Multiple subclassing levels required to use LightningDataModule in LightningCLI

[jasonkena]

Bug description

I get the following error message

  Import path data.snemi.SNEMIDataModule does not correspond to a subclass of LightningDataModule

with yaml config

data:
  class_path: data.snemi.SNEMIDataModule

when defining SNEMIDataModule as follows

class SNEMIDataModule(LightningDataModule):
    ...

this problem is solved by creating a dummy subclass:

class DummyDataModule(LightningDataModule):
    pass
class SNEMIDataModule(DummyDataModule):
    ...

Is this intended behavior?

What version are you seeing the problem on?

v2.2

How to reproduce the bug

No response

Error messages and logs

# Error messages and logs here please

Environment

Current environment

#- Lightning Component (e.g. Trainer, LightningModule, LightningApp, LightningWork, LightningFlow):
#- PyTorch Lightning Version (e.g., 1.5.0):
#- Lightning App Version (e.g., 0.5.2):
#- PyTorch Version (e.g., 2.0):
#- Python version (e.g., 3.9):
#- OS (e.g., Linux):
#- CUDA/cuDNN version:
#- GPU models and configuration:
#- How you installed Lightning(`conda`, `pip`, source):
#- Running environment of LightningApp (e.g. local, cloud):

More info

No response

cc @carmocca @mauvilsa

[jasonkena]

Hmm, the problem seems to be related to using lightning.LightningDataModule instead of pytorch_lightning.LightningDataModule

[awaelchli]

Hi @jasonkena

Mixing imports would definitely not be good. Can you make sure to use consistently only lightning imports everywhere in your program?

For further help, can I ask you to fill out the section of the bug report with the versions of packages please, including jsonargparse?