Update utils/mistral.py

utils/mistral.py

@@ -39,6 +39,16 @@ def Data_Cleaner(text):
         return text  # Return the original text if cleaning goes wrong
 '''
 def Data_Cleaner(text):
+    """
+    Preprocess the JSON string to remove extra spaces, tabs, and newlines.
+    """
+    # Remove leading and trailing whitespace
+    text = text.strip()
+    # Remove unnecessary newlines and tabs
+    text = re.sub(r'\s*\n\s*', ' ', text)  # Replace newlines with a space
+    text = re.sub(r'\s+', ' ', text)  # Replace multiple spaces with a single space
+    text = re.sub(r'\s*([{}:,])\s*', r'\1', text)  # Remove spaces around braces, colons, and commas
+    
     # Use a regex pattern to extract JSON if it exists within ```json and ```
     pattern = r"```json\s*(\{.*?\})\s*```"  # Non-greedy matching inside braces
     match = re.search(pattern, text, re.DOTALL)  # DOTALL to match newlines