Add uniqueness check in doi/recid entry

Adds a uniqueness check in the related DOI and record ID entries in the schema. Also adds testing against this functionality.
HEPData · Aug 11, 2023 · e8064a2 · e8064a2
1 parent 5e1afa0
commit e8064a2
Show file tree

Hide file tree

Showing 5 changed files with 14 additions and 6 deletions.
diff --git a/hepdata_validator/schemas/1.1.1/additional_info_schema.json b/hepdata_validator/schemas/1.1.1/additional_info_schema.json
@@ -19,6 +19,7 @@
     },
     "related_to_hepdata_recids" : {
       "type" : "array",
+      "uniqueItems": true,
       "items" : {"type" : "integer", "minimum": 1}
     },
     "preprintyear": {

diff --git a/hepdata_validator/schemas/1.1.1/submission_schema.json b/hepdata_validator/schemas/1.1.1/submission_schema.json
@@ -8,6 +8,7 @@
     "name": { "type": "string", "maxLength": 64 },
     "related_to_table_dois" : {
       "type" : "array",
+      "uniqueItems": true,
       "items" : {"type" : "string",  "maxLength" : 128, "pattern" : "^10\\.17182\/hepdata\\.\\d+\\.v\\d+\/t\\d+$" }
     },
     "location": { "type": "string", "maxLength": 256 },

diff --git a/testsuite/test_data/TestRelatedSubmissions/invalid_submission_doi.yaml b/testsuite/test_data/TestRelatedSubmissions/invalid_submission_doi.yaml
@@ -11,4 +11,6 @@ related_to_table_dois:
 - "10.17182/hepdata.1"
 - "10.17182/hepdata.1.v1"
 - "10.17182/hepdata.1.v1/"
-- "10.17182/hepdata.1.v1/a2"
+- "10.17182/hepdata.1.v1/a2"
+- "10.17182/hepdata.1.v1/t1"
+- "10.17182/hepdata.1.v1/t1"
diff --git a/testsuite/test_data/TestRelatedSubmissions/invalid_submission_recid.yaml b/testsuite/test_data/TestRelatedSubmissions/invalid_submission_recid.yaml
@@ -1,6 +1,6 @@
 ---
 comment: "Test Text"
-related_to_hepdata_recids: ["a", a, 0]
+related_to_hepdata_recids: ["a", b, 0, 1, 1]
 ---
 name: "Table 1"
 description: Words

diff --git a/testsuite/test_submission_validator.py b/testsuite/test_submission_validator.py
@@ -354,8 +354,9 @@ def test_related_submissions(validator_v1, data_path, capsys):
             "file": "invalid_submission_recid.yaml",
             "errors": [
                 "'a' is not of type 'integer'",
-                "'a' is not of type 'integer'",
-                "0 is less than the minimum of 1"
+                "'b' is not of type 'integer'",
+                "0 is less than the minimum of 1",
+                "has non-unique elements in 'related_to_hepdata_recids'"
         ]},
         {
             "file": "invalid_submission_doi.yaml",
@@ -364,7 +365,8 @@ def test_related_submissions(validator_v1, data_path, capsys):
                 "'10.17182/hepdata.1' does not match",
                 "'10.17182/hepdata.1.v1' does not match",
                 "'10.17182/hepdata.1.v1/' does not match",
-                "'10.17182/hepdata.1.v1/a2'"
+                "'10.17182/hepdata.1.v1/a2' does not match",
+                "has non-unique elements in 'related_to_table_dois'"
         ]}]
 
     for yfile in yaml_files:
@@ -381,9 +383,11 @@ def test_related_submissions(validator_v1, data_path, capsys):
 
             if yfile["errors"]:
                 assert is_valid is False
+                # Should be the same amount of errors as expected
                 assert len(error_list) == len(yfile["errors"])
+                # Then check against the expected errors
                 for error in error_list:
                     assert any(yerrors in error for yerrors in yfile['errors'])
             else:
                 # If there is no expected error, then it should be valid
-                assert is_valid
+                assert is_valid