Initial commit: cc-cursor 全链路量化研究平台

7 Sprints 全部完成:
  Sprint 0: 基础设施 (DataManager + MariaDB)
  Sprint 1: 因子引擎 (34因子/12分类)
  Sprint 2: VectorBT 回测 (5策略+截面)
  Sprint 3: Optuna 优化 (+Walk-Forward)
  Sprint 4: ML 模型 (LightGBM+CatBoost)
  Sprint 5: Qwen 情绪因子 (三源新闻+日期对齐)
  Sprint 6: Agent 系统 (4Agent+日报.md/.html)

生产加固 (15项): Tushare双源fallback, SSH自动恢复, pool_pre_ping,
  save_daily先删后插, load_dotenv绝对路径, 日报5d/20d修复,
  RiskAgent改上证指数, 昨日对比+数据截止, mac_report utf8mb4,
  CLAUDE-*.md 9条已知Bug, demo全参数化, djapi数据源归一化,
  indexDatas API修正

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

finance/data/data_manager.py

@@ -0,0 +1,171 @@
+"""
+DataManager — 统一数据管理层。
+
+策略/模型层通过 DataManager 获取数据，不直接访问 AkShare/Tushare 或数据库。
+优先从 DB 读取，缺失时依次尝试 AkShare → Tushare 拉取并入库。
+"""
+
+import time
+
+import pandas as pd
+
+from config.settings import DEFAULT_START_DATE, DEFAULT_END_DATE
+from data.sources.akshare_source import AkShareSource
+from data.sources.tushare_source import TushareSource
+from database import dao
+from database.models import create_all_tables
+
+
+class DataManager:
+    """统一数据管理。双数据源：AkShare（主）+ Tushare（备）。"""
+
+    def __init__(self):
+        self._ak_source: AkShareSource | None = None
+        self._ts_source: TushareSource | None = None
+
+    @property
+    def ak(self) -> AkShareSource:
+        if self._ak_source is None:
+            self._ak_source = AkShareSource()
+        return self._ak_source
+
+    @property
+    def ts(self) -> TushareSource:
+        if self._ts_source is None:
+            self._ts_source = TushareSource()
+        return self._ts_source
+
+    # ── 初始化 ────────────────────────────────────────────
+
+    def init_db(self) -> None:
+        create_all_tables()
+
+    # ── 内部：双源 try ────────────────────────────────────
+
+    def _try_fetch(self, method_name: str, *args, **kwargs):
+        """
+        依次尝试 AkShare → Tushare 调用同一方法名。
+
+        method_name: 'fetch_daily' | 'fetch_stock_list' | 'fetch_financial'
+        返回: (result_df, source_name) 或 (empty_df, None)
+
+        如果 ts_code 是指数代码，自动路由到 fetch_index_daily。
+        """
+        from data.sources.akshare_source import is_index_code
+
+        # 指数自动路由
+        if method_name == "fetch_daily" and args:
+            ts_code = args[0]
+            if is_index_code(ts_code):
+                method_name = "fetch_index_daily"
+
+        for label, source in [("Tushare", self.ts), ("AkShare", self.ak)]:
+            try:
+                if label == "Tushare" and not source.available:
+                    continue
+                fn = getattr(source, method_name)
+                df = fn(*args, **kwargs)
+                if df is not None and not df.empty:
+                    return df, label
+            except Exception as e:
+                print("  [{}] {} 失败: {}".format(label, method_name, e))
+        return pd.DataFrame(), None
+
+    # ── 股票列表 ──────────────────────────────────────────
+
+    def get_stock_list(self, force_refresh: bool = False) -> pd.DataFrame:
+        if not force_refresh:
+            df = dao.query_stock_list()
+            if not df.empty:
+                print("[DataManager] 从 DB 读取股票列表: {} 只".format(len(df)))
+                return df
+
+        print("[DataManager] 拉取股票列表 (AkShare → Tushare)...")
+        df, src = self._try_fetch("fetch_stock_list")
+        if df.empty:
+            print("[DataManager] 所有数据源均无法获取股票列表")
+            return pd.DataFrame()
+        print("[DataManager] 股票列表已入库 ({}): {} 只".format(src, len(df)))
+        dao.save_stock_list(df)
+        time.sleep(2)
+        return df
+
+    # ── 日线数据 ──────────────────────────────────────────
+
+    def get_daily(
+        self,
+        ts_code: str,
+        start: str | None = None,
+        end: str | None = None,
+        force_refresh: bool = False,
+    ) -> pd.DataFrame:
+        start = start or DEFAULT_START_DATE
+        end = end or DEFAULT_END_DATE or time.strftime("%Y%m%d")
+
+        if not force_refresh:
+            df = dao.query_daily(ts_code, start, end)
+            if not df.empty:
+                return df
+
+        # DB 未命中 → 从数据源拉取
+        df, src = self._try_fetch("fetch_daily", ts_code, start, end)
+        if df.empty:
+            print("[WARN] {} 日线获取失败 (AkShare+Tushare 均不可用)".format(ts_code))
+            return pd.DataFrame()
+
+        # 保存到 DB
+        try:
+            # 筛选 DB 需要的列
+            cols = [c for c in dao._DAILY_COLS if c in df.columns]
+            dao.save_daily(df[cols])
+        except Exception as e:
+            print("[WARN] 日线入库失败: {}".format(e))
+
+        return df
+
+    def sync_daily(self, ts_code: str) -> int:
+        latest = dao.get_latest_trade_date(ts_code)
+        today = time.strftime("%Y%m%d")
+        if latest and latest >= today:
+            print("[DataManager] {} 数据已是最新 ({})".format(ts_code, latest))
+            return 0
+
+        start = latest or DEFAULT_START_DATE
+        df, src = self._try_fetch("fetch_daily", ts_code, start, today)
+        if df.empty:
+            print("[WARN] {} sync_daily 失败 (AkShare+Tushare 均不可用)".format(ts_code))
+            return 0
+
+        cols = [c for c in dao._DAILY_COLS if c in df.columns]
+        dao.save_daily(df[cols])
+        print("[DataManager] {} 同步 {} 条日线 ({})".format(ts_code, len(df), src))
+        return len(df)
+
+    def sync_all_daily(self) -> int:
+        stock_list = self.get_stock_list()
+        total = 0
+        for i, ts_code in enumerate(stock_list.index):
+            try:
+                total += self.sync_daily(ts_code)
+                if (i + 1) % 50 == 0:
+                    print("[DataManager] 进度: {}/{}".format(i + 1, len(stock_list)))
+                    time.sleep(1)
+            except Exception as e:
+                print("[WARN] {} 同步失败: {}".format(ts_code, e))
+        print("[DataManager] 全量同步完成，新增 {} 条".format(total))
+        return total
+
+    # ── 财务数据 ──────────────────────────────────────────
+
+    def get_financial(self, ts_code: str) -> pd.DataFrame:
+        df = dao.query_financial(ts_code)
+        if not df.empty:
+            return df
+        df, src = self._try_fetch("fetch_financial", ts_code)
+        if not df.empty:
+            try:
+                cols = [c for c in dao._FINA_COLS if c in df.columns]
+                dao.save_financial(df[cols])
+            except Exception as e:
+                print("[WARN] 财务数据入库失败: {}".format(e))
+        return df