Update endpoint lgd-publication to support extra data

gene2phenotype_project/gene2phenotype_app/serializers/locus_genotype_disease.py

@@ -573,11 +573,43 @@ def update_mechanism(self, lgd_instance, validated_data):
             is_deleted = 0
         )
 
+        # Save old molecular mechanism to delete it later
+        old_mechanism_obj = lgd_instance.molecular_mechanism
+
+        # Update LGD record
+        # The mechanism has to be updated in the locus_genotype_disease before the evidence is added
+        # Because the evidence is going to be linked to the new lgd.molecular_mechanism
+        lgd_instance.molecular_mechanism = mechanism_obj
+        lgd_instance.date_review = datetime.now()
+        lgd_instance.save()
+
         # Get evidence - the mechanism evidence was validated in the view 'LGDUpdateMechanism'
         # Example: {'pmid': '25099252', 'description': 'text', 'evidence_types': 
         #          [{'primary_type': 'Rescue', 'secondary_type': ['Human', 'Patient Cells']}]}
-        for evidence in mechanism_evidence:
-            pmid = evidence.get("pmid")
+        self.update_mechanism_evidence(lgd_instance, mechanism_evidence)
+
+        # The old molecular mechanism can be deleted - the deletion is stored in the history table
+        # As this method only allows to update 'undetermined' mechanisms, there is no evidence to be deleted
+        old_mechanism_obj.delete()
+
+        return lgd_instance
+
+    def update_mechanism_evidence(self, lgd_obj, validated_data):
+        """
+            Method to only update the evidence of the LGD molecular mechanism.
+
+            'validated_data' example:
+                "mechanism_evidence": [{
+                                        "pmid": "1234",
+                                        "description": "This is new evidence for the existing mechanism evidence.",
+                                        "evidence_types": [ { "primary_type": "Function",
+                                                              "secondary_type": [ "Biochemical" ]}
+                                        ]}]
+        """
+        mechanism_obj = lgd_obj.molecular_mechanism
+
+        for evidence in validated_data:
+            pmid = evidence["pmid"]
 
             # Check if the PMID exists in G2P
             # When updating the mechanism the supporting pmid used as evidence
@@ -588,20 +620,20 @@ def update_mechanism(self, lgd_instance, validated_data):
                 # TODO: improve in future to insert new pmids + link them to the record
                 raise serializers.ValidationError({"message": f"pmid '{pmid}' not found in G2P"})
 
-            if evidence.get("description") != "":
-                description = evidence.get("description")
+            if evidence["description"] != "":
+                description = evidence["description"]
             else:
                 description = None
 
-            evidence_types = evidence.get("evidence_types")
+            evidence_types = evidence["evidence_types"]
             for evidence_type in evidence_types:
                 # primary_type is the evidence subtype ('rescue')
                 primary_type = evidence_type.get("primary_type", None)
                 if not primary_type:
                     raise serializers.ValidationError({"message": f"Empty evidence subtype"})
                 primary_type = primary_type.lower()
                 # secondary_type is the evidence value ('human')
-                secondary_type = evidence_type.get("secondary_type")
+                secondary_type = evidence_type["secondary_type"]
                 for m_type in secondary_type:
                     try:
                         cv_evidence_obj = CVMolecularMechanism.objects.get(
@@ -613,27 +645,28 @@ def update_mechanism(self, lgd_instance, validated_data):
                         raise serializers.ValidationError({"message": f"Invalid mechanism evidence '{m_type}'"})
 
                     # Insert evidence
-                    mechanism_evidence_obj = MolecularMechanismEvidence.objects.create(
+                    try:
+                        mechanism_evidence_obj = MolecularMechanismEvidence.objects.get(
                         molecular_mechanism = mechanism_obj,
-                        description = description,
                         publication = publication_obj,
-                        evidence = cv_evidence_obj,
-                        is_deleted = 0
+                        evidence = cv_evidence_obj
                     )
-
-        # Save old molecular mechanism to delete it later
-        old_mechanism_obj = lgd_instance.molecular_mechanism
-
-        # Update LGD record
-        lgd_instance.molecular_mechanism = mechanism_obj
-        lgd_instance.date_review = datetime.now()
-        lgd_instance.save()
-
-        # The old molecular mechanism can be deleted - the deletion is stored in the history table
-        # As this method only allows to update 'undetermined' mechanisms, there is no evidence to be deleted
-        old_mechanism_obj.delete()
-
-        return lgd_instance
+                    except MolecularMechanismEvidence.DoesNotExist:
+                        mechanism_evidence_obj = MolecularMechanismEvidence.objects.create(
+                            molecular_mechanism = mechanism_obj,
+                            description = description,
+                            publication = publication_obj,
+                            evidence = cv_evidence_obj,
+                            is_deleted = 0
+                        )
+                    else:
+                        if(mechanism_evidence_obj.is_deleted == 1):
+                            mechanism_evidence_obj.is_deleted = 0
+                            mechanism_evidence_obj.save()
+
+                    # Update LGD date_review
+                    lgd_obj.date_review = datetime.now()
+                    lgd_obj.save()
 
     class Meta:
         model = LocusGenotypeDisease

gene2phenotype_project/gene2phenotype_app/serializers/phenotype.py

@@ -131,7 +131,7 @@ class Meta:
 class LGDPhenotypeListSerializer(serializers.Serializer):
     """
         Serializer to accept a list of phenotypes.
-        Called by: LocusGenotypeDiseaseAddPhenotypes()
+        Called by: LocusGenotypeDiseaseAddPhenotypes() and view LGDEditPhenotypes()
     """
     phenotypes = LGDPhenotypeSerializer(many=True)
 

gene2phenotype_project/gene2phenotype_app/serializers/publication.py

@@ -296,20 +296,20 @@ def validate(self, data):
         """
             Overwrite the method to validate the data.
             This validate() method is activated by the curation (draft entry) endpoints and
-            by the LGDPublicationListSerializer as a consequence this method is identical to
-            the validate method in LGDPublicationListSerializer.
+            by the LGDPublicationListSerializer.
+            This method is identical to the validate method in LGDPublicationListSerializer.
         """
         if hasattr(self, 'initial_data'):
             data = self.initial_data
             valid_headers = ["families", "consanguinity", "ancestries", "affected_individuals"]
+            extra_headers = ["phenotypes", "variant_types", "variant_descriptions"]
 
-            publication = self.initial_data.get("publication")
-            # check if 'families' is defined
+            # check if 'families' is defined in the initial data
             # correct structure is: 
             # { "families": 200, "consanguinity": "unknown", "ancestries": "african", 
             #   "affected_individuals": 100 }
-            if "families" in publication and publication.get("families") is not None:
-                families = publication.get("families")
+            if "families" in self.initial_data:
+                families = self.initial_data.get("families") # If 'families' is in initial data then it cannot be null
                 for header in families.keys():
                     if header not in valid_headers:
                         raise serializers.ValidationError(f"Got unknown field in families: {header}")
@@ -335,24 +335,19 @@ def create(self, validated_data):
         comment = None
         families = None
 
-        publication_data = validated_data.get('publication') # includes 'pmid', 'comment', 'families'
+        publication_data = self.initial_data.get('publication') # 'publication': {'pmid': 39385417}
+        comment = self.initial_data.get('comment', None) # 'comment': {'comment': 'this is a comment', 'is_public': 1}
+        families = self.initial_data.get("families", None) # "families": { "families": 200, "consanguinity": "unknown", "ancestries": "african", "affected_individuals": 2 }
 
-        if "comment" in publication_data:
-            comment = publication_data.get("comment")
+        if comment:
+            comment_text = comment["comment"]
 
-            # Check if comment text is invalid or missing
-            if not comment or "comment" not in comment or comment.get("comment") == "":
+            # Check if comment text is empty string
+            if not comment_text or comment_text == "":
                 comment = None
             # If 'is_public' is not defined set it to public (default)
-            elif "is_public" not in comment:
-                comment["is_public"] = 1
-
-        if "families" in publication_data:
-            families = publication_data.get('families') # extra data - families reported in publication
-
-            # Check if family data is invalid or missing
-            if not families:
-                families = None
+            else:
+                comment["is_public"] = comment.get("is_public", 1)
 
         # it is necessary to send the user
         # the publication comment is linked to the user
@@ -379,13 +374,22 @@ def create(self, validated_data):
                 publication = publication_obj,
                 is_deleted = 0
             )
+            # There is a new publication linked to the LGD record, 
+            # the record has to reflect the date of this change
+            # update the date_review of the LGD record
+            lgd.date_review = datetime.now()
+            lgd.save()
 
         # If LGD-publication already exists then returns the existing object
         # New comments and/or family info are added to the existing object
         # If existing LGD-publication is deleted then update to not deleted
         if lgd_publication_obj.is_deleted != 0:
             lgd_publication_obj.is_deleted = 0
             lgd_publication_obj.save()
+            # The lgd-publication is not deleted anymore which is equivallent to
+            # creating a new link, this means we have to update the record 'date_review'
+            lgd.date_review = datetime.now()
+            lgd.save()
 
         return lgd_publication_obj
 
@@ -396,7 +400,8 @@ class Meta:
 class LGDPublicationListSerializer(serializers.Serializer):
     """
         Serializer to accept a list of publications.
-        Called by: LocusGenotypeDiseaseAddPublication()
+        This method only validates the publications, it does not update any data.
+        Called by: LocusGenotypeDiseaseAddPublication() and view LGDEditPublications()
     """
     publications = LGDPublicationSerializer(many=True)
 
@@ -411,7 +416,8 @@ def validate(self, data):
         # by default these fields are not accepted as valid as they are not part of the LGDPublication
         if hasattr(self, 'initial_data'):
             data = self.initial_data
-            valid_headers = ["families", "consanguinity", "ancestries", "affected_individuals"]
+            families_valid_headers = ["families", "consanguinity", "ancestries", "affected_individuals"]
+            extra_headers = ["phenotypes", "variant_types", "variant_descriptions"]
 
             publications = self.initial_data.get("publications")
 
@@ -420,12 +426,11 @@ def validate(self, data):
 
                 # check if 'families' is defined
                 # correct structure is: 
-                # { "families": 200, "consanguinity": "unknown", "ancestries": "african", 
-                #   "affected_individuals": 100 }
+                # "families": { "families": 2, "consanguinity": "unknown", "ancestries": "african", "affected_individuals": 1 }
                 if "families" in publication:
                     families = publication.get("families")
                     for header in families.keys():
-                        if header not in valid_headers:
+                        if header not in families_valid_headers:
                             raise serializers.ValidationError(f"Got unknown field in families: {header}")
 
                 # TODO check if 'comment' is defined

gene2phenotype_project/gene2phenotype_app/views/base.py

@@ -59,6 +59,13 @@ def handle_no_update(self, data, stable_id):
         else:
             raise PermissionDenied(f"Cannot update '{data}' for ID '{stable_id}'")
 
+    def handle_update_exception(self, exception, context_message):
+        if hasattr(exception, 'detail') and 'message' in exception.detail:
+            error_message = exception.detail['message']
+            return Response({"error": f"{context_message}: {error_message}"}, status=status.HTTP_400_BAD_REQUEST)
+        else:
+            error_message = context_message
+            return Response({"error": f"{context_message}"}, status=status.HTTP_400_BAD_REQUEST)
 
 @api_view(['GET'])
 def ListEndpoints(request):