fix: resolve read/unread state persistence issue

[Alam-2U]

Description

This PR fixes an issue where discussion forum threads marked as read were incorrectly showing as unread after a page reload or tab switch.

Cause

Thread read state was being overwritten by requests that did not contain valid user context, causing already-read threads to appear unread.

Fix

Updated the discussion thread read-state handling to ignore requests without valid user context and preserve the correct read/unread status.

Ticket

COSMO2-904

Related PR

edx/frontend-app-discussions#30

[Copilot]

Pull request overview

Fixes incorrect persistence of thread read/unread state by preventing read-state calculation/return when the request lacks valid user context (avoids overwriting a previously-correct UI state on reload/tab switch).

Changes:

• Thread serializer now returns None for read / unread_comments_count when user_id is missing and recalculates read state via backend when present.
• MySQL backend get_read_states now treats user_id=None as “no user context” and returns an empty result.
• Bumped package version to 0.5.9.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
forum/serializers/thread.py	Adjusts read/unread serialization to avoid using stale/invalid context and (currently) recomputes via backend.
forum/backends/mysql/api.py	Treats missing user context (None) as non-computable for read states.
forum/init.py	Version bump reflecting the fix.

Comments suppressed due to low confidence (2)

forum/serializers/thread.py:132

• get_read() now calls backend.get_read_states() per thread, which can create an N+1 query pattern on list endpoints (and get_unread_comments_count() does a second call for the same thread). To avoid a significant performance regression, consider using the precomputed obj['read']/obj['unread_comments_count'] when user_id is present, and/or caching/bulk-fetching read states once for the serializer instead of querying inside each field method.

            course_id = obj["course_id"]
            thread_key = obj["_id"]
            is_read, _ = self.backend.get_read_states(
                [obj["_id"]], user_id, course_id
            ).get(thread_key, (False, obj["comment_count"]))

forum/serializers/thread.py:156

• get_unread_comments_count() repeats the same get_read_states([obj['_id']], ...) lookup done in get_read(), doubling backend/DB work per thread. Consider reusing a single computed read-state result (e.g., memoize per thread in the serializer instance or attach it to obj) so read/unread count are derived from one backend call.

            course_id = obj["course_id"]
            thread_key = obj["_id"]
            _, unread_count = self.backend.get_read_states(
                [obj["_id"]], user_id, course_id
            ).get(thread_key, (False, obj["comment_count"]))

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.