fix: increase ANN search K from 50 to 200 for 72K+ person pool (#7)

With 72K+ stranger persons (~364K face nodes in the HNSW index),
K=50 per prototype query was too small: high-similarity candidates
(e.g. 77%) were being missed because 50+ other stranger face nodes
ranked above them for prominent persons like 胡波 (3505 faces).

Increase annSearchK 50→200 and annHNSWEfSearch 100→200 (must be
>= K per HNSW guidelines) so each of the ~210 per-patrol queries
covers a wider neighborhood, ensuring borderline pairs above the
merge-suggestion threshold (55%) are not silently dropped.

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: davidhoo

backend/internal/service/person_merge_suggestion_ann.go

@@ -8,9 +8,9 @@ import (
 )
 
 const (
-	annSearchK      = 50  // neighbors per prototype query
+	annSearchK      = 200 // neighbors per prototype query; 72K+ person pool needs wider net than 50
 	annHNSWM        = 16  // max neighbors per node; 16 gives better recall at 60K+ scale vs 8
-	annHNSWEfSearch = 100 // search beam width; high value ensures recall near threshold boundary
+	annHNSWEfSearch = 200 // search beam width; must be >= annSearchK
 )
 
 // annIndex is a cached HNSW nearest-neighbor index over all person prototype embeddings.