Update app.py

app.py

@@ -163,36 +163,58 @@ class EventScraper:
         Only return the json. nothing else. no comments."""
         return prompt
     
-    def parse_llm_response(self, response):
-        """Parse LLM's text response into structured events"""
-        try:
-            # Clean the response
-            response = response.strip()
-            
-            # Try parsing as JSON
-            try:
-                events = json.loads(response)
-                
-                # Ensure it's a list
-                if not isinstance(events, list):
-                    events = [events]
-                
-                return events
-            except json.JSONDecodeError:
-                # If direct JSON parsing fails, try extracting JSON
-                json_match = re.search(r'\[.*\]', response, re.DOTALL)
-                if json_match:
-                    try:
-                        events = json.loads(json_match.group(0))
-                        return events
-                    except json.JSONDecodeError:
-                        pass
-                
-                gr.Warning(f"Failed to parse response: {response}")
-                return []
-        except Exception as e:
-            gr.Warning(f"Parsing error: {str(e)}")
+def parse_llm_response(self, response):
+    """Parse LLM's text response into structured events"""
+    try:
+        # Clean the response and handle nested lists
+        response = response.strip()
+        
+        # Try parsing as JSON, handling potential nested structures
+        def flatten_events(data):
+            if isinstance(data, list):
+                flattened = []
+                for item in data:
+                    if isinstance(item, list):
+                        flattened.extend(flatten_events(item))
+                    elif isinstance(item, dict):
+                        flattened.append(item)
+                return flattened
             return []
+
+        try:
+            # First, attempt direct JSON parsing
+            events = json.loads(response)
+            events = flatten_events(events)
+        except json.JSONDecodeError:
+            # If direct parsing fails, try extracting JSON
+            import re
+            json_match = re.search(r'\[.*\]', response, re.DOTALL | re.MULTILINE)
+            if json_match:
+                try:
+                    events = json.loads(json_match.group(0))
+                    events = flatten_events(events)
+                except json.JSONDecodeError:
+                    events = []
+            else:
+                events = []
+        
+        # Clean and validate events
+        cleaned_events = []
+        for event in events:
+            # Ensure each event has at least a name
+            if event.get('name'):
+                # Set default values if missing
+                event.setdefault('date', '')
+                event.setdefault('time', '')
+                event.setdefault('location', '')
+                event.setdefault('description', '')
+                cleaned_events.append(event)
+        
+        return cleaned_events
+    
+    except Exception as e:
+        gr.Warning(f"Parsing error: {str(e)}")
+        return []
     
     def scrape_events(self):
         """Main method to scrape events from all URLs"""