Function create_evaluation_request

RealChart2Code_eval/evaluate_task3.py:561–601

(difficulty: str, category: str, instruction: str, image_path:Path,
                            image_base64: str, evaluation_prompt: str, model_name: str)

def create_evaluation_request(difficulty: str, category: str, instruction: str, image_path:Path,
                            image_base64: str, evaluation_prompt: str, model_name: str) -> Dict:
    
    task_prompt = f"""# Task
    Category: {category}
    """
    # Instruction: {instruction}
    image_chart = encode_image_to_base64(image_path)
    
    user_content = [
        {
            "type": "text",
            "text": f"**Task**:\n{task_prompt}\n\n**Generated Chart**: Please evaluate based on the following generated chart image."
        },
        {
            "type": "image_url",
            "image_url": {
                "url": image_base64
            }
        },
        {
            "type": "text",
            "text": f"**Provided Chart**: Please evaluate based on the following provided chart image. "
        },
        {
            "type": "image_url",
            "image_url": {
                "url": image_chart
            }
        }
    ]

    return {
        "model": model_name,
        "messages": [
            {"role": "system", "content": evaluation_prompt},
            {"role": "user", "content": user_content}
        ],
        "max_tokens": 8192,
        "temperature": 0.0,
    }


def parse_evaluation_response(response_content: str) -> Dict: