research.research_runner

research.research_runner

ResearchRunner

Runs backtest research & computes metrics for a Strategy.

run staticmethod

run(config: RunConfig[TConfig], verbose_iterating: bool = True) -> list[BacktestExperimentResult]

Downloads data and runs experiments with the given config.

Parameters:

Name	Type	Description	Default
config	RunConfig[TConfig]	The configuration object to derive all data from when running experiments.	required
verbose_iterating	bool	Whether to print the experiments that are currently being processed.	True

Returns:

Type	Description
list[BacktestExperimentResult]	A list of experiment result instances for further analysis.

Source code in src\contango\research\research_runner.py

@staticmethod
def run(config: RunConfig[TConfig], verbose_iterating: bool = True) -> list[BacktestExperimentResult]:
    """
    Downloads data and runs experiments with the given config.

    Args:
        config: The configuration object to derive all data from when running experiments.
        verbose_iterating: Whether to print the experiments that are currently being processed.

    Returns:
        A list of experiment result instances for further analysis.
    """
    data = ResearchRunner._load_data(config)
    experiments = ResearchRunner._create_experiments(data, config)
    results = BacktestExperimentRunner.run(experiments, verbose_iterating=verbose_iterating)

    return results

RunConfig

Bases: NamedTuple, Generic[TConfig]

Holds the configuration parameters for a ResearchRunner.

Attributes:

Name	Type	Description
broker	HistoricalBroker[TConfig]	The historical broker to use when polling for OHLCV data.
broker_config	TConfig	The config matching to the broker to determine what type of data to derive.
strategy_factory	Callable[..., Strategy]	A callable that returns the underlying Strategy to test.
param_space	dict[str, list[Any]]	A dictionary mapping each possible parameter in the strategy to a list of the desired parameters to test.
initial_cash	USD	The initial cash for the backtester to start with for every test.
initial_position	units	The initial units for the backtester to start with for every test.
fill_behavior	FillBehavior	The behavior determining when trades will be filled across every backtest.
slippage	percent	The percent difference beteween the expected price & filled price.
commission_per_unit	USD	The price taxed per unit bought.