Fix ordering in (co)domain number accessors for schemas

[bosonbaas]

This PR provides a fix for Issue #520. The error was that the adom_num and acodom_num functions assumed a certain ordering on the codoms field. This is now explicitly aligned to the order of the attrs field.

[epatters]

Thank you, Andrew, for reporting the bug and finding a fix! Following the struct acsets refector, I am less familiar with this part of the codebase than I used to be, but I thought that doms and codoms were named tuples and so I'm surprised that they don't have the correct order.

It also makes me wonder: do the functions dom_nums and codom_nums also have this subtle bug?

@olynch, it would be great if you could carefully review this issue.

[github-actions]

Review Checklist

Does this PR follow the development guidelines? Following is a partial checklist:

Tests

• New features and bug fixes have unit tests
• New modules have tests that are ultimately called by the test runner (test/runtests.jl)
• Existing tests have not been deleted
• Code coverage >= 90% or reduction justified in PR

Documentation

• All exported functions, types, and constants have docstrings, written in complete sentences
• Citations are given for any constructions, algorithms, or code drawn from external sources

Other

• Style guidelines are followed, including indent width 2
• Changes breaking backwards compatibility have been approved

[olynch]

Yes, I see the bug and the fix results in cleaner code, you can go ahead and merge this.

[bosonbaas]

Should we also add the change to the dom_nums codom_nums functions above those?

[epatters]

Let's go ahead and change those functions too. If nothing else, the code will be shorter.

[bosonbaas]

Got those changes added, and it looks like tests are passing!

[epatters]

Great, thanks Andrew!