weibo_signin/backend/task_scheduler/app/main.py

"""
Weibo-HotSign 定时任务调度器 (APScheduler)

启动时从 DB 加载所有 enabled 的 Task，按 cron 精确注册。
每 5 分钟重新同步一次 DB（处理新增/删除/修改的任务）。
签到逻辑直接操作数据库和微博 API，不依赖其他服务。
"""

import os
import sys
import re
import asyncio
import logging
import signal
import time as _time
from datetime import datetime, timedelta

from apscheduler.schedulers.background import BackgroundScheduler
from apscheduler.triggers.cron import CronTrigger
from apscheduler.triggers.interval import IntervalTrigger
from apscheduler.jobstores.memory import MemoryJobStore

# 确保 shared 模块可导入
sys.path.insert(0, os.path.join(os.path.dirname(__file__), "../.."))

from shared.config import shared_settings

logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
    datefmt="%Y-%m-%d %H:%M:%S",
)
logger = logging.getLogger("scheduler")

scheduler = BackgroundScheduler(
    jobstores={"default": MemoryJobStore()},
    timezone="Asia/Shanghai",
)

# 已注册的任务 ID 集合，用于同步时对比
_registered_tasks: dict = {}  # task_id -> cron_expression

# 日志保留天数
SIGNIN_LOG_RETAIN_DAYS = int(os.getenv("SIGNIN_LOG_RETAIN_DAYS", "30"))
CLEANUP_BATCH_SIZE = 1000

# Redis 订阅线程是否运行
_redis_listener_running = False

WEIBO_HEADERS = {
    "User-Agent": (
        "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
        "(KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
    ),
    "Referer": "https://weibo.com/",
    "Accept": "*/*",
    "Accept-Language": "zh-CN,zh;q=0.9,en;q=0.8",
}


def _make_session():
    """每次创建独立的 engine + session，避免 event loop 冲突。"""
    from sqlalchemy.ext.asyncio import AsyncSession, create_async_engine
    from sqlalchemy.orm import sessionmaker

    kwargs = {"echo": False}
    if "sqlite" not in shared_settings.DATABASE_URL:
        kwargs.update(pool_size=5, max_overflow=10, pool_pre_ping=True)
    eng = create_async_engine(shared_settings.DATABASE_URL, **kwargs)
    factory = sessionmaker(eng, class_=AsyncSession, expire_on_commit=False)
    return factory, eng


def _run_async(coro):
    """在新 event loop 中运行 async 函数。"""
    loop = asyncio.new_event_loop()
    asyncio.set_event_loop(loop)
    try:
        return loop.run_until_complete(coro)
    finally:
        loop.close()


def _parse_cookies(cookie_str: str) -> dict:
    cookies = {}
    for pair in cookie_str.split(";"):
        pair = pair.strip()
        if "=" in pair:
            k, v = pair.split("=", 1)
            cookies[k.strip()] = v.strip()
    return cookies


# =============== 任务同步 ===============

def sync_db_tasks():
    """从 DB 同步任务到 APScheduler。新增的加上，删除的移除，cron 变了的更新。"""
    try:
        tasks = _run_async(_load_tasks_from_db())
    except Exception as e:
        logger.error(f"同步任务失败: {e}")
        return

    db_task_ids = set()
    for task_id, account_id, cron_expr in tasks:
        db_task_ids.add(task_id)
        job_id = f"signin_{task_id}"

        # 已存在且 cron 没变 → 跳过
        if task_id in _registered_tasks and _registered_tasks[task_id] == cron_expr:
            continue

        # 新增或 cron 变了 → 添加/替换
        try:
            parts = cron_expr.split()
            if len(parts) != 5:
                logger.warning(f"无效 cron: task={task_id}, expr={cron_expr}")
                continue

            trigger = CronTrigger(
                minute=parts[0], hour=parts[1],
                day=parts[2], month=parts[3], day_of_week=parts[4],
                timezone="Asia/Shanghai",
            )
            scheduler.add_job(
                run_signin,
                trigger=trigger,
                id=job_id,
                args=[task_id, account_id],
                replace_existing=True,
                misfire_grace_time=300,  # 5 分钟内的 misfire 仍然执行
            )
            _registered_tasks[task_id] = cron_expr
            logger.info(f"✅ 注册任务: task={task_id}, account={account_id}, cron={cron_expr}")
        except Exception as e:
            logger.error(f"注册任务失败: task={task_id}, error={e}")

    # 移除 DB 中已删除/禁用的任务
    removed = set(_registered_tasks.keys()) - db_task_ids
    for task_id in removed:
        job_id = f"signin_{task_id}"
        try:
            scheduler.remove_job(job_id)
            logger.info(f"🗑️ 移除任务: task={task_id}")
        except Exception:
            pass
        _registered_tasks.pop(task_id, None)

    logger.info(f"📋 任务同步完成: 活跃 {len(db_task_ids)} 个, 已注册 {len(_registered_tasks)} 个")


async def _load_tasks_from_db():
    """从 DB 读取所有 enabled 且账号 active/pending 的任务。"""
    from sqlalchemy import select
    from shared.models.task import Task
    from shared.models.account import Account

    SessionFactory, eng = _make_session()
    try:
        async with SessionFactory() as session:
            stmt = (
                select(Task.id, Task.account_id, Task.cron_expression)
                .join(Account, Task.account_id == Account.id)
                .where(Task.is_enabled == True)
                .where(Account.status.in_(["pending", "active"]))
            )
            result = await session.execute(stmt)
            return [(str(r[0]), str(r[1]), r[2]) for r in result.all()]
    finally:
        await eng.dispose()


# =============== 签到入口 ===============

def run_signin(task_id: str, account_id: str):
    """APScheduler 调用的签到入口（同步函数，内部跑 async）。"""
    logger.info(f"🎯 开始签到: task={task_id}, account={account_id}")
    try:
        result = _run_async(_async_do_signin(account_id))
        logger.info(f"✅ 签到完成: task={task_id}, result={result}")
    except Exception as e:
        logger.error(f"❌ 签到失败: task={task_id}, error={e}")


async def _async_do_signin(account_id: str):
    """执行单个账号的全量超话签到。"""
    import httpx
    from sqlalchemy import select
    from shared.models.account import Account
    from shared.models.signin_log import SigninLog
    from shared.crypto import decrypt_cookie, derive_key

    SessionFactory, eng = _make_session()
    try:
        async with SessionFactory() as session:
            result = await session.execute(select(Account).where(Account.id == account_id))
            account = result.scalar_one_or_none()
            if not account:
                return {"status": "error", "reason": "account not found"}
            if account.status not in ("pending", "active"):
                return {"status": "skipped", "reason": f"status={account.status}"}

            key = derive_key(shared_settings.COOKIE_ENCRYPTION_KEY)
            try:
                cookie_str = decrypt_cookie(account.encrypted_cookies, account.iv, key)
            except Exception:
                account.status = "invalid_cookie"
                await session.commit()
                return {"status": "failed", "reason": "cookie decryption failed"}

            cookies = _parse_cookies(cookie_str)

            # 获取超话列表
            topics = await _fetch_topics(cookies)
            if not topics:
                return {"status": "completed", "signed": 0, "message": "no topics"}

            # 逐个签到
            signed = already = failed = 0
            for topic in topics:
                await asyncio.sleep(1.5)
                r = await _do_single_signin(cookies, topic)
                if r["status"] == "success":
                    s, signed = "success", signed + 1
                elif r["status"] == "already_signed":
                    s, already = "failed_already_signed", already + 1
                else:
                    s, failed = "failed_network", failed + 1

                session.add(SigninLog(
                    account_id=account.id, topic_title=topic["title"],
                    status=s, reward_info={"message": r["message"]},
                    signed_at=datetime.now(),
                ))

            account.last_checked_at = datetime.now()
            if account.status != "active":
                account.status = "active"
            await session.commit()

            logger.info(
                f"签到结果: account={account_id}, "
                f"signed={signed}, already={already}, failed={failed}"
            )
            return {"signed": signed, "already_signed": already, "failed": failed, "total": len(topics)}
    finally:
        await eng.dispose()


async def _fetch_topics(cookies: dict) -> list:
    """获取关注的超话列表。"""
    import httpx

    topics = []
    async with httpx.AsyncClient(timeout=15, follow_redirects=True) as client:
        await client.get("https://weibo.com/", headers=WEIBO_HEADERS, cookies=cookies)
        xsrf = client.cookies.get("XSRF-TOKEN", "")
        headers = {**WEIBO_HEADERS, "X-Requested-With": "XMLHttpRequest"}
        if xsrf:
            headers["X-XSRF-TOKEN"] = xsrf

        page = 1
        while page <= 10:
            resp = await client.get(
                "https://weibo.com/ajax/profile/topicContent",
                params={"tabid": "231093_-_chaohua", "page": str(page)},
                headers=headers, cookies=cookies,
            )
            data = resp.json()
            if data.get("ok") != 1:
                break
            tlist = data.get("data", {}).get("list", [])
            if not tlist:
                break
            for item in tlist:
                title = item.get("topic_name", "") or item.get("title", "")
                cid = ""
                for field in ("oid", "scheme"):
                    m = re.search(r"100808[0-9a-fA-F]+", item.get(field, ""))
                    if m:
                        cid = m.group(0)
                        break
                if title and cid:
                    topics.append({"title": title, "containerid": cid})
            if page >= data.get("data", {}).get("max_page", 1):
                break
            page += 1
    return topics


async def _do_single_signin(cookies: dict, topic: dict) -> dict:
    """签到单个超话。"""
    import httpx

    try:
        async with httpx.AsyncClient(timeout=15, follow_redirects=True) as client:
            await client.get("https://weibo.com/", headers=WEIBO_HEADERS, cookies=cookies)
            xsrf = client.cookies.get("XSRF-TOKEN", "")
            h = {
                **WEIBO_HEADERS,
                "Referer": f"https://weibo.com/p/{topic['containerid']}/super_index",
                "X-Requested-With": "XMLHttpRequest",
            }
            if xsrf:
                h["X-XSRF-TOKEN"] = xsrf

            resp = await client.get(
                "https://weibo.com/p/aj/general/button",
                params={
                    "ajwvr": "6",
                    "api": "http://i.huati.weibo.com/aj/super/checkin",
                    "texta": "签到", "textb": "已签到",
                    "status": "0",
                    "id": topic["containerid"],
                    "location": "page_100808_super_index",
                    "timezone": "GMT+0800",
                    "lang": "zh-cn",
                    "plat": "Win32",
                    "ua": WEIBO_HEADERS["User-Agent"],
                    "screen": "1920*1080",
                    "__rnd": str(int(_time.time() * 1000)),
                },
                headers=h, cookies=cookies,
            )
            data = resp.json()
            code = str(data.get("code", ""))
            msg = data.get("msg", "")

            if code == "100000":
                tip = ""
                if isinstance(data.get("data"), dict):
                    tip = data["data"].get("alert_title", "") or data["data"].get("tipMessage", "")
                return {"status": "success", "message": tip or "签到成功"}
            elif code == "382004":
                return {"status": "already_signed", "message": msg or f"今天已签到({code})"}
            else:
                return {"status": "failed", "message": f"code={code}, msg={msg}"}
    except Exception as e:
        return {"status": "failed", "message": str(e)}


# =============== 日志清理 ===============

def cleanup_old_signin_logs():
    """清理超过 N 天的签到日志，分批删除避免锁表。"""
    logger.info(f"🧹 开始清理 {SIGNIN_LOG_RETAIN_DAYS} 天前的签到日志...")
    try:
        result = _run_async(_async_cleanup())
        logger.info(f"✅ 日志清理完成: {result}")
    except Exception as e:
        logger.error(f"❌ 日志清理失败: {e}")


async def _async_cleanup():
    from sqlalchemy import select, delete
    from shared.models.signin_log import SigninLog

    cutoff = datetime.now() - timedelta(days=SIGNIN_LOG_RETAIN_DAYS)
    total_deleted = 0

    SessionFactory, eng = _make_session()
    try:
        async with SessionFactory() as session:
            while True:
                stmt = (
                    select(SigninLog.id)
                    .where(SigninLog.signed_at < cutoff)
                    .limit(CLEANUP_BATCH_SIZE)
                )
                result = await session.execute(stmt)
                ids = [row[0] for row in result.all()]

                if not ids:
                    break

                del_stmt = delete(SigninLog).where(SigninLog.id.in_(ids))
                await session.execute(del_stmt)
                await session.commit()
                total_deleted += len(ids)
                logger.info(f"🧹 已删除 {total_deleted} 条...")

                if len(ids) < CLEANUP_BATCH_SIZE:
                    break
    finally:
        await eng.dispose()

    return {"deleted": total_deleted, "cutoff": cutoff.isoformat()}


# =============== Redis 订阅（实时同步任务变更） ===============

def _start_redis_listener():
    """启动 Redis pub/sub 监听线程，收到 task_updates 消息后立即同步。"""
    global _redis_listener_running
    if not shared_settings.USE_REDIS or not shared_settings.REDIS_URL:
        logger.info("⚠️ Redis 未启用，任务变更仅靠 5 分钟轮询同步")
        return

    import threading
    import redis

    def _listen():
        global _redis_listener_running
        _redis_listener_running = True
        logger.info("📡 Redis 订阅线程启动，监听 task_updates 频道")
        while _redis_listener_running:
            try:
                r = redis.from_url(shared_settings.REDIS_URL, decode_responses=True)
                pubsub = r.pubsub()
                pubsub.subscribe("task_updates")
                for message in pubsub.listen():
                    if not _redis_listener_running:
                        break
                    if message["type"] == "message":
                        logger.info(f"📡 收到任务变更通知: {message['data'][:200]}")
                        sync_db_tasks()
                pubsub.close()
                r.close()
            except Exception as e:
                logger.warning(f"Redis 订阅异常，5 秒后重连: {e}")
                _time.sleep(5)

    t = threading.Thread(target=_listen, daemon=True, name="redis-listener")
    t.start()


# =============== 启动入口 ===============

def _shutdown(signum, frame):
    global _redis_listener_running
    logger.info(f"收到信号 {signum}，正在关闭调度器...")
    _redis_listener_running = False
    scheduler.shutdown(wait=False)
    sys.exit(0)


if __name__ == "__main__":
    logger.info("=" * 60)
    logger.info("🚀 Weibo-HotSign 定时调度器启动 (APScheduler)")
    logger.info("=" * 60)

    # 注册信号处理
    signal.signal(signal.SIGTERM, _shutdown)
    signal.signal(signal.SIGINT, _shutdown)

    # 启动调度器
    scheduler.start()

    # 首次同步 DB 任务
    sync_db_tasks()

    # 启动 Redis 订阅，实时接收任务变更通知
    _start_redis_listener()

    # 每 50 分钟重新同步 DB（兜底，正常靠 Redis 实时通知）
    scheduler.add_job(
        sync_db_tasks,
        trigger=IntervalTrigger(minutes=50),
        id="sync_db_tasks",
        replace_existing=True,
        misfire_grace_time=60,
    )

    # 每天凌晨 3 点清理过期签到日志
    scheduler.add_job(
        cleanup_old_signin_logs,
        trigger=CronTrigger(hour=3, minute=0, timezone="Asia/Shanghai"),
        id="cleanup_old_signin_logs",
        replace_existing=True,
        misfire_grace_time=3600,
    )

    logger.info("📋 调度器已启动，等待任务触发...")
    logger.info(f"📋 日志清理: 每天 03:00, 保留 {SIGNIN_LOG_RETAIN_DAYS} 天")

    # 打印当前所有已注册的 job
    jobs = scheduler.get_jobs()
    for job in jobs:
        logger.info(f"  📌 {job.id}: next_run={job.next_run_time}")

    # 主线程保持运行
    try:
        while True:
            _time.sleep(1)
    except (KeyboardInterrupt, SystemExit):
        logger.info("调度器关闭中...")
        scheduler.shutdown(wait=False)