scripts/hackathon_upload.sh

#!/usr/bin/env bash
# Upload the FormScout source tree to the Build Small Hackathon Space, as a PR.
#
# Usage:
#   ./scripts/hackathon_upload.sh                    # message from last git commit
#   ./scripts/hackathon_upload.sh "feat: my change"  # custom message
#
# Pushes (PR) to:
#   spaces/build-small-hackathon/small-functional-movement-screening
#
# The Space already exists, so this script only uploads — it does not create the
# repo. Like hf_upload.sh, `hf upload` ignores .hfignore, so we parse it into
# --exclude globs ourselves. Above LARGE_THRESHOLD files it falls back to
# `hf upload-large-folder` (resumable, but commits to main directly — no PR).
set -euo pipefail

cd "$(dirname "$0")/.."

HACKATHON_OWNER="${FORMSCOUT_HF_HACKATHON_OWNER:-build-small-hackathon}"
REPO_NAME="small-functional-movement-screening"
SPACE_REPO="spaces/$HACKATHON_OWNER/$REPO_NAME"
MSG="${1:-$(git log -1 --pretty=%s)}"
LARGE_THRESHOLD="${FORMSCOUT_HF_LARGE_THRESHOLD:-500}"

# Belt-and-suspenders extras on top of .hfignore. `.cache/` is the resume
# state upload-large-folder writes into the folder being uploaded.
PATTERNS=(
    "*.pdf"
    "**/node_modules/**"
    ".cache/**"
)

# Parse .hfignore into fnmatch-style globs. fnmatch's `*` crosses `/`, but a
# bare name like `.DS_Store` or `dir/` only matches at the root, so emit both
# the rooted and `**/`-prefixed forms.
while IFS= read -r line; do
    line="${line%%#*}"
    line="${line#"${line%%[![:space:]]*}"}"
    line="${line%"${line##*[![:space:]]}"}"
    [[ -z "$line" ]] && continue
    if [[ "$line" == */ ]]; then
        PATTERNS+=("${line}**" "**/${line}**")
    else
        PATTERNS+=("$line" "**/$line")
    fi
done < .hfignore

EXCLUDES=()
for p in "${PATTERNS[@]}"; do
    EXCLUDES+=(--exclude="$p")
done

# Count what would actually be uploaded, using the same filter the hub client
# applies, so the mode decision matches reality.
N_FILES=$(python3 - "${PATTERNS[@]}" <<'EOF'
import sys
from pathlib import Path
from huggingface_hub.utils import filter_repo_objects

patterns = sys.argv[1:]
files = (
    str(p) for p in Path(".").rglob("*")
    if p.is_file() and p.parts[0] != ".git"
)
print(len(list(filter_repo_objects(files, ignore_patterns=patterns))))
EOF
)
echo "── $N_FILES files to upload after .hfignore filtering"

if (( N_FILES == 0 )); then
    echo "✗ nothing to upload — check .hfignore" >&2
    exit 1
fi

if (( N_FILES > LARGE_THRESHOLD )); then
    echo "── $SPACE_REPO: $N_FILES files > $LARGE_THRESHOLD, using upload-large-folder"
    echo "   (resumable; commits directly to main — no PR, no custom message)"
    hf upload-large-folder "$SPACE_REPO" . "${EXCLUDES[@]}"
else
    echo "── uploading (PR) to: $SPACE_REPO"
    hf upload "$SPACE_REPO" . . "${EXCLUDES[@]}" --create-pr --commit-message="$MSG"
fi

echo "✓ done — PR opened on $SPACE_REPO (review & merge in the HF UI)"