issue #1099 - ImportedAlleleInfo store hgvs_resolution

classification/migrations/0145_importedalleleinfo_hgvs_converter_version.py

@@ -0,0 +1,19 @@
+# Generated by Django 4.2.10 on 2024-07-04 08:06
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('classification', '0144_one_off_populate_historical_resolved_variant_hgvsconverter_version'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='importedalleleinfo',
+            name='hgvs_converter_version',
+            field=models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.PROTECT, to='classification.hgvsconverterversion'),
+        ),
+    ]

classification/models/classification_variant_info_models.py

@@ -373,6 +373,7 @@ class CalculatedVariantCoordinate:
     variant_coordinate: Optional[VariantCoordinate]
     genome_build: GenomeBuild
     message: str
+    hgvs_converter_version: Optional[HGVSConverterVersion]
 
     @property
     def variant_coordinate_str(self) -> Optional[str]:
@@ -406,6 +407,9 @@ class ImportedAlleleInfo(TimeStampedModel):
 
     imported_g_hgvs = TextField(null=True, blank=True)
 
+    hgvs_converter_version = ForeignKey(HGVSConverterVersion, null=True, blank=True, on_delete=PROTECT)
+    """ Tool used to resolve hgvs  """
+
     imported_transcript = TextField(null=True, blank=True)
     """
     Only needed if we're using g.hgvs
@@ -680,17 +684,23 @@ def resolved_builds(self) -> list[ResolvedVariantInfo]:
     def calculate_variant_coordinate(self) -> CalculatedVariantCoordinate:
         vc: Optional[VariantCoordinate] = None
         genome_build: Optional[GenomeBuild] = None
-        message: str
+        message: str = ""
+        hgvs_converter_version: Optional[HGVSConverterVersion] = None
         try:
             genome_build = self.imported_genome_build_patch_version.genome_build
             use_hgvs = self.imported_c_hgvs or self.imported_g_hgvs
             hgvs_matcher = HGVSMatcher(genome_build)
+            hgvs_converter_type = hgvs_matcher.hgvs_converter.get_hgvs_converter_type()
+            version = hgvs_matcher.hgvs_converter.get_version()
+
             vc_extra = hgvs_matcher.get_variant_coordinate_used_transcript_kind_method_and_matches_reference(use_hgvs)
-            message = f"HGVS matched by \"{vc_extra.method}\""
+            hgvs_converter_version = HGVSConverterVersion.get(hgvs_converter_type,
+                                                              version=version, method=vc_extra.method)
             vc = vc_extra.variant_coordinate
         except Exception as ex:
             message = str(ex)
-        return CalculatedVariantCoordinate(variant_coordinate=vc, genome_build=genome_build, message=message)
+        return CalculatedVariantCoordinate(variant_coordinate=vc, genome_build=genome_build,
+                                           message=message, hgvs_converter_version=hgvs_converter_version)
 
     def update_variant_coordinate(self):
         """ returns if a valid variant_coordinate could be derived """
@@ -700,6 +710,7 @@ def update_variant_coordinate(self):
         # but it's better to do that in the validation step
         cvc = self.calculate_variant_coordinate()
         self.message = cvc.message
+        self.hgvs_converter_version = cvc.hgvs_converter_version
         self.variant_coordinate = cvc.variant_coordinate_str
         if not cvc.is_valid:
             self.status = ImportedAlleleInfoStatus.FAILED
@@ -779,6 +790,7 @@ def set_matching_failed(self, message: Optional[str] = None):
     def hard_reset_matching_info(self):
         self.status = ImportedAlleleInfoStatus.PROCESSING
         self.matched_variant = None
+        self.hgvs_converter_version = None
         self.allele = None
         for genome_build in [GenomeBuild.grch37(), GenomeBuild.grch38()]:
             self._update_variant(genome_build=genome_build, variant=None)

genes/hgvs/hgvs_matcher.py

@@ -219,7 +219,8 @@ def _set_clingen_allele_registry_missing_transcript(transcript_accession: str):
         cache.set(key, True, timeout=WEEK_SECS)
 
     def get_variant_coordinate(self, hgvs_string: str) -> VariantCoordinate:
-        return self.get_variant_coordinate_used_transcript_kind_method_and_matches_reference(hgvs_string).variant_coordinate
+        vcd = self.get_variant_coordinate_used_transcript_kind_method_and_matches_reference(hgvs_string)
+        return vcd.variant_coordinate
 
     @staticmethod
     def _get_sort_key_transcript_version_and_methods(version, prefer_local=True, closest=False):
@@ -341,7 +342,7 @@ def get_variant_coordinate_used_transcript_kind_method_and_matches_reference(sel
                 # Ensure that ref is always present so we can give warning about provided reference
                 hgvs_string_for_version = hgvs_variant.format(max_ref_length=sys.maxsize)
                 if method == self.HGVS_METHOD_INTERNAL_LIBRARY:
-                    method = self.hgvs_converter.description()
+                    # method = self.hgvs_converter.description(describe_fallback=False)
                     variant_coordinate, matches_reference = self.hgvs_converter.hgvs_to_variant_coordinate_and_reference_match(hgvs_string_for_version, tv)
                 elif method == self.HGVS_METHOD_CLINGEN_ALLELE_REGISTRY:
                     if self._clingen_allele_registry_ok(tv.accession):