scripts/deploy_to_hf_spaces.py

"""Deploy BankMind to a Hugging Face Spaces Docker SDK Space.

Reads `HUGGINGFACE_TOKEN`, `ANTHROPIC_API_KEY`, `QDRANT_URL`, `QDRANT_API_KEY`
from .env. Creates the Space (idempotent: ok if it already exists), sets the
3 runtime secrets, then uploads the repo contents (excluding raw/processed
data, .venv, .git, etc.). Prints the resulting URL.

Usage: python scripts/deploy_to_hf_spaces.py [--space-name USER/REPO]
Default: Arjun10g/bankmind
"""
from __future__ import annotations

import argparse
import os
import sys
from pathlib import Path

from dotenv import load_dotenv

ROOT = Path(__file__).resolve().parents[1]
load_dotenv(ROOT / ".env")

DEFAULT_SPACE = "Arjun10g/bankmind"

# Runtime secrets to set on the Space. These are required for queries that hit
# Qdrant or invoke the LLM. They never leave the user's HF account once set.
SECRETS_TO_SET = ("ANTHROPIC_API_KEY", "QDRANT_URL", "QDRANT_API_KEY")

IGNORE_PATTERNS = [
    ".git/*", ".git/**",
    ".venv/*", ".venv/**",
    ".claude/*", ".claude/**",
    "venv/*", "venv/**",
    "__pycache__/*", "__pycache__/**", "*/__pycache__/*", "*/__pycache__/**",
    "*.pyc", "*.pyo",
    ".env", ".env.local",
    "data/raw/*", "data/raw/**",
    "data/processed/*", "data/processed/**",
    "logs/*", "logs/**",
    ".pytest_cache/*", ".pytest_cache/**",
    ".ruff_cache/*", ".ruff_cache/**",
    ".mypy_cache/*", ".mypy_cache/**",
    ".DS_Store",
    "*.egg-info/*", "*.egg-info/**",
    "build/*", "build/**",
    "dist/*", "dist/**",
    ".uv-cache/*", ".uv-cache/**",
    ".cache/*", ".cache/**",
]


def main() -> int:
    ap = argparse.ArgumentParser()
    ap.add_argument("--space-name", default=DEFAULT_SPACE,
                    help=f"HF Space repo (USER/REPO). Default: {DEFAULT_SPACE}")
    ap.add_argument("--private", action="store_true",
                    help="Create the Space as private (default: public)")
    args = ap.parse_args()

    token = os.environ.get("HUGGINGFACE_TOKEN", "").strip()
    if not token:
        print("ERROR: HUGGINGFACE_TOKEN not set in .env", file=sys.stderr)
        print("       Get one at https://huggingface.co/settings/tokens (Write scope)",
              file=sys.stderr)
        return 1

    from huggingface_hub import HfApi
    from huggingface_hub.utils import RepositoryNotFoundError

    api = HfApi(token=token)

    print(f"\n=== Deploying to {args.space_name} ===")

    # 1. Create Space (idempotent)
    print(f"\n[1/3] Creating Space (Docker SDK, "
          f"{'private' if args.private else 'public'})...")
    try:
        api.repo_info(args.space_name, repo_type="space")
        print(f"  Space {args.space_name} already exists — will overwrite contents.")
    except RepositoryNotFoundError:
        api.create_repo(
            repo_id=args.space_name,
            repo_type="space",
            space_sdk="docker",
            private=args.private,
        )
        print(f"  Created new Space.")

    # 2. Set secrets
    print(f"\n[2/3] Setting Space secrets...")
    n_set = 0
    for key in SECRETS_TO_SET:
        value = os.environ.get(key, "").strip()
        if not value:
            print(f"  SKIP {key} (not in .env)")
            continue
        api.add_space_secret(repo_id=args.space_name, key=key, value=value)
        n_set += 1
        print(f"  SET  {key}")
    print(f"  → {n_set}/{len(SECRETS_TO_SET)} secrets set.")

    # 3. Upload repo contents
    print(f"\n[3/3] Uploading repo contents...")
    api.upload_folder(
        repo_id=args.space_name,
        repo_type="space",
        folder_path=str(ROOT),
        ignore_patterns=IGNORE_PATTERNS,
        commit_message="Deploy BankMind",
    )

    url = f"https://huggingface.co/spaces/{args.space_name}"
    print(f"\n✅ Deployed → {url}")
    print(f"\nThe Space will rebuild from the Dockerfile. First build takes ~10-15 minutes.")
    print(f"Monitor build progress at: {url}/?logs=container")
    return 0


if __name__ == "__main__":
    sys.exit(main())