# audio2text — 双语字幕生成服务
# CPU dev / GPU prod 两套配置，按需切换 service。
#
# 开发（CPU，热重载，改代码零重建）：
#   docker compose --profile dev up -d --build
# CPU 生产：
#   docker compose --profile cpu up -d --build
# GPU 生产（本机有 nvidia runtime 时）：
#   docker compose --profile gpu up -d --build
#
# 模型缓存（/models）跨容器复用，首次启动下载、之后秒起。
# pip wheel 缓存由 BuildKit --mount=type=cache 管理，跨构建复用。

services:
  # ---------------- 开发：CPU + 源码挂载 + uvicorn reload ----------------
  # 改 app/ 下代码无需重建镜像，保存即热重载。
  audio2text-cpu-dev:
    build:
      context: .
      args:
        VARIANT: cpu
      target: dev
    image: audio2text:cpu-dev
    container_name: audio2text-dev
    ports:
      - "8000:8000"
    volumes:
      - ./app:/app/app
      - ./data:/data
      - ./models:/models
      - ./config.yaml:/app/config.yaml:ro
    environment:
      CONFIG_PATH: /app/config.yaml
    restart: unless-stopped
    profiles: ["dev"]

  # ---------------- CPU 生产 ----------------
  audio2text-cpu:
    build:
      context: .
      args:
        VARIANT: cpu
      target: final
    image: audio2text:cpu
    container_name: audio2text
    ports:
      - "8000:8000"
    volumes:
      - ./data:/data
      - ./models:/models
      - ./config.yaml:/app/config.yaml:ro
    environment:
      CONFIG_PATH: /app/config.yaml
    restart: unless-stopped
    profiles: ["cpu", ""]

  # ---------------- GPU 生产 ----------------
  audio2text-gpu:
    build:
      context: .
      args:
        VARIANT: gpu
      target: final
    image: audio2text:gpu
    container_name: audio2text-gpu
    ports:
      - "8001:8000"
    volumes:
      - ./data:/data
      - ./models:/models
      - ./config.gpu.yaml:/app/config.yaml:ro
    environment:
      CONFIG_PATH: /app/config.yaml
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    restart: unless-stopped
    profiles: ["gpu"]