Create summary table of whether omex file runs locally in different simulators

[stellaprins]

Create overview table like the one below for locally ran models (different than the one tested below).

• [ ] run directly with tellurium (maybe other options down the line)
• [ ] run with biosimulators docker locally (on all the different engines)
• [ ] run with biosimulators remotely (on all the different engines) pynml API does this currently (check this)

Use helper functions that Rob wrote.

Down the line:

• [ ] incorperate process (for creating the table) in github actions and automate it as a test.

The resulting minimal omex file SBML\LEMS_NML2_Ex9_FN.omex biosimulations tool was used to run simulations on https://run.biosimulations.org/

Simulator: tellurium 2.2.8 (Sensible output was produced) Simulator: copasi 4.42.284 Failed Simulator: amici 0.18.1 (No output was produced) Simulator: vcell 7.5.0.127 (No output was produced) Simulator: pysces 1.2.1 Failed Simulator: gillespy2 1.6.6 (No sensible output was produced, flat line) Simulator: libsbmlsim 1.4.0 (Sensible output was produced) Simulator: masspy 0.1.6 Failed

[robertvi]

note: tellurium has some support for using combine/omex input files (https://tellurium.readthedocs.io/en/latest/notebooks.html#combine-inline-omex)

[sanjayankur31]

https://github.com/NeuroML/pyNeuroML/blob/development/pyneuroml/biosimulations.py#L167 -> function to simulate an omex archive on biosimulators using its API

[stellaprins]

Github actions is failing, Rob and myself have been looking at it but have not found a solution yet. Any clues what is going on @sanjayankur31 or @pgleeson? We think it may be more related to a conflict in package version than changes to the code.

ImportError: cannot import name 'VisibleDeprecationWarning' from 'numpy' (/opt/hostedtoolcache/Python/3.9.19/x64/lib/python3.9/site-packages/numpy/__init__.py) (see for example here)

ValueError: numpy.dtype size changed, may indicate binary incompatibility. Expected 96 from C header, got 88 from PyObject (see for example here)

[sanjayankur31]

Hello,

Yes, looks like the version of Brian that's being pulled in is 2.5.3, which doesn't seem to have the necessary fixes for numpy2.x support, which were made in 2.6.0:

https://github.com/brian-team/brian2/commit/e8f05c6461d7c530ca26bdbab474a1891d9036b0

OMV uses 2.5.3 as the default version from the looks of it:

https://github.com/OpenSourceBrain/osb-model-validation/blob/e2b060dbcb3d82a1236cd14896b5c1db6a8476ab/omv/engines/getbrian2.py#L6

Numpy2.x was only released yesterday, so we'll have to update omv to work with that.

I think a current workaround is to explicitly install numpy<=1.2.3, as has been done here:

https://github.com/OpenSourceBrain/PotjansDiesmann2014/commit/25059f533e880db4d5aac3ed4b11072ba5f42bcc

See also: https://github.com/OpenSourceBrain/osb-model-validation/issues/91

[sanjayankur31]

The other error is probably also because pytables is not ready for numpy 2.x, so pinning numpy should fix that too.

Requirement already satisfied: numpy in /opt/hostedtoolcache/Python/3.9.19/x64/lib/python3.9/site-packages (from libNeuroML>=0.5.5->pyneuroml) (2.0.0)

[stellaprins]

Your first suggestion appears to have solved the github actions problems! Thanks a lot @sanjayankur31.

[sanjayankur31]

FYI: json files including format information:

• information about simulators: biosimulators.json; the "modelFormat" field for each simulator links to an EDAM id in the ontologies file:
• the ontologies file where the formats etc have uniq EDAM ids: ontologies.json