Integrate gdpr api

[liam-sbhoo]

Development

• Add support for GDPR APIs :)
• Refactoring

Demonstration of GDPR API usage

from tabpfn_client import tabpfn_classifier, UserDataClient

tabpfn_classifier.init(use_server=True)
data_client = UserDataClient()

# all available API(s) are as followed:
data_client.get_data_summary()  # to get summary of uploaded data
data_client.download_all_datasets(save_dir)  # get a zip file of all uploaded data
data_client.delete_dataset(certain_dataset_uid)  # to delete given train set / test set
data_client.delete_all_datasets()  # to delete all uploaded datasets
data_client.delete_user_account()  # to delete the user account (incl. the datasets)

Note

I couldn't think of a cool name for "UserDataClient". Let me know if you've got anything, I will be glad to change it :)

[liam-sbhoo]

Hey Sam and Noah, please also take a look at this client PR when you have the time. Thanks :)

[noahho]

Hey Liam, thanks for the great PR!
I couldn't think of a better word for UserDataClient either..
maybe @SamuelGabriel want to look over it as well, but its looking good for me.

[SamuelGabriel]

Thanks a lot :) This looks pretty gdpr compliant 😆

I only reviewed it superficially, sorry. Maybe we could talk a bit more in depth about the codebase in our meeting next week.

One general point: I think we should very carefully think about a modularization s.t. we can share a lot of code between local and server TabPFN. But how you did it is fine for now, I think.