Fix Lyria API response parsing — audio in parts[1], not parts[0]

The Lyria API returns a candidates[].content.parts[] array where:
- parts[0] = text metadata (e.g. '<instrumental>')
- parts[1] = inlineData with the actual audio payload

Previous code only checked parts[0] for inlineData, which always failed
with 'No audio data in response'. Fixed by iterating all parts to find
the inlineData entry.

Successfully generated first real audio clip (lo-fi hip hop, 30s, 728KB MP3).

Also includes: .env.example, requirements updates, docker-compose fixes,
frontend vite config, logger setup.

docker-compose.yml

@@ -1,12 +1,10 @@
-version: '3.8'
-
 services:
   backend:
     build:
       context: ./backend
       dockerfile: Dockerfile
     ports:
-      - "8000:8000"
+      - "8100:8000"
     volumes:
       - ./data:/app/data
     environment:
@@ -18,7 +16,7 @@ services:
       context: ./frontend
       dockerfile: Dockerfile
     ports:
-      - "3000:3000"
+      - "3100:3000"
     depends_on:
       - backend
     restart: unless-stopped