pycheck tests failing due to missing obsid

[hpearnshaw]

Look into why this obsid assertion isn't working, which was causing the pytest failures while doing the latest PR.

This doesn't seem to be affecting the overall clockfile generation.

Failure details below:

=================================== FAILURES ===================================
_____________________________ test_get_obsid_list ______________________________

    @pytest.mark.remote_data
    def test_get_obsid_list():
        table = get_obsid_list_from_heasarc()
        assert 'met' in table.colnames
>       assert '80002092002' in table['obsid']
E       AssertionError: assert '80002092002' in <Column name='obsid' dtype='str1' length=1>\n

nuclockutils/tests/test_utils.py:10: AssertionError
=========================== short test summary info ============================
FAILED nuclockutils/tests/test_utils.py::test_get_obsid_list - AssertionError: assert '80002092002' in <Column name='obsid' dtype='str1' length=1>\n
============== 1 failed, 35 passed, 1 skipped in 74.52s (0:01:14) ==============

[matteobachetti]

@hpearnshaw there are problems with HEASARC these days, that might have produced this issue. Could you retry?

[hpearnshaw]

Indeed it looks like the tests passed just fine when I actually merged the PR last week (though it took about 5m to run compared to usual 2-3m). I'll see what it does next time I make a PR.

[matteobachetti]

Good, let's keep this open until we are convinced the problem was temporary

[hpearnshaw]

It looks likely that it's issues with data retrieval from HEASARC that is causing this. We can probably disregard test failures that happen for this reason. Matteo, if you agree we can close this issue.

[matteobachetti]

Are the issues very frequent? If so, it might be worth investigating alternative ways to download the data