Filter

class sherpa.data.Filter

Bases: object

A class for representing filters of N-Dimensional datasets.

The filter does not know the size of the dataset or the values of the independent axes.

Attributes:

mask

Mask array for dependent variable

Methods

apply(-> None)	Apply this filter to an array
notice(mins, maxes, axislist[, ignore, ...])	Select a range to notice or ignore (remove).

Attributes Summary

mask	Mask array for dependent variable

Methods Summary

apply(-> None)	Apply this filter to an array
notice(mins, maxes, axislist[, ignore, ...])	Select a range to notice or ignore (remove).

Attributes Documentation

mask

Mask array for dependent variable

Returns:

maskbool or numpy.ndarray

Methods Documentation

apply(array: None) → None

apply(array: Sequence[float] | ndarray) → ndarray

Apply this filter to an array

Parameters:

arrayarray_like or None

Array to be filtered

Returns:

array_likendarray or None

Raises:

sherpa.utils.err.DataErr

The filter has removed all elements or there is a mismatch between the mask and the array argument.

See also

notice

notice(mins: Sequence[float] | ndarray, maxes: Sequence[float] | ndarray, axislist: Sequence[Sequence[float] | ndarray], ignore: bool = False, integrated: bool = False) → None

Select a range to notice or ignore (remove).

The axislist argument is expected to be sent the independent axis of a Data object - so (x, ) for one-dimensional data, (xlo, xhi) for integrated one-dimensional data, (x0, x1) for two-dimensional data, and (x0lo, x1lo, x0hi, x1hi) for integrated two-dimensinal data. The mins and maxes must then be set to match this ordering.

Parameters:

minssequence

The minimum value of the valid range (elements may be None to indicate no lower bound). When not None, it is treated as an inclusive limit, so points >= min are included.

maxessequence

The maximum value of the valid range (elements may be None to indicate no upper bound). It is treated as an inclusive limit (points <= max) when integrated is False, and an exclusive limit (points < max) when integrated is True.

axislist: sequence of arrays

The axis to apply the range to. There must be the same number of elements in mins, maxes, and axislist. The number of elements of each element of axislist must also agree (the cell values do not need to match).

ignorebool, optional

If True the range is to be ignored, otherwise it is included. The default is to include the range.

integratedbool, optional

Is the data integrated (we have low and high bin edges)? The default is False. When True it is expected that axislist contains a even number of rows, where the odd values are the low edges and the even values the upper edges, and that the mins and maxes only ever contain a single value, given in (None, hi) and (lo, None) ordering.

See also

apply

Examples

Select points in xs which are in the range 1.5 <= x <= 4:

>>> f = Filter()
>>> f.mask
True
>>> xs = [1, 2, 3, 4, 5]
>>> f.notice([1.5], [4], (xs, ))
>>> f.mask
array([False,  True,  True,  True, False])

Filter the data to select all points with x0 >= 1.5 and x1 <= 4:

>>> f = Filter()
>>> x0 = [1, 1.4, 1.6, 2, 3]
>>> x1 = [2, 2, 4, 4, 6]
>>> f.notice([1.5, None], [None, 4], (x0, x1))
>>> f.mask
array([False, False,  True,  True, False])

For integrated data sets the lower and upper edges should be sent separately with the max and min limits, along with setting the integrated flag. The following selects the bins that cover the range 2 to 4 and 1.5 to 3.5:

>>> xlo = [1, 2, 3, 4, 5]
>>> xhi = [2, 3, 4, 5, 6]
>>> f = Filter()
>>> f.notice([None, 2], [4, None], (xlo, xhi), integrated=True)
>>> f.mask
array([False,  True,  True,  False, False])
>>> f.notice([None, 1.5], [3.5, None], (xlo, xhi), integrated=True)
>>> f.mask
array([ True,  True,  True, False, False])