You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Several Show2D and Show3D user-facing workflows are being designed or implemented together. Unit tests and typechecks are necessary, but they are not enough for these widgets. A scientist must be able to drive the widget visually in JupyterLab or exported HTML and see the expected behavior with real interactions.
This issue tracks the visual signoff pass for the related user-flow issues:
Run a focused, browser-driven widget review that produces screenshots or an HTML report. Treat visual evidence as required for closing the related issues.
Expected user behavior
A user opens the widget, interacts normally, and sees the scientific view stay honest and stable:
Show2D saved states restore the exact inspection state the user saved.
Show2D padding changes the displayed canvas naturally without changing raw data.
Show3D playback dynamics help a user inspect temporal structure without crowding the UI.
Agent visual verification checklist
Use live JupyterLab for kernel-backed behaviors, not only static HTML.
Use exported HTML where the issue explicitly includes collaborator/report review.
Open the latest target in the in-app browser or Chrome and drive it as a user: clicks, sliders, toggles, save/load, delete, playback, scrub, loop, bounce, FFT, and resize.
Capture screenshots after meaningful interactions.
Attach a single consolidated HTML report when multiple widgets or screenshots are involved.
Include Computer HTML and iPhone HTML links when the report is served for both local and phone review.
Record browser console errors/warnings and widget debug counters where relevant.
Report what was visually verified, what was only unit-tested, and what remains unverified.
Problem
Several Show2D and Show3D user-facing workflows are being designed or implemented together. Unit tests and typechecks are necessary, but they are not enough for these widgets. A scientist must be able to drive the widget visually in JupyterLab or exported HTML and see the expected behavior with real interactions.
This issue tracks the visual signoff pass for the related user-flow issues:
Proposed solution
Run a focused, browser-driven widget review that produces screenshots or an HTML report. Treat visual evidence as required for closing the related issues.
Expected user behavior
A user opens the widget, interacts normally, and sees the scientific view stay honest and stable:
Agent visual verification checklist
Computer HTMLandiPhone HTMLlinks when the report is served for both local and phone review.Suggested review scenes
Show2Dsingle-image ROI state: ROI + FFT + padding + save/load state.Show2D3-panel drift gallery: all-panel padding and selected-panel padding.Show3Dtime-series stack: linear, slow, bounce, focus range, and hold-key-frame playback paths.