You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@lucene.apache.org by ho...@apache.org on 2016/03/01 18:07:23 UTC
[40/50] [abbrv] lucene-solr git commit: Fix tests to randomized BKD
parameters in tests (like TestPointQueries does).
Fix tests to randomized BKD parameters in tests (like TestPointQueries does).
Otherwise, even things like range queries are not tested well unless tests use thousands of documents.
Project: http://git-wip-us.apache.org/repos/asf/lucene-solr/repo
Commit: http://git-wip-us.apache.org/repos/asf/lucene-solr/commit/502b8800
Tree: http://git-wip-us.apache.org/repos/asf/lucene-solr/tree/502b8800
Diff: http://git-wip-us.apache.org/repos/asf/lucene-solr/diff/502b8800
Branch: refs/heads/jira/SOLR-445
Commit: 502b8800fca18e2735d4ecba35af548885bed42b
Parents: dc39fc3
Author: Robert Muir <rm...@apache.org>
Authored: Mon Feb 29 20:29:51 2016 -0500
Committer: Robert Muir <rm...@apache.org>
Committed: Mon Feb 29 20:29:51 2016 -0500
----------------------------------------------------------------------
.../codecs/asserting/AssertingPointFormat.java | 17 ++++++++-
.../org/apache/lucene/index/RandomCodec.java | 37 +++++++++++++++++++-
2 files changed, 52 insertions(+), 2 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/502b8800/lucene/test-framework/src/java/org/apache/lucene/codecs/asserting/AssertingPointFormat.java
----------------------------------------------------------------------
diff --git a/lucene/test-framework/src/java/org/apache/lucene/codecs/asserting/AssertingPointFormat.java b/lucene/test-framework/src/java/org/apache/lucene/codecs/asserting/AssertingPointFormat.java
index e2d1b07..3411c7b 100644
--- a/lucene/test-framework/src/java/org/apache/lucene/codecs/asserting/AssertingPointFormat.java
+++ b/lucene/test-framework/src/java/org/apache/lucene/codecs/asserting/AssertingPointFormat.java
@@ -34,7 +34,22 @@ import org.apache.lucene.util.TestUtil;
*/
public final class AssertingPointFormat extends PointFormat {
- private final PointFormat in = TestUtil.getDefaultCodec().pointFormat();
+ private final PointFormat in;
+
+ /** Create a new AssertingPointFormat */
+ public AssertingPointFormat() {
+ this(TestUtil.getDefaultCodec().pointFormat());
+ }
+
+ /**
+ * Expert: Create an AssertingPointFormat.
+ * This is only intended to pass special parameters for testing.
+ */
+ // TODO: can we randomize this a cleaner way? e.g. stored fields and vectors do
+ // this with a separate codec...
+ public AssertingPointFormat(PointFormat in) {
+ this.in = in;
+ }
@Override
public PointWriter fieldsWriter(SegmentWriteState state) throws IOException {
http://git-wip-us.apache.org/repos/asf/lucene-solr/blob/502b8800/lucene/test-framework/src/java/org/apache/lucene/index/RandomCodec.java
----------------------------------------------------------------------
diff --git a/lucene/test-framework/src/java/org/apache/lucene/index/RandomCodec.java b/lucene/test-framework/src/java/org/apache/lucene/index/RandomCodec.java
index 986dd34..5b04b00 100644
--- a/lucene/test-framework/src/java/org/apache/lucene/index/RandomCodec.java
+++ b/lucene/test-framework/src/java/org/apache/lucene/index/RandomCodec.java
@@ -16,6 +16,7 @@
*/
package org.apache.lucene.index;
+import java.io.IOException;
import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
@@ -27,15 +28,21 @@ import java.util.Random;
import java.util.Set;
import org.apache.lucene.codecs.DocValuesFormat;
+import org.apache.lucene.codecs.PointFormat;
+import org.apache.lucene.codecs.PointReader;
+import org.apache.lucene.codecs.PointWriter;
import org.apache.lucene.codecs.PostingsFormat;
import org.apache.lucene.codecs.asserting.AssertingCodec;
import org.apache.lucene.codecs.asserting.AssertingDocValuesFormat;
+import org.apache.lucene.codecs.asserting.AssertingPointFormat;
import org.apache.lucene.codecs.asserting.AssertingPostingsFormat;
import org.apache.lucene.codecs.blockterms.LuceneFixedGap;
import org.apache.lucene.codecs.blockterms.LuceneVarGapDocFreqInterval;
import org.apache.lucene.codecs.blockterms.LuceneVarGapFixedInterval;
import org.apache.lucene.codecs.blocktreeords.BlockTreeOrdsPostingsFormat;
import org.apache.lucene.codecs.bloom.TestBloomFilteredLucenePostings;
+import org.apache.lucene.codecs.lucene60.Lucene60PointReader;
+import org.apache.lucene.codecs.lucene60.Lucene60PointWriter;
import org.apache.lucene.codecs.memory.DirectDocValuesFormat;
import org.apache.lucene.codecs.memory.DirectPostingsFormat;
import org.apache.lucene.codecs.memory.FSTOrdPostingsFormat;
@@ -80,6 +87,28 @@ public class RandomCodec extends AssertingCodec {
private Map<String,DocValuesFormat> previousDVMappings = Collections.synchronizedMap(new HashMap<String,DocValuesFormat>());
private final int perFieldSeed;
+ // a little messy: randomize the default codec's parameters here.
+ // with the default values, we have e,g, 1024 points in leaf nodes,
+ // which is less effective for testing.
+ // TODO: improve how we randomize this...
+ private final int maxPointsInLeafNode;
+ private final double maxMBSortInHeap;
+
+ @Override
+ public PointFormat pointFormat() {
+ return new AssertingPointFormat(new PointFormat() {
+ @Override
+ public PointWriter fieldsWriter(SegmentWriteState writeState) throws IOException {
+ return new Lucene60PointWriter(writeState, maxPointsInLeafNode, maxMBSortInHeap);
+ }
+
+ @Override
+ public PointReader fieldsReader(SegmentReadState readState) throws IOException {
+ return new Lucene60PointReader(readState);
+ }
+ });
+ }
+
@Override
public PostingsFormat getPostingsFormatForField(String name) {
PostingsFormat codec = previousMappings.get(name);
@@ -121,6 +150,9 @@ public class RandomCodec extends AssertingCodec {
int maxItemsPerBlock = 2*(Math.max(2, minItemsPerBlock-1)) + random.nextInt(100);
int lowFreqCutoff = TestUtil.nextInt(random, 2, 100);
+ maxPointsInLeafNode = TestUtil.nextInt(random, 16, 2048);
+ maxMBSortInHeap = 4.0 + (3*random.nextDouble());
+
add(avoidCodecs,
TestUtil.getDefaultPostingsFormat(minItemsPerBlock, maxItemsPerBlock),
new FSTPostingsFormat(),
@@ -184,6 +216,9 @@ public class RandomCodec extends AssertingCodec {
@Override
public String toString() {
- return super.toString() + ": " + previousMappings.toString() + ", docValues:" + previousDVMappings.toString();
+ return super.toString() + ": " + previousMappings.toString() +
+ ", docValues:" + previousDVMappings.toString() +
+ ", maxPointsInLeafNode=" + maxPointsInLeafNode +
+ ", maxMBSortInHeap=" + maxMBSortInHeap;
}
}