Parsing CellML with libcellml

[MichaelClerx]

Have started wondering what the steps would be to eventually do this

https://github.com/cellml/libcellml/issues/498

Would also involve

• [ ] Making libcellml pip installable (on linux at least) https://github.com/cellml/libcellml/issues/206
• [ ] Transpiling every 1.0/1.1 model to 2.0 with Jon's xslt https://github.com/cellml/cellml1to2
• [ ] We need access to meta data! https://libcellml-tutorials.readthedocs.io/en/combine/tutorials/combine2020/tut6_annotateModel.html

Update Hugh is working on the pip install at the moment.

[MichaelClerx]

Not pip-installable yet, but I think libcellml now has the features we need to read CellML 2.0 models

This would deal with model parsing, validation, and give us an API to extract the equations as MathML

[MichaelClerx]

Here are the tutorials presented at COMBINE 2020: https://libcellml-tutorials.readthedocs.io/en/combine/tutorials/tutorials_index.html

[MichaelClerx]

(There are also "Generator" classes for code generation, but these aren't useful for us as we want to modify things in sympy (after the CellML shell has been discarded) - unless we change it so that the cellml structure is maintained :D )

[MichaelClerx]

And some early work on annotation https://libcellml-tutorials.readthedocs.io/en/combine/tutorials/combine2020/tut6_annotateModel.html

[MichaelClerx]

There is an AST you can get from libcellml, in addition to the raw MathML, which might make transpiling a lot easier

[MichaelClerx]

The plan is for libcellml to stay relatively low level, and for a new libOpenCOR C++ library with Python bindings to be created, which will be able to do higher-level things. (OpenCOR itself will then use libOpenCOR)

[MichaelClerx]

W.r.t. annotations the plan is for libcellml to just provide access to ids, while everything else is handled via https://github.com/sys-bio/libOmexMeta