Make sure observations with status=0 are not flagged suspect

[javierggt]

make sure observations with OK-status in status-override are OK regardless of what the algorithm says

Description

Currently, observations with status != 0 in the obs table of the supplement are excluded from magnitude estimates, but it turns out that observations with status == 0 are not included (Issue #121). The expected behavior is that if status == 0, then they should be marked OK even if it fails the standard cuts. This PR fixes that.

The fix is simple. It mostly follows what is done for excluded observations: there is a list of excluded observations, which are observations in the obs table with status != 0, so I added a corresponding list of included observations. The boolean OK flag considers both.

Most of the changes are for testing:

• a test function
• a test data HDF5 file
• a function to re-create the test data file

Testing

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

Fixes #121

[taldcroft]

LGTM

[jeanconn]

LGTMT