Something that might be useful for developers would be to have the ability to specify a category for splitting up tracing so specific components can be analyzed. Something similar to scope, but that can be applied at the user's discretion outside of scope's particular use with the asynchronous variants of tracing. For example, if I want custom categories to differentiate between methods that interface directly with the scheduler versus methods that write out the necessary scripts:
@critical_path(category="sched_interface")
def submit(script, ...):
sched.submit(script)
@critical_path(category="sched_script")
def write_script(cmd, path, ...):
adapter.write(cmd, path)
Something that might be useful for developers would be to have the ability to specify a
categoryfor splitting up tracing so specific components can be analyzed. Something similar to scope, but that can be applied at the user's discretion outside ofscope's particular use with the asynchronous variants of tracing. For example, if I want custom categories to differentiate between methods that interface directly with the scheduler versus methods that write out the necessary scripts: