Fix IntegrityError on duplicate URN during lead discovery

Author: eracle

crm/models/lead.py

@@ -50,19 +50,20 @@ def get_profile(self, session) -> dict | None:
 
         urn = profile.get("urn") or None
         if urn and self.urn != urn:
-            self.urn = urn
-            self.save(update_fields=["urn"])
+            if Lead.objects.filter(urn=urn).exclude(pk=self.pk).exists():
+                logger.warning("URN %s already owned by another lead — skipping for %s", urn, self.public_identifier)
+            else:
+                self.urn = urn
+                self.save(update_fields=["urn"])
 
         return profile
 
     def get_urn(self, session) -> str:
         """LinkedIn URN. Reads cached column; falls back to a live scrape."""
         if self.urn:
             return self.urn
-        profile = self.get_profile(session)
-        if profile and profile.get("urn"):
-            self.urn = profile["urn"]
-            self.save(update_fields=["urn"])
+        self.get_profile(session)  # sets self.urn as side-effect
+        if self.urn:
             return self.urn
         raise ValueError(f"Lead {self.pk}: could not resolve URN after re-fetch")
 

linkedin/db/leads.py

@@ -33,9 +33,17 @@ def create_enriched_lead(session, url: str, profile: Dict[str, Any]) -> Optional
     public_id = canonical_pid or url_to_public_id(url)
     clean_url = public_id_to_url(public_id)
 
+    urn = profile.get("urn") or None
+
     with transaction.atomic():
         if Lead.objects.filter(public_identifier=public_id).exists():
             return None
+        if urn and Lead.objects.filter(urn=urn).exists():
+            logger.info(
+                "Lead with URN %s already exists — skipping duplicate %s",
+                urn, public_id,
+            )
+            return None
         lead = Lead.objects.create(linkedin_url=clean_url, public_identifier=public_id)
         _cache_urn_from_profile(lead, profile)