Hi team. Happy Friday everyone. First post here, so, first, thanks everyone for the work you put into the project! Quick question about this error:

E   metaflow.decorators.DuplicateFlowDecoratorException: Flow already has a decorator 'named_env_base'. You can specify each decorator only once.

• I have two flows in the same directory (say

flow_A

and

flow_B

). • Both use the same line

@named_env_base(name="my_project:latest")

to set up the conda env for the flow. • In each one of those files I also have a config data class. • Outside of those files, I import those config classes and instantiate those configs to run some tests. When importing the config from

flow_A

, all goes well. That is the flow that first had the

@named_env_base(name="my_project:latest")

decorator. But when importing the config dataclass from

flow_B

, I get the error at the start of this message. ◦ This is not about

flow_B

itself: if I import first from

flow_B

, the same error occurs when tryng to do the second import, now from

flow_A

. Is this intentional? _Are people not supposed to import two files containing flows with the same conda/named_env decorator?_ Why? If this is considered a bug, I'm happy to open an issue. For now I think I'll move the configs out of the flow files, but I really liked having the configs and the flows in the same file.

Did you update to the latest metaflow? I think this was fixed.

Not latest, but recent, 2.12.2. These is my metaflow and netflix extensions versions:

metaflow                  2.12.2             pyhd8ed1ab_0    conda-forge
metaflow-netflixext       1.2.0              pyhd8ed1ab_0    conda-forge

I already did a workaround, but I'll see if I can reproduce the issue in 2.12.7.

This is the fix: https://github.com/Netflix/metaflow/pull/1886; it was first released in 2.12.4

That's awesome. Thank you, @dry-beach-38304!