edX has published a guide for upgrading repos to django 3.2: https://openedx.atlassian.net/wiki/spaces/AC/pages/1704591765/Django+3.2+Upgrade+Steps+to+Upgrade+a+Repository

acceptance criteria

• read the steps to upgrade a repository and update this checklist with the steps that need to be done for this repo.

Django 3.2 migration checklist

• Update django version in tox.ini, github actions flow and there are scripts in https://github.com/edx/repo-tools/tree/master/edx_repo_tools/codemods/django3 that can make code compatible to django3 in most cases
• Upgrade any packages that were being held back only because of missing support for one of the removed Python / django versions
• Install pyupgrade-directories package and run pyup-dirs --py3-plus --py36-plus --py38-plus --recursive . in the repo root folder to clean up code that was only present to support older Python versions. You may want to run each of the --py* options in a separate PR if they make a lot of changes.
• Run the codemod scripts in https://github.com/edx/repo-tools/tree/master/edx_repo_tools/codemods/django3 (such as replace_unicode_with_str.py and script_to_replace_static.sh) to automatically fix some code patterns known to break under Django 3.2.
• Fix any deprecation warnings related to Django 3.0 or 3.1 that occur when the tests are run with Django 2.2
• Fix any remaining Django deprecation warnings or test failures in all the test configurations. But don’t fix deprecation warnings from the Django 3.2 tests yet, fixes for them are likely to break compatibility with Django 2.2 which we still need for now.
• If any of the new test configurations were allowed to fail when added, update the configuration to no longer allow them to do so.
• Update the Trove classifiers in any setup.py, setup.cfg, or pyproject.toml files to reflect the Python and Django versions currently being tested
• For packages already on PyPI, go ahead and make a new release so the Trove classifiers are updated there also. This should be a major version bump if support for older Python releases was dropped.
• If the repository is an IDA, go ahead and update it to deploy on Django 3.2 once all of the new test configurations are passing cleanly.

Description

The plugin is in working form but basic stuff like Testing, Github Actions, and Quality checks are not in place right now. Thus this issue is meant to add all the basic features to the plugin.

Acceptance Criteria

• Enable testing
• Add Quality related plugins
• Enable CI using github actions

Note

It would be better if we inherit the plugin with cookiecutter-django-app

The results that are returned by the openedx comment client are paged, and this app does not account for that paging. As a result, we only inspect the first 20 threads or comments that a user has been involved in. It's possible that the paging was added after we created this app.

Implementation:

These are the functions I used to debug this issue locally and fix a single user. If you use this code, just replace the existing get_authored_threads_and_comments function instead of naming it get_authored_threads_and_comments_v2 as I did here.

course_id = '...'
user = User.objects.get(username='...')
#
def get_involved_threads(course_id, user_id):
  page = 0
  involved_threads = []
  while len(involved_threads) > 0 or page == 0:
    involved_threads = [ 
      Thread.find(id=thread["id"]).retrieve(
        with_responses=True, recursive=True, mark_as_read=False
      )
      for thread in Thread.search(
        {"course_id": course_id, "user_id": comment_user.id, "page": page}
      ).collection
    ]
    yield from involved_threads
    page += 1
#
def get_authored_threads_and_comments_v2(comment_user, enrolled_course_ids):
  for course_id in enrolled_course_ids:
    involved_threads = get_involved_threads(course_id, comment_user.id)
    for thread in involved_threads:
      if thread["user_id"] == comment_user.id:
        yield thread.to_dict()
      children_to_scan = thread["children"]
      while children_to_scan:
         child = children_to_scan.pop(0)
         children_to_scan.extend(child["children"])
         if child["user_id"] == comment_user.id:
            yield child