Update docs for v3.3.0; remove rmessageix

[khaeru]

This PR addresses several issues related to documentation, in preparation for v3.3.0.

One improvement: it adds TeX macros like e.g. \STORAGECHARGE, which MathJax expands to e.g. \text{STORAGE_CHARGE}—these are a little more readable than the default TeX treatment of every single letter (S, T, O, R, etc.) as a distinct symbol rendered in italics. The PR does not apply this improvement to existing equations.

Also:

• Remove rmessageix, per installation instructions for rmessageix package #412 (comment), and update install instructions.
• Re-enable CI builds on Python 3.9 (parallel to Disable Sphinx builds on GHA under Python 3.6 ixmp#414).

How to review

• Build the documentation and look at:
  • Page “Sets and mappings”, new section “Index names”.
  • Page “Install”, section on R.
  • Page “Usage from R via reticulate”.
  • Page “Model core formulation”, equations STORAGE_BALANCE_INIT and STORAGE_EQUIVALENCE.
• Check (per Bars in stacked bar plots in tutorials are too wide #457 (comment)) that stacked bar charts appear reasonable in both Austria and Westeros tutorials.

PR checklist

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

[codecov]

Codecov Report

Merging #473 (549cd57) into master (96abcfd) will not change coverage.
The diff coverage is n/a.

❗ Current head 549cd57 differs from pull request most recent head b651cc9. Consider uploading reports for the commit b651cc9 to get more accurate results

@@          Coverage Diff           @@
##           master    #473   +/-   ##
======================================
  Coverage    93.3%   93.3%           
======================================
  Files          37      37           
  Lines        2819    2819           
======================================
  Hits         2632    2632           
  Misses        187     187           

Impacted Files	Coverage Δ
message_ix/reporting/computations.py	100.0% <ø> (ø)

[khaeru]

The macOS job failures here are due to r-lib/actions#313.

[khaeru]

The macOS job failures here are due to r-lib/actions#313.

This was a duplicate of r-lib/actions#310, already fixed! 🎉

[khaeru]

However, a new failure for the macOS jobs:

Run install.packages(c("reticulate", "dplyr", "IRkernel"))
Installing packages into ‘/Users/runner/work/_temp/Library’
(as ‘lib’ is unspecified)
trying URL 'https://cloud.r-project.org/bin/macosx/contrib/4.1/reticulate_1.20.tgz'
Content type 'application/x-gzip' length 1732040 bytes (1.7 MB)
==================================================
downloaded 1.7 MB

trying URL 'https://cloud.r-project.org/bin/macosx/contrib/4.1/dplyr_1.0.6.tgz'
Content type 'application/x-gzip' length 1256497 bytes (1.2 MB)
==================================================
downloaded 1.2 MB

trying URL 'https://cloud.r-project.org/bin/macosx/contrib/4.1/IRkernel_1.2.tgz'
Content type 'application/x-gzip' length 219794 bytes (214 KB)
==================================================
downloaded 214 KB


The downloaded binary packages are in
	/var/folders/24/8k48jl6d249_n_qfxwsl6xvm0000gn/T//RtmpocXD1z/downloaded_packages
Error in dyn.load(file, DLLpath = DLLpath, ...) : 
  unable to load shared object '/Users/runner/work/_temp/Library/digest/libs/digest.so':
  dlopen(/Users/runner/work/_temp/Library/digest/libs/digest.so, 6): Library not loaded: /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libR.dylib
  Referenced from: /Users/runner/work/_temp/Library/digest/libs/digest.so
  Reason: image not found
Calls: loadNamespace ... namespaceImport -> loadNamespace -> library.dynam -> dyn.load
Execution halted
Error: Process completed with exit code 1.

[khaeru]

However, a new failure for the macOS jobs:

Attempted fixes so far that haven't solved the problem:

1. Remove homebrew R per Issue with installing on MacOS r-lib/bench#109.
2. Split the job steps for install.packages() and IRkernel::installspec().
3. Also install Rcpp.
4. Also install devtools and give devtools::has_devel() per https://rstats.wtf/set-up-an-r-dev-environment.html#verify-system-prep—leads to different errors, not worth debugging.

Other sources that haven't helped: rstudio/rstudio#9134.

[khaeru]

The problem appears to have resolved itself with a rebase—perhaps because the CI cache was discarded.