LongArticleTaskServer/applications/tasks/task_scheduler_v2.py

import asyncio
import json
import time
import traceback
from datetime import datetime
from typing import Awaitable, Callable, Dict

from applications.api import feishu_robot
from applications.utils import task_schedule_response, generate_task_trace_id

from applications.tasks.cold_start_tasks import ArticlePoolColdStart
from applications.tasks.crawler_tasks import CrawlerToutiao
from applications.tasks.data_recycle_tasks import RecycleDailyPublishArticlesTask
from applications.tasks.data_recycle_tasks import CheckDailyPublishArticlesTask
from applications.tasks.data_recycle_tasks import UpdateRootSourceIdAndUpdateTimeTask
from applications.tasks.llm_tasks import TitleRewrite
from applications.tasks.llm_tasks import CandidateAccountQualityScoreRecognizer
from applications.tasks.monitor_tasks import check_kimi_balance
from applications.tasks.monitor_tasks import GetOffVideos
from applications.tasks.monitor_tasks import CheckVideoAuditStatus
from applications.tasks.monitor_tasks import InnerGzhArticlesMonitor
from applications.tasks.monitor_tasks import OutsideGzhArticlesMonitor
from applications.tasks.monitor_tasks import OutsideGzhArticlesCollector
from applications.tasks.monitor_tasks import TaskProcessingMonitor
from applications.tasks.task_mapper import TaskMapper


class TaskScheduler(TaskMapper):
    """统一调度入口：外部只需调用 `await TaskScheduler(data, log_cli, db_cli).deal()`"""

    # ---------- 初始化 ----------
    def __init__(self, data, log_service, db_client):
        self.data = data
        self.log_client = log_service
        self.db_client = db_client
        self.table = "long_articles_task_manager"
        self.trace_id = generate_task_trace_id()

    # ---------- 公共数据库工具 ----------
    async def _insert_or_ignore_task(self, task_name: str, date_str: str) -> None:
        """新建记录（若同键已存在则忽略）"""
        query = (
            f"insert ignore into {self.table} "
            "(date_string, task_name, start_timestamp, task_status, trace_id, data) "
            "values (%s, %s, %s, %s, %s, %s);"
        )
        await self.db_client.async_save(
            query=query,
            params=(
                date_str,
                task_name,
                int(time.time()),
                self.TASK_INIT_STATUS,
                self.trace_id,
                json.dumps(self.data, ensure_ascii=False),
            ),
        )

    async def _try_lock_task(self) -> bool:
        """一次 UPDATE 抢锁；返回 True 表示成功上锁"""
        query = (
            f"update {self.table} "
            "set task_status = %s "
            "where trace_id = %s  and task_status = %s;"
        )
        res = await self.db_client.async_save(
            query=query,
            params=(
                self.TASK_PROCESSING_STATUS,
                self.trace_id,
                self.TASK_INIT_STATUS,
            ),
        )
        return True if res else False

    async def _release_task(self, status: int) -> None:
        query = (
            f"update {self.table} set task_status=%s, finish_timestamp=%s "
            "where trace_id=%s and task_status=%s;"
        )
        await self.db_client.async_save(
            query=query,
            params=(
                status,
                int(time.time()),
                self.trace_id,
                self.TASK_PROCESSING_STATUS,
            ),
        )

    async def _is_processing_overtime(self, task_name) -> bool:
        """检测在处理任务是否超时，或者超过最大并行数，若超时会发飞书告警"""
        query = f"select trace_id from {self.table} where task_status = %s and task_name = %s;"
        rows = await self.db_client.async_fetch(
            query=query, params=(self.TASK_PROCESSING_STATUS, task_name)
        )
        if not rows:
            return False

        processing_task_num = len(rows)
        if processing_task_num >= self.get_task_config(task_name).get(
            "task_max_num", self.TASK_MAX_NUM
        ):
            await feishu_robot.bot(
                title=f"multi {task_name} is processing ",
                detail={"detail": rows},
            )
            return True

        return False

    async def _run_with_guard(
        self, task_name: str, date_str: str, task_coro: Callable[[], Awaitable[int]]
    ):
        """公共：检查、建记录、抢锁、后台运行"""
        # 1. 超时检测
        if await self._is_processing_overtime(task_name):
            return await task_schedule_response.fail_response(
                "5005", "muti tasks with same task_name is processing"
            )

        # 2. 记录并尝试抢锁
        await self._insert_or_ignore_task(task_name, date_str)
        if not await self._try_lock_task():
            return await task_schedule_response.fail_response(
                "5001", "task is processing"
            )

        # 3. 真正执行任务 —— 使用后台协程保证不阻塞调度入口
        async def _wrapper():
            status = self.TASK_FAILED_STATUS
            try:
                status = await task_coro()
            except Exception as e:
                await self.log_client.log(
                    contents={
                        "trace_id": self.trace_id,
                        "function": "cor_wrapper",
                        "task": task_name,
                        "error": str(e),
                    }
                )
                await feishu_robot.bot(
                    title=f"{task_name} is failed",
                    detail={
                        "task": task_name,
                        "err": str(e),
                        "traceback": traceback.format_exc(),
                    },
                )
            finally:
                await self._release_task(status)

        asyncio.create_task(_wrapper(), name=task_name)
        return await task_schedule_response.success_response(
            task_name=task_name,
            data={"code": 0, "message": "task started", "trace_id": self.trace_id},
        )

    # ---------- 主入口 ----------
    async def deal(self):
        task_name: str | None = self.data.get("task_name")
        if not task_name:
            return await task_schedule_response.fail_response(
                "4003", "task_name must be input"
            )

        date_str = self.data.get("date_string") or datetime.now().strftime("%Y-%m-%d")

        # === 所有任务在此注册：映射到一个返回 int 状态码的异步函数 ===
        handlers: Dict[str, Callable[[], Awaitable[int]]] = {
            # 校验kimi余额
            "check_kimi_balance": self._check_kimi_balance_handler,
            # 长文视频发布之后，三天后下架
            "get_off_videos": self._get_off_videos_task_handler,
            # 长文视频发布之后，三天内保持视频可见状态
            "check_publish_video_audit_status": self._check_video_audit_status_handler,
            # 外部服务号发文监测
            "outside_article_monitor": self._outside_monitor_handler,
            # 站内发文监测
            "inner_article_monitor": self._inner_gzh_articles_monitor_handler,
            # 标题重写（代测试）
            "title_rewrite": self._title_rewrite_handler,
            # 每日发文数据回收
            "daily_publish_articles_recycle": self._recycle_article_data_handler,
            # 每日发文更新root_source_id
            "update_root_source_id": self._update_root_source_id_handler,
            # 头条文章，视频抓取
            "crawler_toutiao": self._crawler_toutiao_handler,
            # 文章池冷启动发布
            "article_pool_cold_start": self._article_pool_cold_start_handler,
            # 任务超时监控
            "task_processing_monitor": self._task_processing_monitor_handler,
            # 候选账号质量分析
            "candidate_account_quality_analysis": self._candidate_account_quality_score_handler,
        }

        if task_name not in handlers:
            return await task_schedule_response.fail_response(
                "4001", "wrong task name input"
            )
        return await self._run_with_guard(task_name, date_str, handlers[task_name])

    # ---------- 下面是若干复合任务的局部实现 ----------
    async def _check_kimi_balance_handler(self) -> int:
        response = await check_kimi_balance()
        await self.log_client.log(
            contents={
                "trace_id": self.trace_id,
                "task": "check_kimi_balance",
                "data": response,
            }
        )
        return self.TASK_SUCCESS_STATUS

    async def _get_off_videos_task_handler(self) -> int:
        sub_task = GetOffVideos(self.db_client, self.log_client, self.trace_id)
        return await sub_task.deal()

    async def _check_video_audit_status_handler(self) -> int:
        sub_task = CheckVideoAuditStatus(self.db_client, self.log_client, self.trace_id)
        return await sub_task.deal()

    async def _task_processing_monitor_handler(self) -> int:
        sub_task = TaskProcessingMonitor(self.db_client)
        await sub_task.deal()
        return self.TASK_SUCCESS_STATUS

    async def _inner_gzh_articles_monitor_handler(self) -> int:
        sub_task = InnerGzhArticlesMonitor(self.db_client)
        return await sub_task.deal()

    async def _title_rewrite_handler(self):
        sub_task = TitleRewrite(self.db_client, self.log_client)
        return await sub_task.deal()

    async def _update_root_source_id_handler(self) -> int:
        sub_task = UpdateRootSourceIdAndUpdateTimeTask(self.db_client, self.log_client)
        await sub_task.deal()
        return self.TASK_SUCCESS_STATUS

    async def _outside_monitor_handler(self) -> int:
        collector = OutsideGzhArticlesCollector(self.db_client)
        await collector.deal()
        monitor = OutsideGzhArticlesMonitor(self.db_client)
        return await monitor.deal()  # 应返回 SUCCESS / FAILED

    async def _recycle_article_data_handler(self) -> int:
        date_str = self.data.get("date_string") or datetime.now().strftime("%Y-%m-%d")
        recycle = RecycleDailyPublishArticlesTask(
            self.db_client, self.log_client, date_str
        )
        await recycle.deal()
        check = CheckDailyPublishArticlesTask(self.db_client, self.log_client, date_str)
        await check.deal()
        return self.TASK_SUCCESS_STATUS

    async def _crawler_toutiao_handler(self) -> int:
        sub_task = CrawlerToutiao(self.db_client, self.log_client, self.trace_id)
        method = self.data.get("method", "account")
        media_type = self.data.get("media_type", "article")
        category_list = self.data.get("category_list", [])

        match method:
            case "account":
                await sub_task.crawler_task(media_type=media_type)
            case "recommend":
                await sub_task.crawl_toutiao_recommend_task(category_list)
            case "search":
                await sub_task.search_candidate_accounts()
            case _:
                raise ValueError(f"Unsupported method {method}")
        return self.TASK_SUCCESS_STATUS

    async def _article_pool_cold_start_handler(self) -> int:
        cold_start = ArticlePoolColdStart(
            self.db_client, self.log_client, self.trace_id
        )
        platform = self.data.get("platform", "weixin")
        crawler_methods = self.data.get("crawler_methods", [])
        await cold_start.deal(platform=platform, crawl_methods=crawler_methods)
        return self.TASK_SUCCESS_STATUS

    async def _candidate_account_quality_score_handler(self) -> int:
        task = CandidateAccountQualityScoreRecognizer(
            self.db_client, self.log_client, self.trace_id
        )
        await task.deal()
        return self.TASK_SUCCESS_STATUS