Report CO₂ emissions from transportation

[khaeru]

This PR adds code that mirrors the behaviour of message_data (‘legacy’) reporting, as an exploration of how to exactly replicate these calculations using genno. For this purpose, the code only replaces the portions of message_data.tools.post_processing.default_tables.retr_CO2emi that calculate CO₂ emissions from transportation.

The goals are:

• The magnitudes and labels on the data are the same (to within a small tolerance) as those produced by message_data. (However, per this Slack thread, there does not exist a legacy reporting output for the snapshot, and it would be difficult to produce one due to changes to the units.)
• The code is legible, and it is clear how the genno-based code mirrors the logic of the older code.

Some things specifically omitted, that could be tackled in follow-up PRs:

• Apply units to e.g. out:* earlier, in a single operation for all technologies that share the same units of ACT and output.

Housekeeping

• Adjust usage of genno.Operator.define to avoid mypy complaints in "Code quality" CI job (FYI @glatterf42)

How to review

See:

• The added docs.
• transport-emissions-full-iamc.svg —genno visualization of the calculation of "transport emissions full::iamc".
• transport-emissions-full.csv —output.

PR checklist

• Continuous integration checks all ✅
• Add or expand tests; coverage checks both ✅
• Add, expand, or update documentation.
• Update doc/whatsnew.

[codecov]

Codecov Report

Merging #134 (519bdfc) into main (c9f1638) will increase coverage by 1.8%.
The diff coverage is 99.2%.

@@           Coverage Diff           @@
##            main    #134     +/-   ##
=======================================
+ Coverage   71.5%   73.3%   +1.8%     
=======================================
  Files         80      82      +2     
  Lines       5439    5663    +224     
=======================================
+ Hits        3889    4152    +263     
+ Misses      1550    1511     -39     

Files	Coverage Δ
message_ix_models/model/config.py	100.0% <100.0%> (ø)
message_ix_models/report/__init__.py	80.0% <100.0%> (+2.2%)	⬆️
message_ix_models/report/sim.py	77.9% <ø> (-1.2%)	⬇️
message_ix_models/tests/report/test_compat.py	100.0% <100.0%> (ø)
...essage_ix_models/tests/report/test_computations.py	100.0% <100.0%> (ø)
message_ix_models/tests/test_report.py	92.7% <100.0%> (+0.7%)	⬆️
message_ix_models/util/scenarioinfo.py	100.0% <100.0%> (ø)
message_ix_models/report/compat.py	99.2% <99.2%> (ø)
message_ix_models/report/computations.py	94.7% <90.0%> (+42.7%)	⬆️

... and 2 files with indirect coverage changes