""" dump_db_scripts.py — 把真实库脚本一次性落盘为本地缓存,供离线挖掘/迭代。 复用 run_eval.load_corpus_db 的 SQL 口径(只读),输出: scripts_cache.json — [{id, name, description, source}, ...] 用法: python dump_db_scripts.py --db-host 192.168.20.68 --db-port 5435 \ --db-name postgres --db-user ceshi --db-pass ceshi """ import argparse import json import os import sys sys.path.insert(0, os.path.dirname(__file__)) from run_eval import load_corpus_db def main(): p = argparse.ArgumentParser() p.add_argument("--db-host", default="192.168.20.68") p.add_argument("--db-port", default="5435") p.add_argument("--db-name", default="postgres") p.add_argument("--db-user", default="ceshi") p.add_argument("--db-pass", default="ceshi") p.add_argument("--out", default=os.path.join(os.path.dirname(__file__), "scripts_cache.json")) args = p.parse_args() records = load_corpus_db({ "host": args.db_host, "port": args.db_port, "dbname": args.db_name, "user": args.db_user, "password": args.db_pass, }) # 只留必要字段 slim = [{"id": r["id"], "name": r["name"], "description": r["description"], "source": r["source"]} for r in records] with open(args.out, "w", encoding="utf-8") as f: json.dump(slim, f, ensure_ascii=False, indent=2) print(f"已落盘 {len(slim)} 条脚本 → {args.out}") if __name__ == "__main__": main()