market_sync/app/entrypoints/cli.py

"""CLI 入口：手动触发 / 状态查询。"""
from __future__ import annotations

import argparse
import inspect
import json
import sys
from pathlib import Path

_PROJECT_ROOT = Path(__file__).resolve().parent.parent
if str(_PROJECT_ROOT) not in sys.path:
    sys.path.insert(0, str(_PROJECT_ROOT))


def main():
    parser = argparse.ArgumentParser(
        description="market_data_sync CLI",
    )
    sub = parser.add_subparsers(dest="cmd", required=True)

    # sync <task_id>
    p_sync = sub.add_parser("sync", help="手动执行一次同步任务")
    p_sync.add_argument("task_id", help="dataset_id, 例如 kline_daily")
    p_sync.add_argument("--codes", help="指定股票代码，逗号分隔", default=None)
    p_sync.add_argument("--start", help="起始日期 YYYY-MM-DD", default=None)
    p_sync.add_argument("--end", help="结束日期 YYYY-MM-DD", default=None)
    p_sync.add_argument("--workers", type=int, default=10, help="并发数")
    p_sync.add_argument("--force", action="store_true", help="跳过时间门控（tick_trade 用）")
    p_sync.add_argument("--type2", help="stock_node 用，逗号分隔的 mairui type2 列表(如 '2,3')", default=None)
    p_sync.add_argument("--backfill", help="longhubang 用，单日 YYYYMMDD", default=None)

    sub.add_parser("list", help="列出所有同步任务")
    sub.add_parser("status", help="查看整体状态（调度器 / 数据源 / 同步任务）")
    sub.add_parser("datasources", help="查看数据源健康度")
    sub.add_parser("schedule", help="查看计划任务")
    sub.add_parser("reseed", help="重新 seed 默认计划任务到 config 表")

    args = parser.parse_args()

    if args.cmd == "sync":
        from app.core.utils.logging import setup_logging
        setup_logging()
        from app.tasks import get_task
        from app.core.datasource.registry import build_default_registry
        from app.core.sync.registry import seed_sync_registry
        build_default_registry()
        seed_sync_registry()  # 确保 dataset_registry 有这条任务定义
        try:
            task = get_task(args.task_id)
        except KeyError as e:
            print(f"错误: {e}")
            sys.exit(1)
        kwargs = {}
        if args.codes:
            kwargs["codes"] = [c.strip() for c in args.codes.split(",") if c.strip()]
        if args.start:
            kwargs["start"] = args.start
        if args.end:
            kwargs["end"] = args.end
        if args.workers:
            kwargs["max_workers"] = args.workers
        if args.force:
            kwargs["force"] = True
        if args.type2:
            # 2026-07-07 修复: 不要再 CLI 层 silent-drop 非数字 token,
            # 否则用户敲 "concept" 时 CLI 过滤成 [], 任务层 str-branch 校验
            # 永远不触发。直接把原始 str 传给 task,让 task 负责校验和报错。
            kwargs["type2"] = args.type2
        if args.backfill:
            kwargs["backfill"] = args.backfill
        # 过滤掉 task._run 不接受的 kwargs (2026-07-07 修复):
        # 之前 --type2 / --backfill 加在共享 p_sync parser 上,任何 task 都通过
        # **kwargs 静默吞掉,用户敲错也不会报错。
        try:
            declared = set(inspect.signature(task._run).parameters.keys())
        except (TypeError, ValueError):
            declared = set()
        unknown = [k for k in kwargs if k not in declared]
        if unknown:
            print(f"warning: {args.task_id} 不支持这些参数: {unknown}（已忽略）", file=sys.stderr)
        filtered = {k: v for k, v in kwargs.items() if k in declared}
        result = task.run(trigger_source="cli", **filtered)
        print(json.dumps(result, ensure_ascii=False, indent=2))
        sys.exit(0 if result.get("status") in ("ok", "warning") else 2)

    if args.cmd == "list":
        from app.tasks import TASKS
        print("可用同步任务：")
        for tid, cls in TASKS.items():
            doc = (cls.__doc__ or cls.__name__).split("\n")[0]
            print(f"  {tid:25s}  {doc}")
        return

    if args.cmd == "status":
        from app.core.utils.logging import setup_logging
        setup_logging()
        from app.core.scheduler import get_scheduler_status
        from app.core.datasource.registry import build_default_registry, get_health_status
        from app.core.sync import get_registry_status
        build_default_registry()
        print("\n=== 调度器 ===")
        s = get_scheduler_status()
        print(json.dumps(s, ensure_ascii=False, indent=2, default=str))
        print("\n=== 数据源健康度 ===")
        for r in get_health_status():
            mark = "✓" if r.get("success") else "✗"
            print(f"  {mark} {r.get('name', r.get('key')):30s}  {r.get('message', '')}")
        print("\n=== 同步任务状态 ===")
        for r in get_registry_status():
            print(f"  {r['dataset_id']:25s}  status={r.get('status'):10s}  "
                  f"lastRun={r.get('last_success_at') or r.get('last_failure_at') or '-':20s}  "
                  f"msg={r.get('message', '')[:60]}")
        return

    if args.cmd == "datasources":
        from app.core.utils.logging import setup_logging
        setup_logging()
        from app.core.datasource.registry import build_default_registry, get_health_status, run_health_check
        build_default_registry()
        print("正在跑连通性测试...")
        run_health_check()
        for r in get_health_status():
            mark = "✓" if r.get("success") else "✗"
            print(f"  {mark} {r.get('name', r.get('key')):30s}  provides={r.get('provides', [])}  {r.get('message', '')}")
        return

    if args.cmd == "schedule":
        from app.core.utils.logging import setup_logging
        setup_logging()
        from app.core.scheduler import get_scheduler_status, seed_schedule_configs
        seed_schedule_configs()
        s = get_scheduler_status()
        print(f"调度器运行中: {s['running']}")
        for j in s["jobs"]:
            enabled = "✓" if j.get("enabled") else "✗"
            trigger = j.get("time") or f"every {j.get('intervalSeconds')}s"
            print(f"  {enabled} {j.get('name', j['key']):30s}  {trigger:15s}  "
                  f"condition={j.get('condition', 'always'):12s}  "
                  f"last={j.get('lastStatus', '-')}")
        return

    if args.cmd == "reseed":
        from app.core.utils.logging import setup_logging
        setup_logging()
        from app.core.scheduler.scheduler import seed_schedule_configs
        from app.core.sync.registry import seed_sync_registry
        from app.core.datasource.registry import seed_datasource_configs
        seed_datasource_configs()
        seed_sync_registry()
        seed_schedule_configs()
        print("已重新 seed：datasources + sync_registry + schedules")
        return


if __name__ == "__main__":
    main()