GridCombination#

API documentation for tradeexecutor.backtest.grid_search.GridCombination Python class in Trading Strategy framework.

class GridCombination[source]#

Bases: object

One combination line in grid search.

__init__(index, result_path, parameters, indicators=None)#

Parameters:

index (int) –
result_path (Path) –
parameters (Tuple[GridParameter]) –
indicators (set[tradeexecutor.strategy.pandas_trader.indicator.IndicatorKey] | None) –

Return type:

None

Methods

`__init__`(index, result_path, parameters[, ...])
`as_dict`()	Get as kwargs mapping.
`destructure`()	Open parameters dict.
`get_all_indicators`(combinations)	Get all defined indicators that need to be calculated, across all grid search combinatios.
`get_all_parameters_label`()	Get label which includes single value parameters as well.
`get_compressed_state_file_path`()	Get the state file that is compresed
`get_full_result_path`()	Get the path where the resulting state file is stored.
`get_label`()	Human-readable label for this combination.
`get_metrics_pickle_path`()	Get the filename for joblib pickled results are stored.
`get_parameter`(name)	Get a parameter value.
`get_relative_result_path`()	Get the path where the resulting state file is stored.
`to_strategy_parameters`()
`validate`()	Check arguments can be serialised as fs path.

Attributes

`index`	How many of nth grid combinations this is
`result_path`	In which folder we store the result files of all grid search runs
`parameters`	Alphabetically sorted list of parameters
`indicators`	Indicators for this combination.
`searchable_parameters`	Get all parameters that are searchable.

index: int#: How many of nth grid combinations this is

result_path: Path#

In which folder we store the result files of all grid search runs

Each individual combination will have its subfolder based on its parameter.

parameters: Tuple[GridParameter]#

Alphabetically sorted list of parameters

Each parameter can have 0…n values.] If parameter is not “single”, i.e. single value, then it is searchable.

indicators: set[tradeexecutor.strategy.pandas_trader.indicator.IndicatorKey] | None#

Indicators for this combination.

create_indicators() is called with the parameters and it yields the result of indicators we need to calculate for this grid combination. Only avaiable if trading_strategy_engine_version > 0.5.

One key entry for each trading pair if pair specific indicators are used

property searchable_parameters: List[GridParameter]#

Get all parameters that are searchable.

Searchable parameters have two or more values.

get_relative_result_path()[source]#

Get the path where the resulting state file is stored.

Try to avoid messing with 256 character limit on filenames, thus break down as folders.

Return type:: Path

get_full_result_path()[source]#

Get the path where the resulting state file is stored.

Return type:: Path

get_metrics_pickle_path()[source]#

Get the filename for joblib pickled results are stored.

This file is serialised with joblib and compressed with gzip

Return type:: Path

get_compressed_state_file_path()[source]#

Get the state file that is compresed

Return type:: Path

validate()[source]#: Check arguments can be serialised as fs path.

as_dict()[source]#

Get as kwargs mapping.

Return type:: dict

get_label()[source]#

Human-readable label for this combination.