fix!: use correct representation of Coding object in mappings (#457)

korikuzma · web-flow · commit 14d113e47fc3 · 2025-01-29T16:12:59.000-05:00
close #456 * bump disease normalizer version to pull in updated mappings representations * correct `Coding` representation * `system` MUST use `iriReference`, not a free-text label * `code` MUST use syntax defined by the `system` * `id` will use record `concept_id` * Removes `SYSTEM_URI_TO_NAMESPACE` mapping (since it's no longer needed)
diff --git a/pyproject.toml b/pyproject.toml
@@ -30,13 +30,13 @@ dependencies = [
     "click",
     "boto3",
     "ga4gh.vrs==2.0.0a13",
-    "disease-normalizer~=0.7.0",
+    "disease-normalizer~=0.8.0",
 ]
 dynamic = ["version"]
 
 [project.optional-dependencies]
 etl = [
-    "disease-normalizer[etl]~=0.7.0",
+    #"disease-normalizer[etl]~=0.7.0",
     "owlready2",
     "rdflib",
     "wikibaseintegrator>=0.12.0",
diff --git a/src/therapy/query.py b/src/therapy/query.py
@@ -7,8 +7,7 @@
 from typing import Any, TypeVar
 
 from botocore.exceptions import ClientError
-from disease.schemas import NAMESPACE_TO_SYSTEM_URI as DISEASE_NAMESPACE_TO_SYSTEM_URI
-from disease.schemas import NamespacePrefix as DiseaseNamespacePrefix
+from disease.query import get_concept_mapping as get_disease_concept_mapping
 from ga4gh.core.models import (
     Coding,
     ConceptMapping,
@@ -23,7 +22,6 @@
 from therapy.database import AbstractDatabase
 from therapy.schemas import (
     NAMESPACE_TO_SYSTEM_URI,
-    SYSTEM_URI_TO_NAMESPACE,
     BaseNormalizationService,
     HasIndication,
     MatchesNormalized,
@@ -365,7 +363,7 @@ def _add_merged_meta(self, response: NormalizationService) -> NormalizationServi
 
         sources = []
         for m in therapy.mappings or []:
-            ns = SYSTEM_URI_TO_NAMESPACE.get(m.coding.system)
+            ns = m.coding.id.split(":")[0]
             if ns in PREFIX_LOOKUP:
                 sources.append(PREFIX_LOOKUP[ns])
 
@@ -399,39 +397,42 @@ def _add_therapy(
         :return: completed response object ready to return to user
         """
 
-        def _create_concept_mapping(
+        def _get_concept_mapping(
             concept_id: str,
             relation: Relation,
-            ns_to_system_uri: dict[str, str],
-            ns_prefix: NamespacePrefix | DiseaseNamespacePrefix,
         ) -> ConceptMapping:
-            """Create concept mapping for therapy or disease identifier
+            """Create concept mapping for identifier
 
-            ``system`` will use OBO Foundry persistent URL (PURL), source homepage, or
-            namespace prefix, in that order of preference, if available.
+            ``system`` will use system prefix URL, OBO Foundry persistent URL (PURL), or
+            source homepage, in that order of preference.
 
             :param concept_id: Concept identifier represented as a curie
             :param relation: SKOS mapping relationship, default is relatedMatch
-            :param ns_to_system_uri: Dictionary containing mapping from namespace to
-                system URI
-            :param ns_prefix: Namespace prefix enum
-            :return: Concept mapping for therapy or disease identifier
+            :raises ValueError: If source of concept ID is not a valid
+                ``NamespacePrefix``
+            :return: Concept mapping for identifier
             """
-            source = concept_id.split(":")[0]
+            source, source_code = concept_id.split(":")
 
             try:
-                source = ns_prefix(source)
+                source = NamespacePrefix(source)
             except ValueError:
                 try:
-                    source = ns_prefix(source.upper())
+                    source = NamespacePrefix(source.upper())
                 except ValueError as e:
                     err_msg = f"Namespace prefix not supported: {source}"
                     raise ValueError(err_msg) from e
 
-            system = ns_to_system_uri.get(source, source)
+            if source == NamespacePrefix.CHEBI:
+                source_code = concept_id
 
             return ConceptMapping(
-                coding=Coding(code=code(concept_id), system=system), relation=relation
+                coding=Coding(
+                    id=concept_id,
+                    code=code(source_code),
+                    system=NAMESPACE_TO_SYSTEM_URI[source],
+                ),
+                relation=relation,
             )
 
         therapy_obj = MappableConcept(
@@ -443,20 +444,16 @@ def _create_concept_mapping(
 
         # mappings
         mappings = [
-            _create_concept_mapping(
+            _get_concept_mapping(
                 concept_id=record["concept_id"],
                 relation=Relation.EXACT_MATCH,
-                ns_to_system_uri=NAMESPACE_TO_SYSTEM_URI,
-                ns_prefix=NamespacePrefix,
             )
         ]
         source_ids = record.get("xrefs", []) + record.get("associated_with", [])
         mappings.extend(
-            _create_concept_mapping(
+            _get_concept_mapping(
                 concept_id=source_id,
                 relation=Relation.RELATED_MATCH,
-                ns_to_system_uri=NAMESPACE_TO_SYSTEM_URI,
-                ns_prefix=NamespacePrefix,
             )
             for source_id in source_ids
         )
@@ -490,11 +487,9 @@ def _create_concept_mapping(
 
                 if indication.normalized_disease_id:
                     mappings = [
-                        _create_concept_mapping(
+                        get_disease_concept_mapping(
                             concept_id=indication.normalized_disease_id,
                             relation=Relation.RELATED_MATCH,
-                            ns_to_system_uri=DISEASE_NAMESPACE_TO_SYSTEM_URI,
-                            ns_prefix=DiseaseNamespacePrefix,
                         )
                     ]
                 else:
diff --git a/src/therapy/schemas.py b/src/therapy/schemas.py
@@ -2,6 +2,7 @@
 
 from datetime import datetime
 from enum import Enum, IntEnum
+from types import MappingProxyType
 from typing import Any, Literal
 
 from ga4gh.core.models import MappableConcept
@@ -258,37 +259,39 @@ class NamespacePrefix(Enum):
     WIKIDATA = "wikidata"
 
 
-# Source to URI. Will use OBO Foundry persistent URL (PURL) or source homepage
-NAMESPACE_TO_SYSTEM_URI: dict[NamespacePrefix, str] = {
-    NamespacePrefix.ATC: "https://www.who.int/tools/atc-ddd-toolkit/atc-classification/",
-    NamespacePrefix.CHEBI: "http://purl.obolibrary.org/obo/chebi.owl",
-    NamespacePrefix.CHEMBL: "https://www.ebi.ac.uk/chembl/",
-    NamespacePrefix.CHEMIDPLUS: "https://pubchem.ncbi.nlm.nih.gov/source/ChemIDplus",
-    NamespacePrefix.CASREGISTRY: "https://pubchem.ncbi.nlm.nih.gov/source/ChemIDplus",
-    NamespacePrefix.CVX: "https://www2a.cdc.gov/vaccines/iis/iisstandards/vaccines.asp?rpt=cvx",
-    NamespacePrefix.DRUGBANK: "https://go.drugbank.com",
-    NamespacePrefix.DRUGCENTRAL: "https://drugcentral.org",
-    NamespacePrefix.DRUGSATFDA_ANDA: "https://www.fda.gov/drugs/types-applications/abbreviated-new-drug-application-anda",
-    NamespacePrefix.DRUGSATFDA_NDA: "https://www.fda.gov/drugs/types-applications/new-drug-application-nda",
-    NamespacePrefix.HEMONC: "https://hemonc.org",
-    NamespacePrefix.INCHIKEY: "https://www.chemspider.com",
-    NamespacePrefix.IUPHAR_LIGAND: "https://www.guidetopharmacology.org/GRAC/LigandListForward",
-    NamespacePrefix.GUIDETOPHARMACOLOGY: "https://www.guidetopharmacology.org/GRAC/LigandListForward",
-    NamespacePrefix.MMSL: "https://www.nlm.nih.gov/research/umls/rxnorm/sourcereleasedocs/mmsl.html",
-    NamespacePrefix.MSH: "https://id.nlm.nih.gov/mesh/",
-    NamespacePrefix.NCIT: "http://purl.obolibrary.org/obo/ncit.owl",
-    NamespacePrefix.NDC: "https://dps.fda.gov/ndc",
-    NamespacePrefix.PUBCHEMCOMPOUND: "https://pubchem.ncbi.nlm.nih.gov/docs/compounds",
-    NamespacePrefix.PUBCHEMSUBSTANCE: "https://pubchem.ncbi.nlm.nih.gov/docs/substances",
-    NamespacePrefix.RXNORM: "https://www.nlm.nih.gov/research/umls/rxnorm/index.html",
-    NamespacePrefix.SPL: "https://www.fda.gov/industry/fda-data-standards-advisory-board/structured-product-labeling-resources",
-    NamespacePrefix.UMLS: "https://www.nlm.nih.gov/research/umls/index.html",
-    NamespacePrefix.UNII: "https://precision.fda.gov/uniisearch",
-    NamespacePrefix.UNIPROT: "https://www.uniprot.org",
-    NamespacePrefix.USP: "https://www.usp.org/health-quality-safety/compendial-nomenclature",
-    NamespacePrefix.VANDF: "https://www.nlm.nih.gov/research/umls/sourcereleasedocs/current/VANDF",
-    NamespacePrefix.WIKIDATA: "https://www.wikidata.org",
-}
+# Source to URI. Will use system prefix URL, OBO Foundry persistent URL (PURL) or source homepage
+NAMESPACE_TO_SYSTEM_URI: MappingProxyType[NamespacePrefix, str] = MappingProxyType(
+    {
+        NamespacePrefix.ATC: "https://atcddd.fhi.no/atc_ddd_index/?code=",
+        NamespacePrefix.CHEBI: "https://www.ebi.ac.uk/chebi/searchId.do?chebiId=",
+        NamespacePrefix.CHEMBL: "https://www.ebi.ac.uk/chembl/explore/compound/",
+        NamespacePrefix.CHEMIDPLUS: "https://commonchemistry.cas.org/detail?cas_rn=",
+        NamespacePrefix.CASREGISTRY: "https://commonchemistry.cas.org/detail?cas_rn=",
+        NamespacePrefix.CVX: "https://www2a.cdc.gov/vaccines/iis/iisstandards/vaccines.asp?rpt=cvx",
+        NamespacePrefix.DRUGBANK: "https://go.drugbank.com/drugs/",
+        NamespacePrefix.DRUGCENTRAL: "https://drugcentral.org/drugcard/",
+        NamespacePrefix.DRUGSATFDA_ANDA: "https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=",
+        NamespacePrefix.DRUGSATFDA_NDA: "https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=",
+        NamespacePrefix.HEMONC: "https://hemonc.org",
+        NamespacePrefix.INCHIKEY: "https://www.chemspider.com",
+        NamespacePrefix.IUPHAR_LIGAND: "https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=",
+        NamespacePrefix.GUIDETOPHARMACOLOGY: "https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=",
+        NamespacePrefix.MMSL: "https://www.nlm.nih.gov/research/umls/rxnorm/sourcereleasedocs/mmsl.html",
+        NamespacePrefix.MSH: "https://id.nlm.nih.gov/mesh/",
+        NamespacePrefix.NCIT: "https://ncit.nci.nih.gov/ncitbrowser/ConceptReport.jsp?dictionary=NCI_Thesaurus&code=",
+        NamespacePrefix.NDC: "https://dps.fda.gov/ndc/searchresult?selection=finished_product&content=PRODUCTNDC&type=",
+        NamespacePrefix.PUBCHEMCOMPOUND: "https://pubchem.ncbi.nlm.nih.gov/compound/",
+        NamespacePrefix.PUBCHEMSUBSTANCE: "https://pubchem.ncbi.nlm.nih.gov/substance/",
+        NamespacePrefix.RXNORM: "https://mor.nlm.nih.gov/RxNav/search?searchBy=RXCUI&searchTerm=",
+        NamespacePrefix.SPL: "https://www.fda.gov/industry/fda-data-standards-advisory-board/structured-product-labeling-resources",
+        NamespacePrefix.UMLS: "https://uts.nlm.nih.gov/uts/umls/concept/",
+        NamespacePrefix.UNII: "https://precision.fda.gov/uniisearch/srs/unii/",
+        NamespacePrefix.UNIPROT: "http://purl.uniprot.org/uniprot/",
+        NamespacePrefix.USP: "https://www.usp.org/health-quality-safety/compendial-nomenclature",
+        NamespacePrefix.VANDF: "https://www.nlm.nih.gov/research/umls/sourcereleasedocs/current/VANDF",
+        NamespacePrefix.WIKIDATA: "https://www.wikidata.org/wiki/",
+    }
+)
 
 # URI to source
 SYSTEM_URI_TO_NAMESPACE = {
@@ -540,21 +543,21 @@ class NormalizationService(BaseNormalizationService):
                         {
                             "coding": {
                                 "code": "2555",
-                                "system": "https://www.nlm.nih.gov/research/umls/rxnorm/index.html",
+                                "system": "https://mor.nlm.nih.gov/RxNav/search?searchBy=RXCUI&searchTerm=",
                             },
                             "relation": "exactMatch",
                         },
                         {
                             "coding": {
                                 "code": "C376",
-                                "system": "http://purl.obolibrary.org/obo/ncit.owl",
+                                "system": "https://ncit.nci.nih.gov/ncitbrowser/ConceptReport.jsp?dictionary=NCI_Thesaurus&code=",
                             },
                             "relation": "relatedMatch",
                         },
                         {
                             "coding": {
                                 "code": "15663-27-1",
-                                "system": "https://pubchem.ncbi.nlm.nih.gov/source/ChemIDplus",
+                                "system": "https://commonchemistry.cas.org/detail?cas_rn=",
                             },
                             "relation": "relatedMatch",
                         },
@@ -843,7 +846,7 @@ class SearchService(BaseModel):
                             "data_license": "CC0 1.0",
                             "data_license_url": "https://creativecommons.org/publicdomain/zero/1.0/",
                             "version": "5.1.10",
-                            "data_url": "https://go.drugbank.com/releases/latest#open-data",
+                            "data_url": "https://go.drugbank.com/drugs//releases/latest#open-data",
                             "rdp_url": "http://reusabledata.org/drugbank.html",
                             "data_license_attributes": {
                                 "non_commercial": False,
diff --git a/tests/conftest.py b/tests/conftest.py
@@ -21,6 +21,7 @@ def pytest_collection_modifyitems(items):
     When creating new test modules, be sure to add them here.
     """
     MODULE_ORDER = [  # noqa: N806
+        "test_schemas",
         "test_chembl",
         "test_chemidplus",
         "test_drugbank",
diff --git a/tests/data/fixtures/query_fixtures.json b/tests/data/fixtures/query_fixtures.json
