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

Flyte

Hey all,
is there a way to view the source code of a registered workflow after it is registered?

Haha folks have asked about this, there is a way. Infact <@U0635LYB5PD> just merged something into Flyte devs also maybe <@U05FFAF4FC2> 

I do feel we can add it to the description of a Task though 

Great. Our dev ops guy didn't implement ci registration yet so it's a real life saver :smiley:

If you have ideas let’s put them together about usecases

I mean that the standard way to link between a task and it's code would be to register during ci, which will link the task version with the commit_hash. I assume that's what most users do.

Yes - that is somewhat supported already, if you add some docs link. 

<https://github.com/flyteorg/flytekit/blob/96f2b2aa6cdc68002fde163ff10eee5eb9b8bcf1/flytekit/core/base_task.py#L184|https://github.com/flyteorg/flytekit/blob/96f2b2aa6cdc68002fde163ff10eee5eb9b8bcf1/flytekit/core/base_task.py#L184>

K, thanks!
Hope I'll get an answer soon about retrieving the task's source code

<@U05U6NH0JKW> you can try this source code renderer. <https://docs.flyte.org/en/latest/user_guide/development_lifecycle/decks.html#source-code-renderer>

<@USU6W5ATA> Thx but i'm afraid this doesn't help. I'll explain my case - I have a task version that yielded 88% accuracy, which then changed and started to yield 50% in later versions. I don't have a reference to the source code of the 88% accuracy task. I can only treat this task version as a black box, which I hoped flyte can help me retrieve and deserialize/restore the code from.

Actually, I've just downloaded the tar.gz file with all the task's code, pretty straight forward. Thanks anyways guys!

Yes we should link to the fast registered tarball