update handling ids

cancervariants · Sep 13, 2023 · 04b6dee · 04b6dee
1 parent 1a32958
commit 04b6dee
Show file tree

Hide file tree

Showing 5 changed files with 16 additions and 12 deletions.
diff --git a/README.md b/README.md
@@ -18,7 +18,7 @@ Call the `normalize()` method with a gene term. If available, a rich description
 
 ```
 >>> result = q.normalize("BRAF")
->>> result.gene.id
+>>> result.normalized_id
 "hgnc:1097"
 >>> result.gene.aliases
 ['NS7', 'RAFB1', 'B-raf', 'BRAF-1', 'BRAF1', 'B-RAF1']

diff --git a/docs/source/index.rst b/docs/source/index.rst
@@ -20,7 +20,7 @@ The Gene Normalizer provides tools for resolving ambiguous human gene references
     >>> from gene.database import create_db
     >>> q = QueryHandler(create_db())
     >>> result = q.normalize("BRAF")
-    >>> result.gene.id
+    >>> result.normalized_id
     "hgnc:1097"
     >>> result.gene.aliases
     ['NS7', 'RAFB1', 'B-raf', 'BRAF-1', 'BRAF1', 'B-RAF1']

diff --git a/gene/query.py b/gene/query.py
@@ -374,7 +374,7 @@ def _add_merged_meta(self, response: NormalizeService) -> NormalizeService:
         """
         sources_meta = {}
         gene = response.gene
-        sources = [gene.id.split(":")[0]]
+        sources = [response.normalized_id.split(":")[0]]
         if gene.mappings:
             sources += [m.coding.system for m in gene.mappings]
 
@@ -427,7 +427,7 @@ def _add_gene(
         :return: Response with core Gene
         """
         gene_obj = core_models.Gene(
-            id=record["concept_id"],
+            id=f"normalize.gene.{record['concept_id']}",
             label=record["symbol"],
         )
 
@@ -520,6 +520,7 @@ def _add_gene(
             response = self._add_alt_matches(response, record,
                                              possible_concepts)
 
+        response.normalized_id = record["concept_id"]
         response.gene = gene_obj
         response = self._add_merged_meta(response)
         response.match_type = match_type
@@ -574,7 +575,7 @@ def normalize(self, query: str) -> NormalizeService:
         >>> from gene.database import create_db
         >>> q = QueryHandler(create_db())
         >>> result = q.normalize("BRAF")
-        >>> result.gene.id
+        >>> result.normalized_id
         'hgnc:1097'
         >>> result.aliases
         ['BRAF1', 'RAFB1', 'B-raf', 'NS7', 'B-RAF1']

diff --git a/gene/schemas.py b/gene/schemas.py
@@ -419,6 +419,7 @@ class BaseNormalizationService(BaseModel):
 class NormalizeService(BaseNormalizationService):
     """Define model for returning normalized concept."""
 
+    normalized_id: Optional[constr(pattern=r"^normalize.gene.\w[^:]*:.+$")] = None  # noqa: F722, E501
     gene: Optional[core_models.Gene] = None
     source_meta_: Dict[SourceName, SourceMeta] = {}
 
@@ -428,9 +429,10 @@ class NormalizeService(BaseNormalizationService):
                 "query": "BRAF",
                 "warnings": [],
                 "match_type": 100,
+                "normalized_id": "hgnc:1037",
                 "gene": {
                     "type": "Gene",
-                    "id": "hgnc:1097",
+                    "id": "normalize.gene.hgnc:1097",
                     "label": "BRAF",
                     "mappings": [
                         {

diff --git a/tests/unit/test_query.py b/tests/unit/test_query.py
@@ -32,7 +32,7 @@ def normalized_ache():
     """Return normalized core Gene object for ACHE."""
     params = {
         "type": "Gene",
-        "id": "hgnc:108",
+        "id": "normalize.gene.hgnc:108",
         "label": "ACHE",
         "mappings": [
             {
@@ -198,7 +198,7 @@ def normalized_braf():
     """Return normalized core Gene object for BRAF."""
     params = {
         "type": "Gene",
-        "id": "hgnc:1097",
+        "id": "normalize.gene.hgnc:1097",
         "label": "BRAF",
         "mappings": [
             {
@@ -373,7 +373,7 @@ def normalized_abl1():
     """Return normalized core Gene object for ABL1."""
     params = {
         "type": "Gene",
-        "id": "hgnc:76",
+        "id": "normalize.gene.hgnc:76",
         "label": "ABL1",
         "mappings": [
             {
@@ -554,7 +554,7 @@ def normalized_p150():
     """Return normalized core Gene object for p150."""
     params = {
         "type": "Gene",
-        "id": "hgnc:1910",
+        "id": "normalize.gene.hgnc:1910",
         "label": "CHAF1A",
         "mappings": [
             {
@@ -762,7 +762,7 @@ def normalized_loc_653303():
                 "value": "+"
             }
         ],
-        "id": "ncbigene:653303"
+        "id": "normalize.gene.ncbigene:653303"
     }
     return core_models.Gene(**params)
 
@@ -1094,7 +1094,7 @@ def normalized_ifnr():
     """Return normalized core Gene object for IFNR."""
     params = {
         "type": "Gene",
-        "id": "hgnc:5447",
+        "id": "normalize.gene.hgnc:5447",
         "label": "IFNR",
         "mappings": [
             {
@@ -1185,6 +1185,7 @@ def compare_normalize_resp(resp, expected_query, expected_match_type,
     assert resp.query == expected_query
     compare_warnings(resp.warnings, expected_warnings)
     assert resp.match_type == expected_match_type
+    assert resp.normalized_id == expected_gene.id.split("normalize.gene.")[-1]
     compare_gene(expected_gene, resp.gene)
     if not expected_source_meta:
         assert resp.source_meta_ == {}