https://flyte.org
Join Slack
When registering a custom `TypeTransformer` w/ the `TypeEngine`, does the module need to be explicit...
# flytekit
e
When registering a custom 
TypeTransformer
w/ the 
TypeEngine
, does the module need to be explicitly imported by one of the files containing workflows or tasks in order for it to be picked up? Based on the example of the 
NumpyArrayTransformer
it seemed like it, but not sure where that needs to be declared
For context, I was attempting to add a transformer to support 
collections.abc.Sequence[T]
as a valid type, and was previously hitting this error:
Copy code
ValueError: Generic Type <class 'collections.abc.Sequence'> not supported currently in Flytekit.
t
yes, it does… because we figured if people were writing a type transformer, they’d also be writing a type.
and when the user imported the type, then that would trigger the transformer loading as well.
but this doesn’t include the case where the type and the transformer are in different libraries
👍 1
to mitigate this issue you can use flytekite’s auto loading plugin behavior.
that is, if you add this entrypoint to your setup.py, then any 
import flytekit
would trigger it as well.
does that make sense.
e
Yep that makes sense -- we'll need to figure out a slightly different way to do that since we're using Bazel
https://github.com/flyteorg/flyte/issues/2823 (cc @freezing-airport-6809 @high-accountant-32689 based on our earlier discussion)
h
Great, thank you. We're going to prioritize this for the upcoming release.
🙏 1
159 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