Fixed bug in attribute_type_id detection of attribute_type_id namespaces

NCATSTranslator · Jun 16, 2023 · 111360b · 111360b
1 parent d6d0824
commit 111360b
Show file tree

Hide file tree

Showing 4 changed files with 22 additions and 31 deletions.
diff --git a/docs/validation_codes_dictionary.md b/docs/validation_codes_dictionary.md
@@ -242,14 +242,6 @@
 
 **Description:** Non-Biolink CURIEs are tolerated, but not preferred, as term value for the attribute_type_id properties of edge attributes.
 
-### warning.knowledge_graph.edge.attribute.type_id.unknown_prefix
-
-**Message:** Edge has an attribute_type_id that has a CURIE prefix namespace unknown to Biolink
-
-**Context:** edge_id, identifier
-
-**Description:** The namespaces of 'attribute_type_id' terms may be incomplete with respect to Biolink Model version being used in the knowledge graph.
-
 ### warning.knowledge_graph.edge.attribute.type_id.deprecated
 
 **Message:** Edge has a deprecated attribute_type_id

diff --git a/reasoner_validator/biolink/__init__.py b/reasoner_validator/biolink/__init__.py
@@ -571,20 +571,11 @@ def validate_attributes(self, edge_id: str, edge: Dict):
                                                         infores == kp_source:
                                                     found_kp_knowledge_source = True
 
-                        # if not a Biolink association_slot, at least,
-                        # check if it is an id prefix known to Biolink.
-                        # We won't call it a hard error, but issue a warning
-                        elif not self.bmt.get_element_by_prefix(prefix):
+                        # if not a Biolink 'association_slot', at least, check if the 'attribute_type_id' has a
+                        # namespace (prefix) known to Biolink. We won't call it a hard error, but issue a warning
+                        elif not self.bmt.get_element_by_prefix(attribute_type_id):
                             self.report(
-                                code="warning.knowledge_graph.edge.attribute.type_id.unknown_prefix",
-                                identifier=attribute_type_id,
-                                edge_id=edge_id
-                            )
-                        # TODO: probably need to take a closer look at validating outlier terms here
-                        #       Maybe enumerations will help
-                        else:
-                            self.report(
-                                code="info.knowledge_graph.edge.attribute.type_id.non_biolink_prefix",
+                                code="warning.knowledge_graph.edge.attribute.type_id.non_biolink_prefix",
                                 identifier=attribute_type_id,
                                 edge_id=edge_id
                             )

diff --git a/reasoner_validator/codes.yaml b/reasoner_validator/codes.yaml
@@ -213,12 +213,6 @@ warning:
               - edge_id
               - identifier
             $description: "Non-Biolink CURIEs are tolerated, but not preferred, as term value for the attribute_type_id properties of edge attributes."
-          unknown_prefix:
-            $message: "Edge has an attribute_type_id that has a CURIE prefix namespace unknown to Biolink"
-            $context:
-              - edge_id
-              - identifier
-            $description: "The namespaces of 'attribute_type_id' terms may be incomplete with respect to Biolink Model version being used in the knowledge graph."
           deprecated:
             $message: "Edge has a deprecated attribute_type_id"
             $context:

diff --git a/tests/test_biolink_compliance_validation.py b/tests/test_biolink_compliance_validation.py
@@ -1030,12 +1030,27 @@ def test_pre_trapi_1_4_0_validate_missing_or_empty_attributes(query: Tuple):
                 "attributes": []
             },
             get_ara_test_case(),
-            # "Edge has empty attributes!"
+            # "Edge has empty attributes!" Allowed
             ""
         ),
+        (
+            # Query 3. EDAM-DATA:2526 ought to have an acceptable namespace
+            {
+                "attributes": [
+                    {
+                        "attribute_type_id": "EDAM-DATA:2526",
+                        "value": "some-value"
+                    }
+                ]
+            },
+            get_ara_test_case(),
+            # "Edge has an acceptable namespace prefix
+            ""
+        )
+        # CHEMBL.COMPOUND:CHEMBL112--biolink:occurs_together_in_literature_with->NCBIGene:762
     ]
 )
-def test_post_1_4_0_trapi_validate_missing_or_empty_attributes(query: Tuple):
+def test_post_1_4_0_trapi_validate_attributes(query: Tuple):
     validator = BiolinkValidator(
         graph_type=TRAPIGraphType.Knowledge_Graph,
         biolink_version=LATEST_BIOLINK_MODEL_VERSION,
@@ -2686,8 +2701,7 @@ def test_validate_biolink_curie_in_qualifiers(query: Tuple[str, Dict, str]):
             },
             # f"{KNOWLEDGE_GRAPH_PREFIX}: WARNING - Edge attribute_type_id 'foo:bar' " +
             # f"has a CURIE prefix namespace unknown to Biolink!"
-            # TODO: Code for validating this is commented out pending a BMT repair of the test
-            "warning.knowledge_graph.edge.attribute.type_id.unknown_prefix"
+            "warning.knowledge_graph.edge.attribute.type_id.non_biolink_prefix"
         ),
         (   # Query 26:  # An earlier Biolink Model won't recognize a category not found in its specified release
             "1.8.2",