powerquant-cli 0.2.6__py3-none-any.whl

quant_core/app/data/service/price_adjustment_service.py

@@ -0,0 +1,235 @@
+"""基于 Raw Bar 与复权因子生成显式价格视图。"""
+
+import hashlib
+import json
+from bisect import bisect_right
+from datetime import date
+from decimal import Decimal
+from zoneinfo import ZoneInfo
+
+from quant_core.app.data.exceptions import MarketDataIncompleteException
+from quant_core.app.data.interface.adjustment_factor_query import IAdjustmentFactorQuery
+from quant_core.app.data.schema.adjustment_schema import (
+    AdjustedBarResponse,
+    AdjustedBarSeries,
+    AdjustmentAnchorMode,
+    AdjustmentType,
+    BarViewRequest,
+)
+from quant_core.app.data.schema.adjustment_schema import (
+    AdjustmentFactorResponse as AdjustmentFactor,
+)
+from quant_core.app.data.schema.market_schema import BarResponse
+from quant_core.app.data.schema.pit_daily_window import PITDailyWindow
+
+_CN = ZoneInfo("Asia/Shanghai")
+
+
+class PriceAdjustmentService:
+    """生成研究/策略价格视图；绝不修改 Raw Bar。"""
+
+    def __init__(self, factors: IAdjustmentFactorQuery):
+        self.factors = factors
+
+    @classmethod
+    def pit_daily_windows(
+        cls,
+        raw_bars: list[BarResponse],
+        factor_rows: list[AdjustmentFactor],
+        days: list[date],
+        next_days: list[date],
+        lookback: int,
+    ) -> tuple[PITDailyWindow, ...]:
+        """复用 Market 复权规则，一次装配多日决策时点窗口。"""
+        if lookback < 1 or len(days) != len(next_days):
+            raise ValueError("PIT 窗口的 lookback 和交易日映射非法")
+        by_date = {row.trade_date: row for row in factor_rows}
+        ordered = sorted(raw_bars, key=lambda bar: bar.timestamp)
+        ordered_days = [bar.timestamp.astimezone(_CN).date() for bar in ordered]
+        windows: list[PITDailyWindow] = []
+        for day, anchor_day in zip(days, next_days, strict=True):
+            if anchor_day <= day:
+                raise ValueError("复权锚点必须是目标日之后的交易日")
+            cutoff = bisect_right(ordered_days, day)
+            raw_window = ordered[max(0, cutoff - lookback) : cutoff]
+            if raw_window:
+                anchor = by_date.get(anchor_day)
+                if anchor is None:
+                    raise MarketDataIncompleteException(
+                        dataset="adjustment_factor",
+                        instrument_id=raw_window[-1].instrument_id,
+                        reason=f"缺少决策日锚点因子 {anchor_day}",
+                    )
+                request = BarViewRequest(
+                    adjustment_type=AdjustmentType.PRE,
+                    anchor_mode=AdjustmentAnchorMode.FIXED_DATE,
+                    anchor_date=anchor_day,
+                )
+                adjusted: list[AdjustedBarResponse] = []
+                evidence: list[dict] = []
+                for bar in raw_window:
+                    row = by_date.get(bar.timestamp.astimezone(_CN).date())
+                    if row is None:
+                        raise MarketDataIncompleteException(
+                            dataset="adjustment_factor",
+                            instrument_id=bar.instrument_id,
+                            reason=f"缺少 {bar.timestamp.astimezone(_CN).date()} 复权因子",
+                        )
+                    adjusted.append(cls._adjust_bar(bar, row.factor, anchor.factor, request))
+                    evidence.append(
+                        {
+                            "bar": bar.model_dump(mode="json"),
+                            "factor": str(row.factor),
+                            "revision": row.dataset_revision,
+                        }
+                    )
+                evidence.append(
+                    {
+                        "anchor": anchor_day.isoformat(),
+                        "factor": str(anchor.factor),
+                        "revision": anchor.dataset_revision,
+                    }
+                )
+            else:
+                adjusted = []
+                evidence = [{"day": day.isoformat(), "bars": "absent"}]
+            digest = hashlib.sha256(
+                json.dumps(evidence, sort_keys=True, separators=(",", ":")).encode("utf-8")
+            ).hexdigest()
+            windows.append(
+                PITDailyWindow(
+                    as_of=day,
+                    anchor_date=anchor_day,
+                    bars=tuple(adjusted),
+                    input_digest=digest,
+                )
+            )
+        return tuple(windows)
+
+    async def adjust(self, bars: list[BarResponse], request: BarViewRequest) -> AdjustedBarSeries:
+        if not bars:
+            raise MarketDataIncompleteException(dataset="bar", reason="没有可复权的 Bar")
+        instrument_id = bars[0].instrument_id
+        if request.adjustment_type is AdjustmentType.NONE:
+            return AdjustedBarSeries(
+                instrument_id=instrument_id,
+                adjustment_type=AdjustmentType.NONE,
+                anchor_date=None,
+                factor_revision=None,
+                bars=bars,
+            )
+
+        dates = [bar.timestamp.date() for bar in bars]
+        if request.adjustment_type is AdjustmentType.POST:
+            return await self._adjust_post(bars, request, instrument_id, dates)
+
+        anchor_limit = self._anchor_limit(request)
+        anchor_row = await self.factors.get_adjustment_factor_on_or_before(
+            instrument_id, anchor_limit, request.known_at
+        )
+        if anchor_row is None:
+            raise MarketDataIncompleteException(
+                dataset="adjustment_factor",
+                instrument_id=instrument_id,
+                reason="找不到复权锚点因子",
+            )
+        anchor = anchor_row.trade_date
+        factors = await self.factors.list_adjustment_factors(
+            instrument_id, min(dates), max(dates), request.known_at
+        )
+        by_date = {row.trade_date: row for row in factors}
+        missing = sorted(set(dates) - by_date.keys())
+        if missing:
+            raise MarketDataIncompleteException(
+                dataset="adjustment_factor",
+                instrument_id=instrument_id,
+                start=str(min(missing)),
+                end=str(max(missing)),
+                reason="复权因子缺失，禁止静默使用原始价",
+            )
+        anchor_factor = anchor_row.factor
+        adjusted = [
+            self._adjust_bar(bar, by_date[bar.timestamp.date()].factor, anchor_factor, request)
+            for bar in bars
+        ]
+        revisions = sorted(
+            {by_date[d].dataset_revision for d in set(dates)} | {anchor_row.dataset_revision}
+        )
+        return AdjustedBarSeries(
+            instrument_id=instrument_id,
+            adjustment_type=request.adjustment_type,
+            anchor_date=anchor,
+            factor_revision=",".join(revisions),
+            bars=[*adjusted],
+        )
+
+    @staticmethod
+    def _anchor_limit(request: BarViewRequest) -> date | None:
+        if request.anchor_mode is AdjustmentAnchorMode.FIXED_DATE:
+            assert request.anchor_date is not None
+            return request.anchor_date
+        if request.anchor_mode is AdjustmentAnchorMode.DECISION_TIME:
+            assert request.decision_time is not None
+            return request.decision_time.date()
+        return None
+
+    async def _adjust_post(
+        self,
+        bars: list[BarResponse],
+        request: BarViewRequest,
+        instrument_id: str,
+        dates: list[date],
+    ) -> AdjustedBarSeries:
+        factors = await self.factors.list_adjustment_factors(
+            instrument_id, min(dates), max(dates), request.known_at
+        )
+        by_date = {row.trade_date: row for row in factors}
+        missing = sorted(set(dates) - by_date.keys())
+        if missing:
+            raise MarketDataIncompleteException(
+                dataset="adjustment_factor",
+                instrument_id=instrument_id,
+                start=str(min(missing)),
+                end=str(max(missing)),
+                reason="复权因子缺失，禁止静默使用原始价",
+            )
+        adjusted = [
+            self._adjust_bar(bar, by_date[bar.timestamp.date()].factor, Decimal("1"), request)
+            for bar in bars
+        ]
+        revisions = sorted({by_date[d].dataset_revision for d in set(dates)})
+        return AdjustedBarSeries(
+            instrument_id=instrument_id,
+            adjustment_type=request.adjustment_type,
+            anchor_date=None,
+            factor_revision=",".join(revisions),
+            bars=[*adjusted],
+        )
+
+    @staticmethod
+    def _adjust_bar(
+        bar: BarResponse,
+        factor: Decimal,
+        anchor_factor: Decimal,
+        request: BarViewRequest,
+    ) -> AdjustedBarResponse:
+        ratio = factor if request.adjustment_type is AdjustmentType.POST else factor / anchor_factor
+        if ratio <= 0:
+            raise MarketDataIncompleteException(
+                dataset="adjustment_factor",
+                instrument_id=bar.instrument_id,
+                reason="复权比例必须大于零",
+            )
+        return AdjustedBarResponse(
+            **bar.model_dump(
+                exclude={"open", "high", "low", "close", "previous_close", "volume", "amount"}
+            ),
+            open=bar.open * ratio,
+            high=bar.high * ratio,
+            low=bar.low * ratio,
+            close=bar.close * ratio,
+            previous_close=(bar.previous_close * ratio if bar.previous_close is not None else None),
+            volume=Decimal(bar.volume) / ratio,
+            # 成交额是实际成交事实，不复权。
+            amount=bar.amount,
+        )

quant_core/app/factor/LICENSE.vnpy

@@ -0,0 +1,21 @@
+The MIT License (MIT)
+
+Copyright (c) 2015-present, Xiaoyou Chen
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

quant_core/app/factor/__init__.py

@@ -0,0 +1,13 @@
+"""Factor Domain - 来源无关的 Canonical 因子边界（T1 MVP）
+
+依赖方向（Group TASK-20260907-2308-factor-runtime-mvp 决策）：
+
+    Tushare / Local Calculator / Future Rust or ML
+        → Provider Adapter (T3)
+        → Canonical Factor Store (T2)
+        → Factor Query Runtime (T4)
+        → Strategy / Selection / Agent
+
+本域只拥有 Canonical Contract（T1）与统一查询边界；Provider 专有字段
+止步于 Adapter，Strategy 侧不得感知 Provider / Repository。
+"""