Fix replay state rehydration - #1367
Open
JackWilb wants to merge 6 commits into
Open
Conversation
Contributor
|
A preview of 232f2e8 is uploaded and can be seen here: ✨ https://revisit.dev/study/PR1367 ✨ Changes may take a few minutes to propagate. |
Contributor
There was a problem hiding this comment.
Pull request overview
This PR improves analysis replay fidelity by rehydrating previously recorded component state (sliders, Vega, UpSet, MVNV iframe state, and virtual-chinrest) and ensuring replay remains read-only (no new answers/provenance emitted). It also adds regression coverage to validate deterministic seeking behavior and non-mutating replay.
Changes:
- Add shared Playwright helpers for reading persisted participant/provenance data and seeking within the replay timeline.
- Make replay state deterministic and inert across Vega, UpSet, iframe surfaces, and virtual-chinrest calibrations (rehydration + read-only controls).
- Add focused Vitest + Playwright regression tests for slider rehydration, Vega forward/back/root seeking, and replay immutability.
Reviewed changes
Copilot reviewed 17 out of 17 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| tests/utils.ts | Adds IndexedDB read helpers and a replay seek helper for Playwright tests. |
| tests/replay-slider-rehydration.spec.ts | New Playwright test verifying SMEQ slider value rehydration and no participant-data writes during replay. |
| tests/demo-vega.spec.ts | Extends Vega e2e coverage to verify deterministic seek replay and immutability of stored provenance. |
| tests/demo-trrack.spec.ts | Refactors to use shared IndexedDB/seek helpers for replay verification. |
| src/public/libraries/virtual-chinrest/assets/VirtualChinrestCalibration.tsx | Rehydrates saved card-size calibration in analysis and disables interactive controls. |
| src/public/libraries/virtual-chinrest/assets/ViewingDistanceCalibration.tsx | Selects prior card-size calibration for repeated sequences and rehydrates saved viewing-distance state; blocks input/provenance in analysis. |
| src/public/libraries/virtual-chinrest/assets/tests/VirtualChinrestCalibration.spec.tsx | Adds unit coverage for virtual-chinrest replay rehydration and repeated-sequence lookup behavior. |
| src/public/demo-upset/assets/Upset.tsx | Integrates UpSet provenance state with reVISit replay and makes the surface inert in analysis. |
| src/public/demo-upset/assets/tests/Upset.spec.tsx | Adds unit coverage for UpSet provenance publication and analysis-mode inertness/no emission. |
| src/controllers/VegaController.tsx | Tracks cumulative Vega signals and restores deterministically across seeks; prevents tracking while applying analysis replay and makes the surface inert. |
| src/controllers/tests/IframeController.spec.tsx | Adds unit coverage asserting iframes are inert/non-interactive in analysis replay. |
| src/controllers/tests/ComponentController.spec.tsx | Extends Vega controller unit tests for cumulative signal restoration and analysis-mode non-tracking. |
| src/controllers/IframeController.tsx | Makes website iframes inert/disabled with pointer-events blocked during analysis replay. |
| src/components/response/tests/ResponseInput.spec.tsx | Enhances slider mock and adds tests ensuring replayed slider values rehydrate while disabled. |
| src/components/response/SliderInput.tsx | Rehydrates displayed SMEQ/NASA-TLX slider state on answer changes and blocks updates when disabled. |
| public/example-mvnv/assets/js/loadTasks.js | Rehydrates MVNV iframe task selection state from received answers to support replay. |
| .gitattributes | Adds whitespace handling rule for the MVNV JS asset file. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+205
to
+206
| const fraction = (targetTime - startTime) / (endTime - startTime); | ||
| const x = 20 + Math.min(1, Math.max(0, fraction)) * (timerBounds.width - 40); |
Comment on lines
+18
to
+33
| export function findPreviousCardSizeAnswer( | ||
| answers: Record<string, StoredAnswer>, | ||
| currentTrialOrder: StoredAnswer['trialOrder'] | undefined, | ||
| ) { | ||
| const currentStep = Number.parseInt(currentTrialOrder ?? '', 10); | ||
| if (!Number.isFinite(currentStep)) { | ||
| return undefined; | ||
| } | ||
|
|
||
| return Object.values(answers) | ||
| .filter((answer) => ( | ||
| answer.componentName === cardSizeComponent | ||
| && Number.parseInt(answer.trialOrder, 10) < currentStep | ||
| )) | ||
| .sort((a, b) => Number.parseInt(b.trialOrder, 10) - Number.parseInt(a.trialOrder, 10))[0]; | ||
| } |
jaykim1213
self-requested a review
July 30, 2026 05:40
jaykim1213
requested changes
Jul 30, 2026
jaykim1213
left a comment
jaykim1213
left a comment
Contributor
There was a problem hiding this comment.
Thanks for working on this!
Vega, UpSet, and SMEQ are working nicely now, but three surfaces still don't show the restored state on screen:
- MVNV: the answer data is restored, but the adjMatrix view doesn't update (answerBox checks and row/column selections stay empty).
- Virtual chinrest: the final values is shown in replay, but we can't see the card adjustment or the space-bar measurements replaying
- NASA-TLX: no value is visible in replay.
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
Scope
Check Answer replay and video playback replay remain out of scope and are tracked separately in #1365 and #1366.
Root cause
Several components either kept participant state only in local UI state or replayed only the latest interaction instead of the cumulative tracked state. NASA-TLX's disabled Mantine thumb was hidden, MVNV restored answer text without mapping saved full node names back into its adjacency-matrix state, and virtual-chinrest did not track intermediate calibration state. Some replay paths could also invoke live interaction listeners while applying analysis state, and virtual-chinrest fallback lookup did not distinguish repeated component occurrences.
Validation
yarn unittest --run— 2,042 passed, 1 skippedyarn typecheckyarn lintyarn buildgit diff --checkPlaywright coverage verifies real Vega forward/root/forward replay, inert analysis controls, unchanged provenance storage, exact SMEQ/NASA-TLX slider rehydration, virtual-chinrest card/measurement traversal, and MVNV adjacency-matrix restoration.
The implementation received independent adversarial review; all confirmed correctness, coverage, and minimality findings were addressed.
Closes #1336