You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We already have the technical capabilities in https://github.com/pyiron/conda_subprocess which is integrated in pyiron. If the other frameworks would also support executing Python functions in separate conda environments, then we could add a corresponding parameter in the node definition. Alternatively, for the workflow frameworks which do not provide the option to execute Python functions in a separate conda environment we could dynamically wrap the Python functions to add this functionality, but it might interfere with the decorator of the specific workflow framework.
The text was updated successfully, but these errors were encountered:
@jan-janssen I have been discussing this question here with @gpetretto: Matgenix/jobflow-remote#303
To me, this is again looks like something jobflow would solve at the executing step and not the definition step of the workflow, i.e., we would assign conda environments in jobflow-remote after we have defined the workflow in jobflow.
We already have the technical capabilities in https://github.com/pyiron/conda_subprocess which is integrated in pyiron. If the other frameworks would also support executing Python functions in separate conda environments, then we could add a corresponding parameter in the node definition. Alternatively, for the workflow frameworks which do not provide the option to execute Python functions in a separate conda environment we could dynamically wrap the Python functions to add this functionality, but it might interfere with the decorator of the specific workflow framework.
The text was updated successfully, but these errors were encountered: