docker-compose.yml

services:
  postgres:
    image: postgres:16-alpine
    container_name: cert-study-postgres
    environment:
      POSTGRES_DB: ${POSTGRES_DB:-cert_study}
      POSTGRES_USER: ${POSTGRES_USER:-cert_study}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-cert_study}
    ports:
      - "5432:5432"
    volumes:
      - postgres_data:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-cert_study} -d ${POSTGRES_DB:-cert_study}"]
      interval: 5s
      timeout: 5s
      retries: 20
    restart: unless-stopped

  cert-study-app:
    build:
      context: .
    container_name: cert-study-app
    depends_on:
      postgres:
        condition: service_healthy
    ports:
      - "8501:8501"
    environment:
      STREAMLIT_SERVER_ADDRESS: 0.0.0.0
      STREAMLIT_SERVER_PORT: 8501
      STREAMLIT_BROWSER_GATHER_USAGE_STATS: "false"
      CERT_STUDY_LLM_PROVIDER: ${CERT_STUDY_LLM_PROVIDER:-ollama}
      OLLAMA_MODEL: ${OLLAMA_MODEL:-qwen2.5:14b}
      OLLAMA_FAST_MODEL: ${OLLAMA_FAST_MODEL:-qwen3.5:9b}
      OLLAMA_VISUAL_MODEL: ${OLLAMA_VISUAL_MODEL:-qwen3-vl:8b-instruct-q4_K_M}
      EMBEDDING_MODEL: ${EMBEDDING_MODEL:-BAAI/bge-m3}
      OLLAMA_BASE_URL: ${OLLAMA_BASE_URL:-http://host.docker.internal:11434}
      DATABASE_URL: ${DATABASE_URL:-postgresql+pg8000://cert_study:cert_study@postgres:5432/cert_study}
      AIRFLOW_API_BASE_URL: ${AIRFLOW_API_BASE_URL:-http://host.docker.internal:8080}
      AIRFLOW_API_USERNAME: ${AIRFLOW_API_USERNAME:-admin}
      AIRFLOW_API_PASSWORD: ${AIRFLOW_API_PASSWORD:-admin}
    volumes:
      - ./data:/app/data
      - ./chroma_db:/app/chroma_db
      - ./instance:/app/instance
    restart: unless-stopped

  airflow-init:
    build:
      context: .
      dockerfile: Dockerfile.airflow
    container_name: cert-study-airflow-init
    depends_on:
      postgres:
        condition: service_healthy
    entrypoint: /bin/bash
    command:
      - -c
      - airflow db migrate && airflow users create --username "$${AIRFLOW_ADMIN_USERNAME:-admin}" --password "$${AIRFLOW_ADMIN_PASSWORD:-admin}" --firstname Cert --lastname Study --role Admin --email "$${AIRFLOW_ADMIN_EMAIL:-admin@example.com}" || true
    environment: &airflow-env
      AIRFLOW__CORE__EXECUTOR: SequentialExecutor
      AIRFLOW__CORE__LOAD_EXAMPLES: "false"
      AIRFLOW__CORE__DAGS_ARE_PAUSED_AT_CREATION: "false"
      AIRFLOW__DATABASE__SQL_ALCHEMY_CONN: sqlite:////opt/airflow_db/airflow.db
      AIRFLOW__API__AUTH_BACKENDS: airflow.api.auth.backend.basic_auth
      CERT_STUDY_LLM_PROVIDER: ${CERT_STUDY_LLM_PROVIDER:-ollama}
      OLLAMA_MODEL: ${OLLAMA_MODEL:-qwen2.5:14b}
      OLLAMA_FAST_MODEL: ${OLLAMA_FAST_MODEL:-qwen3.5:9b}
      OLLAMA_VISUAL_MODEL: ${OLLAMA_VISUAL_MODEL:-qwen3-vl:8b-instruct-q4_K_M}
      EMBEDDING_MODEL: ${EMBEDDING_MODEL:-BAAI/bge-m3}
      OLLAMA_BASE_URL: ${OLLAMA_BASE_URL:-http://host.docker.internal:11434}
      DATABASE_URL: ${DATABASE_URL:-postgresql+pg8000://cert_study:cert_study@postgres:5432/cert_study}
      AIRFLOW_ADMIN_USERNAME: ${AIRFLOW_ADMIN_USERNAME:-admin}
      AIRFLOW_ADMIN_PASSWORD: ${AIRFLOW_ADMIN_PASSWORD:-admin}
      AIRFLOW_ADMIN_EMAIL: ${AIRFLOW_ADMIN_EMAIL:-admin@example.com}
      PYTHONPATH: /opt/airflow
    volumes: &airflow-volumes
      - ./dags:/opt/airflow/dags
      - ./cert_study_app:/opt/airflow/cert_study_app
      - ./data:/opt/airflow/data
      - ./chroma_db:/opt/airflow/chroma_db
      - ./airflow_logs:/opt/airflow/logs
      - ./airflow_db:/opt/airflow_db

  airflow-webserver:
    build:
      context: .
      dockerfile: Dockerfile.airflow
    container_name: cert-study-airflow-webserver
    depends_on:
      airflow-init:
        condition: service_started
      postgres:
        condition: service_healthy
    command: webserver
    ports:
      - "8080:8080"
    environment: *airflow-env
    volumes: *airflow-volumes
    restart: unless-stopped

  airflow-scheduler:
    build:
      context: .
      dockerfile: Dockerfile.airflow
    container_name: cert-study-airflow-scheduler
    depends_on:
      airflow-init:
        condition: service_started
      postgres:
        condition: service_healthy
    command: scheduler
    environment: *airflow-env
    volumes: *airflow-volumes
    restart: unless-stopped

volumes:
  postgres_data: