Move Run accession regex to the schema

EBIvariation · Aug 12, 2024 · 6bed4a9 · 6bed4a9
1 parent 75058d3
commit 6bed4a9
Show file tree

Hide file tree

Showing 3 changed files with 4 additions and 9 deletions.
diff --git a/eva_sub_cli/etc/eva_schema.json b/eva_sub_cli/etc/eva_schema.json
@@ -266,7 +266,8 @@
           "runAccessions": {
             "type": "array",
             "items": {
-              "type": "string"
+              "type": "string",
+              "pattern": "^(E|D|S)RR[0-9]{6,}$"
             },
             "description": "List of run accessions linking to the raw data used in this analysis if applicable (e.g. SRR576651, SRR576652)"
           }

diff --git a/eva_sub_cli/semantic_metadata.py b/eva_sub_cli/semantic_metadata.py
@@ -77,11 +77,7 @@ def check_all_analysis_run_accessions(self):
             json_path = f'/{ANALYSIS_KEY}/{idx}/{ANALYSIS_RUNS_KEY}'
             if analysis[ANALYSIS_RUNS_KEY]:
                 for run_acc in analysis[ANALYSIS_RUNS_KEY]:
-                    match = re.match('(E|D|S)RR[0-9]{6,}', run_acc)
-                    if not match:
-                        self.add_error(json_path, f'Invalid run accession format for {run_acc}')
-                    else:
-                        self.check_accession_in_ena(run_acc, 'Run', json_path)
+                    self.check_accession_in_ena(run_acc, 'Run', json_path)
 
     @retry(tries=4, delay=2, backoff=1.2, jitter=(1, 3))
     def check_accession_in_ena(self, ena_accession, accession_type, json_path):

diff --git a/tests/test_semantic_metadata.py b/tests/test_semantic_metadata.py
@@ -163,11 +163,9 @@ def test_check_all_analysis_run_accessions(self):
         checker.check_all_analysis_run_accessions()
         assert checker.errors == []
 
-        metadata["analysis"].append({'runAccessions': ['SRR00000A']})
         metadata["analysis"].append({'runAccessions': ['SRR00000000001']})
 
         checker.check_all_analysis_run_accessions()
         assert checker.errors == [
-            {'property': '/analysis/1/runAccessions', 'description': 'Invalid run accession format for SRR00000A'},
-            {'property': '/analysis/2/runAccessions', 'description': 'Run SRR00000000001 does not exist in ENA or is private'}]
+            {'property': '/analysis/1/runAccessions', 'description': 'Run SRR00000000001 does not exist in ENA or is private'}]