(task_info: Dict, evaluation: Dict, raw_response: str, plot_file: Path,
results_dir: Path, generation_model: str, evaluation_model: str,
execution_success: bool, execution_message: str)
| 735 | |
| 736 | |
| 737 | def save_score_result(task_info: Dict, evaluation: Dict, raw_response: str, plot_file: Path, |
| 738 | results_dir: Path, generation_model: str, evaluation_model: str, |
| 739 | execution_success: bool, execution_message: str) -> Path: |
| 740 | |
| 741 | |
| 742 | task_dir = task_info['data_directory'].name |
| 743 | task_result_dir = results_dir / task_dir |
| 744 | task_result_dir.mkdir(exist_ok=True) |
| 745 | |
| 746 | |
| 747 | difficulty = task_info['difficulty'] |
| 748 | is_multi = task_info['is_multi'] |
| 749 | score_filename = f"score_{difficulty}_mul.json" if is_multi else f"score_{difficulty}.json" |
| 750 | |
| 751 | |
| 752 | overall_score = calculate_overall_score(evaluation) |
| 753 | if task_info['raw_generation_response']==None: |
| 754 | task_info['raw_generation_response']='' |
| 755 | |
| 756 | result = { |
| 757 | "generation_model": generation_model, |
| 758 | "evaluation_model": evaluation_model, |
| 759 | "task_file": str(task_info['task_file']), |
| 760 | "category": task_info['category'], |
| 761 | "instruction": task_info['instruction'], |
| 762 | "difficulty": difficulty, |
| 763 | "is_multi": is_multi, |
| 764 | "timestamp": datetime.now().isoformat(), |
| 765 | "raw_generation_response": task_info['raw_generation_response'], |
| 766 | |
| 767 | "execution_success": execution_success, |
| 768 | "execution_message": execution_message, |
| 769 | "code_generated": bool(task_info.get('code')), |
| 770 | "plot_generated": plot_file is not None, |
| 771 | |
| 772 | "evaluation": evaluation, |
| 773 | "overall_score": overall_score, |
| 774 | "raw_response": raw_response |
| 775 | } |
| 776 | |
| 777 | |
| 778 | if plot_file: |
| 779 | result["plot_file"] = str(plot_file) |
| 780 | |
| 781 | score_file = task_result_dir / score_filename |
| 782 | with open(score_file, 'w', encoding='utf-8') as f: |
| 783 | json.dump(result, f, ensure_ascii=False, indent=2) |
| 784 | |
| 785 | return score_file |
| 786 | |
| 787 | |
| 788 | def prepare_task_info(task_file: Path, max_data_rows: int) -> Optional[Dict]: |
no test coverage detected