use of org.elasticsearch.search.aggregations.bucket.filter.InternalFilter in project elasticsearch by elastic.
the class SignificantTermsSignificanceScoreIT method testBackgroundVsSeparateSet.
// compute significance score by
// 1. terms agg on class and significant terms
// 2. filter buckets and set the background to the other class and set is_background false
// both should yield exact same result
public void testBackgroundVsSeparateSet(SignificanceHeuristic significanceHeuristicExpectingSuperset, SignificanceHeuristic significanceHeuristicExpectingSeparateSets) throws Exception {
SearchResponse response1 = client().prepareSearch(INDEX_NAME).setTypes(DOC_TYPE).addAggregation(terms("class").field(CLASS_FIELD).subAggregation(significantTerms("sig_terms").field(TEXT_FIELD).minDocCount(1).significanceHeuristic(significanceHeuristicExpectingSuperset))).execute().actionGet();
assertSearchResponse(response1);
SearchResponse response2 = client().prepareSearch(INDEX_NAME).setTypes(DOC_TYPE).addAggregation(filter("0", QueryBuilders.termQuery(CLASS_FIELD, "0")).subAggregation(significantTerms("sig_terms").field(TEXT_FIELD).minDocCount(1).backgroundFilter(QueryBuilders.termQuery(CLASS_FIELD, "1")).significanceHeuristic(significanceHeuristicExpectingSeparateSets))).addAggregation(filter("1", QueryBuilders.termQuery(CLASS_FIELD, "1")).subAggregation(significantTerms("sig_terms").field(TEXT_FIELD).minDocCount(1).backgroundFilter(QueryBuilders.termQuery(CLASS_FIELD, "0")).significanceHeuristic(significanceHeuristicExpectingSeparateSets))).execute().actionGet();
StringTerms classes = response1.getAggregations().get("class");
SignificantTerms sigTerms0 = ((SignificantTerms) (classes.getBucketByKey("0").getAggregations().asMap().get("sig_terms")));
assertThat(sigTerms0.getBuckets().size(), equalTo(2));
double score00Background = sigTerms0.getBucketByKey("0").getSignificanceScore();
double score01Background = sigTerms0.getBucketByKey("1").getSignificanceScore();
SignificantTerms sigTerms1 = ((SignificantTerms) (classes.getBucketByKey("1").getAggregations().asMap().get("sig_terms")));
double score10Background = sigTerms1.getBucketByKey("0").getSignificanceScore();
double score11Background = sigTerms1.getBucketByKey("1").getSignificanceScore();
Aggregations aggs = response2.getAggregations();
sigTerms0 = (SignificantTerms) ((InternalFilter) aggs.get("0")).getAggregations().getAsMap().get("sig_terms");
double score00SeparateSets = sigTerms0.getBucketByKey("0").getSignificanceScore();
double score01SeparateSets = sigTerms0.getBucketByKey("1").getSignificanceScore();
sigTerms1 = (SignificantTerms) ((InternalFilter) aggs.get("1")).getAggregations().getAsMap().get("sig_terms");
double score10SeparateSets = sigTerms1.getBucketByKey("0").getSignificanceScore();
double score11SeparateSets = sigTerms1.getBucketByKey("1").getSignificanceScore();
assertThat(score00Background, equalTo(score00SeparateSets));
assertThat(score01Background, equalTo(score01SeparateSets));
assertThat(score10Background, equalTo(score10SeparateSets));
assertThat(score11Background, equalTo(score11SeparateSets));
}
use of org.elasticsearch.search.aggregations.bucket.filter.InternalFilter in project elasticsearch by elastic.
the class TermsShardMinDocCountIT method testShardMinDocCountSignificantTermsTest.
// see https://github.com/elastic/elasticsearch/issues/5998
public void testShardMinDocCountSignificantTermsTest() throws Exception {
String textMappings;
if (randomBoolean()) {
textMappings = "type=long";
} else {
textMappings = "type=text,fielddata=true";
}
assertAcked(prepareCreate(index).setSettings(SETTING_NUMBER_OF_SHARDS, 1, SETTING_NUMBER_OF_REPLICAS, 0).addMapping(type, "text", textMappings));
List<IndexRequestBuilder> indexBuilders = new ArrayList<>();
//high score but low doc freq
addTermsDocs("1", 1, 0, indexBuilders);
addTermsDocs("2", 1, 0, indexBuilders);
addTermsDocs("3", 1, 0, indexBuilders);
addTermsDocs("4", 1, 0, indexBuilders);
//low score but high doc freq
addTermsDocs("5", 3, 1, indexBuilders);
addTermsDocs("6", 3, 1, indexBuilders);
// make sure the terms all get score > 0 except for this one
addTermsDocs("7", 0, 3, indexBuilders);
indexRandom(true, false, indexBuilders);
// first, check that indeed when not setting the shardMinDocCount parameter 0 terms are returned
SearchResponse response = client().prepareSearch(index).addAggregation((filter("inclass", QueryBuilders.termQuery("class", true))).subAggregation(significantTerms("mySignificantTerms").field("text").minDocCount(2).size(2).executionHint(randomExecutionHint()))).execute().actionGet();
assertSearchResponse(response);
InternalFilter filteredBucket = response.getAggregations().get("inclass");
SignificantTerms sigterms = filteredBucket.getAggregations().get("mySignificantTerms");
assertThat(sigterms.getBuckets().size(), equalTo(0));
response = client().prepareSearch(index).addAggregation((filter("inclass", QueryBuilders.termQuery("class", true))).subAggregation(significantTerms("mySignificantTerms").field("text").minDocCount(2).shardMinDocCount(2).size(2).executionHint(randomExecutionHint()))).execute().actionGet();
assertSearchResponse(response);
filteredBucket = response.getAggregations().get("inclass");
sigterms = filteredBucket.getAggregations().get("mySignificantTerms");
assertThat(sigterms.getBuckets().size(), equalTo(2));
}
Aggregations