Refactor variables to use TulipaVariable

[abelsiqueira]

Create field lookup to allow changing the variables without
having to redefine the whole model creation

Related to #923

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.76%. Comparing base (a286c75) to head (d4c426d).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #925      +/-   ##
==========================================
+ Coverage   97.85%   98.76%   +0.90%     
==========================================
  Files          29       30       +1     
  Lines        1214     1210       -4     
==========================================
+ Hits         1188     1195       +7     
+ Misses         26       15      -11     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[github-actions]

Benchmark Results

	a286c75...	d4c426d...	a286c75.../d4c426d86e13ec...
energy_problem/create_model	48.1 ± 1.4 s	46.5 ± 2.2 s	1.03
energy_problem/input_and_constructor	10.1 ± 0.84 s	10.4 ± 0.33 s	0.964
time_to_load	3.98 ± 0.13 s	3.87 ± 0.046 s	1.03

	a286c75...	d4c426d...	a286c75.../d4c426d86e13ec...
energy_problem/create_model	0.66 G allocs: 22.6 GB	0.66 G allocs: 22.6 GB	1
energy_problem/input_and_constructor	0.0327 G allocs: 1.38 GB	0.0327 G allocs: 1.38 GB	1
time_to_load	0.157 k allocs: 11.1 kB	0.157 k allocs: 11.1 kB	1

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).

[abelsiqueira]

@datejada @gnawin I translated most investment variables into SQL, but I can't do the same for assets_decommission_compact_method. If you have an idea, it would be helpful.

The main issue is that the variable apparently is not a subset of the rows of assets_data, so I don't have a good intuition on how to start.

[datejada]

Thanks @abelsiqueira, @gnawin and I reviewed this query and it is almost there. So, two things:

1. We tested it in the multi-year investment data, and it works. However, we want to cover another edge case outside that use case (so, in issue Create a meaningful year data for the Norse Case #684 I added that to the list to improve it). If we change the input to have the asset = ccgt in year = 2030 and commission_year = 2030 with investable = false, then it should not appear in the result of this query. As it is, it seems. Ni and I tried to fix it but couldn't (sorry about that). If we cover that case, then this change is okay.
2. For the indices of the decommission variables, we don't need the investable column (we need the investment integer for creating the variable). So, we were wondering, why do you keep it? Maybe we are missing something. If we don't need it, then we should take it out.

[abelsiqueira]

1. Can we add that to the actual tests? Otherwise there is guarantee that it works as expected.
   I also don't see the relation between this new data row and the compact formulation. There are other rows that result in investable = false in the query, so investable is not taken into account

2. I was just using for checking that things propagate

[datejada]

Hi @abelsiqueira, thanks for the PR. We have a comment on the decommission variable indices. Please let us know if you need to discuss it, and we can comment on Teams.

[datejada]

LGTM. Thanks!