make_data.py

import json
import random
import os

# 文件路径
file_path = os.path.join(os.path.dirname(__file__), 'Qwen2.5-7B-Instruct_data', 'Qwen2.5-7B-Instruct_data.jsonl')

success_samples = []
fail_samples = []

with open(file_path, 'r', encoding='utf-8') as f:
    for line in f:
        try:
            entry = json.loads(line)
            label = entry.get("label")
            if label == "success":
                success_samples.append(entry)
            elif label == "fail":
                fail_samples.append(entry)
        except Exception:
            continue

num = len(success_samples)
# 随机选取等量fail
fail_samples_selected = random.sample(fail_samples, num) if len(fail_samples) >= num else fail_samples

def split_data(samples, train_ratio=0.6):
    random.shuffle(samples)
    split_idx = int(len(samples)*train_ratio)
    return samples[:split_idx], samples[split_idx:]

success_train, success_test = split_data(success_samples)
fail_train, fail_test = split_data(fail_samples_selected)

output_dir = os.path.join(os.path.dirname(__file__), 'Qwen2.5-7B-Instruct_data')
os.makedirs(output_dir, exist_ok=True)

def save_jsonl(lst, path):
    with open(path, 'w', encoding='utf-8') as f:
        for item in lst:
            f.write(json.dumps(item, ensure_ascii=False) + '\n')

save_jsonl(success_train, os.path.join(output_dir, 'success_train.jsonl'))
save_jsonl(success_test, os.path.join(output_dir, 'success_test.jsonl'))
save_jsonl(fail_train, os.path.join(output_dir, 'fail_train.jsonl'))
save_jsonl(fail_test, os.path.join(output_dir, 'fail_test.jsonl'))

print(f"数据已保存到 {output_dir}\nSuccess总数={len(success_samples)}, Fail选中={len(fail_samples_selected)}")
print(f"Success: train={len(success_train)}, test={len(success_test)}")
print(f"Fail: train={len(fail_train)}, test={len(fail_test)}")