提交

backend/app/api.py

@@ -3,8 +3,8 @@
   GET  /api/health               健康检查
   GET  /api/candles/{sym}        取 K 线（支持 1d/1w/1M/1y 周期，日线为基底聚合）
   GET  /api/stocks               全市场股票列表（基本信息 + 最新行情 + 缓存条数）
-  GET  /api/stock/{code}/chips   个股筹码峰（cyq_chips/cyq_perf，按复权口径换算）
   GET  /api/market/overview      主页大盘总览（A 股/港美指数 + 两市市值成交统计）
+  GET  /api/market/index-candles 上证指数全量 K 线（1d/1w/1M/1y 聚合）
   POST /api/backtest             跑回测，返回 K线+指标+买卖点+净值+绩效
   POST /api/screener/run         智能选股：自然语言 -> 条件 -> 全市场筛选
   POST /api/screener/sync        启动全市场数据同步（后台任务）
@@ -18,6 +18,7 @@ import json
 from datetime import datetime
 
 import pandas as pd
+from pydantic import TypeAdapter
 from fastapi import APIRouter, Depends, File, HTTPException, Response, UploadFile
 from fastapi.responses import StreamingResponse
 from sqlalchemy import delete, func, select, text
@@ -30,8 +31,9 @@ from .auth import require_user
 from .backtest.events import EventEngineError, run_event_backtest
 from .backtest.strategies import build_strategy
 from .config import settings
-from .data import fetcher, repository, tushare_provider
+from .data import fetcher, repository
 from .data.aggregation import bars_per_year, resample_bars
+from .data.index_series import SH_INDEX, get_index_daily
 from .data.market_overview import MarketOverviewError, fetch_overview
 from .data.symbols import plain_code
 from .db import async_session, get_session
@@ -53,8 +55,6 @@ from .schemas import (
     BacktestRequest,
     BacktestResponse,
     CandleOut,
-    ChipRowOut,
-    ChipsResponse,
     EquityPoint,
     EventBacktestRequest,
     EventBacktestResponse,
@@ -133,6 +133,8 @@ def _rows_to_bars(rows) -> list[Bar]:
 _ADJUST_MODES = ("bfq", "qfq", "hfq")
 # MA 全量集合（前端已改为本地计算 MA，后端始终返回此集合以保证缓存一致）
 _FULL_MA_SET = (5, 10, 20, 30, 60, 120, 250)
+# 指数 K 线支持的周期（日线基底聚合）
+_INDEX_TIMEFRAMES = ("1d", "1w", "1M", "1y")
 
 # 信息卡一条 SQL 拿全：stock_basic 基本信息 + 「优先与行情同日、缺则最新日」的 daily_snapshot
 # （LATERAL 单条替换原两条查询，语义不变：target 为 NULL 时全按最新日兜底）
@@ -409,6 +411,32 @@ async def get_market_overview(session: AsyncSession = Depends(get_session)) -> M
     return MarketOverviewResponse(**data)
 
 
+@router.get("/market/index-candles", response_model=list[CandleOut])
+async def get_index_candles(timeframe: str = "1d") -> Response:
+    """上证指数全量 K 线：日线为基底（tushare index_daily，进程内+Redis 缓存），
+    聚合到 1d/1w/1M/1y。收盘口径（数据随 EOD 更新，与总览 spark 一致）。"""
+    if timeframe not in _INDEX_TIMEFRAMES:
+        raise HTTPException(status_code=400, detail=f"timeframe 仅支持 {'/'.join(_INDEX_TIMEFRAMES)}")
+    key = f"idxkj:{cache.digest('idxc', SH_INDEX, timeframe)}"
+    cached = await _cached_json_response(key)
+    if cached is not None:
+        return cached
+    try:
+        bars = resample_bars(await get_index_daily(), timeframe)
+    except Exception as e:  # noqa: BLE001
+        raise HTTPException(status_code=502, detail=f"指数数据获取失败: {e}")
+    outs = [
+        CandleOut(ts=b.ts, open=b.open, high=b.high, low=b.low, close=b.close,
+                  volume=b.volume, amount=b.amount, turnover=None)
+        for b in bars
+    ]
+    # pydantic-core 序列化（与 _raw_json 同款），历史不可变、TTL 兜到当日更新
+    raw = TypeAdapter(list[CandleOut]).dump_json(outs).decode()
+    cache.local_set(key, raw, ttl=300)
+    await cache.cache_set(key, raw, ttl=7200)
+    return Response(content=raw, media_type="application/json")
+
+
 @router.post("/backtest", response_model=BacktestResponse)
 async def backtest(
     req: BacktestRequest,
@@ -1112,84 +1140,3 @@ async def screener_preview(
     cache.local_set(f"pvj:{cache_key}", raw, ttl=120)
     asyncio.create_task(cache.cache_set(f"pvj:{cache_key}", raw, ttl=600))
     return Response(content=raw, media_type="application/json")
-
-
-@router.get("/stock/{ts_code}/chips", response_model=ChipsResponse)
-async def stock_chips(
-    ts_code: str, date: str | None = None, adjust: str = "qfq",
-    session: AsyncSession = Depends(get_session),
-) -> Response:
-    """个股筹码峰（Tushare cyq_chips + cyq_perf，数据自 2018 年起）。
-
-    date=YYYY-MM-DD 为参考日（日 K 传当日；周/月 K 由前端传周期末）：返回
-    <=date 的最近有筹码数据的交易日截面；缺省取最新。
-    价格/成本均按 adjust（bfq/qfq/hfq）用 adj_factor 本地换算，与 K 线同口径。
-    """
-    if adjust not in _ADJUST_MODES:
-        raise HTTPException(status_code=400, detail=f"adjust 仅支持 {'/'.join(_ADJUST_MODES)}")
-    ref: str | None = None
-    if date:
-        try:
-            ref = datetime.strptime(date.strip()[:10], "%Y-%m-%d").strftime("%Y%m%d")
-        except ValueError:
-            raise HTTPException(status_code=400, detail="date 格式应为 YYYY-MM-DD")
-
-    # 历史截面不可变；键带 candles 版本号（adj_factor 随同步更新后旧缓存失效）。
-    # 同 preview：存序列化 JSON 直返（j: 前缀），跳过 pydantic 校验/序列化。
-    cache_key = cache.digest("chips", ts_code, ref or "latest", adjust, await cache.get_version("candles"))
-    cached = await _cached_json_response(f"chipsj:{cache_key}")
-    if cached is not None:
-        return cached
-
-    try:
-        perf, rows = await asyncio.to_thread(tushare_provider.fetch_chips, ts_code, ref)
-    except Exception as e:  # noqa: BLE001
-        raise HTTPException(status_code=502, detail=f"筹码数据获取失败: {e}")
-    if not perf:
-        resp = ChipsResponse(
-            ts_code=ts_code, trade_date=None, adjust=adjust,
-            error="无筹码数据（cyq 数据自 2018 年起，或参考日早于数据起点）",
-        )
-        raw = _raw_json(resp)
-        cache.local_set(f"chipsj:{cache_key}", raw, ttl=120)
-        await cache.cache_set(f"chipsj:{cache_key}", raw, ttl=3600)
-        return Response(content=raw, media_type="application/json")
-
-    d = datetime.strptime(str(perf["trade_date"]), "%Y%m%d")
-
-    # 复权换算（与 _adjust_bars 同口径）：qfq=f(d)/f_latest，hfq=f(d)，bfq=1
-    mult = 1.0
-    if adjust != "bfq":
-        factors = (await session.execute(
-            select(AdjFactor.trade_date, AdjFactor.adj_factor)
-            .where(AdjFactor.ts_code == ts_code, AdjFactor.trade_date <= d)
-            .order_by(AdjFactor.trade_date)
-        )).all()
-        if factors:
-            latest_f = (await session.execute(
-                select(AdjFactor.trade_date, AdjFactor.adj_factor).where(AdjFactor.ts_code == ts_code)
-                .order_by(AdjFactor.trade_date.desc()).limit(1)
-            )).first()
-            f_at = float(factors[-1][1])  # <=d 的最近因子（因子是阶梯函数）
-            f_latest = float(latest_f[1]) if latest_f else f_at
-            mult = f_at / f_latest if adjust == "qfq" else f_at
-
-    def _px(v) -> float | None:
-        return None if v is None or v != v else round(float(v) * mult, 3)
-
-    resp = ChipsResponse(
-        ts_code=ts_code,
-        trade_date=str(perf["trade_date"]),
-        adjust=adjust,
-        rows=[ChipRowOut(price=round(p * mult, 3), percent=v) for p, v in rows],
-        his_low=_px(perf.get("his_low")), his_high=_px(perf.get("his_high")),
-        cost_5pct=_px(perf.get("cost_5pct")), cost_15pct=_px(perf.get("cost_15pct")),
-        cost_50pct=_px(perf.get("cost_50pct")), cost_85pct=_px(perf.get("cost_85pct")),
-        cost_95pct=_px(perf.get("cost_95pct")),
-        weight_avg=_px(perf.get("weight_avg")),
-        winner_rate=_px(perf.get("winner_rate")),
-    )
-    raw = _raw_json(resp)
-    cache.local_set(f"chipsj:{cache_key}", raw, ttl=120)
-    await cache.cache_set(f"chipsj:{cache_key}", raw, ttl=21600)
-    return Response(content=raw, media_type="application/json")

backend/app/data/tushare_provider.py

@@ -66,32 +66,3 @@ def fetch_daily(code: str, start: str = "20200101", end: str | None = None,
             )
         )
     return bars
-
-
-def fetch_chips(ts_code: str, end: str | None = None):
-    """拉取筹码分布截面：cyq_perf（成本分位/获利比例/平均成本）+ cyq_chips（价位→占比）。
-
-    end 为 YYYYMMDD 参考日：取 <=end 的最近有数据交易日（周/月 K 线由前端换算成
-    周期末传入，这里只需向前吸附到实际数据日）；None 取最新。
-    数据自 2018 年起，早于此返回 (None, [])。返回 (perf_dict|None, [(price, percent)])。
-    """
-    from datetime import timedelta
-
-    pro = get_pro()
-    e = end or datetime.now().strftime("%Y%m%d")
-    if e < "20180101":
-        return None, []
-    # 吸附余量 60 个自然日：覆盖春节等长假 + 月初参考日（如月 K 传上月末）
-    s = (datetime.strptime(e, "%Y%m%d") - timedelta(days=60)).strftime("%Y%m%d")
-    s = max(s, "20180101")
-
-    perf = pro.cyq_perf(ts_code=ts_code, start_date=s, end_date=e)
-    if perf is None or perf.empty:
-        return None, []
-    perf = perf.sort_values("trade_date").iloc[-1]  # <=end 的最近一条
-    d = str(perf["trade_date"])
-    chips = pro.cyq_chips(ts_code=ts_code, trade_date=d)
-    rows: list[tuple[float, float]] = []
-    if chips is not None and not chips.empty:
-        rows = [(float(p), float(v)) for p, v in zip(chips["price"], chips["percent"])]
-    return perf.to_dict(), rows

backend/app/schemas.py

@@ -413,25 +413,3 @@ class MarketOverviewResponse(BaseModel):
     amount_history: list[AmountBarOut] = []  # 近 N 交易日两市成交额（旧 -> 新，末根可能盘中）
     errors: list[str] = []           # 部分来源失败的说明（透明但不阻塞展示）
 
-
-# ---------- 筹码峰（个股详情） ----------
-class ChipRowOut(BaseModel):
-    price: float            # 已按 adjust 换算的价位
-    percent: float          # 该价位筹码占比（0~1，全价位合计 ≈1）
-
-
-class ChipsResponse(BaseModel):
-    ts_code: str
-    trade_date: str | None = None       # 筹码截面所在交易日（YYYYMMDD）
-    adjust: str                         # rows[].price 的复权口径 bfq/qfq/hfq
-    rows: list[ChipRowOut] = Field(default_factory=list)
-    his_low: float | None = None        # 历史最低/最高价（换算后）
-    his_high: float | None = None
-    cost_5pct: float | None = None      # 成本分位（换算后），5/95 即 90% 筹码成本区间
-    cost_15pct: float | None = None
-    cost_50pct: float | None = None
-    cost_85pct: float | None = None
-    cost_95pct: float | None = None
-    weight_avg: float | None = None     # 平均成本（换算后）
-    winner_rate: float | None = None    # 获利比例 %
-    error: str | None = None            # 无数据时的提示（2018 前无数据等）