Very large models (10.000+ variables, or ones using very complex update functions) are represented using very large SBML files which our parser can't handle at the moment.
The solution is to use a different parser that does not need to store the whole abstract-syntax-tree of the XML document in memory at the same time. However, this requires a substantial change in biodivine-lib-param-bn.
Correction: It seems that the SBML parser itself is mostly ok (it's not super fast, but its fine). However, static analysis is the real bottleneck in the pipeline.
Very large models (10.000+ variables, or ones using very complex update functions) are represented using very large SBML files which our parser can't handle at the moment.
See for example: https://www.ebi.ac.uk/biomodels/MODEL2007020001#Overview (this one is "only" 3000 variables, but still quite large)
The solution is to use a different parser that does not need to store the whole abstract-syntax-tree of the XML document in memory at the same time. However, this requires a substantial change in
biodivine-lib-param-bn
.