Running tasks in subprocesses

The python module bandsaw.advices.subprocess contains the SubprocessAdvice class, which allows running task computations in a subprocess. Additionally, the subprocess can use a different python interpreter than the one, that is used for running the rest of the recipe. This can be used for bringing together tasks with conflicting dependencies or different python versions.

Configuration

directory (str)

The directory, where the intermediate results and temporary data to communicate with the subprocess will be written to. Can be None, in which case a temporary directory will be used.

interpreter (bandsaw.interpreter.Interpreter)

The python interpreter to use for executing the task. If None, the same interpreter will be used, that is executing the workflow.

Example configuration

import bandsaw.advices.subprocess
import bandsaw.interpreter

configuration = bandsaw.Configuration().add_advice_chain(
    bandsaw.advices.subprocess.SubprocessAdvice(
        directory='/my/subprocess/directory',
        interpreter=bandsaw.interpreter.Interpreter(
            path=['/my/other/interpreter', '/python/path'],
            executable='/path/to/the/other/python',
        )
    )
)

---

Last update: 2021-10-28