Parsing of simple mathematical models - from Model Manipulation

[skeating]

Remaining items on this section of #21

• See existing pure MathML tests - stick these in components
• With single connection
• Include a purely algebraic system as well as ODEs
• Ensure variables have units and annotations associated
  • Single 'bqbiol:is' annotation
  • Multiple 'bqbiol:isVersionOf' annotations
  • Be able to access the annotations of a variable (subject = cmeta:id, property, object)
• Ensure numbers have units associated (just to be able to support getting the units for an equation, below)
• Units conversions at connections - add new equations to model
• Ensure encapsulation is treated correctly
  • Two 'global' components can connect
  • A parent and child can connect in both directions
  • Accessing a hidden component gives error

@jonc125 will need you to expand on what this means. I'm assuming you mean create CellML models with the MathML inside.

[jonc125]

Indeed, that was the idea. Up to you how much of this you think we need, or whether we can just reuse @MichaelClerx 's test cases. To some degree basic math support (at least for model manip & code gen) will be tested just by running that protocol once everything connects up!

[MichaelClerx]

Probably better to have some unit tests as well though, no?

[jonc125]

Indeed, it's just a question of how much to prioritise now. Enough that you're confident it's doing what's needed...