Describe the bug
This is a bit of an edge case I uncovered while pursuing a different bug. Currently, QC presumes that each antenna is associated with exactly one row of the FEED table. This is not necessarily true as the FEED table has a SPECTRAL_WINDOW_ID column i.e. there may be more rows than anticipated. This can make the application of the parallactic angle fail spectacularly.
To Reproduce
Simply run QC on the above data and set input_model.apply_p_jones=True. Note that this is not in any way sensible as this is single correlation data, but it does expose the problem.
Expected behavior
Parallactic angle (which includes the receptor angles) should function regardless of the structure of the FEED table.
Describe the bug This is a bit of an edge case I uncovered while pursuing a different bug. Currently, QC presumes that each antenna is associated with exactly one row of the FEED table. This is not necessarily true as the FEED table has a SPECTRAL_WINDOW_ID column i.e. there may be more rows than anticipated. This can make the application of the parallactic angle fail spectacularly.
Data description (if applicable) The data comes from the casa imaging tutorial: https://casaguides.nrao.edu/index.php?title=VLA_CASA_Imaging-CASA6.5.2
To Reproduce Simply run QC on the above data and set
input_model.apply_p_jones=True. Note that this is not in any way sensible as this is single correlation data, but it does expose the problem.Expected behavior Parallactic angle (which includes the receptor angles) should function regardless of the structure of the FEED table.
Version main