Fix issue #20: when sanity_check_hierarchy is called with only Emissions|CO2 input data

[jkikstra]

Note: follows from #19 and accompanying PR #22 - so let's first merge that one (if not based on that, it is not possible to reproduce the original error that we address here).

For full description of error, see #20.

• Tests added
• Documentation added/changed if necessary
• Example added (in the documentation, to an existing notebook, or in a new notebook)
• Description in CHANGELOG.rst added (single line such as: (`#XX <https://github.com/iiasa/climate-assessment/pull/XX>`_) Added feature which does something)

[jkikstra]

Wrote an ad-hoc fix, but not so sure if this works in the desired way for all cases.

Using input file , we get the input co2_infill_db as object passed to sanity_check_hierarchy(): ar6_minimum_emissions_co2_infill_db.csv
This has "Emissions|CO2" for both scenarios, even though the input GCAM 5.3 - NGFS2_Current Policies scenario only has "Emissions|CO2|Energy and Industrial Processes"

Maybe @znicholls would like to have a look at the desired behaviour here?

[znicholls]

Maybe @znicholls would like to have a look at the desired behaviour here?

I've had a quick look. I'd say let's sort out #19 first, then we can add the test you have added here and work out how to pass it. It is a confusing maze of stuff in all these checks (e.g. I have no idea why we return total CO2 from the infilling database in run_infilling) so I think we'll have to just do something fairly ad-hoc to workaround it, then we deal with it properly as part of a larger refactoring (raises again the question of whether to invest more serious time into such an effort).