upload_to_hf.sh

#!/usr/bin/env bash
# upload_to_hf.sh
# ----------------
# Upload the pipeline/ folder to a HuggingFace Space.
#
# Usage:
#   ./upload_to_hf.sh [HF_SPACE]
#
# Examples:
#   ./upload_to_hf.sh your-username/SpatialBench
#   HF_TOKEN=hf_xxx ./upload_to_hf.sh your-username/SpatialBench
#
# The script uploads only the files needed to run the Space:
#   app.py, pipeline/, configs/, requirements.txt, README.md
# It never uploads .env, __pycache__, or result/log directories.
#
# Requirements:
#   pip install huggingface_hub

set -euo pipefail

# ---------------------------------------------------------------------------
# Configuration
# ---------------------------------------------------------------------------

HF_SPACE="${1:-}"
if [[ -z "$HF_SPACE" ]]; then
    echo "Usage: $0 <your-username/SpaceName>"
    echo "Example: $0 weijiang/SpatialBench"
    exit 1
fi

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"

# ---------------------------------------------------------------------------
# Auth
# ---------------------------------------------------------------------------

if [[ -z "${HF_TOKEN:-}" ]]; then
    # Try reading from huggingface-cli login cache
    HF_TOKEN_FILE="${HOME}/.cache/huggingface/token"
    if [[ -f "$HF_TOKEN_FILE" ]]; then
        HF_TOKEN="$(cat "$HF_TOKEN_FILE")"
        export HF_TOKEN
    fi
fi

if [[ -z "${HF_TOKEN:-}" ]]; then
    echo "No HuggingFace token found."
    echo "Either:"
    echo "  1. Run:  huggingface-cli login"
    echo "  2. Set:  export HF_TOKEN=hf_xxx"
    exit 1
fi

echo "Uploading to HuggingFace Space: $HF_SPACE"
echo "Source directory: $SCRIPT_DIR"
echo ""

# ---------------------------------------------------------------------------
# Upload via Python (huggingface_hub)
# ---------------------------------------------------------------------------

python - <<PYEOF
import os
import sys
from pathlib import Path
from huggingface_hub import HfApi, upload_folder

api = HfApi(token=os.environ["HF_TOKEN"])
space_id = "$HF_SPACE"
local_dir = Path("$SCRIPT_DIR")

# Ensure the Space exists (type=gradio); if it already exists this is a no-op.
try:
    api.create_repo(
        repo_id=space_id,
        repo_type="space",
        space_sdk="gradio",
        exist_ok=True,
        private=True,
    )
    print(f"Space ready: https://huggingface.co/spaces/{space_id}")
except Exception as e:
    print(f"[WARN] Could not create/verify Space: {e}")
    print("Proceeding with upload anyway...")

# Files/dirs to ignore (never upload secrets, cache, or large result dirs)
IGNORE_PATTERNS = [
    ".env",
    "*.pyc",
    "__pycache__",
    "*.egg-info",
    ".git",
    ".gitignore",
    # Large result/log directories that live outside pipeline/ anyway
    "spatial-reasoning-results*",
    "spatial_reasoning_logs",
    "llm-maze-evaluation-results",
    "eval_llm_logs",
]

print(f"\nUploading files from: {local_dir}")
print("Ignoring: " + ", ".join(IGNORE_PATTERNS))
print("")

url = upload_folder(
    repo_id=space_id,
    repo_type="space",
    folder_path=str(local_dir),
    path_in_repo=".",
    ignore_patterns=IGNORE_PATTERNS,
    commit_message="Update SpatialBench pipeline",
    token=os.environ["HF_TOKEN"],
)

print(f"\nUpload complete!")
print(f"Space URL: https://huggingface.co/spaces/{space_id}")
PYEOF