Refactor get_duos() to return the DUOS ID

DataBiosphere · Nov 27, 2024 · 94f2099 · 94f2099
1 parent 7ea685b
commit 94f2099
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 6 deletions.
diff --git a/src/azul/plugins/repository/tdr_anvil/__init__.py b/src/azul/plugins/repository/tdr_anvil/__init__.py
@@ -479,8 +479,11 @@ def _supplementary_bundle(self, bundle_fqid: TDRAnvilBundleFQID) -> TDRAnvilBund
 
     def _duos_bundle(self, bundle_fqid: TDRAnvilBundleFQID) -> TDRAnvilBundle:
         assert not bundle_fqid.is_batched, bundle_fqid
-        duos_info = self.tdr.get_duos(bundle_fqid.source)
-        description = None if duos_info is None else duos_info.get('studyDescription')
+        duos_id, duos_info = self.tdr.get_duos(bundle_fqid.source)
+        if duos_info is None:
+            description = None
+        else:
+            description = duos_info.get('studyDescription')
         ref, row = self._get_dataset(bundle_fqid.source.spec)
         expected_entity_id = change_version(bundle_fqid.uuid,
                                             self.bundle_uuid_version,

diff --git a/src/azul/terra.py b/src/azul/terra.py
@@ -646,19 +646,26 @@ def for_registered_user(cls, authentication: OAuth2) -> 'TDRClient':
     def drs_client(self) -> DRSClient:
         return DRSClient(http_client=self._http_client)
 
-    def get_duos(self, source: TDRSourceRef) -> Optional[MutableJSON]:
+    def get_duos(self,
+                 source: TDRSourceRef
+                 ) -> tuple[str | None, MutableJSON | None]:
         response = self._retrieve_source(source)
         try:
             duos_id = response['duosFirecloudGroup']['duosId']
         except (KeyError, TypeError):
             log.warning('No DUOS ID available for %r', source.spec)
-            return None
+            return None, None
         else:
             url = self._duos_endpoint('dataset', 'registration', duos_id)
             response = self._request('GET', url)
             if response.status == 404:
                 log.warning('No DUOS dataset registration with ID %r from %r',
                             duos_id, source.spec)
-                return None
+                return None, None
             else:
-                return self._check_response(url, response)
+                response = self._check_response(url, response)
+                assert response['consentGroups']
+                for consent_group in response['consentGroups']:
+                    assert duos_id == consent_group['datasetIdentifier'], \
+                        (duos_id, consent_group['datasetIdentifier'])
+                return duos_id, response