Allowing time_utils.select_time_period to work on arrays containing NaNs

[dougiesquire]

This PR extends time_utils.select_time_period to work on arrays containing NaNs.

Note this builds off the branch requested for merge in #9, so best to merge that first

Closes #11

[codecov-commenter]

Codecov Report

Merging #12 (7487854) into master (b38788e) will increase coverage by 1.70%.
The diff coverage is 93.82%.

@@            Coverage Diff             @@
##           master      #12      +/-   ##
==========================================
+ Coverage   16.52%   18.23%   +1.70%     
==========================================
  Files          11       15       +4     
  Lines         599      894     +295     
==========================================
+ Hits           99      163      +64     
- Misses        500      731     +231     

Impacted Files	Coverage Δ
unseen/array_handling.py	41.79% <ø> (ø)
unseen/tests/conftest.py	91.89% <91.89%> (ø)
unseen/time_utils.py	33.89% <92.30%> (+18.85%)	⬆️
unseen/tests/test_time_utils.py	95.00% <95.00%> (ø)
unseen/tests/test_array_handling.py	100.00% <100.00%> (+3.70%)	⬆️
unseen/fileio.py	0.00% <0.00%> (ø)
unseen/indices.py	0.00% <0.00%> (ø)
unseen/bias_correction.py	0.00% <0.00%> (ø)
unseen/independence.py	0.00% <0.00%> (ø)
unseen/similarity.py	0.00% <0.00%> (ø)
... and 1 more

---

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 b38788e...7487854. Read the comment docs.

[DamienIrving]

Thanks for adding this functionality, @dougiesquire.

Before merging we just need to decide on a few general principles that apply to many/all functions we write:

• Google or Numpy style docstrings?
• Assumed dimension names or keyword argument defaults?
• Dual xarray Dataset/DataArray functionality (and thus whether to use ds or da as a generic variable name)

[dougiesquire]

Thanks for the review Damien. I've implemented what we discussed in person:

• Use numpy-style doc strings
• Rename da to ds
• Include options args "time_name" etc and specify in the docstring whether these are dims, coords, variables or can be multiple
• Test functions using both DataArrays and Datasets

I'll go ahead and merge if you have no issues

[DamienIrving]

Thanks, @dougiesquire. This looks great.