private static String[] getStopWords(String path) {
try {
List<String> list = FileUtils.readLines(new File(path), "utf-8");
System.out.println("==============================================================");
System.out.println("getStopWords " + list.get(159));
System.out.println("==============================================================");
return list.toArray(new String[list.size()]);
} catch (FileNotFoundException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
} finally {
}
return null;
}
/**
* Search a picture
*
* @param imageBytes
* @param maxHits
* @param feature
* @param pictures
* @return
* @throws IOException
*/
public static List<LirePictureSortable> search(
byte[] imageBytes, int maxHits, FeatureEnumerate feature, List<LirePictureSortable> pictures)
throws IOException {
File path = getPath(feature.getText());
log.debug("reading from indexed path " + path.getAbsolutePath());
List<LirePictureSortable> result = new ArrayList<>();
try {
IndexReader ir = DirectoryReader.open(FSDirectory.open(path));
ImageSearcher searcher = new GenericFastImageSearcher(maxHits, feature.getValueClass());
// searching with a image file ...
InputStream in = new ByteArrayInputStream(imageBytes);
ImageSearchHits hits = searcher.search(in, ir);
float score = 0.0F;
for (int i = 0; i < hits.length(); i++) {
score = hits.score(i);
LirePictureSortable lp =
new LirePictureSortable(
UUID.fromString(hits.doc(i).getValues(DocumentBuilder.FIELD_NAME_IDENTIFIER)[0]),
score,
feature);
/* check is its a picture was repeated by picture UUID */
if (pictures.contains(lp)) {
lp = pictures.get(pictures.indexOf(lp));
lp.addDescriptor(feature, score);
lp.addScore(score);
} else {
result.add(lp);
}
}
} catch (Exception e) {
e.printStackTrace();
}
return result;
}
@Test
public void testRecoveryDiff() throws IOException, InterruptedException {
int numDocs = 2 + random().nextInt(100);
List<Document> docs = new ArrayList<>();
for (int i = 0; i < numDocs; i++) {
Document doc = new Document();
doc.add(
new StringField("id", "" + i, random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new TextField(
"body",
TestUtil.randomRealisticUnicodeString(random()),
random().nextBoolean() ? Field.Store.YES : Field.Store.NO));
doc.add(
new SortedDocValuesField(
"dv", new BytesRef(TestUtil.randomRealisticUnicodeString(random()))));
docs.add(doc);
}
long seed = random().nextLong();
Store.MetadataSnapshot first;
{
Random random = new Random(seed);
IndexWriterConfig iwc =
new IndexWriterConfig(new MockAnalyzer(random)).setCodec(actualDefaultCodec());
iwc.setMergePolicy(NoMergePolicy.INSTANCE);
iwc.setUseCompoundFile(random.nextBoolean());
iwc.setMaxThreadStates(1);
final ShardId shardId = new ShardId(new Index("index"), 1);
DirectoryService directoryService = new LuceneManagedDirectoryService(random);
Store store =
new Store(
shardId,
ImmutableSettings.EMPTY,
directoryService,
randomDistributor(random, directoryService),
new DummyShardLock(shardId));
IndexWriter writer = new IndexWriter(store.directory(), iwc);
final boolean lotsOfSegments = rarely(random);
for (Document d : docs) {
writer.addDocument(d);
if (lotsOfSegments && random.nextBoolean()) {
writer.commit();
} else if (rarely(random)) {
writer.commit();
}
}
writer.commit();
writer.close();
first = store.getMetadata();
assertDeleteContent(store, directoryService);
store.close();
}
long time = new Date().getTime();
while (time == new Date().getTime()) {
Thread.sleep(10); // bump the time
}
Store.MetadataSnapshot second;
Store store;
{
Random random = new Random(seed);
IndexWriterConfig iwc =
new IndexWriterConfig(new MockAnalyzer(random)).setCodec(actualDefaultCodec());
iwc.setMergePolicy(NoMergePolicy.INSTANCE);
iwc.setUseCompoundFile(random.nextBoolean());
iwc.setMaxThreadStates(1);
final ShardId shardId = new ShardId(new Index("index"), 1);
DirectoryService directoryService = new LuceneManagedDirectoryService(random);
store =
new Store(
shardId,
ImmutableSettings.EMPTY,
directoryService,
randomDistributor(random, directoryService),
new DummyShardLock(shardId));
IndexWriter writer = new IndexWriter(store.directory(), iwc);
final boolean lotsOfSegments = rarely(random);
for (Document d : docs) {
writer.addDocument(d);
if (lotsOfSegments && random.nextBoolean()) {
writer.commit();
} else if (rarely(random)) {
writer.commit();
}
}
writer.commit();
writer.close();
second = store.getMetadata();
}
Store.RecoveryDiff diff = first.recoveryDiff(second);
assertThat(first.size(), equalTo(second.size()));
for (StoreFileMetaData md : first) {
assertThat(second.get(md.name()), notNullValue());
// si files are different - containing timestamps etc
assertThat(second.get(md.name()).isSame(md), equalTo(false));
}
assertThat(diff.different.size(), equalTo(first.size()));
assertThat(
diff.identical.size(),
equalTo(0)); // in lucene 5 nothing is identical - we use random ids in file headers
assertThat(diff.missing, empty());
// check the self diff
Store.RecoveryDiff selfDiff = first.recoveryDiff(first);
assertThat(selfDiff.identical.size(), equalTo(first.size()));
assertThat(selfDiff.different, empty());
assertThat(selfDiff.missing, empty());
// lets add some deletes
Random random = new Random(seed);
IndexWriterConfig iwc =
new IndexWriterConfig(new MockAnalyzer(random)).setCodec(actualDefaultCodec());
iwc.setMergePolicy(NoMergePolicy.INSTANCE);
iwc.setUseCompoundFile(random.nextBoolean());
iwc.setMaxThreadStates(1);
iwc.setOpenMode(IndexWriterConfig.OpenMode.APPEND);
IndexWriter writer = new IndexWriter(store.directory(), iwc);
writer.deleteDocuments(new Term("id", Integer.toString(random().nextInt(numDocs))));
writer.commit();
writer.close();
Store.MetadataSnapshot metadata = store.getMetadata();
StoreFileMetaData delFile = null;
for (StoreFileMetaData md : metadata) {
if (md.name().endsWith(".liv")) {
delFile = md;
break;
}
}
Store.RecoveryDiff afterDeleteDiff = metadata.recoveryDiff(second);
if (delFile != null) {
assertThat(
afterDeleteDiff.identical.size(), equalTo(metadata.size() - 2)); // segments_N + del file
assertThat(afterDeleteDiff.different.size(), equalTo(0));
assertThat(afterDeleteDiff.missing.size(), equalTo(2));
} else {
// an entire segment must be missing (single doc segment got dropped)
assertThat(afterDeleteDiff.identical.size(), greaterThan(0));
assertThat(afterDeleteDiff.different.size(), equalTo(0));
assertThat(afterDeleteDiff.missing.size(), equalTo(1)); // the commit file is different
}
// check the self diff
selfDiff = metadata.recoveryDiff(metadata);
assertThat(selfDiff.identical.size(), equalTo(metadata.size()));
assertThat(selfDiff.different, empty());
assertThat(selfDiff.missing, empty());
// add a new commit
iwc = new IndexWriterConfig(new MockAnalyzer(random)).setCodec(actualDefaultCodec());
iwc.setMergePolicy(NoMergePolicy.INSTANCE);
iwc.setUseCompoundFile(
true); // force CFS - easier to test here since we know it will add 3 files
iwc.setMaxThreadStates(1);
iwc.setOpenMode(IndexWriterConfig.OpenMode.APPEND);
writer = new IndexWriter(store.directory(), iwc);
writer.addDocument(docs.get(0));
writer.close();
Store.MetadataSnapshot newCommitMetaData = store.getMetadata();
Store.RecoveryDiff newCommitDiff = newCommitMetaData.recoveryDiff(metadata);
if (delFile != null) {
assertThat(
newCommitDiff.identical.size(),
equalTo(
newCommitMetaData.size()
- 5)); // segments_N, del file, cfs, cfe, si for the new segment
assertThat(newCommitDiff.different.size(), equalTo(1)); // the del file must be different
assertThat(newCommitDiff.different.get(0).name(), endsWith(".liv"));
assertThat(
newCommitDiff.missing.size(), equalTo(4)); // segments_N,cfs, cfe, si for the new segment
} else {
assertThat(
newCommitDiff.identical.size(),
equalTo(newCommitMetaData.size() - 4)); // segments_N, cfs, cfe, si for the new segment
assertThat(newCommitDiff.different.size(), equalTo(0));
assertThat(
newCommitDiff.missing.size(),
equalTo(
4)); // an entire segment must be missing (single doc segment got dropped) plus the
// commit is different
}
store.deleteContent();
IOUtils.close(store);
}