search

scidash / neuronunit

A package for data-driven validation of neuron and ion channel models using SciUnit
http://neuronunit.scidash.org
38 stars 24 forks source link

Use dynamic binding of methods, to make methods available, while avoiding re-classing pyNN backends as NU backends. #193

Open russelljjarvis opened 5 years ago

russelljjarvis commented 5 years ago

Proof of concept is demonstrated here: https://github.com/russelljjarvis/neuronunit/blob/dev/neuronunit/unit_test/dynamic_bind.ipynb

Any ideas about how to bind methods (and capabilities), from the reduced model class into the binding method in the notebook.

I think this scheme gives users of NU a lot more flexibility, when they are starting with off the shelf rigorous/standardised models.

russelljjarvis commented 5 years ago

This approach seems to work okay for non parallel tests. PyNN contains NEURON, which is unpickable, so may as well read from a backends file.

russelljjarvis commented 5 years ago

Is there some way we can enhance or extend their product? Perhaps it's more limited than ours ... ?

Actually, I think it might be limited to certain types of publications too. So I think there is worth in our approach.

Maybe we can extend our work, by enabling a query of their database: https://github.com/scottyih/SemanticScholarAPI/blob/master/citingPapers.py

rgerkin commented 5 years ago

@russelljjarvis Your last reply is in the wrong issue thread. Can you delete and move it to the one for the COA tool?