Fix fuzziness parsing in multi_match function. Update tests. (#668)…

… (#669) Signed-off-by: Yury Fridlyand <[email protected]> (cherry picked from commit a91c194) Co-authored-by: Yury-Fridlyand <[email protected]>
opensearch-project · Jun 29, 2022 · 864732a · 864732a
1 parent 5f00b9f
commit 864732a
Show file tree

Hide file tree

Showing 6 changed files with 9 additions and 9 deletions.
diff --git a/integ-test/src/test/java/org/opensearch/sql/ppl/RelevanceFunctionIT.java b/integ-test/src/test/java/org/opensearch/sql/ppl/RelevanceFunctionIT.java
@@ -42,11 +42,11 @@ public void test_multi_match_all_params() throws IOException {
     String query = "SOURCE=" + TEST_INDEX_BEER
         + " | WHERE multi_match(['Body', Tags], 'taste beer', operator='and', analyzer=english,"
         + "auto_generate_synonyms_phrase_query=true, boost = 0.77, cutoff_frequency=0.33,"
-        + "fuzziness = 14, fuzzy_transpositions = false, lenient = true, max_expansions = 25,"
+        + "fuzziness = 'AUTO:1,5', fuzzy_transpositions = false, lenient = true, max_expansions = 25,"
         + "minimum_should_match = '2<-25% 9<-3', prefix_length = 7, tie_breaker = 0.3,"
         + "type = most_fields, slop = 2, zero_terms_query = 'ALL') | fields Id";
     var result = executeQuery(query);
-    assertEquals(507, result.getInt("total"));
+    assertEquals(424, result.getInt("total"));
   }
 
   @Test

diff --git a/integ-test/src/test/java/org/opensearch/sql/sql/MultiMatchIT.java b/integ-test/src/test/java/org/opensearch/sql/sql/MultiMatchIT.java
@@ -39,11 +39,11 @@ public void test_all_params() throws IOException {
     String query = "SELECT Id FROM " + TEST_INDEX_BEER
         + " WHERE multi_match(['Body', Tags], 'taste beer', operator='and', analyzer=english,"
         + "auto_generate_synonyms_phrase_query=true, boost = 0.77, cutoff_frequency=0.33,"
-        + "fuzziness = 14, fuzzy_transpositions = false, lenient = true, max_expansions = 25,"
+        + "fuzziness = 'AUTO:1,5', fuzzy_transpositions = false, lenient = true, max_expansions = 25,"
         + "minimum_should_match = '2<-25% 9<-3', prefix_length = 7, tie_breaker = 0.3,"
         + "type = most_fields, slop = 2, zero_terms_query = 'ALL');";
     var result = new JSONObject(executeQuery(query, "jdbc"));
-    assertEquals(507, result.getInt("total"));
+    assertEquals(424, result.getInt("total"));
   }
 
   @Test

diff --git a/...org/opensearch/sql/opensearch/storage/script/filter/lucene/relevance/MultiMatchQuery.java b/...org/opensearch/sql/opensearch/storage/script/filter/lucene/relevance/MultiMatchQuery.java
@@ -29,7 +29,7 @@ public MultiMatchQuery() {
             b.autoGenerateSynonymsPhraseQuery(Boolean.parseBoolean(v.stringValue())))
         .put("boost", (b, v) -> b.boost(Float.parseFloat(v.stringValue())))
         .put("cutoff_frequency", (b, v) -> b.cutoffFrequency(Float.parseFloat(v.stringValue())))
-        .put("fuzziness", (b, v) -> b.fuzziness(Integer.parseInt(v.stringValue())))
+        .put("fuzziness", (b, v) -> b.fuzziness(v.stringValue()))
         .put("fuzzy_transpositions", (b, v) ->
             b.fuzzyTranspositions(Boolean.parseBoolean(v.stringValue())))
         .put("lenient", (b, v) -> b.lenient(Boolean.parseBoolean(v.stringValue())))

diff --git a/...test/java/org/opensearch/sql/opensearch/storage/script/filter/FilterQueryBuilderTest.java b/...test/java/org/opensearch/sql/opensearch/storage/script/filter/FilterQueryBuilderTest.java
@@ -506,7 +506,7 @@ void should_build_multi_match_query_with_custom_parameters() {
             + "    \"operator\" : \"AND\",\n"
             + "    \"analyzer\" : \"keyword\",\n"
             + "    \"slop\" : 1,\n"
-            + "    \"fuzziness\" : \"2\",\n"
+            + "    \"fuzziness\" : \"AUTO:2,4\",\n"
             + "    \"prefix_length\" : 1,\n"
             + "    \"max_expansions\" : 3,\n"
             + "    \"minimum_should_match\" : \"3\",\n"
@@ -527,7 +527,7 @@ void should_build_multi_match_query_with_custom_parameters() {
                 dsl.namedArgument("analyzer", literal("keyword")),
                 dsl.namedArgument("auto_generate_synonyms_phrase_query", literal("false")),
                 dsl.namedArgument("cutoff_frequency", literal("4.3")),
-                dsl.namedArgument("fuzziness", literal("2")),
+                dsl.namedArgument("fuzziness", literal("AUTO:2,4")),
                 dsl.namedArgument("fuzzy_transpositions", literal("false")),
                 dsl.namedArgument("lenient", literal("false")),
                 dsl.namedArgument("max_expansions", literal("3")),

diff --git a/.../test/java/org/opensearch/sql/opensearch/storage/script/filter/lucene/MultiMatchTest.java b/.../test/java/org/opensearch/sql/opensearch/storage/script/filter/lucene/MultiMatchTest.java
@@ -74,7 +74,7 @@ static Stream<List<Expression>> generateValidData() {
         List.of(
             dsl.namedArgument("fields", fields_value),
             dsl.namedArgument("query", query_value),
-            dsl.namedArgument("fuzziness", DSL.literal("4"))
+            dsl.namedArgument("fuzziness", DSL.literal("AUTO:2,4"))
         ),
         List.of(
             dsl.namedArgument("fields", fields_value),

diff --git a/sql/src/test/java/org/opensearch/sql/sql/antlr/SQLSyntaxParserTest.java b/sql/src/test/java/org/opensearch/sql/sql/antlr/SQLSyntaxParserTest.java
@@ -187,7 +187,7 @@ public void can_parse_multi_match_relevance_function() {
     assertNotNull(parser.parse(
         "SELECT id FROM test WHERE"
             + " multi_match([\"Tags\" ^ 1.5, Title, `Body` 4.2], 'query', analyzer=keyword,"
-            + "operator='AND', tie_breaker=0.3, type = \"most_fields\", fuzziness = 4)"));
+            + "operator='AND', tie_breaker=0.3, type = \"most_fields\", fuzziness = \"AUTO\")"));
   }
 
   @Test