API¶

The supported Python API centers on the compiled closed e-BH and closed BY wrappers. The functions below are the public, documented entry points intended for direct use from installed packages and wheels.

Closed testing procedures¶

`closedeBH`¶

closedeBH ¶

closedeBH(evalues, subset=None, alpha=0.05, approximate=False)

Run closed e-BH in discovery mode or set-checking mode.

Parameters:

Name	Type	Description	Default
`evalues`	`Sequence[float]`	One-dimensional sequence of e-values.	required
`subset`	`Iterable[int] \| None`	If `None`, return the number of discoveries. Otherwise check whether the specified subset is mean consistent. A subset may be given either as a boolean mask or as 0-based integer indices.	`None`
`alpha`	`float`	Target FDR level in `(0, 1]`.	`0.05`
`approximate`	`bool`	If `True`, use the approximate discovery-mode routine.	`False`

Returns:

Type	Description
`int \| bool`	An integer discovery count when `subset is None`, otherwise a single
`int \| bool`	boolean.

Examples:

>>> closedeBH([1.0, 100.0], alpha=0.05)
1
>>> closedeBH([1.0, 100.0], subset=[1], alpha=0.05)
True

`closedBY`¶

closedBY ¶

closedBY(pvalues, subset=None, alpha=0.05, approximate=False)

Run closed BY in discovery mode or set-checking mode.