Fix search by email index keyword (#10698)

* Fix search by email index keyword

* Fix search by email index keyword

ingestion/src/metadata/ingestion/ometa/mixins/user_mixin.py

@@ -35,7 +35,7 @@ class OMetaUserMixin:
     client: REST
 
     email_search = (
-        "/search/query?q=email:{email}&from={from_}&size={size}&index="
+        "/search/query?q=email.keyword:{email}&from={from_}&size={size}&index="
         + ES_INDEX_MAP[User.__name__]
     )
 
@@ -44,7 +44,7 @@ class OMetaUserMixin:
         self,
         email: Optional[str],
         from_count: int = 0,
-        size: int = 10,
+        size: int = 1,
         fields: Optional[list] = None,
     ) -> Optional[User]:
         """

ingestion/src/metadata/ingestion/sink/elasticsearch_mapping/user_search_index_mapping.py

@@ -56,7 +56,13 @@ USER_ELASTICSEARCH_INDEX_MAPPING = textwrap.dedent(
         "type": "text"
       },
       "email": {
-        "type": "text"
+        "type": "text",
+        "fields": {
+          "keyword": {
+            "type": "keyword",
+            "ignore_above": 256
+          }
+        }
       },
       "isAdmin": {
         "type": "text"

ingestion/tests/integration/ometa/test_ometa_user_api.py

@@ -78,6 +78,10 @@ class OMetaUserTest(TestCase):
             data=CreateUserRequest(name="Levy", email="user2.1234@getcollate.io"),
         )
 
+        cls.user_3: User = cls.metadata.create_or_update(
+            data=CreateUserRequest(name="Lima", email="random.lima@getcollate.io"),
+        )
+
         # Leave some time for indexes to get updated, otherwise this happens too fast
         cls.check_es_index()
 
@@ -112,6 +116,13 @@ class OMetaUserTest(TestCase):
             self.metadata.get_user_by_email(email="idonotexist@random.com")
         )
 
+        # Non existing email returns, even if they have the same domain
+        # To get this fixed, we had to update the `email` field in the
+        # index as a `keyword` and search by `email.keyword` in ES.
+        self.assertIsNone(
+            self.metadata.get_user_by_email(email="idonotexist@getcollate.io")
+        )
+
         # I can get User 1, who has the name equal to its email
         self.assertEqual(
             self.user_1.id,

openmetadata-service/src/main/resources/elasticsearch/en/user_index_mapping.json

@@ -39,7 +39,13 @@
         "type": "text"
       },
       "email": {
-        "type": "text"
+        "type": "text",
+        "fields": {
+          "keyword": {
+            "type": "keyword",
+            "ignore_above": 256
+          }
+        }
       },
       "isAdmin": {
         "type": "boolean"