Fix 2 reporting bugs

[khaeru]

• Close Calling Reporter.from_scenario() causes MissingKeyError #988.
• Close Add out_hist and in_hist to Reporter #989.

How to review

• Read the diff, in particular the added tests and docs (section on "Historical and reference values").
• Note that the CI checks all pass.

PR checklist

• Continuous integration checks all ✅
• Add or expand tests; coverage checks both ✅
• Add, expand, or update documentation.
• Update release notes.

[codecov]

Codecov Report

❌ Patch coverage is 95.78947% with 8 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.3%. Comparing base (ae12261) to head (a7440e2).
⚠️ Report is 17 commits behind head on main.

Files with missing lines	Patch %	Lines
message_ix/testing/__init__.py	71.4%	6 Missing ⚠️
message_ix/tests/test_report.py	98.4%	2 Missing ⚠️

Additional details and impacted files

@@          Coverage Diff           @@
##            main    #991    +/-   ##
======================================
  Coverage   92.2%   92.3%            
======================================
  Files         60      60            
  Lines       5155    5260   +105     
======================================
+ Hits        4756    4855    +99     
- Misses       399     405     +6     

Files with missing lines	Coverage Δ
message_ix/common.py	100.0% <100.0%> (ø)
message_ix/report/__init__.py	100.0% <100.0%> (ø)
message_ix/report/operator.py	97.6% <100.0%> (+0.3%)	⬆️
message_ix/tests/test_tutorials.py	96.4% <ø> (ø)
message_ix/tests/test_report.py	99.0% <98.4%> (+0.3%)	⬆️
message_ix/testing/__init__.py	98.3% <71.4%> (-1.7%)	⬇️

[khaeru]

@Tyler-lc @junukitashepard FYI this PR will fix the two reporting issues we discussed last week. Per the PR checklist, I still need to update/expand the documentation to cover the features. You can wait until I do that and request review, or if curious you can check the tests to see the expected behaviour.

[Tyler-lc]

I had a look at the diff and now it seems that historical and reference activities will be calculated with yv==ya right? But no choice regarding the method, at least for now. Which is fine by me. Just wanted to make sure that is the case.

[khaeru]

I had a look at the diff and now it seems that historical and reference activities will be calculated with yv==ya right? But no choice regarding the method, at least for now. Which is fine by me. Just wanted to make sure that is the case.

Actually there is a choice. I've now updated the docs; please see the preview build here which is supposed to explain. In brief:

1. in:*:historical+current is the key to use for yV==yA.
2. in:*:historical+weighted is the key to use a weighted average across all historical vintages yV<=yA.
   • The docs explain that these shares have to be provided by the user.

The test (that was already on the branch) checks that (a) an exception is raised if the user does not provide share data and (b) the calculation succeeds if they do.

[junukitashepard]

Thanks @khaeru. I have tested the updates for historical out and it works as intended.

[junukitashepard]

I tested this historical output for my work and the update works as intended.