compiler: Concretize SubDimensions to same object across repeated calls to concretize_subdims

[ZoeLeibowitz]

Ensure that SubDimensions are consistently concretized to the same object across multiple calls to the function. This is important when using rcompile on equations with SubDimensions.

TODO: Enhance SymbolRegistry by making it rebuildable, enabling operations such as rcompile on specific subsets.

[codecov]

Codecov Report

Attention: Patch coverage is 89.47368% with 2 lines in your changes missing coverage. Please review.

Project coverage is 87.29%. Comparing base (71e7eda) to head (6835b2b).

Files with missing lines	Patch %	Lines
tests/test_builtins.py	50.00%	2 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2509   +/-   ##
=======================================
  Coverage   87.28%   87.29%           
=======================================
  Files         238      238           
  Lines       45703    45716   +13     
  Branches     4057     4057           
=======================================
+ Hits        39892    39906   +14     
- Misses       5126     5127    +1     
+ Partials      685      683    -2     

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

[mloubout]

Not completely sure carrying around an extra mapper is the best way but don't have a better option top of my head without lots of rework

[mloubout]

If I remember correctly this one behaves a bit differently and was leading to issues in the tests but if ci is green that's fine. The requirements need to be update though to scipy version with that import

[EdCaunt]

CI now fails with the old version. See here

[EdCaunt]

Looks good to me. As @mloubout mentioned, we should probably revisit and rework this at some point in the future, but it is a relatively minor intervention for now

[mloubout]

I think it's too restrictive, it's a very new version, would wrap the import in try/except to handle different version import path

[georgebisbas]

ah sorry jsut saw that

[georgebisbas]

I think it is better to add an upper bound:

e.g. <=

and then (if needed) do what @mloubout did here (?):

dee09dd

[georgebisbas]

keep lower bound ofc

[FabioLuporini]

Not completely sure carrying around an extra mapper is the best way but don't have a better option top of my head without lots of rework

We discussed this in yesterday's meeting, @mloubout .

My suggestion was to rather use sregistry. And, in fact, we could (should) use sregistry also in place of that mapper we are currently carrying along throughout the whole pass