v1.1.9

• Supports Python 3.11
• Adds fix for regression in Numba 0.57.0
  numba/numba#8940

v1.1.7

• Adds __version__ field for getting version number of PyBroker.
• Adds cover_fill_price, cover_shares, and cover_limit_price properties to ExecContext. These alias buy_fill_price, buy_shares, and buy_limit_price respectively. However, setting the cover* properties guarantees that the resulting buy orders are placed before any sell orders. This is useful for covering short positions when max_short_positions is set, since normally all sell orders are placed before all buy orders.
• Adds optional cash argument to calc_target_shares to override using portfolio equity by default.

v1.1.6

• Allows setting None values with param.
• Adds config member to BaseContext.
• Changes calc_bootstrap parameter to default to False.

v1.1.4

Adds param function for setting and getting global parameters.

v1.1.3

Fixes duplicate stop ID error when there are multiple symbols.

v1.1.1

• Adds set_before_exec and set_after_exec functions for processing all ExecContexts before and after all execution functions.
• Adds sharpe_length configuration option to annualize the Sharpe Ratio.
• Exposes sessions field on PosSizeContext.
• Replaces Log Profit Factor with exponentiated Profit Factor.
• Allows setting buy_fill_price and sell_fill_price to np.floating values.
• Adds win_rate and loss_rate to BaseContext for tracking the running win/loss rates.
• Adds largest_win_pct and largest_loss_pct metrics.
• Exposes profit_factor, sharpe, and drawdown fields for accessing confidence interval objects.

v1.1.0

• Adds support for the following stops:

1. Stop loss
2. Trailing stop loss
3. Take profit

• Allows canceling pending orders.

• Upgrades alpaca-trade-api-python to alpaca-py package.

• Adds bars field to Position class.

v1.0.21

• Adds exit_on_last_bar configuration option to exit any open positions on the last bar of data available for a symbol. (See #6)
• Adds PriceScope to improve performance when retrieving price data.

v1.0.20

• Fixes bug where test results were being repeated for each walkforward window.

v1.0.19

• Fixes bug where incorrect market value was being calculated if bar dates were not shared by multiple symbols (#7)
• Changes PnL in TestResult#portfolio to be running total of realized + unrealized PnL. Removes unrealized_pnl column.
• Adds unrealized_pnl to EvalMetrics.
• Adds better error messaging if "predict" method is not found on a trained model instance when running predictions.
• Updates the project's license to Apache 2.0 with Commons Clause. PyBroker is free for non-commercial use.