Consistent validation with pydantic & vocpub (disallow None)

src/voc4cat/models.py

@@ -56,7 +56,7 @@ def reset_curies(curies_map: dict) -> None:
 def split_curie_list(cls, v):
     if v is None:
         return []
-    if type(v) is list:
+    if isinstance(v, list):
         return v
     return (p.strip() for p in v.split(","))
 
@@ -149,11 +149,11 @@ class ConceptScheme(BaseModel):
     title: str
     description: str
     created: datetime.date
-    modified: datetime.date = None
+    modified: datetime.date
     creator: Ror | Orcid | str
     publisher: Ror | str
     provenance: str
-    version: str = None
+    version: str = "automatic"
     custodian: str = None
     pid: str = None
     vocab_name: str = Field("", exclude=True)
@@ -170,8 +170,8 @@ def creator_must_be_from_list(cls, v):
 
     @validator("modified")
     def set_modified_date_if_missing(cls, v):
-        if os.getenv("CI") is not None:  # Don't track modified date in GitHub.
-            v = None
+        if os.getenv("VOC4CAT_VERSION") is not None:
+            v = datetime.datetime.now(datetime.timezone.utc).strftime("%Y-%m-%d")
         return v
 
     @validator("publisher")
@@ -187,7 +187,7 @@ def publisher_must_be_from_list(cls, v):
     @validator("version")
     def version_from_env(cls, v):
         if os.getenv("CI") is not None:  # Don't track version in GitHub.
-            v = None
+            v = "automatic"
         version_from_env = os.getenv("VOC4CAT_VERSION")
         if version_from_env is not None:
             if not version_from_env.startswith("v"):

src/voc4cat/transform.py

@@ -78,12 +78,12 @@ def autoversion_cs(graph: Graph) -> Graph:
         date_modified = os.getenv("VOC4CAT_MODIFIED")
         graph.add((cs, DCTERMS.modified, Literal(date_modified, datatype=XSD.date)))
     if os.getenv("VOC4CAT_VERSION") is not None:
-        graph.remove((None, OWL.versionInfo, None))
         version = os.getenv("VOC4CAT_VERSION")
         if version is not None and not version.startswith("v"):
             msg = 'Invalid environment variable VOC4CAT_VERSION "%s". Version must start with letter "v".'
             logger.error(msg, version)
             raise Voc4catError(msg % version)
+        graph.remove((None, OWL.versionInfo, None))
         graph.add(
             (cs, OWL.versionInfo, Literal(version)),
         )

tests/test_models.py

@@ -119,8 +119,8 @@ def test_vocabulary_valid_in_ci():
         custodian="Vance Kelly",
         pid="http://pid.geoscience.gov.au/dataset/ga/114541",
     )
-    assert cs.modified is None
-    assert cs.version is None
+    assert cs.modified == datetime.date(2020, 4, 4)
+    assert cs.version == "automatic"
 
 
 @mock.patch.dict(os.environ, {"CI": "", "VOC4CAT_VERSION": "v2023-08-15"})
@@ -138,7 +138,9 @@ def test_vocabulary_valid_version_via_envvar():
         custodian="Vance Kelly",
         pid="http://pid.geoscience.gov.au/dataset/ga/114541",
     )
-    assert cs.modified is None
+    assert cs.modified == datetime.datetime.now(datetime.timezone.utc).strftime(
+        "%Y-%m-%d"
+    )
     assert cs.version == "v2023-08-15"
 
 

tests/test_template043.py

@@ -18,7 +18,7 @@ def test_empty_template():
             test_file,
             output_type="file",
         )
-    assert "7 validation errors for ConceptScheme" in str(e)
+    assert "9 validation errors for ConceptScheme" in str(e)
 
 
 def test_simple():