Updating CASA data in docker images regularly

[Mulan-94]

When running plot-elevation-tracks in the caracal obsconf worker, part of the current output we get is

SEVERE  gaincal::MeasTable::dUTC(Double) (file measures/Measures/MeasTable.cc, line 6307Leap second table TAI_UTC seems out-of-date. 
SEVERE  gaincal::MeasTable::dUTC(Double) (file measures/Measures/MeasTable.cc, line 6307)+       Until table is updated (see aips++ manager) times and coordinates
SEVERE  gaincal::MeasTable::dUTC(Double) (file measures/Measures/MeasTable.cc, line 6307)+       derived from UTC could be wrong by 1s or more.

This casaguide wiki page suggests that the solution is updating CASA data. I currently solve this by adding

RUN pip install --no-cache-dir --upgrade  --extra-index-url https://go.nrao.edu/pypi casadata

to the owlcat base image in stimela. However, the image swells up by about an extra GB and it is suggested here that this data should be updated ever so often. How can we go about doing this regularly @gijzelaerr ? @o-smirnov suggested having an image solely containing this updated data pulled into the owlcat base, or updating the stimela base altogether

[bennahugo]

To be honest It is only necessary for meerkat when recomputing uvw coordinates, where it could cause noticable rotational offsets in the astrometry off axis. As this is not a very common operation I wouldn't worry too much about it for now. It is also possible to mount in an RC file and custom user data directory - which would be much better than bloating the images and uploading regular large images. See the CASA documentation for that - you can write a script to rsync it from NRAO directly rather than a pip package dependency.

…

On Mon, 13 Jun 2022, 18:24 Lexy Andati, ***@***.***> wrote: When running plot-elevation-tracks in the caracal obsconf worker, part of the current output we get is SEVERE gaincal::MeasTable::dUTC(Double) (file measures/Measures/MeasTable.cc, line 6307Leap second table TAI_UTC seems out-of-date. SEVERE gaincal::MeasTable::dUTC(Double) (file measures/Measures/MeasTable.cc, line 6307)+ Until table is updated (see aips++ manager) times and coordinates SEVERE gaincal::MeasTable::dUTC(Double) (file measures/Measures/MeasTable.cc, line 6307)+ derived from UTC could be wrong by 1s or more. This casaguide wiki page <https://casaguides.nrao.edu/index.php/Fixing_out_of_date_TAI_UTC_tables_(missing_information_on_leap_seconds)>suggests that the solution is updating CASA data. I currently solve this by adding RUN pip install --no-cache-dir --upgrade --extra-index-url https://go.nrao.edu/pypi casadata to the owlcat base image in stimela. However, the image swells up by about an extra GB and it is suggested here <https://casaguides.nrao.edu/index.php?title=Updating_the_CASA_Data_Repository> that this data should be updated ever so often. How can we go about doing this regularly @gijzelaerr <https://github.com/gijzelaerr> ? @o-smirnov <https://github.com/o-smirnov> suggested having an image solely containing this updated data pulled into the owlcat base, or updating the stimela base altogether — Reply to this email directly, view it on GitHub <#756>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4RE6U3VBK3UCSVYTVDNSLVO5OFLANCNFSM5YUZDBKQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Mulan-94]

It is also possible to mount in an RC file and custom user data directory -

ok thanks! looking into this. Silly question, what exactly is an RC file?

[bennahugo]

Sorry I missed this -- too many tickets coming through

> $ cat ${HOME}/.casarc                                                                                     
measures.directory: /usr/share/casacore/data

The directory listed here then needs mounting inside the container