Improve handling of scalars in NetCDFTimeConverter.convert

[spencerkclark]

This provides a simple fix for #45.

[coveralls]

Coverage remained the same at 91.228% when pulling 8c9362f on spencerkclark:convert-scalar into 531dd0d on SciTools:master.

[bjlittle]

@spencerkclark Once #66 lands (hopefully), you'll need to rebase, then I'll review and merge 👍

[bjlittle]

@spencerkclark Time to rebase 👍

Then let's get this finally banked 😃

[codecov-commenter]

Codecov Report

Merging #46 (fe10e4c) into master (865dc22) will decrease coverage by 0.06%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #46      +/-   ##
==========================================
- Coverage   89.14%   89.08%   -0.07%     
==========================================
  Files           1        1              
  Lines         175      174       -1     
  Branches       42       42              
==========================================
- Hits          156      155       -1     
  Misses         13       13              
  Partials        6        6              

Impacted Files	Coverage Δ
nc_time_axis/__init__.py	89.08% <100.00%> (-0.07%)	⬇️

---

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 865dc22...fe10e4c. Read the comment docs.

[spencerkclark]

Awesome, should be all set @bjlittle -- I also removed an errant print statement that seemed to have crept in during #66.

[bjlittle]

I'll get to this, this evening 👍

Awesome!

[bjlittle]

@spencerkclark Perfect, thanks 🍻