Avoid duplicated opencv/ort fetches#6494
Open
jpggvilaca wants to merge 2 commits into
Open
Conversation
jpggvilaca
commented
May 15, 2026
| // before build/init resolve, so we don't leak the in-flight worker. | ||
| signal.addEventListener('abort', worker.terminate, { once: true }); | ||
| // | ||
| // CRITICAL: pass an arrow function, NOT `worker.terminate` directly. |
Contributor
Author
There was a problem hiding this comment.
this is the fix for the duplication.
jpggvilaca
commented
May 15, 2026
| headers: { | ||
| 'Cross-Origin-Embedder-Policy': 'credentialless', | ||
| 'Cross-Origin-Opener-Policy': 'same-origin', | ||
| 'Cache-Control': 'public, max-age=31536000, immutable', |
Contributor
Author
There was a problem hiding this comment.
this should be done in production (already asked backend) but it also improves DX. Basically it tells the browser that these files will never change (opencv, wasm, etc) so it's all good to cache them, even if hashed (e.g. ort23478234823.wasm)
Contributor
There was a problem hiding this comment.
Pull request overview
This PR adjusts annotator web-worker lifecycle management to avoid repeated ORT/OpenCV/model fetches by keeping workers alive across intra-dataset navigation and terminating them only when leaving the dataset section.
Changes:
- Move annotator-tool worker teardown from annotator unmount to the dataset-section boundary via a new
useTerminateAnnotatorWorkersOnUnmounthook. - Keep the Segment Anything worker query alive indefinitely (
gcTime: Infinity) and fix its abort handler to reliably terminate on cancellation. - Add a global
Cache-Controlheader in the Rsbuild dev server config.
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
application/ui/src/router.tsx |
Introduces DatasetSection wrapper to own worker lifetime across dataset routes. |
application/ui/src/features/annotator/tools/use-preload-webworkers.hook.ts |
Removes per-annotator unmount teardown so workers can persist within the dataset section. |
application/ui/src/features/annotator/tools/terminate-annotator-workers-on-unmount.hook.ts |
Adds dataset-section unmount cleanup to terminate workers and remove worker queries. |
application/ui/src/features/annotator/tools/segment-anything-tool/use-segment-anything.hook.ts |
Improves worker caching and abort termination behavior; sets gcTime: Infinity. |
application/ui/rsbuild.config.ts |
Adds a global Cache-Control response header for the dev server. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
🐳 Docker image sizes
|
📊 Test coverage report
|
dwesolow
reviewed
May 15, 2026
| // for no functional gain (encoder/decoder always run sequentially anyway). | ||
| // | ||
| // `gcTime: Infinity` is critical: the default 5-min gc would evict the worker | ||
| // entry whenever SAM is unmounted (switching tools/projects), causing the |
Contributor
There was a problem hiding this comment.
unmounted and not used for 5 min I guess, not right away after unmunting
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Explanation:
onnx spawns X workers, where X depends on the amount of cpu cores your machine has, and each worker imports ort, hence why we would see all these imports. Same goes for opencv. We could limit the threads but concurrency is what we want, so we have a tradeoff between bandwith/memory and performance.
Before:
After:
How to test
Checklist