Function extract_json

src/utils/utils.py:48–79

Extract and parse JSON content from a text response. Attempts to parse the response as JSON directly, then tries to extract JSON from code blocks if direct parsing fails. Args: response (str): The text response containing JSON content Returns: dict: The parsed JSON

(response: str)

    return code 

def extract_json(response: str) -> dict:
    """Extract and parse JSON content from a text response.

    Attempts to parse the response as JSON directly, then tries to extract JSON from code blocks
    if direct parsing fails.

    Args:
        response (str): The text response containing JSON content

    Returns:
        dict: The parsed JSON content as a dictionary, or empty list if parsing fails

    Note:
        Will attempt to parse content between ```json markers first, then between generic ``` markers
    """
    try:
        evaluation_json = json.loads(response)
    except json.JSONDecodeError:
        # If JSON parsing fails, try to extract the content between ```json and ```
        match = re.search(r'```json\n(.*?)\n```', response, re.DOTALL)
        if not match:
            # If no match for ```json, try to extract content between ``` and ```
            match = re.search(r'```\n(.*?)\n```', response, re.DOTALL)
        
        if match:
            evaluation_content = match.group(1)
            evaluation_json = json.loads(evaluation_content)
        else:
            # return empty list
            evaluation_json = []
            print(f"Warning: Failed to extract valid JSON content from {response}")
    return evaluation_json

def _fix_unicode_to_latex(text: str, parse_unicode: bool = True) -> str:
    """Convert Unicode symbols to LaTeX source code.