEquals("dim=a path=[] value=3 childCount=3\n foo1 (1)\n foo2 (1)\n foo3 (1)\n", results.get(0).toString());
assertEquals("dim=b path=[] value=2 childCount=2\n bar1 (1)\n bar2 (1)\n", results.get(1).toString());
assertEquals("dim=c path=[] value=1 childCount=1\n baz1 (1)\n", results.get(2).toString());
writer.close();
IOUtils.close(taxoWriter, searcher.getIndexReader(), taxoReader, taxoDir, dir);
}
use of org.apache.lucene.facet.taxonomy.directory.DirectoryTaxonomyReader in project lucene-solr by apache.
the class TestTaxonomyFacetSumValueSource method testRollupValues.
public void testRollupValues() throws Exception {
Directory indexDir = newDirectory();
Directory taxoDir = newDirectory();
DirectoryTaxonomyWriter taxoWriter = new DirectoryTaxonomyWriter(taxoDir);
IndexWriter iw = new IndexWriter(indexDir, newIndexWriterConfig(new MockAnalyzer(random())));
FacetsConfig config = new FacetsConfig();
config.setHierarchical("a", true);
for (int i = 0; i < 4; i++) {
Document doc = new Document();
doc.add(new NumericDocValuesField("price", (i + 1)));
doc.add(new FacetField("a", Integer.toString(i % 2), "1"));
iw.addDocument(config.build(taxoWriter, doc));
}
DirectoryReader r = DirectoryReader.open(iw);
DirectoryTaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoWriter);
FacetsCollector sfc = new FacetsCollector();
newSearcher(r).search(new MatchAllDocsQuery(), sfc);
Facets facets = new TaxonomyFacetSumValueSource(taxoReader, config, sfc, DoubleValuesSource.fromLongField("price"));
assertEquals("dim=a path=[] value=10.0 childCount=2\n 1 (6.0)\n 0 (4.0)\n", facets.getTopChildren(10, "a").toString());
iw.close();
IOUtils.close(taxoWriter, taxoReader, taxoDir, r, indexDir);
}
Aggregations