Auto-deploy backend from GitHub Actions

app/api/v1/endpoints/learn.py

@@ -31,6 +31,8 @@ async def create_learning_deck(
 
         return deck
 
+    except HTTPException:
+        raise
     except RuntimeError as e:
         logger.error(f"Deck Generation Error: {e}")
         raise HTTPException(

app/services/llm_service.py

@@ -127,7 +127,14 @@ async def generate_learning_deck(
             detail="Learning Deck AI timed out generating the lesson. Please try again.",
         )
     except Exception as e:
+        error_str = str(e)
+        if "429" in error_str or "RESOURCE_EXHAUSTED" in error_str:
+            raise HTTPException(
+                status_code=status.HTTP_429_TOO_MANY_REQUESTS,
+                detail="Tuklascope AI is currently analyzing too many anomalies globally. Please wait a few seconds and try again.",
+            )
+        # Handle Safety blocks in the future here
         raise HTTPException(
             status_code=status.HTTP_422_UNPROCESSABLE_ENTITY,
-            detail=f"Learning Deck AI Failed: {str(e)}",
+            detail=f"Learning Deck AI Failed: {error_str}",
         )