Yearbook opus

[cornhundred]

No description provided.

[cornhundred]

Overview of yearbook_opus Branch Changes

Files Changed

File	Changes	Description
js/viz/yearbook.js	+850 lines (new)	Main visualization logic
js/ui/yearbook_ui.js	+650 lines (new)	Control panel UI with pagination
js/deck-gl/core/yearbook_viewports.js	+208 lines (new)	Multi-viewport grid & discontiguous tile loading
src/celldega/viz/widget.py	+206 lines	Python Yearbook class
js/deck-gl/layers/image_layers.js	+71 lines	Per-portrait image layer support
js/widget.js	+59 lines	Widget routing for Yearbook
src/celldega/__init__.py	+2 lines	Export Yearbook
src/celldega/viz/__init__.py	+2 lines	Export Yearbook

Key Features Implemented

1. Yearbook Widget - Portrait grid visualization of selected cells

   • Configurable num_rows, num_cols, portrait_size_um
   • Centers each portrait on a specific cell

2. Multi-Viewport Architecture - Each portrait is a separate OrthographicView

   • Synced zoom across all portraits (no panning)
   • layerFilter ensures per-portrait image layers render only in their viewport

3. Discontiguous Tile Loading - Loads only tiles needed for visible portraits

   • Shared cell/transcript/path layers across all viewports
   • Per-portrait image layers with extent restrictions

4. Pagination - Navigate through cells when more than grid capacity

   • Debounced to prevent rapid-click issues
   • UI stays in sync with actual displayed content

5. Control Panel - Similar to Landscape widget

   • Gene bars and cell bars based on visible transcripts/cells
   • Image layer toggles, cell/transcript visibility sliders

[cornhundred]

ready for testing

[Copilot]

Pull request overview

This PR introduces the "Yearbook" widget, a new visualization component for displaying cell portraits in a grid layout. The widget provides a synchronized, multi-viewport interface where each cell gets its own zoomed-in view, enabling detailed comparative analysis across multiple cells simultaneously.

Key Changes:

• Added new Yearbook widget class supporting configurable grid layouts with synchronized zoom
• Implemented pagination for navigating through large cell collections
• Extracted reusable scale bar utilities for shared use across visualization components

Reviewed changes

Copilot reviewed 13 out of 14 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/celldega/viz/widget.py	Added Yearbook widget class with grid configuration, pagination, and cell metadata handling
src/celldega/viz/init.py	Exported Yearbook widget for public API
src/celldega/init.py	Added Yearbook to top-level package exports
pyproject.toml	Bumped version to 0.15.0a4
package.json	Updated JavaScript package version to match
notebooks/Yearbook_HCRC.ipynb	Added example notebook demonstrating Yearbook usage with Xenium data
js/widget.js	Added render_yearbook function and integrated Yearbook component routing
js/viz/yearbook.js	Implemented core yearbook visualization logic with multi-viewport rendering and zoom synchronization
js/viz/landscape_ist.js	Refactored to use extracted scale_bar utility
js/utils/scale_bar.js	Extracted scale bar creation into reusable utility module
js/ui/yearbook_ui.js	Created yearbook-specific UI with pagination controls and bar graph updates
js/deck-gl/layers/image_layers.js	Added make_yearbook_image_layers for per-portrait image layer creation
js/deck-gl/core/yearbook_viewports.js	Implemented viewport management utilities for grid layout and discontiguous tile loading

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[cornhundred]

This pull request can be tested using the notebook Yearbook_HCRC.ipynb

[huanlity]

LGTM ✅. Works as expected.

What I really like

1. The ability to zoom in/out from any viewport
2. Reducing rows/columns automatically results in larger viewports
3. Overall responsiveness

Things to consider

1. Only the first (upper-left) viewport currently supports panning. I find this quite useful when users want to quickly inspect surrounding cells, so it might be worth extending this capability to other viewports as well.
2. Adding a minimal label to each viewport could be helpful for displaying basic information about the focused cell. This can be part of a new PR and would be better implemented once we define the criteria/strategy for selecting representative cells.
3. The viewport divider could be more visually distinct. Right now it’s a thin white space, which works in most cases, but when the image is turned off it becomes difficult to see viewport boundaries and relative transcript (trx) locations. A light grey divider might improve clarity.

Minor, but nice-to-have

1. Automatically setting the cell marker size so users don’t need to worry about it. At minimum, we could limit it so it doesn’t exceed ¼ of the viewport area; similarly, limit the trx marker size to ⅛ of the viewport area.
2. I assume the zoom threshold isn’t constrained yet but nice to have.

[cornhundred]

Thanks @huanlity, I actually intended to not allow panning in any viewport but we can fix that later. I want to add cell labels in another pull request.3 sounds good too for a subsequent pull request. Agree on the nice to haves for a future pull request.