Not converting lumped to distributed kinetics breaks many things

moraru commented 1 year ago

While fixing orphan overrides related issues in PR #607 coercing lumped kinetics into distributed kinetics was dropped during import. This creates several problems:

(minor) import behavior and variable naming is inconsistent between GUI and CLI
(major) every single biomodel that has both a compartmental and a spatial application will no longer roundtrip into a single biomodel, since the reactions in the underlying model and reactions are no longer the same
(major) import and execution will fail for many stochastic models originated from vcell and for most stochastic models originating from other

jcschaff commented 1 year ago

The trick is to apply the distributed transformation right after the sedml override application... I think.

moraru commented 1 year ago

Because successful application of math overrides can only happen after simcontexts are created, coercing must be done ahead of time (unless the entire sedml processing is being overhauled).

moraru commented 1 year ago

coercing while in the import phase will break the fix for resolving constant rate parameters as implemented in #607 this now has to be pushed onto the sbml importer

moraru commented 1 year ago

Fixed by 6de113f6631e7708081fc163a98f5ee2f67f0139 and d4e52b9923ca157b14277073760663a35843e1b0 and be6bc17c502110d7e26b049051a17d4b63ba0901

moraru commented 1 year ago

deployed by PR #619

virtualcell / vcell

Not converting lumped to distributed kinetics breaks many things #615