What's New
Semantic coloring now works automatically — no frontmatter needed.
Colors are derived from k-means clustering of your note embeddings:
- Each note gets its top-2 nearest clusters + a distance-based weight
- The same clustering used for zones drives the color assignments
- Changing zone granularity updates semantic colors accordingly
Changes
kMeans()now returns centroids alongside assignments- New
computeSemanticAssignments()function - Removed
sem_k/sem_a/sem_b/sem_wfrontmatter parsing - Centroids cached in zone cache for fast recomputation
- "Semantic (sem_k)" dropdown label simplified to "Semantic"