GONRG-1700: fix id composing

src/dags/libs/search_record_ids.py

@@ -155,7 +155,7 @@ class ExtendedSearchId(SearchId):
 
     def _extract_id_from_response(self, response: dict):
         results = response.get("results")
-        record_ids = ["".join([elem.get("id"), str(elem.get("version"))]) for elem in results]
+        record_ids = [":".join([elem.get("id"), str(elem.get("version", ""))]) for elem in results]
         record_ids.extend([elem.get("id") for elem in results])
         logger.debug(f"response ids: {record_ids}")
         return record_ids

tests/plugin-unit-tests/data/other/ExtractedIds.json

+[
+    "some_test_id",
+    "some_test_id:12345"
+]

tests/plugin-unit-tests/data/other/SearchResponseValid.json

 {
     "results": [
+        {
+            "id": "some_test_id",
+            "version": 12345
+        }
     ],
     "aggregations": null,
     "totalCount": 45

tests/plugin-unit-tests/file_paths.py

@@ -43,6 +43,7 @@ TRAVERSAL_MANIFEST_EMPTY_PATH = f"{DATA_PATH_PREFIX}/invalid/TraversalEmptyManif
 
 SEARCH_VALID_RESPONSE_PATH = f"{DATA_PATH_PREFIX}/other/SearchResponseValid.json"
 SEARCH_INVALID_RESPONSE_PATH = f"{DATA_PATH_PREFIX}/other/SearchResponseInvalid.json"
+SEARCH_EXTRACTED_IDS_PATH = f"{DATA_PATH_PREFIX}/other/ExtractedIds.json"
 
 
 SURROGATE_MANIFEST_WELLBORE = f"{DATA_PATH_PREFIX}/surrogate/Wellbore.0.3.0.json"

tests/plugin-unit-tests/test_search_record_id.py

@@ -12,8 +12,7 @@
 #  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
-
-
+import json
 import os
 import sys
 import http
@@ -26,13 +25,13 @@ sys.path.append(f"{os.getenv('AIRFLOW_SRC_DIR')}/dags")
 from mock_providers import get_test_credentials
 from file_paths import (
     SEARCH_INVALID_RESPONSE_PATH,
-    SEARCH_VALID_RESPONSE_PATH
-)
+    SEARCH_VALID_RESPONSE_PATH,
+    SEARCH_EXTRACTED_IDS_PATH)
 from libs.exceptions import RecordsNotSearchableError
 from libs.context import Context
 from libs.refresh_token import AirflowTokenRefresher
 from tenacity import stop_after_attempt
-from libs.search_record_ids import SearchId
+from libs.search_record_ids import SearchId, ExtendedSearchId
 from mock_responses import MockSearchResponse
 
 
@@ -133,3 +132,27 @@ class TestManifestProcessor:
         with pytest.raises(ValueError):
             SearchId("http://test", record_ids, AirflowTokenRefresher(get_test_credentials()),
                      Context(app_key="", data_partition_id=""))
+
+    @pytest.mark.parametrize(
+        "record_ids,search_response_path,extracted_ids_path",
+        [
+            pytest.param(
+                ["test"],
+                SEARCH_VALID_RESPONSE_PATH,
+                SEARCH_EXTRACTED_IDS_PATH
+            )
+        ]
+    )
+    def test_search_found_all_records(self, monkeypatch, record_ids: list,
+                                      search_response_path: str,
+                                      extracted_ids_path: str):
+        self.mock_storage_response(monkeypatch, search_response_path, total_count=len(record_ids))
+        with open(search_response_path) as f:
+            response = json.load(f)
+        id_searcher = ExtendedSearchId("http://test", record_ids,
+                                       AirflowTokenRefresher(get_test_credentials()),
+                                       Context(app_key="", data_partition_id=""))
+        record_ids = id_searcher._extract_id_from_response(response)
+        with open(extracted_ids_path) as f:
+            extracted_ids = json.load(f)
+        assert set(record_ids) == set(extracted_ids)