docker-compose.dev.yml

x-app: &default-app
  image: radis_dev:latest
  volumes:
    - .:/app
    - vscode-server:/root/.vscode-server/extensions
    - vscode-server-insiders:/root/.vscode-server-insiders/extensions
  environment:
    DJANGO_INTERNAL_IPS: ${DJANGO_INTERNAL_IPS:?}
    DJANGO_SETTINGS_MODULE: "radis.settings.development"
    FORCE_DEBUG_TOOLBAR: ${FORCE_DEBUG_TOOLBAR:-true}
    REMOTE_DEBUGGING_ENABLED: ${REMOTE_DEBUGGING_ENABLED:-false}
    REMOTE_DEBUGGING_PORT: ${REMOTE_DEBUGGING_PORT:-5678}
  pull_policy: never

x-llamacpp: &llamacpp
  environment:
    HTTP_PROXY: ${HTTP_PROXY:-}
    HTTPS_PROXY: ${HTTPS_PROXY:-}
    LLAMA_CACHE: "/models"
    NO_PROXY: ${NO_PROXY:-}
  hostname: llamacpp.local
  volumes:
    - models_data:/models

services:
  init:
    <<: *default-app
    profiles:
      - never

  web:
    <<: *default-app
    build:
      target: development
    ports:
      - "${WEB_DEV_PORT:-8000}:8000"
      - "${REMOTE_DEBUGGING_PORT:-5678}:5678"
    command: >
      bash -c "
        wait-for-it -s postgres.local:5432 -t 60 && 
        ./manage.py migrate &&
        ./manage.py create_superuser &&
        ./manage.py create_example_users &&
        ./manage.py create_example_groups &&
        ./manage.py populate_example_reports --lng ${EXAMPLE_REPORTS_LANGUAGE:-en} &&
        wait-for-it -s llamacpp.local:8080 -t 60 &&
        ./manage.py runserver 0.0.0.0:8000
      "

  default_worker:
    <<: *default-app
    command: >
      bash -c "
        wait-for-it -s postgres.local:5432 -t 60 &&
        ./manage.py bg_worker -l debug -q default --autoreload
      "

  llm_worker:
    <<: *default-app
    command: >
      bash -c "
        wait-for-it -s postgres.local:5432 -t 60 &&
        ./manage.py bg_worker -l debug -q llm --autoreload
      "

  postgres:
    environment:
      POSTGRES_PASSWORD: "postgres"

  llamacpp_cpu:
    <<: *llamacpp
    image: ghcr.io/ggerganov/llama.cpp:server
    entrypoint: []
    command: >
      bash -c "
        /llama-server \\
          --model-url ${LLM_MODEL_URL} \\
          --host 0.0.0.0 \\
          --port 8080 \\
          --ctx-size 4096
      "
    profiles: ["cpu"]

  llamacpp_gpu:
    <<: *llamacpp
    image: ghcr.io/ggerganov/llama.cpp:server-cuda
    entrypoint: []
    command: >
      bash -c "
        /llama-server \\
          --model-url ${LLM_MODEL_URL} \\
          --host 0.0.0.0 \\
          --port 8080 \\
          --ctx-size 4096 \\
          --gpu-layers 99
      "
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    profiles: ["gpu"]

volumes:
  models_data:
  vscode-server:
  vscode-server-insiders: