Improve the curation level on BioSimulators

[draeger]

The website biosimulators.org collects various simulation engine and stand-alone tools for simulation in systems biology. SBSCL is among them but currently with suboptimal curation level indication.

The curation level is calculated based on how much information is provided and whether that information has been verified.

• 5: Containerized command-line interface that is consistent with SED-ML, KiSAO, COMBINE, and other conventions for simulation results, logs of simulations, command-line interfaces, Docker images, etc. Compatibility is verified by our test suite.
• 4: Docker image which has not been verified or is not consistent with these conventions.
• 3: Algorithm parameters are annotated.
• 2: Algorithms are annotated.
• 1: A record in the database.

Level 3 is easy to reach, and doesn't require any code. The easiest way to achieve level 5 is to use our Python library to build a consistent Python API and command-line program and then containerize this command-line program. This would entail implementing a callback function which can execute a single SED-ML task and return the results of the required observables as a dictionary of numpy arrays. This function could be implemented using a Python-Java bridge. All other SED-ML, KiSAO, COMBINE, etc. functionality would be handled by our library. For Java, Ion has also incorporated some of the same functionality into a fork of jLibSEDML.

Once a containerized command-line interface is submitted to BioSimulators, it would be available for simulation through runBioSimulations.

A contact person for working on this is @jonrkarr.

[jonrkarr]

If you'd like to use the improvements to jLibSedml, reach out to @moraru who's also leading BioSimulators/BioSimulations.