Stages and configurations

The descriptions of each stage are copied from the docstrings of stages.

Each stage has parameters, which are defined in the YAML config file passed to the pipeline.

load_data

Reads processing set, selects one partition, and returns xarray dataset to be used by further stage. The proceessing set path is passed through --input option from cli.

Name	Type	Default	Description	Nullable	Allowed Values
msv2_chunks	ChunkConfig	ChunkConfig(frequency=10, polarization=1)	Chunking configurations for reading the measurement set in MSv2 format	False
selected_obs	list | str	'all'		False

vis_stokes_conversion

Converts visibilities to expected output polarizations. The visibilities are taken from processing set in the _upstream_output_.

Name	Type	Default	Description	Nullable	Allowed Values
output_polarizations	list	['I']	List of desired polarization codes. Allowed values are: ["I"] , ["I","Q"] , ["I","V"] , ["I","Q","U","V"] , ["XX","YY"] , ["XX","YY","XY","YX"] , ["XX","XY","YX","YY"] , ["RR","LL"] , ["RR","RL","LR","LL"].	False

read_model

Read model image(s) from FITS file(s). Supports reading from continuum or spectral FITS images.

Please refer to "Getting Started" in the documentation (or "README.md" in the repository) to understand the requirements of the model image.

If do_power_law_scaling is True, this function can scale model image across channels. This is only applicable for images with single frequency channel. The formula for power law scaling is as follows

\[scaled\_channel = current\_channel * (\frac{channel\_frequency} {reference\_frequency})^{-\alpha}\]

Where:

• \({channel\_frequency}\): Frequency of the current channel

• \({reference\_frequency}\): Reference frequency

• \({\alpha}\): Spectral index

Name	Type	Default	Description	Nullable	Allowed Values
image	str	'/path/to/wsclean-%s-image.fits'	Path to the image file. The value must have a %s placeholder to fill-in polarization values. The polarization values are taken from the polarization coordinate present in the processing set in upstream_output. For example, if polarization coordinates are ['I', 'Q'], and image param is /data/wsclean-%s-image.fits, then the read_model stage will try to read /data/wsclean-I-image.fits and /data/wsclean-Q-image.fits images. Please refer README to understand the requirements of the model image.	False
do_power_law_scaling	bool	False	Whether to perform power law scaling to scale model image across channels. Only applicable for continuum images.	False
spectral_index	float	0.75	Spectral index (alpha) to perform power law scaling. Note: The ratio of frequencies is raised to -spectral_index Please refer read_model stage for information on the formula.	False

predict_stage

Predicts visibilities from model image data using ducc0.wgridder.

Name	Type	Default	Description	Nullable	Allowed Values
epsilon	float	0.0001	Floating point accuracy for ducc gridder. Equivalent to the epsilon parameter in ducc0.gridder library functions.	False
cell_size	float	60.0	Cell size in arcsecond. For better results, this should match the cell size of the continuum image to be predicted.	False
export_model	bool	False	Whether to export the predicted model as zarr file	False
psout_name	str	'vis_model'	Name of the exported model visibilities.	False

continuum_subtraction

Perform subtraction of visibilities. The "VISIBILITY" and "VISIBILITY_MODEL" are taken from _upstream_output_.

Name	Type	Default	Description	Nullable	Allowed Values
export_residual	bool	False	Export the residual visibilities	False
psout_name	str	'vis_residual'	Output file name prefix of residual data	False
report_poly_fit	bool	False	Whether to report extent of continuum subtraction	False

flagging

Perfoms flagging on visibilities using strategies and existing flags.

Name	Type	Default	Description	Nullable	Allowed Values
strategy_configs	StrategyConfig	StrategyConfig(base_threshold=2.0, iteration_count=3, threshold_factor_step=4.0, transient_threshold_factor=5.0, threshold_timestep_rms=3.0, threshold_channel_rms=99.0, keep_outliers=True, keep_original_flags=True, low_pass_filter=LowPassFilter(do_low_pass=False, window_size=[11, 21], time_sigma=6.0, freq_sigma=7.0))		False
strategy_file	str | NoneType	None		True
export_flags	bool	False		False
psout_name	str	'flags'		False

imaging

Performs clean algorithm on the visibilities present in processing set. Processing set is present in from the _upstream_output_.

For detailed parameter info, please refer to "Stages and configurations" section in the documentation.

Name	Type	Default	Description	Nullable	Allowed Values
gridding_params	GriddingParams	GriddingParams(cell_size=None, scaling_factor=3.0, epsilon=0.0001, image_size=256)		False
deconvolution_params	DeconvolutionParams	DeconvolutionParams(algorithm='msclean', gain=0.7, threshold=0.0, fractional_threshold=0.01, scales=[0, 3, 10, 30], niter=100, use_radler=False)		False
n_iter_major	int	1		False
psf_image_path	str | NoneType	None		True
beam_info	dict	{'bmaj': None, 'bmin': None, 'bpa': None}		False
image_name	str	'spectral_cube'		False
export_format	str	'fits'		False
export_model_image	bool	False		False
export_psf_image	bool	False		False
export_residual_image	bool	False		False