One pattern we're seeing emerge is we often have a task before our map task that prepares the list of inputs like:

inputs = prepare_inputs(input_list=some_list, static_input_1 = 'value')
results = map_task(task_to_map)(i=inputs)

Where

prepare_inputs

basically just zips up the input_list with one or more static values. In non-flyte python, we use functools.partial for this. We could keep writing a helper task like

prepare_inputs

, but we were wondering if there was some more paradigmatic way of doing this in flyte?

.with_overrides

is clearly focused on execution parameters (e.g., its use with the spark task) so it doesn't seem like a good target for extending for this use case. It almost seems you like you could do something like functools.partial on the task itself in the workflow, and then add this to flytekit (would be interested in doing this if something like this doesn't already exist or is in the works, couldn't see anything in open issues)

(Lmk if I should move this to onboarding?)

(and the static input comes from another task)

i think we ran into some issue at implementation, and we did the more restrictive thing to start

but we should add it I think.

the partial concept extends to workflows kinda naturally because launch plans can have fixed inputs too

this is something we’ll need to do carefully i think but yeah, we’ve also talked about adding default inputs to tasks recently and i think this will be related.