ardg-nrao / plumber Goto Github PK

I think we should create a simple PB along the lines of having one zernike term in it that we can generate and then test against. This needs to happen after we finish #1 . I am thinking a simple pytest framework setup.

Crashes here: https://github.com/ARDG-NRAO/plumber/blob/master/plumber/image.py#L46

I think it needs to be _stokes_data i.e. with leading underscore.

CASA has built in functions to generate the hour angle latitude and time of observation from which we can derive the parallactic angle. The goal is to be able to compare it with the astropy method which is producing an offset with the parang generated within the casa c++ code.

When the input image (currently only single freq images work) has > 1 Stokes plane plumber crashes. The default behaviour so far has been for plumber to read in a single plane Stokes image and generate 4 separate images containing the Stokes -I, -Q, -U, -V beams respectively.

The idea is now to generate a Stokes cube if the input image is also a Stokes cube.

plumber currently uses CASA 6 and the CASA toolkit to read and write the input images. Switch to using spectral-cube for the IO.

This also opens up the ability to more easily generate leakage cubes, rather than only single frequency images.

Plumber currently does a linear averaging over the given parallactic angle range. This is not reflective of real observations, since parallactic angle is not a linear function of HA.

Implement a switch in scripts/parang_finder.py to output a file with a list of parallactic angles and associated weights, so scripts/plumber.py can perform more accurate averaging.

As the Plumber-data repository is updated now, I tried installing the recent version of Plumber by cloning the Git repository to a local area and then performing "python setup.py install". The installation reports no error, but when I run it, e.g., parang_finder on an MS to find the parallactic angles, it returns angle values which are completely different from what was returned by the older version of Plumber. In addition, the telescope name is now returned as 'M', whereas earlier it used to be 'MeerKAT'. The position of the telescope is also now different from the stored value in the MS which we checked using msmd() task in CASA. Earlier this position value was reported to be the same as in the MS. Here are some screenshots of running older vs newer version of Plumber on an MS. First image is from running the older version, whereas second one is from the newer version.

The logging prints to console at the moment, but it does not write the logs into the default logfile. There is also no way to specify a custom name/path for a logfile.

Moving the zernike polynomial generation to C++ . The python wrapping is done using pybind11. The goal here is to speed up the production of the convolution function grid to optionally then produce convolution function for gridding directly from plumber.