spey.get_backend

spey.get_backend(name: str) → Callable[[Any], StatisticalModel]

Statistical model backend retreiver. Available backend names can be found via AvailableBackends() function.

Parameters:

name (Text) – backend identifier. This backend refers to different packages that prescribes likelihood function.

Raises:

• PluginError – If the backend is not available or the available backend requires different version of spey.

• AssertionError – If the backend does not have necessary metadata.

Returns:

A callable function that takes backend specific arguments and two additional keyword arguments analysis (which is a unique identifier of analysis name as str) and xsection (which is cross section value with a.u.). Details about the function can be found in statistical_model_wrapper(). This wrapper returns a StatisticalModel object.

Return type:

Callable[[Any, ...], StatisticalModel]

Example:

>>> import spey; import numpy as np
>>> stat_wrapper = spey.get_backend("default_pdf.uncorrelated_background")

>>> data = np.array([1])
>>> signal = np.array([0.5])
>>> background = np.array([2.])
>>> background_unc = np.array([1.1])

>>> stat_model = stat_wrapper(
...     signal_yields=signal,
...     background_yields=background,
...     data=data,
...     covariance_matrix=background_unc,
...     analysis="simple_sl",
...     xsection=0.123
... )
>>> stat_model.exclusion_confidence_level()

Note

The documentation of the stat_wrapper defined above includes the docstring of the backend as well. Hence typing stat_wrapper? in terminal will result with complete documentation for the statistical_model_wrapper() and the backend it self which is in this particular example SimplifiedLikelihoodInterface.