Decouple core functionality from use with Jupyter notebooks

[jonrkarr]

CoLoMoTo Docker is a start to providing access to multiple logical modeling tools through Python. The usability of the tools outside Jupyter (i.e. other Python codes) would be enhanced by factoring the Jupyter notebook functionality from the core simulation capabilities.

[pauleve]

This was the idea indeed. Most of the Python modules have a Jupyter sub-modules which is loaded only within IPython for extra functionalities. Do you have examples where the behavior is not appropriate outside a Jupyter/IPython console?

[jonrkarr]

No need to respond to all of these issues. I'm starting with the GINsim Python package and using GitHub issues to share things I notice.

One benefit of focusing on a more standardized interface to logical modeling tools (through SED-ML/KiSAO or otherwise) is that this will lessen the need for code documentation for each unique API. Instead, you could provide higher-level domain-specific documentation (e.g., what KiSAO terms are needed to specific which algorithms).

The GINsim package seems to works fine outside of Jupyter. But it loads extra modules that aren't needed. It also prints out extra status information related to Jupyter.

With more standardization around simulation capabilities, the Jupyter notebook funcationality and visualization could be factored out such that it could be coupled with any standardized simulation tool.