Ops with scalars and vectors

[znicholls]

Pull request

Please confirm that this pull request has done the following:

• Tests added
• Documentation added (where applicable)
• Example added (either to an existing notebook or as a new notebook, where applicable)
• Description in CHANGELOG.rst added

Adding to CHANGELOG.rst

Please add a single line in the changelog notes similar to one of the following:

- (`#XX <https://github.com/openscm/scmdata/pull/XX>`_) Added feature which does something
- (`#XX <https://github.com/openscm/scmdata/pull/XX>`_) Fixed bug identified in (`#YY <https://github.com/openscm/scmdata/issues/YY>`_)

[znicholls]

@lewisjared making this work with pint will require some thinking so let's release 0.6.2 without it and add later

[codecov]

Codecov Report

Merging #82 into master will increase coverage by 0.17%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #82      +/-   ##
==========================================
+ Coverage   94.72%   94.89%   +0.17%     
==========================================
  Files          15       15              
  Lines        1648     1665      +17     
  Branches      362      366       +4     
==========================================
+ Hits         1561     1580      +19     
+ Misses         59       58       -1     
+ Partials       28       27       -1     

Impacted Files	Coverage Δ
src/scmdata/run.py	93.43% <100.00%> (+0.45%)	⬆️
src/scmdata/timeseries.py	90.42% <100.00%> (+1.26%)	⬆️

---

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 b9d03f6...f63bc0b. Read the comment docs.

[znicholls]

Alright @lewisjared turns out this was easier than I thought. Can you take a look? Not sure how to better document the Pint support...?

[znicholls]

Also it turns out xarray has a bit of unit support already (e.g. pydata/xarray#3643) so we should be able to switch Timeseries across at some point in future without too many headaches.

[znicholls]

Other question, do the errors which come from using wrong shape data for ops make sense?

[lewisjared]

Brilliant!

Is it simple to add pow? I guess that is a different case as only a unitless scalar make sense. x ^ 2 [GtC / a] doesn't really make sense??

[znicholls]

Is it simple to add pow?

I don't know. I don't think it will be as simple because it should be dimensionless as you say. One for another PR/new issue?

[lewisjared]

Other question, do the errors which come from using wrong shape data for ops make sense?

Maybe not. I don't think that the ops follow broadcasting rules. The case where other.shape == (1, 3) doesn't work (with the length check removed) so maybe the message about broadcasting is a lie. Perhaps the length check should be testing for identical shapes? Technically I think a array of shape (n_times) works, but it might be simpler to only support scalars or identically shaped 2d arrays?

[znicholls]

Technically I think a array of shape (n_times) works, but it might be simpler to only support scalars or identically shaped 2d arrays?

I think so too

[znicholls]

@lewisjared this should be good to go, outstanding issues:

• remaining threads
• docs (do you want anything more?)

[lewisjared]

I think the docs could be tricky so that's a future us problem