Df ops

[gidden]

Description of PR

Originally in #276, @znicholls wrote:

This PR supercedes #229 and adds a subtraction operation for IamDataFrames. It is specifically very limited in scope as there are likely many kinks to iron out. Further PRs can add extra operations, other types etc.

This PR starts as rebased on #276 and adds additional ops.

Please confirm that this PR has done the following:

• Tests Added
• Documentation Added
• Description in RELEASE_NOTES.md Added

Adding to RELEASE_NOTES.md (remove section after adding to RELEASE_NOTES.md)

Please add a single line in the release notes similar to the following:

- (#XX)[http://link-to-pr.com] Added feature which does something

[coveralls]

Coverage increased (+0.2%) to 85.804% when pulling c0f8770 on df-ops into 7d46b2c on master.

[coveralls]

Coverage increased (+0.2%) to 85.857% when pulling a2dcf61 on df-ops into 7d46b2c on master.

[danielhuppmann]

thanks for starting another push at that @gidden and @znicholls - but in light of the comments by @gaurav-ganti in #332, I'd revisit the approach of subtracting one IamDataFrame from another, and instead have the work with one IamDataFrame. Should simplify things...

[gidden]

Ok, everyone. Please see here a revised implementation. Notably, all ops specific code now lives in ops.py. There are corresponding (minimal) tests in test_ops.py. I tried to keep this interface as general as possible so we can separate the discussion to how IamDataFrame.subtract() should look vs. "how should we do subtraction".

This is hopefully flexible enough that we can implement both ways and debate the relative benefits over beer =)... noting that have at the moment chosen to implement my desired way!

[danielhuppmann]

I (still) don't agree with this approach, but don't want to stand in the way of you moving forward on this. So I change my review to Comment and will step from further discussions.

[danielhuppmann]

a few comments on the implementation (my general reservation about the within-vs.-across-dataframe operation notwithstanding)...

[gidden]

back to _inplace per offline discussions @danielhuppmann

[znicholls]

This is how we handle ops in scmdata if helpful https://github.com/openscm/scmdata/blob/master/src/scmdata/ops.py

[gidden]

Finally (and happily) closing in favor of #527