IndicatorSet#

API documentation for tradeexecutor.strategy.pandas_trader.indicator.IndicatorSet Python class in Trading Strategy framework.

class IndicatorSet[source]#

Bases: object

Define the indicators that are needed by a trading strategy.

  • For backtesting, indicators are precalculated

  • For live trading, these indicators are recalculated for the each decision cycle

  • Indicators are calculated for each given trading pair, unless specified otherwise, and a separate IndicatorKey is generated by generate_combinations()

  • Indicators that have multiple instances of itself with different parameters in the same backtest are called variations - mainly useful for strategies where parameters are rollingly adjusted.

See CreateIndicatorsProtocolV2 for usage.

__init__(indicators=<factory>, variation_cache=<factory>, variation_lookup=<factory>)#
Parameters:
Return type:

None

Methods

__init__([indicators, variation_cache, ...])

add(name, func[, parameters, source, order, ...])

Add a new indicator to this indicator set.

from_indicator_keys(indicator_keys)

Reconstruct the original indicator set from keys.

generate_combinations(strategy_universe)

Create all individual indicator (per pair) we need to calculate for this trading universe.

get_count()

How many indicators we have

get_indicator(name)

Get a named indicator definition.

get_indicator_by_name_and_parameters(name, ...)

Get a variation indicator definition.

get_label()

get_variations(name)

Get all parameter variations for an indicator.

has_indicator(name)

is_varying_indicator(name)

iterate()

make_parameter_key(name, parameters)

Attributes

indicators

variation_cache

variation_lookup

get_count()[source]#

How many indicators we have

Return type:

int

get_indicator(name)[source]#

Get a named indicator definition.

Parameters:

name (str) –

Return type:

tradeexecutor.strategy.pandas_trader.indicator.IndicatorDefinition | None

get_variations(name)[source]#

Get all parameter variations for an indicator.

Return type:

list[dict]

get_indicator_by_name_and_parameters(name, parameters)[source]#

Get a variation indicator definition.

Parameters:

  • name (str) –

  • parameters (dict) –

Return type:

tradeexecutor.strategy.pandas_trader.indicator.IndicatorDefinition | None

add(name, func, parameters=None, source=IndicatorSource.close_price, order=1, variations=False)[source]#

Add a new indicator to this indicator set.

Builds an indicator set for the trading strategy, called from create_indicators.

See CreateIndicatorsProtocol for usage.

Parameters:

  • name (str) –

    Name of the indicator.

    Human-readable name. If the same function is calculated multiple times, e.g. EMA, you can have names like ema_short and ema_long.

  • func (Callable) –

    Python function to be called.

    Function takes arguments from parameters dict. It must return either pd.DataFrame or pd.Series.

  • parameters (dict | None) –

    Parameters to be passed to the Python function.

    Raw func Python arguments.

    You can pass parameters as is from StrategyParameters.

  • source (IndicatorSource) –

    Data source on this indicator is calculated.

    Defaults to the close price for each trading pair. To calculate universal indicators set to IndicatorSource.strategy_universe.

  • order –

    Dependency resolution order.

    See tradingstrategy.strategy.pandas_trader.indicator.IndicatorKey.order parameter.

  • variations –

    Set for rolling indicators that need to have their values calculated with several combinations for a backtest run.

    You can add the same variation multiple times and the sequent adds are ignored.

generate_combinations(strategy_universe)[source]#

Create all individual indicator (per pair) we need to calculate for this trading universe.

  • Because most indicators are per pair, we need to combine the trading pair universe with wanted indicators to get the final set of indicators we need to calcuate

  • Note that some indicator are calculated over the whole universe or an external source and may not have a trading pair defined

Returns:

IndicatorKey iterator.

(Trading pair, indicator) iterator

Parameters:

strategy_universe (TradingStrategyUniverse) –

Return type:

Iterable[IndicatorKey]

static from_indicator_keys(indicator_keys)[source]#

Reconstruct the original indicator set from keys.

  • Used when grid search passes data around processes

Parameters:

indicator_keys (set[tradeexecutor.strategy.pandas_trader.indicator.IndicatorKey]) –

Return type:

IndicatorSet

__init__(indicators=<factory>, variation_cache=<factory>, variation_lookup=<factory>)#
Parameters:
Return type:

None