Several discussions have highlighted the need to execute SED-ML tasks in parallel -- see #54. This has led to clarifying the meaning of subtask/@order -- see #95.
To facilitate parallel execution, SED-ML needs more information. To avoid complicating the core of SED-ML, this could be placed in a separate namespace similar to SBML packages.
Additional classes/attributes could be added for this information. This would enable parallel execution, for example by conversion into a CWL file and execution with a CWL engine such as Airflow or Toil or conversion to Nextflow.
Tasks
CPU requirements
Memory requirements
Walltime requirements
Preferred tool to execute the task (e.g., a URL for a Docker image)
Environment variables (e.g., for licenses to software)
Subtasks
Dependency on other subtasks
Outputs
Plot (creation) in particular needs similar information to tasks
Several discussions have highlighted the need to execute SED-ML tasks in parallel -- see #54. This has led to clarifying the meaning of subtask/@order -- see #95.
To facilitate parallel execution, SED-ML needs more information. To avoid complicating the core of SED-ML, this could be placed in a separate namespace similar to SBML packages.
Additional classes/attributes could be added for this information. This would enable parallel execution, for example by conversion into a CWL file and execution with a CWL engine such as Airflow or Toil or conversion to Nextflow.