docs: token classification tutorial

[sdiazlor]

Pull Request Template

Closes #5173

Type of change

• Documentation update

How Has This Been Tested

Checklist

• I added relevant documentation

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[sdiazlor]

@davidberenstein1957 After querying a dataset, the best way to iterate over the retrieved records is using .to_list. I think it's useful to change it in the how-to guides instead of the current list(). WDYT?

[github-actions]

Docs for this PR have been deployed hidden from versioning: https://argilla-io.github.io/argilla/docs_update-admonitions-token-tutorial

[davidberenstein1957]

@sdiazlor I see there is something wrong with the doc workflow. I will check what is going wrong here. I think it might be the GH actions extension we are using because the message seems to have been printed correctly within the actions workflow.

[davidberenstein1957]

@sdiazlor, thanks for handling this. I think it will offer a good basis for people to get started.

Some high level comments, will potentially leave some more in the notebook :)

• not 100% sure what you mean by list vs to_list. Ideally, we should unify communication on how we intend the API to be used. In that case the Records.to_list would be the preferred way and we should try to use this in the rest of the guides too.
• I think we should try to format cells in the notebook to adhere to line width etc, this can be done by right-clicking on a notebook and then selecting "format". Perhaps you can add ipynb files to the ruff pre.commit formatter?
• I think we can remove "Task" from the naming of both the tutorials.

[sdiazlor]

@davidberenstein1957 Thanks for your comments!

I meant that when doing this, it is difficult to iterate over the records and retrieve the needed data(fields, suggestions, responses, etc.).
filtered_records = list(dataset.records(status_filter))

Instead, as in the tutorials, it's more helpful to use.
filtered_records = dataset.records(status_filter).to_list(flatten=True)

But your answer solved the doubt, so I'll update it.

Thank you for the formatting reminder, ipynb files were not included.

[sdiazlor]

@davidberenstein1957 I applied the changes. However, it doesn't let me update the length too much, especially for the scripts due to pre-commit. We should update the length in pyproject.toml from 120 to the default 88.

[davidberenstein1957]

@sdiazlor, feel free to change it to line-width=88 in another PR :)

#5211