# Runs the pipeline as a long-lived HTTP API for the Android app, on top
# of the same tooling-container image the numbered host scripts use.
# `docker compose up` alone is enough: the api service waits for Ollama
# and pulls the model itself (see tooling-container/src/server.py).
#
# The numbered host scripts (0_run_ollama.sh, 1/2/3_*.sh) and
# test_batch.sh are untouched and remain the way to inspect/test the
# pipeline stage by stage; this is the additive "serve it for real" path.
#
# Usage: docker compose up -d
# API reachable at http://<this-host>:8080 from any device on the LAN.
services:
  ollama:
    image: ollama/ollama
    volumes:
      - fuelish-ollama-data:/root/.ollama
    restart: unless-stopped

  api:
    build:
      context: ./tooling-container
    restart: unless-stopped
    depends_on:
      - ollama
    environment:
      OLLAMA_URL: http://ollama:11434/api/chat
      OLLAMA_BASE_URL: http://ollama:11434
      OLLAMA_MODEL: qwen2.5:3b
    command: ["uvicorn", "server:app", "--app-dir", "/opt/app/src", "--host", "0.0.0.0", "--port", "8000"]
    ports:
      - "8080:8000"

volumes:
  fuelish-ollama-data:
    name: fuelish-ollama-data