chore: 重构前基线 — 9 个 sync task 全部 ok + akshare 移除 + mairui 资金流接入

状态：
- 9 个 sync task（stock_basic / kline_daily / kline_index / kline_5min /
  moneyflow / industry_sector / sector_features / share_snapshot / market_regime）
- 数据源：baostock + mairui + 雪球(pysnowball) + 新浪（4 个）
- 项目级约束：永远不用 akshare（已落实）
- kline_5min 改用 DB 快照统一全量/增量逻辑
- 零后端 Chrome 扩展 xueqiu_sync（独立项目）

app/core/scheduler/__init__.py

@@ -0,0 +1,20 @@
+"""scheduler 子包入口。"""
+from app.core.scheduler.scheduler import (
+    start_scheduler,
+    stop_scheduler,
+    get_scheduler_status,
+    seed_schedule_configs,
+    register_job,
+    is_trading_day,
+    next_trading_day,
+)
+
+__all__ = [
+    "start_scheduler",
+    "stop_scheduler",
+    "get_scheduler_status",
+    "seed_schedule_configs",
+    "register_job",
+    "is_trading_day",
+    "next_trading_day",
+]

app/core/scheduler/scheduler.py

@@ -0,0 +1,392 @@
+"""轻量级定时任务调度器（照搬 dashboard/api/services/scheduler.py 模式）。
+
+后台线程每 N 秒扫描 config 表中 category='schedule' 的任务，到达触发时间时
+执行对应的 job 函数。执行记录写回 config 表。
+
+任务定义字段（存在 config 表，value 是 JSON）：
+  {
+    "name": "...",
+    "time": "15:40",                  # 触发时间（HH:MM），intervalSeconds=0 时必填
+    "intervalSeconds": 0,             # > 0 时按秒触发
+    "condition": "trading_day",       # always | trading_day
+    "job": "sync_kline_daily",        # 已注册的 job 名
+    "enabled": true,
+    "lastRun": "", "lastStatus": "", "lastMessage": ""
+  }
+"""
+from __future__ import annotations
+
+import json
+import threading
+import time
+from datetime import date, datetime, timedelta
+from typing import Any, Callable, Optional
+
+from app.core.config import settings
+from app.core.db import ops as db_ops
+from app.core.utils.logging import get_logger
+
+logger = get_logger("scheduler")
+
+
+# ── 交易日判断 ─────────────────────────────────────────────────────────
+
+_KNOWN_HOLIDAYS: set[str] = set()
+
+
+def _init_holidays() -> None:
+    """从 config 表读出休市日（key=trading_calendar_holidays, category=general）。"""
+    global _KNOWN_HOLIDAYS
+    r = db_ops.fetch_config_by_key("trading_calendar_holidays")
+    if r and r.get("category") == "general":
+        try:
+            holidays = json.loads(r.get("value", "") or "[]")
+            if isinstance(holidays, list):
+                _KNOWN_HOLIDAYS = {h for h in holidays if isinstance(h, str)}
+        except (json.JSONDecodeError, TypeError):
+            pass
+
+
+def is_trading_day(d: Optional[date] = None) -> bool:
+    """判断 A 股交易日（简化：周一至周五 + 非 _KNOWN_HOLIDAYS）。"""
+    if d is None:
+        d = date.today()
+    if d.weekday() >= 5:
+        return False
+    if d.isoformat() in _KNOWN_HOLIDAYS:
+        return False
+    return True
+
+
+def next_trading_day(after: Optional[date] = None) -> date:
+    d = after or date.today()
+    while not is_trading_day(d):
+        d += timedelta(days=1)
+    return d
+
+
+# ── Job registry ──────────────────────────────────────────────────────
+
+_job_registry: dict[str, Callable[[], dict]] = {}
+
+
+def register_job(name: str):
+    """装饰器：注册一个 job 函数。函数应返回 {status, message, ...}。"""
+    def decorator(fn):
+        _job_registry[name] = fn
+        return fn
+    return decorator
+
+
+# ── Scheduler 状态 ────────────────────────────────────────────────────
+
+_lock = threading.Lock()
+_running = False
+_thread: Optional[threading.Thread] = None
+
+# 配置缓存：避免每次 tick 全量查 config 表
+_config_cache_ts: float = 0.0
+_config_cache: list[dict] = []
+_CONFIG_CACHE_TTL = 10.0  # 缓存 10 秒，减少 DB 查询
+
+
+def get_scheduler_status() -> dict[str, Any]:
+    with _lock:
+        return {
+            "running": _running,
+            "jobs": _load_job_statuses(),
+        }
+
+
+def _load_schedule_configs() -> list[dict]:
+    """从 config 表读出所有 category='schedule' 的任务定义（带缓存）。"""
+    global _config_cache_ts, _config_cache
+    now = time.time()
+    if now - _config_cache_ts < _CONFIG_CACHE_TTL and _config_cache:
+        return _config_cache
+    result = []
+    for r in db_ops.fetch_configs_by_category("schedule"):
+        try:
+            obj = json.loads(r.get("value", "") or "{}")
+            obj["_key"] = r["key"]
+            result.append(obj)
+        except (json.JSONDecodeError, TypeError):
+            continue
+    _config_cache = result
+    _config_cache_ts = now
+    return result
+
+
+def _load_job_statuses() -> list[dict]:
+    cfgs = _load_schedule_configs()
+    return [
+        {
+            "key": c["_key"],
+            "name": c.get("name", ""),
+            "time": c.get("time", ""),
+            "intervalSeconds": c.get("intervalSeconds", 0),
+            "condition": c.get("condition", "always"),
+            "job": c.get("job", ""),
+            "enabled": c.get("enabled", True),
+            "lastRun": c.get("lastRun", ""),
+            "lastStatus": c.get("lastStatus", ""),
+            "lastMessage": c.get("lastMessage", ""),
+        }
+        for c in cfgs
+    ]
+
+
+def update_job_status(key: str, status: str, message: str) -> None:
+    """更新 job 执行状态 — 使用 fetch_config_by_key 高效查找。"""
+    r = db_ops.fetch_config_by_key(key)
+    if r is None or r.get("category") != "schedule":
+        return
+    try:
+        obj = json.loads(r.get("value", "") or "{}")
+    except (json.JSONDecodeError, TypeError):
+        obj = {}
+    obj["lastRun"] = datetime.now().isoformat(timespec="seconds")
+    obj["lastStatus"] = status
+    obj["lastMessage"] = message
+    db_ops.upsert_config(
+        key, json.dumps(obj, ensure_ascii=False),
+        category="schedule", description=r.get("description", ""),
+    )
+    # 更新后立即刷新缓存
+    global _config_cache_ts
+    _config_cache_ts = 0.0
+
+
+# ── 调度循环 ──────────────────────────────────────────────────────────
+
+def _loop() -> None:
+    global _running
+    _init_holidays()
+
+    last_fired: dict[str, str] = {}      # key → date string
+    last_fired_ts: dict[str, float] = {} # key → timestamp
+
+    while _running:
+        try:
+            now = datetime.now()
+            today_str = now.strftime("%Y-%m-%d")
+            current_time = now.strftime("%H:%M")
+            now_ts = time.time()
+
+            for cfg in _load_schedule_configs():
+                if not cfg.get("enabled", True):
+                    continue
+
+                key = cfg["_key"]
+                job_time = cfg.get("time", "")
+                condition = cfg.get("condition", "always")
+                job_name = cfg.get("job", "")
+                interval = int(cfg.get("intervalSeconds", 0) or 0)
+                persisted_last_run = str(cfg.get("lastRun", "") or "").strip()
+
+                # 启动时记住今天已跑过的时间触发任务
+                if not interval and persisted_last_run and key not in last_fired:
+                    try:
+                        persisted_date = datetime.fromisoformat(persisted_last_run).strftime("%Y-%m-%d")
+                    except ValueError:
+                        persisted_date = persisted_last_run[:10]
+                    if persisted_date == today_str:
+                        last_fired[key] = today_str
+
+                if interval > 0:
+                    if now_ts - last_fired_ts.get(key, 0) < interval:
+                        continue
+                else:
+                    if current_time < job_time:
+                        continue
+                    if last_fired.get(key) == today_str:
+                        continue
+
+                if condition == "trading_day" and not is_trading_day():
+                    continue
+
+                # 触发
+                fn = _job_registry.get(job_name)
+                if fn is None:
+                    update_job_status(key, "error", f"未注册的 job: {job_name}")
+                    continue
+
+                logger.info(f"[scheduler] 触发 {cfg.get('name', key)} @ {current_time}")
+                try:
+                    result = fn() or {}
+                    status = result.get("status", "ok")
+                    message = result.get("message", "")
+                    update_job_status(key, status, message)
+                    logger.info(f"[scheduler] 完成 {cfg.get('name', key)} → {status}: {message}")
+                except Exception as e:
+                    err = f"{type(e).__name__}: {e}"
+                    update_job_status(key, "error", err[:200])
+                    logger.error(f"[scheduler] 失败 {cfg.get('name', key)}: {err}")
+
+                last_fired[key] = today_str
+                last_fired_ts[key] = now_ts
+        except Exception as e:
+            logger.error(f"[scheduler] loop 异常: {e}")
+
+        time.sleep(max(5, settings.scheduler_tick_seconds))
+
+
+def start_scheduler() -> None:
+    global _running, _thread
+    with _lock:
+        if _running:
+            return
+        _running = True
+        _thread = threading.Thread(target=_loop, daemon=True, name="market-data-sync-scheduler")
+        _thread.start()
+        logger.info("[scheduler] 已启动")
+
+
+def stop_scheduler() -> None:
+    global _running
+    with _lock:
+        _running = False
+        logger.info("[scheduler] 已停止")
+
+
+# ── 默认任务 seed ────────────────────────────────────────────────────
+
+DEFAULT_SCHEDULES: list[tuple[str, dict, str]] = [
+    (
+        "schedule_stocks_basic",
+        {
+            "name": "开盘前同步股票信息",
+            "time": "09:00",
+            "condition": "trading_day",
+            "job": "stock_basic",
+            "enabled": True,
+        },
+        "每个交易日 09:00 拉取全市场股票基础信息 + 股本快照",
+    ),
+    (
+        "schedule_kline_index",
+        {
+            "name": "盘后指数日K线",
+            "time": "15:30",
+            "condition": "trading_day",
+            "job": "kline_index",
+            "enabled": True,
+        },
+        "每个交易日 15:30 拉取六大指数日K线",
+    ),
+    (
+        "schedule_kline_daily",
+        {
+            "name": "盘后日K线增量",
+            "time": "15:40",
+            "condition": "trading_day",
+            "job": "kline_daily",
+            "enabled": True,
+        },
+        "每个交易日 15:40 增量同步全市场日K线（多源交叉校验）",
+    ),
+    (
+        "schedule_kline_5min",
+        {
+            "name": "盘后5分钟K线",
+            "time": "16:00",
+            "condition": "trading_day",
+            "job": "kline_5min",
+            "enabled": True,
+        },
+        "每个交易日 16:00 拉取 5 分钟 K 线（数据量大，可选启用）",
+    ),
+    (
+        "schedule_moneyflow",
+        {
+            "name": "盘后资金流",
+            "time": "16:30",
+            "condition": "trading_day",
+            "job": "moneyflow",
+            "enabled": True,
+        },
+        "每个交易日 16:30 拉取个股资金流",
+    ),
+    (
+        "schedule_industry_sector",
+        {
+            "name": "周一行业映射",
+            "time": "09:30",
+            "condition": "trading_day",
+            "job": "industry_sector",
+            "enabled": True,
+        },
+        "每个交易日 09:30 全量更新股票-行业映射（开销大，可改为 weekly）",
+    ),
+    (
+        "schedule_market_regime",
+        {
+            "name": "盘后市场情绪",
+            "time": "16:15",
+            "condition": "trading_day",
+            "job": "market_regime",
+            "enabled": True,
+        },
+        "每个交易日 16:15 聚合市场情绪（基于本地日K线）",
+    ),
+    (
+        "schedule_share_snapshot",
+        {
+            "name": "盘后股本快照",
+            "time": "16:20",
+            "condition": "trading_day",
+            "job": "share_snapshot",
+            "enabled": True,
+        },
+        "每个交易日 16:20 刷新个股总股本/流通股本快照",
+    ),
+]
+
+
+def seed_schedule_configs() -> None:
+    """写入默认计划任务配置到 config 表。"""
+    existing = {r["key"]: r for r in db_ops.fetch_configs_by_category("schedule")}
+    for key, obj, desc in DEFAULT_SCHEDULES:
+        full = {
+            **obj,
+            "lastRun": "",
+            "lastStatus": "",
+            "lastMessage": "",
+        }
+        if key in existing:
+            # 保留 lastRun / lastStatus 等
+            try:
+                old = json.loads(existing[key].get("value", "") or "{}")
+                for k in ("lastRun", "lastStatus", "lastMessage"):
+                    if k in old:
+                        full[k] = old[k]
+            except Exception:
+                pass
+        db_ops.upsert_config(
+            key, json.dumps(full, ensure_ascii=False), category="schedule", description=desc,
+        )
+
+
+# ── 注册 job（延迟导入避免循环依赖）────────────────────────────────────
+
+
+def register_sync_jobs() -> None:
+    """把 8 个 sync 任务注册成 job。
+
+    注：job 函数不接受任何参数，trigger_source 默认为 'schedule'。
+    """
+    from app.tasks import get_task
+
+    for dataset_id in [
+        "stock_basic", "kline_daily", "kline_index", "kline_5min",
+        "moneyflow", "industry_sector", "share_snapshot", "market_regime",
+    ]:
+
+        def _make_job(did=dataset_id):
+            def _job() -> dict:
+                task = get_task(did)
+                return task.run(trigger_source="schedule")
+            _job.__name__ = f"job_{did}"
+            return _job
+
+        register_job(dataset_id)(_make_job())