[Phase 4] Add the virtualized table projection #23

New issue

Closed

opened 2026-07-25 21:55:19 +00:00 by lost-rob0t · 0 comments

lost-rob0t commented

2026-07-25 21:55:19 +00:00

(Migrated from github.com)

Parent: #2
Depends on: #7, #12, #16

Scope

Build a TanStack Table projection for nodes, edges, and bounded query/action results.
Add sorting, filtering, grouping, column visibility, row selection, and expandable details.
Use TanStack Virtual for large row and column windows.
Preserve selection by stable identifier when rows leave the render window.
Keep projection state separate from canonical graph data.

Acceptance criteria

Graph selections can focus matching rows.
Table selections can publish stable graph identifiers.
A 100,000-row fixture remains usable within the defined performance budget.
Empty, partial, truncated, and error states are distinct.
Exporting the current bounded table projection does not mutate the graph.

Parent: #2 Depends on: #7, #12, #16 ## Scope - Build a TanStack Table projection for nodes, edges, and bounded query/action results. - Add sorting, filtering, grouping, column visibility, row selection, and expandable details. - Use TanStack Virtual for large row and column windows. - Preserve selection by stable identifier when rows leave the render window. - Keep projection state separate from canonical graph data. ## Acceptance criteria - Graph selections can focus matching rows. - Table selections can publish stable graph identifiers. - A 100,000-row fixture remains usable within the defined performance budget. - Empty, partial, truncated, and error states are distinct. - Exporting the current bounded table projection does not mutate the graph.