akquant 0.2.28__tar.gz → 0.2.30__tar.gz

{akquant-0.2.28 → akquant-0.2.30}/Cargo.lock

@@ -30,7 +30,7 @@ dependencies = [
 
 [[package]]
 name = "akquant"
-version = "0.2.28"
+version = "0.2.30"
 dependencies = [
  "anyhow",
  "chrono",

{akquant-0.2.28 → akquant-0.2.30}/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "akquant"
-version = "0.2.28"
+version = "0.2.30"
 edition = "2024"
 description = "High-performance quantitative trading framework based on Rust and Python"
 license = "MIT"

{akquant-0.2.28 → akquant-0.2.30}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: akquant
-Version: 0.2.28
+Version: 0.2.30
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy

{akquant-0.2.28 → akquant-0.2.30}/docs/zh/guide/quant_basics.md

@@ -206,6 +206,7 @@ class DualMovingAverageStrategy(Strategy):
         # 定义策略参数：短期窗口5，长期窗口20
         self.short_window = 5
         self.long_window = 20
+        self.warmup_period = self.long_window + 1
 
     def on_bar(self, bar: Bar):
         # 获取历史收盘价数据
@@ -213,7 +214,7 @@ class DualMovingAverageStrategy(Strategy):
         hist = self.get_history(count=self.long_window + 1, field="close")
 
         # 如果数据不足，无法计算均线，直接返回
-        if len(hist) < self.long_window:
+        if len(hist) < self.long_window + 1:
             return
 
         # 计算短期和长期均线

{akquant-0.2.28 → akquant-0.2.30}/examples/03_parameter_optimization_advanced.py

@@ -25,6 +25,7 @@ class DualMovingAverageStrategy(Strategy):
         # 定义策略参数：短期窗口5，长期窗口20
         self.short_window = short_window
         self.long_window = long_window
+        self.warmup_period = long_window + 1
 
     def on_bar(self, bar: Bar) -> None:
         """Handle new bar data.
@@ -37,7 +38,7 @@ class DualMovingAverageStrategy(Strategy):
         hist = self.get_history(count=self.long_window + 1, field="close")
 
         # 如果数据不足，无法计算均线，直接返回
-        if len(hist) < self.long_window:
+        if len(hist) < self.long_window + 1:
             return
 
         # 计算短期和长期均线

{akquant-0.2.28 → akquant-0.2.30}/examples/12_wfo_integrated.py

@@ -24,6 +24,7 @@ class DualMovingAverageStrategy(Strategy):
         """
         self.short_window = short_window
         self.long_window = long_window
+        self.warmup_period = long_window + 1
 
     def on_bar(self, bar: Bar) -> None:
         """处理 Bar 数据.
@@ -33,7 +34,7 @@ class DualMovingAverageStrategy(Strategy):
         """
         # 获取历史收盘价
         hist = self.get_history(count=self.long_window + 1, field="close")
-        if len(hist) < self.long_window:
+        if len(hist) < self.long_window + 1:
             return
 
         closes = hist

{akquant-0.2.28 → akquant-0.2.30}/examples/textbook/ch05_strategy.py

@@ -64,8 +64,9 @@ class MyFirstStrategy(Strategy):
         self.entry_price = 0.0  # 记录开仓价格
 
         # 设置预热期 (Warmup Period)
-        # 引擎会在正式回测前预加载数据，确保 get_history 能获取到足够的数据
-        self.warmup_period = long_window
+        # 本示例会请求 long_window + 1 根数据，并用 [:-1] 排除当前 Bar，
+        # 因此预热期也需要同步加 1，避免首次回调拿到前导 NaN。
+        self.warmup_period = long_window + 1
 
     # --------------------------------------------------------------------------
     # 2. 启动回调 (On Start)

{akquant-0.2.28 → akquant-0.2.30}/examples/textbook/ch10_analysis.py

@@ -48,7 +48,8 @@ class AnalysisStrategy(Strategy):
         super().__init__()
         self.short_window = short_window
         self.long_window = long_window
-        self.warmup_period = long_window
+        # 本示例会请求 long_window + 1 根数据，并用 [:-1] 排除当前 Bar。
+        self.warmup_period = long_window + 1
 
     def on_bar(self, bar: Bar) -> None:
         """收到 Bar 事件的回调."""

{akquant-0.2.28 → akquant-0.2.30}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "maturin"
 
 [project]
 name = "akquant"
-version = "0.2.28"
+version = "0.2.30"
 description = "High-performance quantitative trading framework based on Rust and Python"
 readme = "README.md"
 license = {text = "MIT License"}

{akquant-0.2.28 → akquant-0.2.30}/python/akquant/backtest/engine.py

@@ -370,6 +370,40 @@ def _index_to_local_trading_days(
     return cast(pd.DatetimeIndex, local_index.tz_convert(timezone))
 
 
+def _build_trading_day_metadata(
+    data_map_for_indicators: Dict[str, pd.DataFrame], timezone: str
+) -> Tuple[List[pd.Timestamp], Dict[str, Tuple[int, int]]]:
+    """Build sorted trading days and per-day nanosecond bounds."""
+    all_dates: set[pd.Timestamp] = set()
+    day_bounds: Dict[str, Tuple[int, int]] = {}
+
+    for df in data_map_for_indicators.values():
+        if df.empty or not isinstance(df.index, pd.DatetimeIndex):
+            continue
+
+        local_index = _index_to_local_trading_days(
+            cast(pd.DatetimeIndex, df.index), timezone
+        )
+        normalized_index = cast(pd.DatetimeIndex, local_index.normalize())
+        all_dates.update(normalized_index.unique())
+
+        grouped = df.groupby(normalized_index, sort=False)
+        for day_ts, day_df in grouped:
+            day_key = pd.Timestamp(day_ts).date().isoformat()
+            start_ns = int(day_df.index.min().value)
+            end_ns = int(day_df.index.max().value)
+            if day_key in day_bounds:
+                prev_start, prev_end = day_bounds[day_key]
+                day_bounds[day_key] = (
+                    min(prev_start, start_ns),
+                    max(prev_end, end_ns),
+                )
+            else:
+                day_bounds[day_key] = (start_ns, end_ns)
+
+    return sorted(all_dates), day_bounds
+
+
 BacktestDataInput = Union[
     pd.DataFrame, Dict[str, pd.DataFrame], List[Bar], DataFeed, DataFeedAdapter
 ]
@@ -2912,36 +2946,13 @@ def run_backtest(
     # Inject trading days to strategy (for add_daily_timer)
     all_strategy_instances = [strategy_instance, *slot_strategy_instances.values()]
     if data_map_for_indicators:
-        all_dates: set[pd.Timestamp] = set()
-        day_bounds: Dict[str, Tuple[int, int]] = {}
-        for df in data_map_for_indicators.values():
-            if not df.empty and isinstance(df.index, pd.DatetimeIndex):
-                local_index = _index_to_local_trading_days(
-                    cast(pd.DatetimeIndex, df.index), timezone
-                )
-                normalized_index = cast(pd.DatetimeIndex, local_index.normalize())
-                dates = normalized_index.unique()
-                all_dates.update(dates)
-                for raw_day_ts in dates:
-                    day_ts = pd.Timestamp(raw_day_ts)
-                    day_df = df[normalized_index == day_ts]
-                    if day_df.empty:
-                        continue
-                    day_key = day_ts.date().isoformat()
-                    start_ns = int(day_df.index.min().value)
-                    end_ns = int(day_df.index.max().value)
-                    if day_key in day_bounds:
-                        prev_start, prev_end = day_bounds[day_key]
-                        day_bounds[day_key] = (
-                            min(prev_start, start_ns),
-                            max(prev_end, end_ns),
-                        )
-                    else:
-                        day_bounds[day_key] = (start_ns, end_ns)
+        all_dates, day_bounds = _build_trading_day_metadata(
+            data_map_for_indicators, timezone
+        )
 
         for current_strategy in all_strategy_instances:
             if hasattr(current_strategy, "_trading_days") and all_dates:
-                current_strategy._trading_days = sorted(list(all_dates))
+                current_strategy._trading_days = all_dates
             if hasattr(current_strategy, "_trading_day_bounds"):
                 current_strategy._trading_day_bounds = day_bounds
 
@@ -4641,32 +4652,13 @@ def run_warm_start(
         warnings.warn(warning_message, RuntimeWarning, stacklevel=2)
         logger.warning(warning_message)
     if data_map_for_indicators:
-        all_dates: set[pd.Timestamp] = set()
-        day_bounds: Dict[str, Tuple[int, int]] = {}
-        for df in data_map_for_indicators.values():
-            if not df.empty and isinstance(df.index, pd.DatetimeIndex):
-                local_index = _index_to_local_trading_days(
-                    cast(pd.DatetimeIndex, df.index), timezone_name
-                )
-                dates = local_index.normalize().unique()
-                all_dates.update(dates)
-                grouped = df.groupby(local_index.normalize())
-                for day_ts, day_df in grouped:
-                    day_key = pd.Timestamp(day_ts).date().isoformat()
-                    start_ns = int(day_df.index.min().value)
-                    end_ns = int(day_df.index.max().value)
-                    if day_key in day_bounds:
-                        prev_start, prev_end = day_bounds[day_key]
-                        day_bounds[day_key] = (
-                            min(prev_start, start_ns),
-                            max(prev_end, end_ns),
-                        )
-                    else:
-                        day_bounds[day_key] = (start_ns, end_ns)
+        all_dates, day_bounds = _build_trading_day_metadata(
+            data_map_for_indicators, timezone_name
+        )
 
         for current_strategy in all_strategy_instances:
             if all_dates and hasattr(current_strategy, "_trading_days"):
-                current_strategy._trading_days = sorted(list(all_dates))
+                current_strategy._trading_days = all_dates
             if hasattr(current_strategy, "_trading_day_bounds"):
                 current_strategy._trading_day_bounds = day_bounds
             setattr(current_strategy, "_engine", engine)

{akquant-0.2.28 → akquant-0.2.30}/tests/test_examples_regression.py

@@ -132,6 +132,24 @@ def test_textbook_futures_strategy_uses_short_for_bearish_signal(
     assert captured["buy"] is None
 
 
+def test_textbook_dual_ma_examples_request_enough_warmup_bars() -> None:
+    """Dual-MA textbook examples should align warmup with N+1 history windows."""
+    ch05 = _load_example_module(
+        "example_textbook_ch05_strategy",
+        "examples/textbook/ch05_strategy.py",
+    )
+    ch10 = _load_example_module(
+        "example_textbook_ch10_analysis",
+        "examples/textbook/ch10_analysis.py",
+    )
+
+    ch05_strategy = ch05.MyFirstStrategy(short_window=5, long_window=20)
+    ch10_strategy = ch10.AnalysisStrategy(short_window=5, long_window=20)
+
+    assert ch05_strategy.warmup_period == 21
+    assert ch10_strategy.warmup_period == 21
+
+
 def test_textbook_futures_example_documents_fill_policy_and_bps_slippage() -> None:
     """Textbook futures example should retain safer fill/slippage configuration."""
     root = Path(__file__).resolve().parents[1]

{akquant-0.2.28 → akquant-0.2.30}/tests/test_strategy_extras.py

@@ -29,6 +29,7 @@ from akquant.akquant import (
 )
 from akquant.backtest import FunctionalStrategy
 from akquant.backtest.engine import (
+    _build_trading_day_metadata,
     _prime_framework_boundary_timers,
     _prime_framework_pre_open_timers,
 )
@@ -3798,6 +3799,40 @@ def test_collect_boundary_timers_and_prime_globally_once() -> None:
     ]
 
 
+def test_build_trading_day_metadata_merges_multi_symbol_day_bounds() -> None:
+    """Trading day metadata should merge per-day bounds across symbols."""
+    data_map = {
+        "AAA": pd.DataFrame(
+            {"close": [10.0, 10.5]},
+            index=pd.DatetimeIndex(
+                [
+                    pd.Timestamp("2023-01-03 01:30:00", tz="UTC"),
+                    pd.Timestamp("2023-01-03 07:00:00", tz="UTC"),
+                ]
+            ),
+        ),
+        "BBB": pd.DataFrame(
+            {"close": [20.0, 21.0]},
+            index=pd.DatetimeIndex(
+                [
+                    pd.Timestamp("2023-01-03 02:00:00", tz="UTC"),
+                    pd.Timestamp("2023-01-03 06:00:00", tz="UTC"),
+                ]
+            ),
+        ),
+    }
+
+    trading_days, day_bounds = _build_trading_day_metadata(data_map, "Asia/Shanghai")
+
+    assert trading_days == [pd.Timestamp("2023-01-03", tz="Asia/Shanghai")]
+    assert day_bounds == {
+        "2023-01-03": (
+            pd.Timestamp("2023-01-03 01:30:00", tz="UTC").value,
+            pd.Timestamp("2023-01-03 07:00:00", tz="UTC").value,
+        )
+    }
+
+
 def test_precise_boundary_hooks_delay_after_trading_until_day_end() -> None:
     """Precise boundary hooks should not emit after_trading during same-day timers."""