version: '3.9'

services:
  web:
    build:
      context: .
      dockerfile: apps/web/Dockerfile
      args:
        - NEXT_PUBLIC_DEMO_TOKEN=${DEMO_TOKEN}
    ports:
      - "3000:3000"
    environment:
      - SQLITE_PATH=/data/hermaquette.db
      - NODE_ENV=production
      - NEXT_PUBLIC_DEMO_TOKEN=${DEMO_TOKEN}
      - STRIPE_SECRET_KEY=${STRIPE_SECRET_KEY}
      - STRIPE_PUBLISHABLE_KEY=${STRIPE_PUBLISHABLE_KEY}
      - PUBLIC_BASE_URL=${PUBLIC_BASE_URL}
      - DEMO_TOKEN=${DEMO_TOKEN}
      - DEMO_ALLOW_PENDING_QUOTE=${DEMO_ALLOW_PENDING_QUOTE:-true}
    volumes:
      - sqlite_data:/data
      - artifacts:/artifacts
    networks:
      - hermaquette_net
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "wget", "-q", "-O-", "http://localhost:3000/api/health"]
      interval: 30s
      timeout: 10s
      retries: 3

  hermes-agent:
    build:
      context: .
      dockerfile: services/hermes-agent/Dockerfile
    environment:
      - SQLITE_PATH=/data/hermaquette.db
      - OPENAI_API_KEY=${OPENAI_API_KEY}
      - NEMOTRON_API_KEY=${NEMOTRON_API_KEY}
      - NEMOTRON_MODEL=${NEMOTRON_MODEL:-nvidia/llama-3.1-nemotron-70b-instruct}
      - NEMOTRON_BASE_URL=${NEMOTRON_BASE_URL:-https://integrate.api.nvidia.com/v1}
      - STRIPE_SECRET_KEY=${STRIPE_SECRET_KEY}
      - STRIPE_ISSUING_ENABLED=${STRIPE_ISSUING_ENABLED:-false}
      - NANOBANANA_API_KEY=${NANOBANANA_API_KEY}
      - FAL_KEY=${FAL_KEY}
      - FAL_BUDGET_USD=${FAL_BUDGET_USD:-10}
      - FAL_DEV_BUDGET_USD=${FAL_DEV_BUDGET_USD:-7}
      - DEMO_MODE=${DEMO_MODE:-false}
      - SCULPTEO_API_KEY=${SCULPTEO_API_KEY}
      - SCULPTEO_API_URL=${SCULPTEO_API_URL:-https://www.sculpteo.com/api}
      - AGENTMAIL_API_KEY=${AGENTMAIL_API_KEY}
      - CAD_DFM_URL=http://cad-dfm:8000
      - ARTIFACTS_DIR=/artifacts
      - DEMO_TOKEN=${DEMO_TOKEN}
      - HAPPY_PATH=${HAPPY_PATH:-off}
      - TERMINAL_CWD=/app
      # Hermes gateways (started by start.sh inside this container)
      # Port 8642: primary (GPT-5.5/ChatGPT OAuth)  Port 8643: Nemotron (NVIDIA)
      - HERMES_API_KEY=${HERMES_API_KEY:-hermaquette-local}
      - HERMES_LLM_PROVIDER=${HERMES_LLM_PROVIDER:-openai-codex}
      - HERMES_LLM_MODEL=${HERMES_LLM_MODEL:-gpt-5.5}
      - HERMES_LLM_MODEL_API_FALLBACK=${HERMES_LLM_MODEL_API_FALLBACK:-gpt-4o}
      - HERMES_REASONING_EFFORT=${HERMES_REASONING_EFFORT:-high}
      - HERMES_GATEWAY_URL=http://127.0.0.1:8642/v1
      - HERMES_NEMOTRON_GATEWAY_URL=http://127.0.0.1:8643/v1
      # ChatGPT OAuth: copy contents of ~/.hermes/auth.json after `hermes setup --portal`
      - HERMES_AUTH_JSON=${HERMES_AUTH_JSON:-}
      # DEMO flags
      - DEMO_ALLOW_PENDING_QUOTE=${DEMO_ALLOW_PENDING_QUOTE:-true}
    volumes:
      - sqlite_data:/data
      - artifacts:/artifacts
      # NOTE: do NOT bind-mount ./hermes here — the image already contains a fully
      # populated /hermes (SOUL.md, AGENTS.md, skills/) via `COPY hermes/`. On Coolify
      # the host ./hermes is near-empty, so a bind mount shadows the image and breaks
      # the persona/skill copies (cp /hermes/SOUL.md fails → crash-loop). MEMORY.md
      # lesson persistence is sacrificed (fine for the demo).
    depends_on:
      - cad-dfm
    networks:
      - hermaquette_net
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "curl", "-sf", "http://127.0.0.1:8642/health"]
      interval: 30s
      timeout: 10s
      retries: 3

  cad-dfm:
    build:
      context: ./services/cad-dfm
      dockerfile: Dockerfile
    # port 8000 intentionally not published to host — internal to hermaquette_net only
    volumes:
      - artifacts:/artifacts
    networks:
      - hermaquette_net
    restart: unless-stopped
    healthcheck:
      test: ["CMD", "wget", "-q", "-O-", "http://localhost:8000/health"]
      interval: 30s
      timeout: 10s
      retries: 3

volumes:
  sqlite_data:
  artifacts:

networks:
  hermaquette_net:
    driver: bridge