Jug Shell

The jug shell subcommand opens up a shell within the environment of the Jugfile. It can be used for debugging and exploration of the task structure.

To obtain a list of all tasks seen by jug you can run tasks = get_tasks(). As a convenience function, you can also use get_filtered_tasks() to obtain filtered versions:

all_tasks = get_tasks()

loadable = get_filtered_tasks(loadable=True) failed = get_filtered_tasks(failed=True)

This returns only tasks that fulfil the filtering criteria used

Added in version 2.3: get_filtered_tasks was added in version 2.3. In previous versions, you had to filter the output of get_tasks() yourself

Inside the environment, you can use the value(task) function to load results (if available; otherwise, an exception is thrown).

The invalidate(task) function invalidates the results of its argument and all dependents, recursively, much like the jug invalidate subcommand.

You can call methods on Task objects directly as well:

  • run(): run the task and return the result even if the task has run before.

  • can_run(): whether all dependencies are available

  • can_load(): whether the task has already run

  • invalidate(): remove the result of this task (not recursive, unlike the invalidate subcommand, unlike the invalidate function).

You can access the original function with the .f attribute as well.


Jug shell requires IPython to be installed.