migrate travis-ci to cirrus-ci

[bjlittle]

🚀 Pull Request

Description

This PR shows some much need ❤️ to nc-time-axis.

For fuller details see the associated Issue #63.

Note that, in order to reboot the CI, I've folding in the following awesome contributions:

• PR Remove cftime.utime #59 by @aulemahal, and
• PR use object instead of np.object #56 by @mathause.

This PR will also close:

• PRs mods needed for cftime PR #202 #51 (which is addressed by Remove cftime.utime #59), and
• PR migrate to GitHub Actions #57.

Closes #63

[codecov-commenter]

Codecov Report

❗ No coverage uploaded for pull request base (master@0ec3b51). Click here to learn what that means.
The diff coverage is n/a.

@@            Coverage Diff            @@
##             master      #66   +/-   ##
=========================================
  Coverage          ?   89.14%           
=========================================
  Files             ?        1           
  Lines             ?      175           
  Branches          ?       42           
=========================================
  Hits              ?      156           
  Misses            ?       13           
  Partials          ?        6           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0ec3b51...08caec3. Read the comment docs.

[lbdreyer]

Looking good so far! A really great improvement!

I have a few minor things to look through (I'll do these tomorrow) but I thought I'd submit the few minor comments I had so far

[lbdreyer]

Looks like this is good to go!

Thanks for working on this @bjlittle !

[bjlittle]

@lbdreyer Awesome! 🥳 🍻

[ocefpaf]

Do we need an env file for each Python version? That will be an extra burden to maintain :-/

[trexfeathers]

Sorry not to have caught this before merge. has_year_zero was only introduced as a kwarg in Unidata/cftime#234, which means the cftime requirement should be pinned: cftime>=1.5.

(I think this ought to fix the nc-time-axis failures on SciTools/iris#4194, although it's hard to tell if it's the root cause).