Suggestion: Separate GINsim and biolqm conda/PyPI packages

[jonrkarr]

Typically separate software are distributed as separate packages.

Because the two are distributed as one conda/PyPI package and the GINsim Jupyter tutorial uses biolqm, the boundary between the two packages is unclear to me. Is biolqm implemented using GINsim? Or are these separate tools?

Part of my confusion stems from the fact that the Python packages expose differently functionality. GINsim exposes model converters and biolqm exposes a simulation method. This makes it feel like biolqm is a simulation tool built upon GINsim, which I don't believe is the case. This would be less confusing if the two Python packages were separated and the GINsim package also exposed GINsim's simulation capabilities.

[aurelien-naldi]

It's the opposite: bioLQM is used as an internal library of GINsim and we use a java to python bridge in GINsim to expose the whole Java API. The Python bindings thus both rely on the same GINsim jar, which contains bioLQM.

I know that it's not an ideal setup but it allows to get a bioLQM model directly from a GINsim model by exposing an internal java object

[jonrkarr]

I understand that's how its implemented, but the Python bindings and examples seem to complicate the presentation of this.