Download 4-HSLS/parse_codebook.py from yijiexu/LLMValues: direct link, hf CLI and curl.
- Browser
- Download file 5.94 kB
-
https://huggingface.co/datasets/yijiexu/LLMValues/resolve/main/4-HSLS/parse_codebook.py
- Command line
-
hf download hf://datasets/yijiexu/LLMValues/4-HSLS/parse_codebook.py
-
curl -L -o parse_codebook.py https://huggingface.co/datasets/yijiexu/LLMValues/resolve/main/4-HSLS/parse_codebook.py
5.94 kB
| import re | |
| import json | |
| import sys | |
| from tqdm import tqdm | |
| def parse_codebook_to_jsonl(input_text_file, output_jsonl_file): | |
| """ | |
| 将一个特定格式的 codebook 文本文件解析并转换为 JSONL 文件。 | |
| 此版本经过修正,能够将所有非结构化字段(如 Description, Please note, | |
| SAS Logic 等)的内容统一合并到 'description' 键下,以满足特定需求。 | |
| Args: | |
| input_text_file (str): 输入的 codebook 文本文件路径。 | |
| output_jsonl_file (str): 输出的 JSONL 文件路径。 | |
| """ | |
| try: | |
| with open(input_text_file, 'r', encoding='utf-8') as f: | |
| content = f.read() | |
| except FileNotFoundError: | |
| print(f"❌ 错误:输入文件 '{input_text_file}' 未找到。请确保文件存在且路径正确。") | |
| sys.exit(1) | |
| except Exception as e: | |
| print(f"❌ 读取文件时发生错误: {e}") | |
| sys.exit(1) | |
| # 1. 分割记录 | |
| # 使用正则表达式按双虚线分隔符来分割整个文件内容。 | |
| blocks = re.split(r'\n-+\n-+\s*\n', content) | |
| # 初始化统计计数器和失败记录列表 | |
| success_count = 0 | |
| failure_count = 0 | |
| failed_records = [] | |
| # 定义标准、结构化的字段键名 | |
| STANDARD_KEYS = {'file', 'name', 'position', 'length', 'label'} | |
| print(f"🚀 开始解析文件 '{input_text_file}'...") | |
| with open(output_jsonl_file, 'w', encoding='utf-8') as outfile: | |
| # 定义用于匹配 "Key: Value" 格式的正则表达式 | |
| key_pattern = re.compile(r'^([\w\s/()-]+?):\s*(.*)$') | |
| for block in tqdm(blocks, desc="解析变量记录", unit="条"): | |
| block = block.strip() | |
| # 跳过文件头、结尾或空的文本块 | |
| if not block or not block.startswith("File:"): | |
| continue | |
| try: | |
| # --- 步骤 1: 初步解析为临时字典 --- | |
| temp_record = {} | |
| current_key = None | |
| value_buffer = [] | |
| lines = block.split('\n') | |
| for line in lines: | |
| match = key_pattern.match(line) | |
| if match: | |
| if current_key: | |
| temp_record[current_key] = '\n'.join( | |
| value_buffer).strip() | |
| key = match.group(1).strip().lower().replace(' ', '_') | |
| value = match.group(2).strip() | |
| current_key = key | |
| value_buffer = [value] if value else [] | |
| elif current_key: | |
| value_buffer.append(line) | |
| if current_key: | |
| temp_record[current_key] = '\n'.join(value_buffer).strip() | |
| if not temp_record: | |
| continue | |
| # --- 步骤 2: 后处理,将非标准字段合并到 description --- | |
| final_record = {} | |
| description_parts = [] | |
| # 遍历临时记录的键,为了保持原始顺序,可以对键进行排序(可选,但推荐) | |
| # 这里我们假设原始顺序不重要,或者temp_record(Python 3.7+)保持插入顺序 | |
| for key, value in temp_record.items(): | |
| if key in STANDARD_KEYS: | |
| # 如果是标准键,直接复制 | |
| final_record[key] = value | |
| else: | |
| # 如果是非标准键,将其值存入 description 列表 | |
| # 为了保持内容干净,我们只合并值,不添加原始键名 | |
| description_parts.append(value) | |
| # 合并所有描述性内容 | |
| if description_parts: | |
| final_record['description'] = '\n\n'.join( | |
| description_parts).strip() | |
| # 转换数值类型 | |
| for field in ['position', 'length']: | |
| if field in final_record: | |
| try: | |
| final_record[field] = int(final_record[field]) | |
| except (ValueError, TypeError): | |
| # 保留原始值并在需要时添加错误标记 | |
| pass | |
| # 写入文件 | |
| json.dump(final_record, outfile, ensure_ascii=False) | |
| outfile.write('\n') | |
| success_count += 1 | |
| except Exception as e: | |
| failure_count += 1 | |
| failed_records.append({ | |
| "error": str(e), | |
| "block_content": block | |
| }) | |
| # 3. 打印最终的统计报告 | |
| print("\n" + "=" * 50) | |
| print("🎉 解析完成!") | |
| print("=" * 50) | |
| print(f"📊 统计报告:") | |
| print(f" - 总共处理记录块: {len(blocks)}") | |
| print(f" - ✅ 成功解析并写入: {success_count} 条") | |
| print(f" - ❌ 解析失败: {failure_count} 条") | |
| print("=" * 50) | |
| if failure_count > 0: | |
| print("\n🔍 以下是解析失败的记录详情:\n") | |
| for i, failure in enumerate(failed_records, 1): | |
| print(f"--- 失败记录 {i} ---") | |
| print(f"错误原因: {failure['error']}") | |
| print("原始文本块:") | |
| print("-" * 20) | |
| print(failure['block_content']) | |
| print("-" * 20 + "\n") | |
| print(f"✨ 结果已保存至 '{output_jsonl_file}'。") | |
| # --- 如何使用 --- | |
| # 1. 将此脚本保存为 .py 文件(例如 parser.py)。 | |
| # 2. 确保你的数据文件 'Codebook_HSLS_17_PETS.txt' 与脚本在同一目录。 | |
| # 3. 运行脚本: python parser.py | |
| if __name__ == "__main__": | |
| # 定义输入和输出文件名 | |
| input_file = 'Codebook_HSLS_17_PETS.txt' | |
| output_file = 'codebook.jsonl' | |
| # 调用主函数 | |
| parse_codebook_to_jsonl(input_file, output_file) | |