Python API - copy dataset

[smidm]

How to copy or clone a dataset? I would like to use Dataset.filter, which is done in-place and continue working with the original dataset.

[vinnamkim]

Hi @smidm, You can use deepcopy such as

import datumaro as dm
from copy import deepcopy

dataset = dm.Dataset.import_from(...)
cloned = deepcopy(dataset)

[smidm]

This is how I did it. I'm bit uncertain of the outcome when using copy module on complex objects. In many cases it fails when the object has db connections, file handles, ... I'd propose to add copy() or clone() method to the Dataset object to make it easier for users.

[vinnamkim]

Thanks, we will work on your suggestion.

[smidm]

this is chatgpt generated content without context of Dataset

[vinnamkim]

Sorry but I think this answer can confuse our community users, so I hide it.