提交

backend/app/data/etf_provider.py

@@ -0,0 +1,109 @@
+"""东方财富 ETF 快照（免 token 直连 HTTP，仅用于 etf_basic 列表与规模字段）。
+
+K 线数据不走这里：统一走 Tushare fund_daily（见 etf_sync / fetcher），
+东财在链路里只承担一件 Tushare quicksync 镜像做不到的事——
+全市场 ETF 名单 + 总市值/流通市值/换手率（镜像上 fund_etf_basic 不存在）。
+
+接口注意：clist 实测 pz 上限 100（传 50000 也只回 100），必须按 pn 翻页
+拿全 ~1600 只；push2 主站短连发几次会直接断连，push2delay 镜像稳，
+列表/市值用延迟值无妨（价格另有 K 线）。
+"""
+from __future__ import annotations
+
+import asyncio
+
+import httpx
+
+from .symbols import is_etf_symbol
+
+_HEADERS = {
+    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
+                  "(KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36",
+}
+
+# 东财 ETF 板块（akshare fund_etf_spot_em 同款集合：股票/混合/债券/货币/商品/跨境等）
+_SPOT_HOSTS = [
+    "https://push2delay.eastmoney.com",
+    "https://push2.eastmoney.com",
+]
+_SPOT_PARAMS = {
+    "pn": "1",
+    "pz": "100",
+    "po": "1",
+    "np": "1",
+    "ut": "bd1d9ddb04089700cf9c27f6f7426281",
+    "fltt": "2",
+    "invt": "2",
+    "fid": "f12",
+    "fs": "b:MK0021,b:MK0022,b:MK0023,b:MK0024,b:MK0026,b:MK0027,b:MK0028,b:MK0029",
+    # f12 代码 / f13 市场(1沪0深) / f14 名称 / f8 换手% / f20 总市值 / f21 流通市值（元）
+    "fields": "f12,f13,f14,f8,f20,f21",
+}
+
+
+def _num(v) -> float | None:
+    """东财 fltt=2 下停牌/缺数据的字段是 '-' 字符串。"""
+    if v is None or isinstance(v, str):
+        return None
+    f = float(v)
+    return None if f != f else f
+
+
+def new_client() -> httpx.AsyncClient:
+    """统一构造（超时/UA）；同步工厂，`async with new_client() as c` 使用。"""
+    return httpx.AsyncClient(timeout=httpx.Timeout(15.0), headers=_HEADERS)
+
+
+async def _fetch_spot_page(client: httpx.AsyncClient, host: str, pn: int):
+    params = {**_SPOT_PARAMS, "pn": str(pn)}
+    resp = await client.get(f"{host}/api/qt/clist/get", params=params)
+    resp.raise_for_status()
+    data = resp.json().get("data") or {}
+    return data.get("total") or 0, data.get("diff") or []
+
+
+async def fetch_etf_spot(client: httpx.AsyncClient) -> list[dict]:
+    """全市场场内 ETF 快照（翻页拿全 ~1600 只）-> [{ts_code, symbol, name, exchange,
+    turnover_rate, total_mv, circ_mv}]。空结果视为异常（接口改版/被拦截时宁可不覆盖表）。"""
+    diff: list[dict] = []
+    for host in _SPOT_HOSTS:
+        try:
+            total, first = await _fetch_spot_page(client, host, 1)
+            diff = first
+            pn = 2
+            while total and len(diff) < total:
+                await asyncio.sleep(0.15)  # 翻页间隔，礼貌控频
+                _, page = await _fetch_spot_page(client, host, pn)
+                if not page:
+                    break
+                diff.extend(page)
+                pn += 1
+            break  # 首个可用 host 拿完即止
+        except Exception:  # noqa: BLE001 —— 主镜像抖动换备用镜像整重来
+            diff = []
+            continue
+    rows: list[dict] = []
+    for d in diff:
+        code = str(d.get("f12") or "").strip()
+        name = str(d.get("f14") or "").strip()
+        market = d.get("f13")
+        if not code or not name or market is None:
+            continue
+        if not is_etf_symbol(code):
+            continue  # 板块返回里混进的 LOF/封基（16/50/57 开头）不进 ETF 表
+        exchange = "SH" if int(market) == 1 else "SZ"
+        rows.append({
+            "ts_code": f"{code}.{exchange}",
+            "symbol": code,
+            "name": name,
+            "exchange": exchange,
+            "turnover_rate": _num(d.get("f8")),
+            "total_mv": _num(d.get("f20")),
+            "circ_mv": _num(d.get("f21")),
+        })
+    if not rows:
+        raise RuntimeError("东财 ETF 快照为空（接口可能改版或被限流）")
+    return rows
+
+
+__all__ = ["fetch_etf_spot", "new_client"]