More Like This Query: values of a multi-value fields are compared at …

…the same level.

Previously, More Like This would create a new mlt query for each value of a
multi-value field. This could result in all the values of the field to be
selected, which defeats the purpose of More Like This. Instead, the correct
behavior is to generate only one mlt query for all the values of the field.
This commit provides the correct behavior for More Like This DSL. The fix for
More Like This API will be coming in another commit.

Closes #6310

src/main/java/org/elasticsearch/common/lucene/search/MoreLikeThisQuery.java

@@ -21,7 +21,6 @@
 
 import org.apache.lucene.analysis.Analyzer;
 import org.apache.lucene.index.IndexReader;
-import org.apache.lucene.queries.mlt.MoreLikeThis;
 import org.apache.lucene.search.BooleanClause;
 import org.apache.lucene.search.BooleanQuery;
 import org.apache.lucene.search.Query;
@@ -31,6 +30,7 @@
 import org.elasticsearch.common.io.FastStringReader;
 
 import java.io.IOException;
+import java.io.Reader;
 import java.util.Arrays;
 import java.util.Set;
 
@@ -43,18 +43,18 @@ public class MoreLikeThisQuery extends Query {
 
     private TFIDFSimilarity similarity;
 
-    private String likeText;
+    private String[] likeText;
     private String[] moreLikeFields;
     private Analyzer analyzer;
     private float percentTermsToMatch = DEFAULT_PERCENT_TERMS_TO_MATCH;
-    private int minTermFrequency = MoreLikeThis.DEFAULT_MIN_TERM_FREQ;
-    private int maxQueryTerms = MoreLikeThis.DEFAULT_MAX_QUERY_TERMS;
-    private Set<?> stopWords = MoreLikeThis.DEFAULT_STOP_WORDS;
-    private int minDocFreq = MoreLikeThis.DEFAULT_MIN_DOC_FREQ;
-    private int maxDocFreq = MoreLikeThis.DEFAULT_MAX_DOC_FREQ;
-    private int minWordLen = MoreLikeThis.DEFAULT_MIN_WORD_LENGTH;
-    private int maxWordLen = MoreLikeThis.DEFAULT_MAX_WORD_LENGTH;
-    private boolean boostTerms = MoreLikeThis.DEFAULT_BOOST;
+    private int minTermFrequency = XMoreLikeThis.DEFAULT_MIN_TERM_FREQ;
+    private int maxQueryTerms = XMoreLikeThis.DEFAULT_MAX_QUERY_TERMS;
+    private Set<?> stopWords = XMoreLikeThis.DEFAULT_STOP_WORDS;
+    private int minDocFreq = XMoreLikeThis.DEFAULT_MIN_DOC_FREQ;
+    private int maxDocFreq = XMoreLikeThis.DEFAULT_MAX_DOC_FREQ;
+    private int minWordLen = XMoreLikeThis.DEFAULT_MIN_WORD_LENGTH;
+    private int maxWordLen = XMoreLikeThis.DEFAULT_MAX_WORD_LENGTH;
+    private boolean boostTerms = XMoreLikeThis.DEFAULT_BOOST;
     private float boostTermsFactor = 1;
 
 
@@ -63,7 +63,7 @@ public MoreLikeThisQuery() {
     }
 
     public MoreLikeThisQuery(String likeText, String[] moreLikeFields, Analyzer analyzer) {
-        this.likeText = likeText;
+        this.likeText = new String[]{likeText};
         this.moreLikeFields = moreLikeFields;
         this.analyzer = analyzer;
     }
@@ -72,7 +72,7 @@ public MoreLikeThisQuery(String likeText, String[] moreLikeFields, Analyzer anal
     public int hashCode() {
         int result = boostTerms ? 1 : 0;
         result = 31 * result + Float.floatToIntBits(boostTermsFactor);
-        result = 31 * result + likeText.hashCode();
+        result = 31 * result + Arrays.hashCode(likeText);
         result = 31 * result + maxDocFreq;
         result = 31 * result + maxQueryTerms;
         result = 31 * result + maxWordLen;
@@ -99,7 +99,7 @@ public boolean equals(Object obj) {
             return false;
         if (boostTermsFactor != other.boostTermsFactor)
             return false;
-        if (!likeText.equals(other.likeText))
+        if (!(Arrays.equals(likeText, other.likeText)))
             return false;
         if (maxDocFreq != other.maxDocFreq)
             return false;
@@ -132,7 +132,7 @@ public boolean equals(Object obj) {
 
     @Override
     public Query rewrite(IndexReader reader) throws IOException {
-        MoreLikeThis mlt = new MoreLikeThis(reader, similarity == null ? new DefaultSimilarity() : similarity);
+        XMoreLikeThis mlt = new XMoreLikeThis(reader, similarity == null ? new DefaultSimilarity() : similarity);
 
         mlt.setFieldNames(moreLikeFields);
         mlt.setAnalyzer(analyzer);
@@ -145,10 +145,15 @@ public Query rewrite(IndexReader reader) throws IOException {
         mlt.setStopWords(stopWords);
         mlt.setBoost(boostTerms);
         mlt.setBoostFactor(boostTermsFactor);
+
+        Reader[] readers = new Reader[likeText.length];
+        for (int i = 0; i < readers.length; i++) {
+            readers[i] = new FastStringReader(likeText[i]);
+        }
         //LUCENE 4 UPGRADE this mapps the 3.6 behavior (only use the first field)
-        BooleanQuery bq = (BooleanQuery) mlt.like(new FastStringReader(likeText), moreLikeFields[0]);
-        BooleanClause[] clauses = bq.getClauses();
+        BooleanQuery bq = (BooleanQuery) mlt.like(moreLikeFields[0], readers);
 
+        BooleanClause[] clauses = bq.getClauses();
         bq.setMinimumNumberShouldMatch((int) (clauses.length * percentTermsToMatch));
 
         bq.setBoost(getBoost());
@@ -157,14 +162,22 @@ public Query rewrite(IndexReader reader) throws IOException {
 
     @Override
     public String toString(String field) {
-        return "like:" + likeText;
+        return "like:" + Arrays.toString(likeText);
     }
 
     public String getLikeText() {
+        return (likeText == null ? null : likeText[0]);
+    }
+
+    public String[] getLikeTexts() {
         return likeText;
     }
 
     public void setLikeText(String likeText) {
+        this.likeText = new String[]{likeText};
+    }
+
+    public void setLikeText(String... likeText) {
         this.likeText = likeText;
     }