ph4ser

The module ph4ser is a self-calibration pipeline that uses WSClean as imager. This is an improved version of the self-calibration module contained in the morphen_dev repository. ph4ser is still in development, so expect bugs and issues. Feel free to send any feedback and idea that you may have.

Install dependencies

To install the dependencies, just follow the instructions found in https://github.com/lucatelli/morphen.

Then, clone the ph4ser repository.

git clone https://github.com/lucatelli/ph4ser.git

Running the code

The module needs the configuration file ph4ser_config.py. In there, you can manually input the information needed for ph4ser.

Set visibility name/path.

The first step is to set the path to the visibility data:

visibility_info = {'path':"/data/C_band/MCG+12-02-001/autoselfcal/",
                   'field':'MCG12-02-001',
                   'vis_name':'MCG12-02-001.calibrated.avg',
                   'savename':'_A_C_sf'
       }

1. Copy the splited calibrated visibility into path.
2. Provide the visibility filename to vis_name (without the .ms extensions)
3. savename is just a convenient name so that at the end self-calibrated files will be saved with that prefix.

Set additional info

The second step is to tell ph4ser the data's instrument/receiver information.

receiver = 'C'
instrument = 'EVLA' # 'EVLA' or 'eM'

Complementary information, such as the image size can be specified within the dictionary init_parameters. To set the image size in x and y direction to 2048 pixels, you can use the keys imsize and imsizey, respectively. For example:

init_parameters = {'fov_image': {'imsize': 1024*8,
                                'cell': '0.4arcsec',
                                'basename': 'FOV_phasecal_image',
                                'FIELD_SHIFT': None,
                                'niter': 1000,
                                'robust': 0.5},
                  'test_image': {'imsize': int(1024*2),
                                 'imsizey': int(1024*2),
                                 'FIELD_SHIFT': None,
                                 'cell': cell_size,
                                 'prefix': 'test_image',
                                 'uvtaper': [''],
                                 'niter': 10000,
                                 'robust': 0.5 if receiver in ('K', 'Ka') or instrument == 'eM' else 0.0}
                  }

Selecting steps that will be run.

The number of self-calibration steps that are going to be run depends on the collective flux density of all sources within the region to be imaged. If the total flux density of all sources is high, more steps of self-calibration will be performed (e.g. p -> p -> p -> ap). If the value is low, we must avoid multiple self-calibration steps, so usually a single phase step will be executed (and an amplitude will be attempted).

The general configuration is to set all steps, and the exact number will be determined during runtime, when a test image will be generated.

The default configuration contained within ph4ser_config is:

steps = [
    'startup',          # create directory structure, start variables and clear visibilities.
#     'fov_image',      # create a FOV image
    'save_init_flags',  # save (or restore) the initial flags
    'statwt',           # run statwt on the initial data
    'autoflag_init',    # run rflag on the initial data -- use for L and S band data.
    'test_image',       # create a test image
    'select_refant',    # select reference antenna
    'p0',               # initial test  of selfcal, phase only (p)
    'p1',               # redo phase-only selfcal (if enough flux density); ignores p0
    'p2',               # continue phase-only selfcal (incremental)
    'ap1',              # amp-selfcal (ap); uses p0 or (p1 and p2)
    'split_trial_1',    # split the data after first trial (and run wsclean)
#     'autoflag_final',   # run rflag on the final data (use for L and S band data)
    'report_results'    # report results of first trial
]

Set the path to the WSClean singularity container.

The imaging routines with be run with the help of the script imaging_with_wsclean.py. The only change in that file is to set where your singularity container with WSClean is saved. Check line 219, just edit the variable wsclean_dir. In my case, I have:

    if running_container == 'singularity':
        mount_dir = root_dir_sys + ':/mnt'
        root_dir = '/mnt/'
        wsclean_dir = '/media/sagauga/xfs_evo/wsclean3.4-idg-everybeam-eMERLIN_portable.sif'

Running the code

The code ph4ser can be run on a terminal or on a Jupyter notebook. If run in a notebook, additional plots generated during runtime will stay there. This is useful to easily check how self-calibration improves the visibilities.

If you prefer to run ph4ser on the terminal, after making the changes to the configuration file, you can simply do

python ph4ser.py

Remember to activate your conda environment in order to run ph4ser.

An alternative, is to run the entire pipeline within a Jupyter notebook. See below.

Example notebook.

The Jupyter notebook examples/ph4ser_example.ipynb is an example containing the required set of commands to run self-calibration on a visibility.