feat: enhance JSON validation to support both model versions

Updated validate_json function to handle both:
- Original model version: outputs plain JSON format
- New fine-tuned version: outputs JSON wrapped in code blocks

Key changes:
- Added support for extracting JSON from code blocks (```json\n{...}\n```)
- Maintained backward compatibility with plain JSON format
- Preserved JSON formatting fixes for phone numbers and trailing commas
- Improved error handling for both formats

This ensures the demo interface works seamlessly with both old and new versions of the fine-tuned Gemma-3 model, providing robust parsing regardless of the output format.

app.py

@@ -44,14 +44,29 @@ def load_model():
         return None, None
 
 def validate_json(output: str) -> tuple:
-    """Validate and extract JSON from model output"""
+    """Validate and extract JSON from model output - supports both plain JSON and code block formats"""
     try:
-        json_match = re.search(r'\{[\s\S]*\}', output)
-        if not json_match:
-            return False, None, "No JSON found / 未找到JSON"
+        # First, try to extract JSON from code blocks (new model version)
+        json_match = re.search(r'```(?:json)?\s*(\{[\s\S]*?\})\s*```', output)
+        if json_match:
+            json_str = json_match.group(1)
+        else:
+            # If no code block, look for JSON directly (old model version)
+            json_match = re.search(r'\{[\s\S]*\}', output)
+            if not json_match:
+                return False, None, "No JSON found / 未找到JSON"
+            json_str = json_match.group(0)
+        
+        # Fix common JSON issues for both formats
+        # 1. Add quotes around phone numbers (they often start with 0)
+        json_str = re.sub(r'("phone_num":\s*)(\d[-\d]*)', r'\1"\2"', json_str)
         
-        json_str = json_match.group(0)
+        # 2. Add quotes around num_people if it's a number
+        json_str = re.sub(r'("num_people":\s*)(\d+)', r'\1"\2"', json_str)
+        
+        # 3. Fix trailing commas
         json_str = re.sub(r',\s*\}', '}', json_str)
+        
         parsed = json.loads(json_str)
         return True, parsed, "Valid JSON / 有效的JSON"
     except json.JSONDecodeError: