honeymoose
/
OpenSearch
mirror of https://github.com/honeymoose/OpenSearch.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Fix for highlighting when number_of_fragments is set to 0 and term_vector is not set, closes #873

This commit is contained in:
Lukáš Vlček 2011-04-21 14:55:21 +02:00
parent dedbedbb77
commit adc3dc0e99
2 changed files with 73 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
modules/elasticsearch/src/main/java/org/elasticsearch/search/highlight/HighlightPhase.java
View File

@ -119,13 +119,15 @@ public class HighlightPhase implements SearchHitPhase {
textsToHighlight = lookup.source().getValues(mapper.names().fullName());
}
// a HACK to make highlighter do highlighting, even though its using the single frag list builder
int numberOfFragments = field.numberOfFragments() == 0 ? 1 : field.numberOfFragments();
ArrayList<TextFragment> fragsList = new ArrayList<TextFragment>();
try {
for (Object textToHighlight : textsToHighlight) {
String text = textToHighlight.toString();
Analyzer analyzer = context.mapperService().documentMapper(hitContext.hit().type()).mappers().indexAnalyzer();
TokenStream tokenStream = analyzer.reusableTokenStream(mapper.names().indexName(), new FastStringReader(text));
TextFragment[] bestTextFragments = highlighter.getBestTextFragments(tokenStream, text, false, field.numberOfFragments());
TextFragment[] bestTextFragments = highlighter.getBestTextFragments(tokenStream, text, false, numberOfFragments);
for (TextFragment bestTextFragment : bestTextFragments) {
if (bestTextFragment != null && bestTextFragment.getScore() > 0) {
fragsList.add(bestTextFragment);
@ -142,11 +144,22 @@ public class HighlightPhase implements SearchHitPhase {
}
});
}
int numberOfFragments = fragsList.size() < field.numberOfFragments() ? fragsList.size() : field.numberOfFragments();
String[] fragments = new String[numberOfFragments];
for (int i = 0; i < fragments.length; i++) {
fragments[i] = fragsList.get(i).toString();
String[] fragments;
// number_of_fragments is set to 0 but we have a multivalued field
if (field.numberOfFragments() == 0 && textsToHighlight.size() > 1) {
fragments = new String[1];
for (int i = 0; i < fragsList.size(); i++) {
fragments[0] = (fragments[0] != null ? (fragments[0]+" ") : "") + fragsList.get(i).toString();
}
} else {
// refine numberOfFragments if needed
numberOfFragments = fragsList.size() < numberOfFragments ? fragsList.size() : numberOfFragments;
fragments = new String[numberOfFragments];
for (int i = 0; i < fragments.length; i++) {
fragments[i] = fragsList.get(i).toString();
}
}
if (fragments.length > 0) {
HighlightField highlightField = new HighlightField(field.field(), fragments);
highlightFields.put(highlightField.name(), highlightField);

57
modules/test/integration/src/test/java/org/elasticsearch/test/integration/search/highlight/HighlighterSearchTests.java
View File

@ -68,7 +68,60 @@ public class HighlighterSearchTests extends AbstractNodesTests {
return client("server1");
}
@Test public void testSourceLookupHighlighting() throws Exception {
@Test public void testSourceLookupHighlightingUsingPlainHighlighter() throws Exception {
try {
client.admin().indices().prepareDelete("test").execute().actionGet();
} catch (Exception e) {
// ignore
}
client.admin().indices().prepareCreate("test").setSettings(ImmutableSettings.settingsBuilder().put("number_of_shards", 2))
.addMapping("type1", jsonBuilder().startObject().startObject("type1").startObject("properties")
// we don't store title and don't use term vector, now lets see if it works...
.startObject("title").field("type", "string").field("store", "no").field("term_vector", "no").endObject()
.startObject("attachments").startObject("properties").startObject("body").field("type", "string").field("store", "no").field("term_vector", "no").endObject().endObject().endObject()
.endObject().endObject().endObject())
.execute().actionGet();
for (int i = 0; i < 5; i++) {
client.prepareIndex("test", "type1", Integer.toString(i))
.setSource(XContentFactory.jsonBuilder().startObject()
.field("title", "This is a test on the highlighting bug present in elasticsearch")
.startArray("attachments").startObject().field("body", "attachment 1").endObject().startObject().field("body", "attachment 2").endObject().endArray()
.endObject())
.setRefresh(true).execute().actionGet();
}
SearchResponse search = client.prepareSearch()
.setQuery(fieldQuery("title", "bug"))
.addHighlightedField("title", -1, 0)
.execute().actionGet();
assertThat(search.hits().totalHits(), equalTo(5l));
assertThat(search.hits().hits().length, equalTo(5));
assertThat(search.getFailedShards(), equalTo(0));
for (SearchHit hit : search.hits()) {
assertThat(hit.highlightFields().get("title").fragments()[0], equalTo("This is a test on the highlighting <em>bug</em> present in elasticsearch"));
}
search = client.prepareSearch()
.setQuery(fieldQuery("attachments.body", "attachment"))
.addHighlightedField("attachments.body", -1, 0)
.execute().actionGet();
System.out.println(search);
assertThat(search.hits().totalHits(), equalTo(5l));
assertThat(search.hits().hits().length, equalTo(5));
assertThat(search.getFailedShards(), equalTo(0));
for (SearchHit hit : search.hits()) {
assertThat(hit.highlightFields().get("attachments.body").fragments()[0], equalTo("<em>attachment</em> 1 <em>attachment</em> 2"));
}
}
@Test public void testSourceLookupHighlightingUsingFastVectorHighlighter() throws Exception {
try {
client.admin().indices().prepareDelete("test").execute().actionGet();
} catch (Exception e) {
@ -79,7 +132,7 @@ public class HighlighterSearchTests extends AbstractNodesTests {
.addMapping("type1", jsonBuilder().startObject().startObject("type1").startObject("properties")
// we don't store title, now lets see if it works...
.startObject("title").field("type", "string").field("store", "no").field("term_vector", "with_positions_offsets").endObject()
.startObject("attachments").startObject("properties").startObject("body").field("type", "string").field("term_vector", "with_positions_offsets").endObject().endObject().endObject()
.startObject("attachments").startObject("properties").startObject("body").field("type", "string").field("store", "no").field("term_vector", "with_positions_offsets").endObject().endObject().endObject()
.endObject().endObject().endObject())
.execute().actionGet();