feat: make Llama.cpp context size configurable via environment

- Add LLAMA_ARG_CTX_SIZE environment variable support in llamacpp service
- Increase default context size from 4096 to 8192 tokens for better performance
- Allow overriding via .env file for different deployment scenarios
- Maintains backward compatibility with existing setups

Author: chris-stinemetz

docker-compose.yml

@@ -412,7 +412,7 @@ services:
     container_name: llama-decoder-dev-remote
     environment:
       - LLAMA_ARG_MODEL=/models/model.gguf
-      - LLAMA_ARG_CTX_SIZE=4096
+      - LLAMA_ARG_CTX_SIZE=${LLAMA_ARG_CTX_SIZE:-8192}
       - LLAMA_ARG_HOST=0.0.0.0
       - LLAMA_ARG_PORT=8080
     ports: