Cannot read leap second table TAI_UTC - casadata dependency for parallactic angles

[JSKenyon]

Describe the bug
QuartiCal will fail if parallactic angles are enabled on a system which doesn't have the casadata files and a properly configured .casarc as the data files will not be found. @AKHughes1994 reported this issue in #328.

To Reproduce
Run QC on a system without/with a misconfigured .casarc file.

Workaround
Download the measures data and add/update .casarc. The following is an example but be careful if .casarc already exists. You could also adjust the paths accordingly if you didn't want the data to live in ~/measures.

QuartiCal/.github/workflows/ci.yaml

Lines 81 to 83 in 0982990

	mkdir -p ~/measures
	curl ftp://ftp.astron.nl/outgoing/Measures/WSRT_Measures.ztar | tar xvzf - -C ~/measures
	echo "measures.directory: ~/measures" > ~/.casarc

Expected behavior
QuartiCal should work/provide an informative error message.

Version
Python3.10, main

[bennahugo]

The casadata from pip I think is meant for casatools 6. Unfortunately casacore and casatools seems to be using a different mechanism for specifying data paths (essentially .casarc has been replaced by a python configuration script for some reason I don't quite fathom). We should probably completely ditch casa tooling - parang is dead simple to compute with e.g. ephem.

…

On Wed, Feb 28, 2024 at 10:28 AM JSKenyon ***@***.***> wrote: *Describe the bug* QuartiCal will fail if parallactic angles are enabled on a system which doesn't have the casadata files and a properly configured .casarc as the data files will not be found. @AKHughes1994 <https://github.com/AKHughes1994> reported this issue in #328 <#328>. *To Reproduce* Run QC on a system without/with a misconfigured .casarc file. *Workaround* Download the measures data and add/update .casarc. The following is an example but be careful if .casarc already exists. You could also adjust the paths accordingly if you didn't want the data to live in ~/measures. https://github.com/ratt-ru/QuartiCal/blob/09829908551ce18211e546a9053023c50b5bce7b/.github/workflows/ci.yaml#L81-L83 *Expected behavior* QuartiCal should work/provide an informative error message. *Version* Python3.10, main — Reply to this email directly, view it on GitHub <#330>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4RE6RJWMVBDY5XJHPC4STYV3TEVAVCNFSM6AAAAABD5R2MYWVHI2DSMVQWIX3LMV43ASLTON2WKOZSGE2TQMZYHAZTQNA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

-- -- Benjamin Hugo

[JSKenyon]

If you have an example of this lying around I would be very interested. The measures stuff is the source of many headaches. I did look at both skyfield and astropy at some point, but I didn't really know how to evaluate their relative accuracy. I like skyfield but the coordinate conversion was fiddly.

[o-smirnov]

We should probably completely ditch casa tooling

Yes! Yes! Yes!

[JSKenyon]

See skyfielders/python-skyfield#819 for reference.

[bennahugo]

See updating custom location: https://casadocs.readthedocs.io/en/stable/notebooks/external-data.html

Why they ditched a simple configuration file for a full blown python script is quite beyond me. casatools doesn't even detect standard casacore-data installed by debian anymore from my experiences updating docker tests (unless you provide the config file).

[bennahugo]

E.g.: https://github.com/ratt-ru/montblanc/blob/65e959599b7a1e831442259713a26179019b67f2/.ci/withoutnvcc.py3.10.docker#L125-L126