Allow keeping y boundary cells at second divertor

[johnomotani]

Enables keep_yboundaries option to work in double-null topology.

Should be merged after #39 and #30.

[pep8speaks]

Hello @johnomotani! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2019-08-06 13:09:36 UTC

[codecov-io]

Codecov Report

Merging #40 into master will increase coverage by 2.14%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master      #40      +/-   ##
==========================================
+ Coverage    49.7%   51.85%   +2.14%     
==========================================
  Files           8        8              
  Lines         344      351       +7     
  Branches       63       63              
==========================================
+ Hits          171      182      +11     
+ Misses        155      153       -2     
+ Partials       18       16       -2

Impacted Files	Coverage Δ
xbout/load.py	80% <100%> (+4.39%)	⬆️

---

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 9502b3d...94b16e7. Read the comment docs.

[TomNicholas]

This is great! Easier to fix than I had imagined.

Only comment I can think of is whether it would be more internally consistent to use the _get_seps() function rather than ds['jxseps1_2'] for example? Possibly not if we haven't yet deciphered those...

[TomNicholas]

Also might want to merge the guard cells branch into this one

[johnomotani]

Only comment I can think of is whether it would be more internally consistent to use the _get_seps() function rather than ds['jxseps1_2'] for example? Possibly not if we haven't yet deciphered those...

It's a good point. After thinking a bit, it seems sensible to me to keep separate the changes here (which affect pre-processing files before they are loaded) and _get_seps (which calculates indices for slicing up the collected arrays) - even though they do very similar things. What _get_seps ends up doing (in cases with y-boundaries, in some changes that I haven't made a PR for yet) depends on what was done in the pre-processing (i.e. were y-boundary cells kept or not).

[TomNicholas]

What _get_seps ends up doing (in cases with y-boundaries, in some changes that I haven't made a PR for yet) depends on what was done in the pre-processing (i.e. were y-boundary cells kept or not).

This seems reasonable to me.

I just merged in the final changes to the keep_guard_cells branch (not entirely sure if that was necessary but oh well) - so if someone could give this a final double-check before merging that would be great.