Command Line Interface to run the pipeline

The main function of the pipeline can be found in imaging_prototype.py.

Prototype imaging pipeline

usage: imaging_prototype.py [-h] --input_ms INPUT_MS [--output_dir OUTPUT_DIR]
                            --input_nchan INPUT_NCHAN
                            [--nchan_per_vis NCHAN_PER_VIS]
                            [--n_major N_MAJOR]
                            [--imaging_context IMAGING_CONTEXT]
                            [--imaging_npixel IMAGING_NPIXEL]
                            [--imaging_cellsize IMAGING_CELLSIZE]
                            [--component_threshold COMPONENT_THRESHOLD]
                            [--clean_threshold CLEAN_THRESHOLD]
                            [--processing_func PROCESSING_FUNC]
                            [--use_dask USE_DASK]

Named Arguments

--input_ms

MeasurementSet to be read (including path to directory)

--output_dir

Directory where output files should go.Default is the one where program is executed from.

Default: “.”

--input_nchan

Number of channels in a single data descriptor in the MS.Note: we don’t allow specifying a list of data descriptor,instead we always load [0] from the MS.

--nchan_per_vis

How many channels per Visibility to read in

Default: 1

--n_major

How many major cycles to run

Default: 1

--imaging_context

What gridding context to use: ng (nifty-gridder) | 2d | awprojection | wg (WAGG for GPU only)

Default: “ng”

--imaging_npixel

Number of pixels in ra, dec: Should be a composite of 2, 3, 5

Default: 512

--imaging_cellsize

Cellsize (radians). Default is to calculate.

--component_threshold

Sources with absolute flux > this level (Jy) are fitted using SkyComponents

Default: 1.0

--clean_threshold

Clean stopping threshold (Jy/beam), note that this is different from component_threshold

Default: 1.0

--processing_func

Which processing functions to use: ‘rascil’: RASCIL-based processing components’proc_func’: accelerated functions from the Processing Function Library

Default: “rascil”

--use_dask

Whether to run the computation with Dask.delayed or not

Default: “False”