Sync mobile graph recovery and scale guards #9

Merged

lost-rob0t merged 1 commit from agent/sync-mobile-graph-recovery into main

2026-08-08 19:57:56 +00:00

lost-rob0t commented

2026-08-08 19:56:14 +00:00

(Migrated from github.com)

What changed

sync the benchmarked graph-scale guard and adaptive layout cutoff into the Auto-Dig Quasar fork
scope datasets before graph construction and prevent oversized document loads
restore dataset/document graph navigation and search-result imports
fix hidden mobile dataset filters and make the recovery action name the exact filter
keep the graph count clear of the mobile agent bubble and reduce empty-state action width
bring the graph editor/research-node dependencies required by the updated GraphPage into the fork

Validation

58 unit files, 273 tests passed
integration test passed
formatting, lint, typecheck, package boundaries, and static checks passed
production build passed
graph benchmark completed through 8,000 nodes / 15,999 elements

The force layout crossed the practical cutoff between 250 and 500 nodes; the fork now switches large graphs to grid and refuses documents/elements beyond the benchmarked ceilings.

## What changed - sync the benchmarked graph-scale guard and adaptive layout cutoff into the Auto-Dig Quasar fork - scope datasets before graph construction and prevent oversized document loads - restore dataset/document graph navigation and search-result imports - fix hidden mobile dataset filters and make the recovery action name the exact filter - keep the graph count clear of the mobile agent bubble and reduce empty-state action width - bring the graph editor/research-node dependencies required by the updated GraphPage into the fork ## Validation - 58 unit files, 273 tests passed - integration test passed - formatting, lint, typecheck, package boundaries, and static checks passed - production build passed - graph benchmark completed through 8,000 nodes / 15,999 elements The force layout crossed the practical cutoff between 250 and 500 nodes; the fork now switches large graphs to grid and refuses documents/elements beyond the benchmarked ceilings.