Skip to content

feat(ratings): "Didn't use it" feedback in the rating UI [Rating page for assets (My Downloads)] - #2292

Merged
Tweekazoid merged 2 commits into
mainfrom
claude/didnt-use
Sep 1, 2026
Merged

feat(ratings): "Didn't use it" feedback in the rating UI [Rating page for assets (My Downloads)]#2292
Tweekazoid merged 2 commits into
mainfrom
claude/didnt-use

Conversation

@PetrDlouhy

Copy link
Copy Markdown
Contributor

What

The My downloads page's "I didn't use this asset" feedback, now in the add-on: a "Didn't use it" dropdown at the bottom of every ratings UI (fast-rate popup, sidebar ratings panel, asset detail card). Server API: Blendkit-server#3674; Client proxy: bk_client#49 (submodule bumped to that commit).

  • One menu for every state — pick a reason to flag (the server's shared admin-managed choice set, fetched once per session), the current reason wears a checkmark, "Undo" lives at the bottom.
  • Mutual exclusivity, drawn not explained — a rated asset's menu warns "This replaces your rating"; the pick deletes the scores server-side (replace_rating) and remembers them session-locally. The hint over the now-disabled rating rows names the numbers ("Your 8★ / 4 h rating was cleared - Undo restores it") and "Undo - restore my rating" re-rates, which also drops the flag server-side (rating wins). Rating a flagged asset clears the flag; the local store mirrors it.
  • Errors land where the eyes are — server refusals ("Only downloaded assets can be flagged.") show as red text inline under the control instead of only the corner report overlay; the server's detail sentence is extracted even from a Client that passes the raw JSON body through.
  • Bugfix: handle_get_rating_task skips unknown rating types (historic competition-2022-votes / nodevember-votes) instead of crashing the task-handling timer — hit on devel with a validator account.

Testing

tests/test_didnt_use.py (23 tests): task handlers, dispatch routing, operator flows (flag/undo/replace/restore, partial memory), inline-error extraction incl. the raw-JSON regression, and the unknown-rating-type crash reproduction. Full suite in headless Blender 5.2: 513 tests pass (one pre-existing test_init bl_info failure exists on clean main in local extension-mode runs). Exercised manually end-to-end against devel.

Release note

Real users need a bk_client release + client version pin bump after bk_client#49 merges; dev builds (--client-source local) work as-is.

🤖 Generated with Claude Code

The My downloads page's "I didn't use this asset" flag, now also in the
add-on - a dropdown at the bottom of every ratings UI (fast-rate popup,
sidebar panel, asset detail):

- One menu for every state: pick a reason to flag (choices are the
  server's shared admin-managed set, fetched once per session), the
  current reason wears a checkmark, undo lives at the bottom.
- Rating and flag are mutually exclusive (server-enforced). A rated
  asset's menu warns "This replaces your rating"; the pick then deletes
  the scores server-side (replace_rating) and the cleared numbers are
  remembered session-locally - the hint over the disabled rating rows
  names them and "Undo - restore my rating" re-rates, which also drops
  the flag server-side. Rating a flagged asset clears the flag, locally
  mirrored.
- Server refusals (undownloaded asset, conflicts) land as red text
  inline under the control - the corner report overlay is out of the
  popup's sight. The server's detail sentence is extracted even from a
  Client that passes the raw JSON body through.
- handle_get_rating_task now skips unknown rating types (historic
  competition/nodevember votes) instead of crashing the task timer.

bk_client submodule bumped to the endpoints' commit
(BlenderKit/bk_client#49).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@PetrDlouhy PetrDlouhy self-assigned this Sep 1, 2026
@Tweekazoid
Tweekazoid merged commit 95ee70c into main Sep 1, 2026
29 checks passed
@Tweekazoid
Tweekazoid deleted the claude/didnt-use branch September 1, 2026 21:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants