docker-compose.yml

version: "3.9"

# ─────────────────────────────────────────────────────────────────────────────
# Job Discovery System — Full Stack
#
# Services:
#   redis        — shared message broker
#   postgres     — shared database
#   crawler      — Scrapy job crawler (one-shot)
#   aggregator   — FastAPI stream consumer + jobs REST API  (:8000)
#   scraper      — Platform scraper (Naukri/LinkedIn/Internshala)  (:8001)
#   contact      — Contact discovery pipeline  (:8002)
#   email-gen    — Cold email generator  (:8003)
#   gateway      — API gateway + dashboard  (:8080)
#   scheduler    — APScheduler pipeline automation (background)
# ─────────────────────────────────────────────────────────────────────────────

x-common-env: &common-env
  REDIS_HOST: redis
  REDIS_PORT: "6379"
  DATABASE_URL: postgresql://${POSTGRES_USER:-jobuser}:${POSTGRES_PASSWORD:-jobpass}@postgres:5432/${POSTGRES_DB:-jobsdb}

services:

  # ── Redis ──────────────────────────────────────────────────────────────────
  redis:
    image: redis:7-alpine
    ports:
      - "6379:6379"
    command: redis-server --maxmemory 512mb --maxmemory-policy allkeys-lru
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      retries: 5
    networks: [jobnet]

  # ── PostgreSQL ─────────────────────────────────────────────────────────────
  postgres:
    image: postgres:16-alpine
    environment:
      POSTGRES_USER:     ${POSTGRES_USER:-jobuser}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-jobpass}
      POSTGRES_DB:       ${POSTGRES_DB:-jobsdb}
    ports:
      - "5433:5432"
    volumes:
      - pgdata:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-jobuser} -d ${POSTGRES_DB:-jobsdb}"]
      interval: 5s
      retries: 10
    networks: [jobnet]

  # ── Crawler (Scrapy, one-shot) ─────────────────────────────────────────────
  crawler:
    build:
      context: ./crawler-service
      dockerfile: Dockerfile
    environment:
      <<: *common-env
      JOBSEEKER_ROLE:  ${JOBSEEKER_ROLE:-Backend Engineer}
      JOBSEEKER_STACK: ${JOBSEEKER_STACK:-Python,FastAPI,PostgreSQL,Redis,Go}
    depends_on:
      redis:
        condition: service_healthy
    restart: "no"
    command: ["scrapy", "crawl", "remotive"]
    networks: [jobnet]

  # ── Aggregator ─────────────────────────────────────────────────────────────
  aggregator:
    build:
      context: ./aggregator-service
      dockerfile: Dockerfile
    environment:
      <<: *common-env
    depends_on:
      redis:
        condition: service_healthy
      postgres:
        condition: service_healthy
    networks: [jobnet]
    restart: unless-stopped
    healthcheck:
      test: ["CMD-SHELL", "curl -sf http://localhost:8000/health || exit 1"]
      interval: 10s
      retries: 5
      start_period: 15s

  # ── Platform Scraper ───────────────────────────────────────────────────────
  scraper:
    build:
      context: ./scraper-service
      dockerfile: Dockerfile
    environment:
      <<: *common-env
    depends_on:
      redis:
        condition: service_healthy
    networks: [jobnet]
    restart: unless-stopped
    healthcheck:
      test: ["CMD-SHELL", "curl -sf http://localhost:8000/health || exit 1"]
      interval: 10s
      retries: 5
      start_period: 20s

  # ── Contact Discovery ──────────────────────────────────────────────────────
  contact:
    build:
      context: ./contact-discovery-service
      dockerfile: Dockerfile
    environment:
      <<: *common-env
      GITHUB_TOKEN: ${GITHUB_TOKEN:-}
    depends_on:
      postgres:
        condition: service_healthy
    networks: [jobnet]
    restart: unless-stopped
    healthcheck:
      test: ["CMD-SHELL", "curl -sf http://localhost:8000/health || exit 1"]
      interval: 10s
      retries: 5
      start_period: 25s

  # ── Email Generator ────────────────────────────────────────────────────────
  email-gen:
    build:
      context: ./email-generator-service
      dockerfile: Dockerfile
    environment:
      <<: *common-env
      OLLAMA_BASE_URL:    ${OLLAMA_BASE_URL:-http://host.docker.internal:11434}
      GMAIL_ADDRESS:      ${GMAIL_ADDRESS:-}
      GMAIL_APP_PASSWORD: ${GMAIL_APP_PASSWORD:-}
      YOUR_NAME:          ${YOUR_NAME:-Applicant}
      YOUR_GITHUB_URL:    ${YOUR_GITHUB_URL:-}
    depends_on:
      postgres:
        condition: service_healthy
    networks: [jobnet]
    restart: unless-stopped
    healthcheck:
      test: ["CMD-SHELL", "curl -sf http://localhost:8000/health || exit 1"]
      interval: 10s
      retries: 5
      start_period: 15s

  # ── API Gateway + Dashboard ────────────────────────────────────────────────
  gateway:
    build:
      context: ./gateway
      dockerfile: Dockerfile
    ports:
      - "${GATEWAY_PORT:-8080}:8080"
    environment:
      AGGREGATOR_URL: http://aggregator:8000
      SCRAPER_URL:    http://scraper:8000
      CONTACT_URL:    http://contact:8000
      EMAIL_GEN_URL:  http://email-gen:8000
      SUMMARY_PATH:   /data/run_summary.json
    volumes:
      - scheduler_data:/data
    depends_on:
      aggregator:
        condition: service_healthy
      scraper:
        condition: service_healthy
      contact:
        condition: service_healthy
      email-gen:
        condition: service_healthy
    networks: [jobnet]
    restart: unless-stopped
    healthcheck:
      test: ["CMD-SHELL", "curl -sf http://localhost:8080/api/health || exit 1"]
      interval: 15s
      retries: 5
      start_period: 30s

  # ── Scheduler (APScheduler background process) ───────────────────────────────
  scheduler:
    build:
      context: ./scheduler
      dockerfile: Dockerfile
    environment:
      GATEWAY_URL:            http://gateway:8080
      JOBSEEKER_ROLE:         ${JOBSEEKER_ROLE:-Backend Engineer}
      JOBSEEKER_STACK:        ${JOBSEEKER_STACK:-Python,FastAPI,PostgreSQL,Redis,Go}
      CRAWL_INTERVAL_HOURS:   ${CRAWL_INTERVAL_HOURS:-8}
      DISCOVER_INTERVAL_HOURS: ${DISCOVER_INTERVAL_HOURS:-24}
      SCRAPER_WAIT_SECS:      ${SCRAPER_WAIT_SECS:-60}
      DISCOVER_DELAY_SECS:    ${DISCOVER_DELAY_SECS:-30}
      SUMMARY_PATH:           /data/run_summary.json
    volumes:
      - scheduler_data:/data   # shared with gateway for GET /api/summary
    depends_on:
      gateway:
        condition: service_healthy
    networks: [jobnet]
    restart: unless-stopped

volumes:
  pgdata:
  scheduler_data:    # shared between scheduler (writer) and gateway (reader)

networks:
  jobnet:
    driver: bridge