Document optional Maxima integration backend#118
Conversation
|
That looks really cool! Thanks for sharing. Would you be interested to migrate it to JuliaSymbolics? I am just a little scared of single contributor repos from folks that don't have any other Github history: those tend to get work up front and then just sit without updates after the initial creation. If you would like help with the long term maintenance, we'd be happy to help if moved in here. There probably isn't too much but language bindings are one thing that can be nasty for long term. |
|
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #118 +/- ##
===========================================
+ Coverage 13.91% 50.98% +37.07%
===========================================
Files 22 23 +1
Lines 4291 4309 +18
===========================================
+ Hits 597 2197 +1600
+ Misses 3694 2112 -1582 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
|
cool thank you @Amin-El-Sayed ! you could copy the file test/rundifficulttests.jl in your repo and sligtly modify it to benchmark the Maxima method against the rule based and risch methods. should be a minimal change but give us a lot of insights |
Thanks, that sounds great. I would definitely be interested in moving it under JuliaSymbolics. What transfer process would you prefer? If you want, I can transfer the existing SymbolicIntegrationMaxima.jl repository to JuliaSymbolics and then update links/docs/registry metadata as needed. |
|
Making it a part of the lib folder in SymbolicIntegration, like how OrdinaryDiffEq.jl does its sublibraries, is the way to keep it alive the most, because then anyone who works on this repo is forced to look at it from time to time. That's what I'd recommend above anything else, but migrating the repo is a second option. |
|
That sounds like the best path to me. I am happy to move it into SymbolicIntegration.jl as a lib subpackage. Before I prepare the PR, could you confirm the intended layout? I would assume: lib/SymbolicIntegrationMaxima/ with the existing package name and UUID kept unchanged. If that is the right structure, I can prepare the migration PR and adapt the difficult-test comparison there so it reuses SymbolicIntegration.jl's existing test corpus instead of duplicating benchmark data in the external repo. |
|
Just one runtests.jl. https://github.com/SciML/OrdinaryDiffEq.jl/tree/master/lib/DelayDiffEq is an example |
This adds a short documentation note for external integration methods, with SymbolicIntegrationMaxima.jl as an example optional backend.
The change intentionally keeps the built-in/default behavior unchanged: