setup.py

"""
Quick-start setup script:
  1. Creates virtual environment
  2. Installs dependencies
  3. Creates .env from template
  4. Initialises the database
  5. Optionally downloads spaCy model
"""
import subprocess
import sys
import os
import shutil
from pathlib import Path

BASE = Path(__file__).parent


def run(cmd, **kw):
    print(f"\n▶  {' '.join(cmd) if isinstance(cmd, list) else cmd}")
    subprocess.run(cmd, check=True, **kw)


def main():
    # Virtual environment
    venv = BASE / "venv"
    py = sys.executable

    if not venv.exists():
        print("Creating virtual environment…")
        run([py, "-m", "venv", str(venv)])

    # Determine venv python/pip
    if sys.platform == "win32":
        venv_py  = str(venv / "Scripts" / "python.exe")
        venv_pip = str(venv / "Scripts" / "pip.exe")
    else:
        venv_py  = str(venv / "bin" / "python")
        venv_pip = str(venv / "bin" / "pip")

    # Install dependencies
    print("\nInstalling Python dependencies…")
    run([venv_pip, "install", "--upgrade", "pip"])
    run([venv_pip, "install", "-r", str(BASE / "requirements.txt")])

    # .env
    env_example = BASE / ".env.example"
    env_file    = BASE / ".env"
    if not env_file.exists() and env_example.exists():
        shutil.copy(env_example, env_file)
        print(f"\n✅  Created .env  – edit it to set your TESSERACT_CMD and other settings.")
    else:
        print(f"\nℹ️   .env already exists – skipping.")

    # Create directories
    for d in ["uploads", "logs"]:
        (BASE / d).mkdir(exist_ok=True)

    # DB init
    print("\nInitialising database…")
    env = os.environ.copy()
    env["PYTHONPATH"] = str(BASE)
    run([venv_py, "-c", "from app.database import init_db; init_db(); print('DB ready.')"],
        cwd=str(BASE), env=env)

    # spaCy model (optional)
    try:
        run([venv_py, "-m", "spacy", "download", "en_core_web_sm"], cwd=str(BASE))
    except Exception:
        print("⚠️  Could not download spaCy model – NER will be skipped. Run manually:")
        print(f"    {venv_py} -m spacy download en_core_web_sm")

    print("\n" + "="*60)
    print("✅  Setup complete!")
    print(f"   Start the server: {venv_py} run.py")
    print("   Dashboard:        http://127.0.0.1:8000")
    print("   API Docs:         http://127.0.0.1:8000/api/docs")
    print("="*60)


if __name__ == "__main__":
    main()