Bugfixes

[javierggt]

Description

This is an omnibus PR to fix a collection of small bugs:

• Fixes Consistency in data types in supplement tables #59
• Fixes obs table initialized with incorrect types #63
• Fixes Question about number of stars being processed #66

Testing

• Passes unit tests on MacOS, linux, Windows (at least one required)
• Functional testing

  cp $SKA/data/agasc/* .
  export AGASC_DIR=`pwd`
  agasc-supplement-obs-status --obs-status-file status[_obs|_bad].yml 
  agasc-supplement-update --start 2021:011 --stop 2021:012 [--report]
  agasc-supplement-update # (to check --start==two weeks ago)
  

For functional testing, I have a file called status.yml which contains this:

bad:
  77073552: 11
  23434: 10
obs:
  - obsid: 56311
    ok: false
  - obsid: 56308
    ok: true
    agasc_id: [806750112]
  - obsid: 11849
    ok: false
    agasc_id: [1019348536, 1019350904]
    comments: just removed them

I also made copies of the file including only the obs/bad part. I ran the scripts mentioned. Opened the HDF5 file in ipython and checked that the dtypes were correct, that the values in obs/bad were the ones in the file, and that the mags table times correspond to what was requested.

[taldcroft]

Let me know when this is ready for review.

[taldcroft]

For functional testing, do you have some canonical test data files to start from? In other words, how can I replicate and verify your functional testing?

[taldcroft]

And not that I don't believe you, just want a starting point of something that is expected to work and help me learn about running the code.

[javierggt]

Actually... for that, you can do it on PR #78, which includes these changes and you see the progress bar.

[taldcroft]

In general, incompatible command line arguments should raise an exception, not warn. Otherwise the program is guessing what the user really wanted.

[javierggt]

Fixed

[taldcroft]

Looks good to me!