LOT 1 - Optimizare script build -Instalare mono comanda

ai_platform/modules/forensic_features/README.md

@@ -56,12 +56,12 @@ Output: text formatat + 12-15 imagini PNG base64 + scoruri raw structurate.
 # Build și pornește container
 docker compose up -d
 
-# Verifică
-curl http://localhost:8080/health
-curl http://localhost:8080/api/forensic-modules
+# Verifică (8085 = portul host; containerul expune intern 8080)
+curl http://localhost:8085/health
+curl http://localhost:8085/api/forensic-modules
 
 # Test end-to-end
-curl -X POST http://localhost:8080/api/forensic-evidence \
+curl -X POST http://localhost:8085/api/forensic-evidence \
      -F "video=@test_video.mp4" \
      -F "encode_images=1"
 ```
@@ -99,7 +99,7 @@ forensic_features/
 ├── .env.example
 ├── requirements.txt             # Lean: numpy, scipy, opencv, mediapipe, aiohttp
 │
-├── api.py                       # REST API (270 linii, doar 5 endpoints)
+├── api.py                       # REST API (418 linii, 6 endpoints — inclusiv /metrics)
 ├── preprocessing.py             # Frame extraction prin ffmpeg
 ├── face_landmarker.task         # MediaPipe model (~3.6 MB)
 │