Metrics

systematica.portfolio.metrics

Variables

• run_from_portfolio_metric: systematica.portfolio.metrics.PortfolioMetric: Initiated portfolio metric class with default metric config.

check_metrics

check_metrics(
    metrics: str | List[str],
)

Validate metrics.

Parameters:

Name	Type	Default	Description
metrics	str	--	| tp.List[str] Performance metrics.

Raises:

Type	Description
TypeError	If metrics is not a string or list.
ValueError	If metric has not been configured and is not tunable.
ValueError	If mixing tunable and non-tunable metrics.

get_total_returns_nb

get_total_returns_nb(
    close: numpy.ndarray,
    sim_out: vectorbtpro.portfolio.enums.SimulationOutput,
)

Calculate the total returns.

Parameters:

Name	Type	Default	Description
close	tp.Array2d	--	Array of close prices.
sim_out	SimulationOutputT	--	Simulation output containing order records.

Returns:

Type	Description
float	Total returns.

get_expectancy_nb

get_expectancy_nb(
    close: numpy.ndarray,
    sim_out: vectorbtpro.portfolio.enums.SimulationOutput,
)

Calculate the expectancy.

Parameters:

Name	Type	Default	Description
close	tp.Array2d	--	Array of close prices.
sim_out	SimulationOutputT	--	Simulation output containing order records.

Returns:

Type	Description
float	Expectancy.

get_sharpe_nb

get_sharpe_nb(
    close: numpy.ndarray,
    sim_out: <function NamedTuple at 0x10535c180>,
    ann_factor: int,
) ‑> float

Calculate the Sharpe ratio.

Note on ddof:

• Use ddof=0 for population variance, where you have data for the entire population.
• Use ddof=1 for sample variance, to correct for bias when estimating the variance of a population from a sample.

Parameters:

Name	Type	Default	Description
close	tp.Array2d	--	Array of close prices.
sim_out	tp.NamedTuple	--	Simulation output containing returns.
ann_factor	int	--	Annualization factor.

Returns:

Type	Description
float	Sharpe ratio.

PortfolioMetric

PortfolioMetric(
    registry: <function NamedTuple at 0x10535c180> = None,
)

Portfolio metric class to manage and execute portfolio metrics.

Method generated by attrs for class PortfolioMetric.

Instance variables

• all_non_tunable_metrics: List[str]:

• all_tunable_metrics: List[str]:

• registry: <function NamedTuple at 0x10535c180>: Metric vbt.HybridConfig registry. This registry should have a retrivable metric name as key and schema. if None, fallbacks to METRIC_REGISTRY.

Methods

get_all_metrics

get_all_metrics(
    self,
    include_non_tunable: bool = False,
) ‑> List[str]

Get a list of metric keys from the registry. Optionally include non-tunable metrics based on user choice.

Parameters:

Name	Type	Default	Description
include_non_tunable	bool	False	Whether to include non-tunable metrics. Defaults to False.

Returns:

Type	Description
List[str]	List of metric keys with valid (non-NaNs) rolling_func values.

get_all_rolling_metrics

get_all_rolling_metrics(
    self,
    include_non_tunable: bool = False,
) ‑> List[str]

Get a list of metric keys from the registry that have a valid rolling_func. Optionally include non-tunable metrics based on user choice.

Parameters:

Name	Type	Default	Description
include_non_tunable	bool	False	Whether to include non-tunable metrics. Defaults toFalse.

Returns:

Type	Description
List[str]	List of metric keys with valid (non-NaNs) rolling_func values.

get_all_pf_metrics

get_all_pf_metrics(
    self,
    include_non_tunable: bool = False,
) ‑> List[str]

Get a list of metric keys from the registry that have a valid pf_func. Optionally include non-tunable metrics based on user choice.

Parameters:

Name	Type	Default	Description
include_non_tunable	bool	False	Whether to include non-tunable metrics. Defaults to False.

Returns:

Type	Description
List[str]	List of metric keys with valid (non-NaNs) pf_func values.

run

run(
    self,
    pf: vectorbtpro.portfolio.base.Portfolio,
    metrics: str | List[str],
    key: str,
    **kwargs,
) ‑> Dict[str, float | pandas.core.series.Series | pandas.core.frame.DataFrame]

Retrieve a dictionary of metric keys and their function values from the registry and execute it using vbt.deep_get_attr. Only includes metrics with non-None and string-type values.

• Metrics with None or missing are excluded.
• The function stops iterating once all requested metrics are found.
• Input validation ensures all provided metrics exist in the registry.

Parameters:

Name	Type	Default	Description
pf	vbt.Portfolio	--	Portfolio instance.
metrics	str	--	or list of str A single metric key or a list of metric keys to look up in the registry.
key	str	--	Type of key to fetch.
kwargs	tp.Kwargs	--	Additional key-word arguments.

Raises:

Type	Description
KeyError	If any provided metric key is not found in the registry.

Returns:

Type	Description
tp.Dict[str, float | tp.SeriesFrame]	Dictionary mapping valid metric keys to their corresponding output values.

run_rolling_metric

run_rolling_metric(
    self,
    pf: vectorbtpro.portfolio.base.Portfolio,
    metrics: str | List[str],
    window: int,
    minp: int = None,
    to_numpy: bool = False,
    **kwargs,
) ‑> pandas.core.frame.DataFrame | numpy.ndarray

Retrieve a dictionary of metric keys and their rolling function values from the registry and execute it using vbt.deep_get_attr. Only includes metrics with non-None and string-type values.

Parameters:

Name	Type	Default	Description
pf	vbt.Portfolio	--	Portfolio instance.
metrics	str	--	or list of str A single metric key or a list of metric keys to look up in the registry.
window	int	--	Rolling window size.
minp	int	None	Minimum number of observations required.
to_numpy	bool	False	If False, returns a pandas DataFrame. Otherwise a numpy array.
kwargs	tp.Kwargs	--	Additional key-word arguments.

Returns:

Type	Description
pd.DataFrame | tp.Array	Rolling output.

run_pf_metric

run_pf_metric(
    self,
    pf: vectorbtpro.portfolio.base.Portfolio,
    metrics: str | List[str],
    to_dict: bool = False,
    to_numpy: bool = False,
    **kwargs,
) ‑> collections.OrderedDict | numpy.ndarray | pandas.core.series.Series | pandas.core.frame.DataFrame

Retrieve a dictionary of metric keys and their pf function values from the registry and execute it using vbt.deep_get_attr. Only includes metrics with non-None and string-type values.

Parameters:

Name	Type	Default	Description
pf	vbt.Portfolio	--	Portfolio instance.
metrics	str	--	or list of str A single metric key or a list of metric keys to look up in the registry.
to_dict	bool	False	Change frame format output to dict type if set to True. Raw pd.DataFrame object data otherwise. Frames are retrieve when group_by is False in the vbt.PF configuration and/or when rolling calculation is triggered. Defaults to False.
to_numpy	bool	False	If False, returns a pandas DataFrame. Otherwise a numpy array.
kwargs	tp.Kwargs	--	Additional key-word arguments.

Returns:

Type	Description
OrderedDict	Float or int if metrics is standalone. OrderedDict mapping valid metric keys to their corresponding output.