Fix/ Venue: withdrawn and desk-rejected visibility

[celestemartinez]

Fixes #1992

I am always adding readers to authors/authorids in the content and readers to withdrawn/desk-rejected submissions.
I then check withdrawn_submission_public/desk_rejected_submission_public to see if the readers should be everyone instead. I don't think we should check what the readers are in the submission_stage because these can change at any time.

This PR also creates a /-/Deletion invitation so authors can delete their submissions between the abstract and the submission deadline.

[melisabok]

it seems the tests don't finish because one test is waiting for the process NeurIPS.cc/2023/Track/Datasets_and_Benchmarks/-/Withdrawal-0-0 to finish, probably this is not being run.

[celestemartinez]

it seems the tests don't finish because one test is waiting for the process NeurIPS.cc/2023/Track/Datasets_and_Benchmarks/-/Withdrawal-0-0 to finish, probably this is not being run.

Thanks for the heads-up! This should be fixed.

Note that now the default will be to have withdrawn/desk-rejected and their author names hidden (so I adding readers to notes and authors/authorids by default). Only if PCs explicitly choose in the request form to make public will they be public. This means that papers that were initially single blind could potentially have their author names hidden (this was not allowed in API1 because in this case we wouldn't create blind submissions). If this sounds good, then I can remove this check from the request form preprocess:

if 'Double-blind' not in request_form.content['Author and Reviewer Anonymity']:
    if 'No' in note.content.get('withdrawn_submissions_author_anonymity', ''):
        raise openreview.OpenReviewException('Author identities of withdrawn submissions can only be anonymized for double-blind submissions')

    if 'No' in note.content.get('desk_rejected_submissions_author_anonymity', ''):
        raise openreview.OpenReviewException('Author identities of desk-rejected submissions can only be anonymized for double-blind submissions')

[melisabok]

I agree that the default option should be hiding the data including single blind venues. Let's remove that check.

[melisabok]

do we still need this change?

[celestemartinez]

Removed.

[celestemartinez]

I am checking if the invitation has a content here, if not then the date process does not try to update the note.

[melisabok]

looks good

[melisabok]

Adding Emilia as a reviewer so she can be aware of these new changes.