Inspect final-motion failures and keep released stress packs verifiable - #51
Merged
Merged
Conversation
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Stress Lab previously exposed the new failure taxonomy only as JSON, and requiring that file made the verifier reject already published experiment packs. Add a classification selector and direct jumps to the measured final-motion window, while preserving all trial counts and the full replay matrix. Accept the exact legacy inventory and continue recomputing taxonomy for new packs. Missing input frames or camera hashes can no longer count as identical policy input.
Retain the outside changes and their measured repeat; independently verify the actual saved 30-trial rerun. Clarify that continued movement does not prove progress or future success. Prepare v0.12.0, with bilingual documentation and an actual interface screenshot.
Validation: 202 Python tests (three optional Rerun skips; the one initial stale-page failure passed after complete page regeneration), 25 focused stress checks, all 49 browser tests, and the JavaScript template check. The independently installed v0.12 wheel verifies the actual GitHub v0.11 archive with all 30 trials. Desktop, mobile and local-file tests cover filtering, empty states and source-sample jumps.