Fix: Invalid columns in compact manifest for AnVIL (#6110)

DataBiosphere · Sep 30, 2024 · c957158 · c957158
1 parent a66625b
commit c957158
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 12 deletions.
diff --git a/src/azul/plugins/metadata/anvil/__init__.py b/src/azul/plugins/metadata/anvil/__init__.py
@@ -280,8 +280,13 @@ def manifest_config(self) -> ManifestConfig:
         # the desired manifest column name, or None to omit the column from the
         # manifest.
         custom_field_names = {
+            ('bundles', 'uuid'): 'files.bundle_uuid',
+            ('bundles', 'version'): 'files.bundle_version',
+            ('contents', 'activities', 'activity_table'): None,
             ('contents', 'files', 'uuid'): None,
             ('contents', 'files', 'version'): None,
+            ('sources', 'id'): 'sources.source_id',
+            ('sources', 'spec'): 'sources.source_spec',
         }
 
         def recurse(mapping: MetadataPlugin._FieldMapping, path: FieldPath):
@@ -304,7 +309,7 @@ def recurse(mapping: MetadataPlugin._FieldMapping, path: FieldPath):
         # The file URL is synthesized from the `uuid` and `version` fields.
         # Above, we already configured these two fields to be omitted from the
         # manifest since they are not informative to the user.
-        result[('contents', 'files')]['file_url'] = 'files.file_url'
+        result[('contents', 'files')]['file_url'] = 'files.azul_file_url'
         return result
 
     def verbatim_pfb_schema(self,

diff --git a/test/service/test_manifest.py b/test/service/test_manifest.py
@@ -1705,29 +1705,30 @@ def bundles(cls) -> list[SourcedBundleFQID]:
         ]
 
     def test_compact_manifest(self):
+        self.maxDiff = None
         response = self._get_manifest(ManifestFormat.compact, filters={})
         self.assertEqual(200, response.status_code)
         expected = [
             (
-                'bundle_uuid',
+                'files.bundle_uuid',
                 '6b0f6c0f-5d80-a242-accb-840921351cd5',
                 '826dea02-e274-affe-aabc-eb3db63ad068',
                 '826dea02-e274-affe-aabc-eb3db63ad068'
             ),
             (
-                'bundle_version',
+                'files.bundle_version',
                 '2022-06-01T00:00:00.000000Z',
                 '2022-06-01T00:00:00.000000Z',
                 '2022-06-01T00:00:00.000000Z'
             ),
             (
-                'source_id',
+                'sources.source_id',
                 '6c87f0e1-509d-46a4-b845-7584df39263b',
                 '6c87f0e1-509d-46a4-b845-7584df39263b',
                 '6c87f0e1-509d-46a4-b845-7584df39263b'
             ),
             (
-                'source_spec',
+                'sources.source_spec',
                 'tdr:bigquery:gcp:test_anvil_project:anvil_snapshot:/2',
                 'tdr:bigquery:gcp:test_anvil_project:anvil_snapshot:/2',
                 'tdr:bigquery:gcp:test_anvil_project:anvil_snapshot:/2'
@@ -1966,12 +1967,6 @@ def test_compact_manifest(self):
                 '18b3be87-e26b-4376-0d8d-c1e370e90e07',
                 'a60c5138-3749-f7cb-8714-52d389ad5231'
             ),
-            (
-                'activities.activity_table',
-                '',
-                'sequencingactivity',
-                'sequencingactivity'
-            ),
             (
                 'activities.activity_type',
                 '',
@@ -2075,7 +2070,7 @@ def test_compact_manifest(self):
                 self._drs_uri('v1_6c87f0e1-509d-46a4-b845-7584df39263b_8b722e88-8103-49c1-b351-e64fa7c6ab37')
             ),
             (
-                'files.file_url',
+                'files.azul_file_url',
                 self._file_url('6b0f6c0f-5d80-4242-accb-840921351cd5', self.version),
                 self._file_url('15b76f9c-6b46-433f-851d-34e89f1b9ba6', self.version),
                 self._file_url('3b17377b-16b1-431c-9967-e5d01fc5923f', self.version)