Sam Isaacson

Alternatively we could just use a flat vector and linear search given we don’t expect lots of metadata.

The reason for Symbolic's interface doesn't really matter here though -- what matters is that now we are using a completely different interface. (I'm not talking about the internal implementation...

@TorkelE once these are addressed we can merge.

Other things I mentioned today - consider switching the order of unknowns i.e. varying space then species instead of species then space. - using a view of a slice of...

We need to see how MOL is representing PDE solutions under the hood and tie into that ideally.

Support accessing variables via the system `lrs.A` and such, and support `getproperty`. Then fix `_symbol_to_var` to work like MTK.

Hi, thanks for the PR! We can't merge this at this time as currently any updates to master will break doc builds unfortunately (due to bugs in ModelingToolkit's V8 releases...

Happy to hear you found a workaround! I would strongly recommend you change your workflow to creating a second system with any additional reactions and then using `ModelingToolkit.extend` to merge...

See this tutorial: https://docs.sciml.ai/Catalyst/stable/catalyst_functionality/compositional_modeling/#Compositional-modeling-tooling

We've removed these mutating functions for Catalyst V14 so I'm going to close this.