scripts/entrypoint.sh

#!/usr/bin/env bash
# Demarrage du Space : telecharge l'index (dataset HF prive) si absent, puis lance l'app.
set -e

INDEX_DIR="/app/pcp/data/index"
if [ ! -f "${INDEX_DIR}/embeddings.npy" ] && [ -n "${PCP_INDEX_DATASET}" ]; then
  echo "[entrypoint] telechargement de l'index depuis ${PCP_INDEX_DATASET}..."
  python - <<'PY'
import os
from huggingface_hub import snapshot_download
repo = os.environ["PCP_INDEX_DATASET"]
token = os.environ.get("HF_TOKEN") or None
dst = "/app/pcp/data/index"
os.makedirs(dst, exist_ok=True)
snapshot_download(repo_id=repo, repo_type="dataset", local_dir=dst, token=token,
                  allow_patterns=["chunks.json", "embeddings.npy", "meta.json",
                                  "catalogue_vecs.npy", "catalogue_vecs_meta.json",
                                  "example_vecs.npy", "example_blocks.json"])
print("[entrypoint] index telecharge ->", os.listdir(dst))
PY
fi

exec uvicorn pcp.server:app --host 0.0.0.0 --port 7860