|
|
import json |
|
|
|
|
|
|
|
|
with open('ms-swift/matched_scores_2_1.json', 'r', encoding='utf-8') as f: |
|
|
allcorrect_data = json.load(f) |
|
|
|
|
|
|
|
|
with open('/root/autodl-tmp/600_train/merged_shuffled_train.json', 'r', encoding='utf-8') as f: |
|
|
merged_data = json.load(f) |
|
|
|
|
|
|
|
|
for entry in allcorrect_data: |
|
|
|
|
|
key = entry.get('key') |
|
|
if key: |
|
|
|
|
|
if key in merged_data: |
|
|
|
|
|
error_type = merged_data[key].get('error_type') |
|
|
entry['error_type'] = error_type |
|
|
|
|
|
|
|
|
output_file = 'ms-swift/allcorrect_with_error_type.json' |
|
|
with open(output_file, 'w', encoding='utf-8') as f: |
|
|
json.dump(allcorrect_data, f, ensure_ascii=False, indent=2) |
|
|
|
|
|
print(f"处理完成,结果已保存到 {output_file}") |
|
|
|
|
|
|
|
|
error_type_stats = {} |
|
|
for entry in allcorrect_data: |
|
|
error_type = entry.get('error_type') |
|
|
if error_type: |
|
|
error_type_stats[error_type] = error_type_stats.get(error_type, 0) + 1 |
|
|
else: |
|
|
error_type_stats['no_error_type'] = error_type_stats.get('no_error_type', 0) + 1 |
|
|
|
|
|
print("\nError Type 统计:") |
|
|
for error_type, count in error_type_stats.items(): |
|
|
print(f"{error_type}: {count}") |
|
|
|