WtowersInvert

class ska_sdp_func_python.xradio.imaging.WtowersInvert(subgrid_size: int = 256, support: int = 8, w_support: int = 8, oversampling: int = 16384, w_oversampling: int = 16384, subgrid_frac: float = 0.6666666666666666, padding_factor: float = 1.2, nthreads: int | None = 1, verbosity: int = 0)

Bases: InvertBackend

Validate and store user options for the “end-to-end” W-towers gridder. Most parameters should be left to their default values, unless there is a very good reason not to. subgrid_size is the main performance parameter that can be tweaked.

Parameters:

subgrid_size – Sub-grid size in pixels. Must be even and at least 16.
support – Kernel support size along u and v. Must be even and at least 4.
w_support – Kernel support size along w. Must be even and at least 4.
oversampling – The oversampling factor for the uv-kernel.
w_oversampling – The oversampling factor for the w-kernel.
subgrid_frac – Fraction of the subgrid that will be used. Must not exceed 1. A value of 0.66 is recommended.
padding_factor – Image padding factor; in practice, image a field of view padding_factor times larger than requested, then discard the edges. Higher values increase image accuracy and compute cost.
nthreads – Number of threads used for parallel processing. None means use all threads available on the CPU. Be careful when using values other than 1 in a dask context; the scheduler assumes that each task requires 1 thread unless told otherwise using e.g. dask resources.
verbosity – Print additional output to stdout based on value. 0 means print nothing; 1 or higher means print a timing report; 2 or higher means print the number of visibilities per subgrid

Attributes Summary

`l_step_sign`	Sign of the l coordinate step to use when tagging the image array axes returned by the gridder.
`m_step_sign`	Sign of the m coordinate step to use when tagging the image array axes returned by the gridder.
`nthreads`
`output_dim_order`	Order (in C convention) of "l" and "m" dimensions in the output 2D images.
`oversampling`
`padding_factor`
`subgrid_frac`
`subgrid_size`
`support`
`verbosity`
`w_oversampling`
`w_support`

Attributes Documentation

l_step_sign

m_step_sign

nthreads: int | None = 1

output_dim_order

oversampling: int = 16384

padding_factor: float = 1.2

subgrid_frac: float = 0.6666666666666666

subgrid_size: int = 256

support: int = 8

verbosity: int = 0

w_oversampling: int = 16384

w_support: int = 8