quantlab.backtest.config.WeightsBacktestConfig#

class quantlab.backtest.config.WeightsBacktestConfig(*, price_dataset, start_date, end_date, rebalance_periods, output_dir, model=None, model_mode=None, checkpoint=None, cv_project_dir=None, fees=0.0005, slippage=0.0005, init_cash=1000000.0, sizing_basis='fill', benchmark_dataset=None, risk_model=None, tracker=<factory>, name=None, fill_price_column, valuation_price_column, trading_days_per_year, session_minutes_per_day, direction=None, top_n=None)[source]#

Bases: BacktestConfig

Config of a backtest of given target weights, market conventions included.

The backtester reading it (WeightsVectorBt) has no fixed market: the fill and valuation columns and the annualization are fields here instead of a class constant, so one class serves any frame a caller brings. It takes no model; its weights come to run_weights(). direction and top_n record the top-N selection that built the weights from scores, when one did; they are shown in the report and never used to select.

Examples

>>> import pandas as pd
>>> from quantlab.backtest.config import WeightsBacktestConfig
>>> from quantlab.dataset.memory import FrameDataset
>>> price_dataset = FrameDataset(pd.DataFrame({
...     "timestamp": pd.to_datetime(["2024-01-02", "2024-01-03"]),
...     "symbol": ["AAA", "AAA"], "open": [10.0, 11.0], "close": [10.5, 11.5],
... }))
>>> cfg = WeightsBacktestConfig(
...     price_dataset=price_dataset,
...     start_date="2024-01-01",
...     end_date="2024-06-28",
...     output_dir=None,
...     rebalance_periods=1,
...     fill_price_column="open",
...     valuation_price_column="close",
...     trading_days_per_year=252,
...     session_minutes_per_day=390,
... )
>>> cfg.model, cfg.top_n
(None, None)

direction = None#: The selection side the weights were built with, or None for weights given as such. A record only.

fill_price_column#: Price variable orders fill at.

session_minutes_per_day#: Minutes of one trading session, to annualize intraday bars.

top_n = None#: The names per side the weights were built with, or None. A record only.

trading_days_per_year#: Trading days in a year, to annualize daily and longer bars.

valuation_price_column#: Price variable the portfolio is valued at after each bar.