feat: Phase 2.6 edges migration to Qdrant (TKG-only architecture)

Phase 2.6.1: co_occurrence_edges migration
- build_co_occurrence_edges_from_qdrant()
- Qdrant embeddings → frame grouping → YOLO objects
- Result: 6679 edges (vs 6701 PostgreSQL)

Phase 2.6.2: face_face_edges migration
- build_face_face_edges_from_qdrant()
- Qdrant embeddings → frame grouping → face pairs
- mutual_gaze detection preserved
- Result: 6 edges (exact match)

Phase 2.6.3: speaker_face_edges migration
- build_speaker_face_edges_from_qdrant()
- Qdrant embeddings → trace_id frame ranges
- SPEAKS_AS edge creation

Architecture:
- All edges use Qdrant payload (no face_detections queries)
- PostgreSQL fallback for empty Qdrant
- Estimated 3.6x performance improvement

Testing:
- Playground (3003): ✓ All Phase 2.6 logs verified
- Edge counts: ✓ Close match with PostgreSQL
- Fallback: ✓ Working

Docs:
- docs_v1.0/DESIGN/TKG_PHASE2_6_EDGES_MIGRATION.md
- docs_v1.0/M4_workspace/2026-06-21_phase2_6_test.md

scripts/start_gemma2_9b_mlx.sh

@@ -0,0 +1,24 @@
+#!/bin/bash
+# Start Gemma 2 9B with MLX
+
+PORT=8083
+MODEL="mlx-community/gemma-2-9b-it-4bit"
+LOG_FILE="/Users/accusys/momentry_core/logs/gemma2_9b_mlx_8083.log"
+
+# Kill existing process if running
+if lsof -i :$PORT > /dev/null 2>&1; then
+    OLD_PID=$(lsof -i :$PORT | awk 'NR>1 {print $2}')
+    kill $OLD_PID
+    sleep 2
+fi
+
+# Start MLX server
+nohup python3 -m mlx_lm.server \
+    --model $MODEL \
+    --host 0.0.0.0 \
+    --port $PORT \
+    > $LOG_FILE 2>&1 &
+
+echo "Gemma 2 9B MLX started on port $PORT"
+echo "Log file: $LOG_FILE"
+echo "Health check: http://localhost:$PORT/health"