https://flyte.org
Join Slack
I'm getting a typing complaint while writing an integration test for a workflow: :thread:
# announcements
f
I'm getting a typing complaint while writing an integration test for a workflow: 🧵
This is broadly the test case:
Copy code
from our.codebase.task import task, OurParameterType, OurTaskResult

def test_our_task():
    @workflow
    def do_workflow(arg1: FlyteDirectory, arg2: OurParameterType) -> OurTaskResult:
        return task(arg1, arg2)

    arg1 = foo
    arg2 = bar

    ret: OurTaskResult = do_workflow(arg1, arg2)
Two typing issues occur: 1. 
no parameter named arg2
2. 
VoidPromise is incompatible with OurTaskResult
In the case of #1, is this some byproduct of the decorator? in the case of #2 should the return type of 
do_work
be something else?
OurTaskResult
is essentially a:
Copy code
@dataclass_json
@dataclass
class FlyteResult:
  result: <our dc>
Where 
our dc
is passed during module evaluation
h
@faint-smartphone-23356. in the case of #1, workflows and tasks only accept named arguments. Can you try replacing the invocation of 
do_workflow
with:
Copy code
do_workflow(arg1=arg1, arg2=arg2)
As for #2, this is a byproduct of the flytekit decorators. I think we can do something there, but will require a few changes.
f
Thanks @high-accountant-32689 - we actually are passing with keyword args - sorry, that was just me not being precise in my example
message has been deleted
and the definition:
h
got it, thanks for clarifying. And just so I understand, the test fails or this is about the types not matching (and mypy complaining)?
f
mypy compaints
or, in this case, pyright
test passes just fine
h
got it. Yeah, we'll need to make some changes to the workflow decorator to ensure mypy/pyright is happy.
Can you file an issue?
f
Will do
h
thank you!
f
@high-accountant-32689 is this a bug or housekeeping? Feels like housekeeping honestly because things work
h
I agree
👍 1
f
h
Amazing, thanks. Will ping here when we figure out priorities. In the meantime, if you want to take a crack at it, please feel free.
👍 1
166 Views
Open in Slack
PreviousNext
Powered by

Flyte

Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Powered by