# compose.yml — Peregrine by Circuit Forge LLC
# Profiles: remote | cpu | single-gpu | dual-gpu
services:

  app:
    build: .
    ports:
      - "${STREAMLIT_PORT:-8501}:8501"
    volumes:
      - ./config:/app/config
      - ./data:/app/data
      - ${DOCS_DIR:-~/Documents/JobSearch}:/docs
    environment:
      - STAGING_DB=/app/data/staging.db
      - DOCS_DIR=/docs
      - ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY:-}
      - OPENAI_COMPAT_URL=${OPENAI_COMPAT_URL:-}
      - OPENAI_COMPAT_KEY=${OPENAI_COMPAT_KEY:-}
    depends_on:
      searxng:
        condition: service_healthy
    extra_hosts:
      - "host.docker.internal:host-gateway"
    restart: unless-stopped

  searxng:
    image: searxng/searxng:latest
    ports:
      - "${SEARXNG_PORT:-8888}:8080"
    volumes:
      - ./docker/searxng:/etc/searxng:ro
    healthcheck:
      test: ["CMD", "wget", "-q", "--spider", "http://localhost:8080/"]
      interval: 10s
      timeout: 5s
      retries: 3
    restart: unless-stopped

  ollama:
    image: ollama/ollama:latest
    ports:
      - "${OLLAMA_PORT:-11434}:11434"
    volumes:
      - ${OLLAMA_MODELS_DIR:-~/models/ollama}:/root/.ollama
      - ./docker/ollama/entrypoint.sh:/entrypoint.sh
    environment:
      - OLLAMA_MODELS=/root/.ollama
      - DEFAULT_OLLAMA_MODEL=${OLLAMA_DEFAULT_MODEL:-llama3.2:3b}
    entrypoint: ["/bin/bash", "/entrypoint.sh"]
    profiles: [cpu, single-gpu, dual-gpu]
    restart: unless-stopped

  vision:
    build:
      context: .
      dockerfile: scripts/vision_service/Dockerfile
    ports:
      - "${VISION_PORT:-8002}:8002"
    environment:
      - VISION_MODEL=${VISION_MODEL:-vikhyatk/moondream2}
      - VISION_REVISION=${VISION_REVISION:-2025-01-09}
    profiles: [single-gpu, dual-gpu]
    restart: unless-stopped

  vllm:
    image: vllm/vllm-openai:latest
    ports:
      - "${VLLM_PORT:-8000}:8000"
    volumes:
      - ${VLLM_MODELS_DIR:-~/models/vllm}:/models
    command: >
      --model /models/${VLLM_MODEL:-Ouro-1.4B}
      --trust-remote-code
      --max-model-len 4096
      --gpu-memory-utilization 0.75
      --enforce-eager
      --max-num-seqs 8
      --cpu-offload-gb ${CPU_OFFLOAD_GB:-0}
    profiles: [dual-gpu]
    restart: unless-stopped

  finetune:
    build:
      context: .
      dockerfile: Dockerfile.finetune
    volumes:
      - ${DOCS_DIR:-~/Documents/JobSearch}:/docs
      - ${OLLAMA_MODELS_DIR:-~/models/ollama}:/ollama-models
      - ./config:/app/config
    environment:
      - DOCS_DIR=/docs
      - OLLAMA_URL=http://ollama:11434
      - OLLAMA_MODELS_MOUNT=/ollama-models
      - OLLAMA_MODELS_OLLAMA_PATH=/root/.ollama
    depends_on:
      ollama:
        condition: service_started
    profiles: [cpu, single-gpu, dual-gpu]
    restart: "no"