Add delete and list methods for FeedbackDataset in Argilla via the Python client

[frascuchon]

Since we can now delete workspaces, we need to tackle two main goals to improve this feature when datasets are linked to the workspace:

• Support for removing a feedback dataset from the Python client

...
ds = FeedbackDataset.from_argilla(...)
ds.delete()
# or just (since the from_argilla will try to fetch all records locally
FeedbackDataset.delete(dataset_id)

• Fetch the infor about which datasets are linked to the workspace

As a first approach, we could return the needed info as part of the delete error response. The name for "old" datasets, and the id for feedback ones

Listing and deleting old task datasets

import argilla as rg

ws = rg.Workspace.from_name("ws")

# listing old datasets for a workspace
datasets = rg.list_datasets(workspace=ws)
# or for all workspaces
datasets = rg.list_datasets()

# deleting old datasets
for dataset in datasets:
    rg.delete(name=dataset.name, workspace=datase.workspace)

Listing and deleting new feedback datasets

from argilla.feedback import FeedbackDataset

datasets = FeedbackDataset.list(workspace=ws)
# or all workspaces
datasets = FeedbackDataset.list()
for dataset in datasets:
    dataset.delete()

[alvarobartt]

Now users can call delete() over an existing FeedbackDataset in Argilla, as well as deleting records from Argilla via delete_records or in-place deletion of records via record.delete().

But listing FeedbackDataset datasets is still pending, we just need to decide whether a list() method is the best approach to tackle it or not.

[alvarobartt]

Re-opened, as the FeedbackDataset.list() method is still missing