233 lines
9.9 KiB
Python
233 lines
9.9 KiB
Python
|
#!/usr/bin/env python3
|
|||
|
|
# -*- coding: utf-8 -*-
|
|||
|
|
"""同步活动报名「真实姓名」→ 用户昵称（增量脚本，手动不定时运行）。
|
|||
|
|
|
|||
|
|
背景
|
|||
|
|
----
|
|||
|
|
活动报名时多数用户未在 bookings.name 填真实姓名（默认占位「微信用户」），
|
|||
|
|
真实姓名存于报名表单 form_data_json 的「姓名」字段（events.form_fields_json
|
|||
|
|
中 label=「姓名」的字段）。本脚本按 bookings.username = users.username
|
|||
|
|
关联用户，把报名真实姓名写入 users.nickname。
|
|||
|
|
|
|||
|
|
增量机制
|
|||
|
|
--------
|
|||
|
|
- state 文件（默认 scripts/.sync_event_name_state.json）记录每个活动上次
|
|||
|
|
同步到的最大 created_at；本次只处理 created_at 大于游标的新增报名。
|
|||
|
|
- 同一用户多条报名只取最新一条。
|
|||
|
|
- 幂等：昵称已等于目标姓名的不重复写。
|
|||
|
|
- --full：忽略游标全量重扫（新增历史报名补同步）。
|
|||
|
|
|
|||
|
|
覆盖策略
|
|||
|
|
--------
|
|||
|
|
- 默认只填充「空 / 占位昵称」（空串、等于 username、等于『微信用户』）的
|
|||
|
|
用户，不覆盖用户已设置的真实昵称。
|
|||
|
|
- --overwrite：强制把昵称改为报名姓名（覆盖已有昵称）。
|
|||
|
|
|
|||
|
|
生产连接
|
|||
|
|
--------
|
|||
|
|
优先用命令行参数；否则依次读取以下 .env 的
|
|||
|
|
PINEAGENTS_DEMO_DATABASE_URL（或 PINEAGENTS_MYSQL_* 分字段）：
|
|||
|
|
1. serverrun/.env（生产部署目录，TenXun-Ubuntu2）
|
|||
|
|
2. .env（项目根）
|
|||
|
|
|
|||
|
|
用法
|
|||
|
|
----
|
|||
|
|
python scripts/sync_event_name_to_nickname.py --dry-run
|
|||
|
|
python scripts/sync_event_name_to_nickname.py
|
|||
|
|
python scripts/sync_event_name_to_nickname.py --full --overwrite --dry-run
|
|||
|
|
"""
|
|||
|
|
from __future__ import annotations
|
|||
|
|
|
|||
|
|
import argparse
|
|||
|
|
import asyncio
|
|||
|
|
import json
|
|||
|
|
import os
|
|||
|
|
import re
|
|||
|
|
import sys
|
|||
|
|
from datetime import datetime, timezone
|
|||
|
|
from pathlib import Path
|
|||
|
|
|
|||
|
|
try:
|
|||
|
|
from dotenv import load_dotenv
|
|||
|
|
except ImportError: # 无 dotenv 时跳过（连接参数必须显式给出）
|
|||
|
|
def load_dotenv(*_a, **_k): # noqa: ARG001
|
|||
|
|
return False
|
|||
|
|
|
|||
|
|
try:
|
|||
|
|
import asyncmy
|
|||
|
|
except ImportError:
|
|||
|
|
print("缺少依赖 asyncmy，请先安装：pip install asyncmy", file=sys.stderr)
|
|||
|
|
raise SystemExit(1)
|
|||
|
|
|
|||
|
|
SCRIPT_DIR = Path(__file__).resolve().parent
|
|||
|
|
PROJECT_ROOT = SCRIPT_DIR.parent
|
|||
|
|
DEFAULT_EVENT_ID = "ETL33LOE" # OPC青年数字电商创训营
|
|||
|
|
DEFAULT_STATE_FILE = SCRIPT_DIR / ".sync_event_name_state.json"
|
|||
|
|
|
|||
|
|
PLACEHOLDER_NAMES = {"微信用户", "游客", "未填写", ""}
|
|||
|
|
|
|||
|
|
|
|||
|
|
def _load_db_config(args: argparse.Namespace) -> dict:
|
|||
|
|
"""解析生产数据库连接：参数 > serverrun/.env > .env。"""
|
|||
|
|
if args.host:
|
|||
|
|
return {"host": args.host, "port": args.port, "user": args.user,
|
|||
|
|
"password": args.password, "db": args.db}
|
|||
|
|
|
|||
|
|
for env_path in (PROJECT_ROOT / "serverrun" / ".env", PROJECT_ROOT / ".env"):
|
|||
|
|
if not env_path.exists():
|
|||
|
|
continue
|
|||
|
|
load_dotenv(env_path)
|
|||
|
|
url = os.environ.get("PINEAGENTS_DEMO_DATABASE_URL", "")
|
|||
|
|
m = re.match(r"^\w+\+?\w*://([^:]+):([^@]+)@([^:/]+):(\d+)/([^?]+)", url)
|
|||
|
|
if m:
|
|||
|
|
return {"host": m.group(3), "port": int(m.group(4)),
|
|||
|
|
"user": m.group(1), "password": m.group(2), "db": m.group(5)}
|
|||
|
|
host = os.environ.get("PINEAGENTS_MYSQL_HOST", "")
|
|||
|
|
if host:
|
|||
|
|
return {"host": host,
|
|||
|
|
"port": int(os.environ.get("PINEAGENTS_MYSQL_PORT", "3306")),
|
|||
|
|
"user": os.environ.get("PINEAGENTS_MYSQL_USER", ""),
|
|||
|
|
"password": os.environ.get("PINEAGENTS_MYSQL_PASSWORD", ""),
|
|||
|
|
"db": os.environ.get("PINEAGENTS_MYSQL_DBNAME", "opc")}
|
|||
|
|
break # 只在第一个存在的 .env 中取配置
|
|||
|
|
|
|||
|
|
print("未找到生产数据库配置：请用 --host/--port/--user/--password/--db 指定，"
|
|||
|
|
"或确认 serverrun/.env / .env 存在。", file=sys.stderr)
|
|||
|
|
raise SystemExit(1)
|
|||
|
|
|
|||
|
|
|
|||
|
|
def _parse_iso(s: str) -> datetime:
|
|||
|
|
return datetime.fromisoformat(s.replace("Z", "+00:00"))
|
|||
|
|
|
|||
|
|
|
|||
|
|
def _is_placeholder_name(name: str) -> bool:
|
|||
|
|
return name.strip() in PLACEHOLDER_NAMES
|
|||
|
|
|
|||
|
|
|
|||
|
|
async def sync(args: argparse.Namespace) -> int:
|
|||
|
|
cfg = _load_db_config(args)
|
|||
|
|
print(f"[连接] 生产库 mysql://{cfg['user']}@{cfg['host']}:{cfg['port']}/{cfg['db']}")
|
|||
|
|
conn = await asyncmy.connect(host=cfg["host"], port=cfg["port"], user=cfg["user"],
|
|||
|
|
password=cfg["password"], db=cfg["db"], charset="utf8mb4")
|
|||
|
|
try:
|
|||
|
|
cur = await conn.cursor()
|
|||
|
|
|
|||
|
|
# 1. 活动表单：定位「姓名」字段
|
|||
|
|
await cur.execute("SELECT form_fields_json FROM events WHERE id=%s", (args.event_id,))
|
|||
|
|
row = await cur.fetchone()
|
|||
|
|
if not row:
|
|||
|
|
print(f"活动不存在: {args.event_id}", file=sys.stderr)
|
|||
|
|
return 2
|
|||
|
|
fields = json.loads(row[0] or "[]")
|
|||
|
|
name_field = next((f.get("id") for f in fields if f.get("label") == "姓名"), None)
|
|||
|
|
if not name_field:
|
|||
|
|
print(f"活动 {args.event_id} 未定义「姓名」字段", file=sys.stderr)
|
|||
|
|
return 2
|
|||
|
|
print(f"[字段] 姓名字段: {name_field}")
|
|||
|
|
|
|||
|
|
# 2. 增量游标
|
|||
|
|
state = {}
|
|||
|
|
if args.state_file.exists():
|
|||
|
|
try:
|
|||
|
|
state = json.loads(args.state_file.read_text(encoding="utf-8"))
|
|||
|
|
except (OSError, json.JSONDecodeError):
|
|||
|
|
state = {}
|
|||
|
|
cursor_ts = "" if args.full else (state.get(args.event_id) or {}).get("last_created_at", "")
|
|||
|
|
if cursor_ts:
|
|||
|
|
print(f"[增量] 游标 last_created_at = {cursor_ts}（--full 可全量重扫）")
|
|||
|
|
else:
|
|||
|
|
print("[增量] 无历史游标，全量扫描")
|
|||
|
|
|
|||
|
|
# 3. 取报名（只取该活动、created_at > 游标）
|
|||
|
|
sql = ("SELECT id, username, name, form_data_json, created_at FROM bookings "
|
|||
|
|
"WHERE event_id=%s")
|
|||
|
|
params = [args.event_id]
|
|||
|
|
if cursor_ts:
|
|||
|
|
sql += " AND created_at > %s"
|
|||
|
|
params.append(cursor_ts)
|
|||
|
|
sql += " ORDER BY created_at ASC"
|
|||
|
|
await cur.execute(sql, params)
|
|||
|
|
rows = await cur.fetchall()
|
|||
|
|
print(f"[扫描] 待处理报名 {len(rows)} 条")
|
|||
|
|
|
|||
|
|
# 4. 同一用户多条报名只保留最新一条
|
|||
|
|
latest: dict[str, dict] = {}
|
|||
|
|
for bid, username, name, form_json, created_at in rows:
|
|||
|
|
if not username:
|
|||
|
|
continue
|
|||
|
|
fd = json.loads(form_json or "{}")
|
|||
|
|
real_name = str(fd.get(name_field) or "").strip()
|
|||
|
|
latest[username] = {"booking_id": bid, "name": name or "",
|
|||
|
|
"real_name": real_name, "created_at": created_at}
|
|||
|
|
|
|||
|
|
# 5. 关联用户并执行
|
|||
|
|
updated, skipped_no_name, skipped_placeholder, skipped_missing, skipped_same = 0, 0, 0, 0, 0
|
|||
|
|
max_created = cursor_ts
|
|||
|
|
for username, info in latest.items():
|
|||
|
|
if info["created_at"] > max_created:
|
|||
|
|
max_created = info["created_at"]
|
|||
|
|
real = info["real_name"]
|
|||
|
|
if not real:
|
|||
|
|
skipped_no_name += 1
|
|||
|
|
continue
|
|||
|
|
if _is_placeholder_name(real) or real == username or len(real) < 2:
|
|||
|
|
skipped_placeholder += 1
|
|||
|
|
continue
|
|||
|
|
|
|||
|
|
await cur.execute("SELECT username, nickname FROM users WHERE username=%s", (username,))
|
|||
|
|
u = await cur.fetchone()
|
|||
|
|
if not u:
|
|||
|
|
skipped_missing += 1
|
|||
|
|
print(f" [跳过] 用户不存在 {username}")
|
|||
|
|
continue
|
|||
|
|
_u, nickname = u[0], u[1] or ""
|
|||
|
|
if nickname == real:
|
|||
|
|
skipped_same += 1
|
|||
|
|
continue
|
|||
|
|
if not args.overwrite and not (
|
|||
|
|
not nickname.strip()
|
|||
|
|
or nickname.strip() == username
|
|||
|
|
or _is_placeholder_name(nickname.strip())
|
|||
|
|
):
|
|||
|
|
skipped_same += 1 # 已有真实昵称且未开 --overwrite，不动
|
|||
|
|
print(f" [保留] {username} 已有昵称「{nickname}」，不覆盖")
|
|||
|
|
continue
|
|||
|
|
print(f" [{'DRY' if args.dry_run else 'SET '}] {username} 昵称: "
|
|||
|
|
f"{nickname or '(空)'} -> {real} (报名 {info['booking_id']})")
|
|||
|
|
if not args.dry_run:
|
|||
|
|
await cur.execute("UPDATE users SET nickname=%s WHERE username=%s",
|
|||
|
|
(real, username))
|
|||
|
|
updated += 1
|
|||
|
|
|
|||
|
|
if not args.dry_run:
|
|||
|
|
await conn.commit()
|
|||
|
|
# 更新游标（幂等，重复跑安全）
|
|||
|
|
state.setdefault(args.event_id, {})["last_created_at"] = max_created
|
|||
|
|
args.state_file.parent.mkdir(parents=True, exist_ok=True)
|
|||
|
|
args.state_file.write_text(
|
|||
|
|
json.dumps(state, ensure_ascii=False, indent=2), encoding="utf-8")
|
|||
|
|
print(f"[游标] 已更新 -> {max_created}")
|
|||
|
|
print(f"[结果] {'（dry-run 未写库）' if args.dry_run else '已写库'}"
|
|||
|
|
f" 更新 {updated}；跳过：无姓名 {skipped_no_name}、占位/无效 {skipped_placeholder}、"
|
|||
|
|
f"用户不存在 {skipped_missing}、无需变更 {skipped_same}")
|
|||
|
|
return 0
|
|||
|
|
finally:
|
|||
|
|
conn.close()
|
|||
|
|
|
|||
|
|
|
|||
|
|
if __name__ == "__main__":
|
|||
|
|
parser = argparse.ArgumentParser(description="同步活动报名真实姓名到用户昵称（增量）")
|
|||
|
|
parser.add_argument("--event-id", default=DEFAULT_EVENT_ID, help="活动 ID")
|
|||
|
|
parser.add_argument("--dry-run", action="store_true", help="只预览不写库")
|
|||
|
|
parser.add_argument("--full", action="store_true", help="忽略增量游标全量重扫")
|
|||
|
|
parser.add_argument("--overwrite", action="store_true", help="覆盖用户已有昵称")
|
|||
|
|
parser.add_argument("--state-file", type=Path, default=DEFAULT_STATE_FILE,
|
|||
|
|
help="增量游标文件路径")
|
|||
|
|
parser.add_argument("--host", help="MySQL 主机（默认读 .env）")
|
|||
|
|
parser.add_argument("--port", type=int, default=0, help="MySQL 端口")
|
|||
|
|
parser.add_argument("--user", help="MySQL 用户")
|
|||
|
|
parser.add_argument("--password", help="MySQL 密码")
|
|||
|
|
parser.add_argument("--db", default="opc", help="MySQL 库名")
|
|||
|
|
args = parser.parse_args()
|
|||
|
|
raise SystemExit(asyncio.run(sync(args)))
|