fetch_drivaernet.py

"""Download the small DrivAerNet++ aerodynamic-coefficient CSVs.

DrivAerNet++ (Elrefaie et al., MIT, CC-BY-NC-4.0) ships its full 39 TB of
meshes + CFD volume fields on Harvard Dataverse (use Globus for that).
But the per-design DRAG coefficients and frontal areas are published as
two small CSVs via the authors' Dropbox — a few hundred KB total — and
that is all we need for:
  * drivaernet_data.retrieve_by_cd()  (real CFD retrieval)
  * dataset statistics / calibration
  * labels for a parametric (26-param) drag surrogate

This script fetches just those two CSVs into ./DrivAerNet/.

Full dataset (meshes, point clouds, pressure / wall-shear / volume fields)
is NOT downloaded here — see:
  https://dataverse.harvard.edu/dataverse/DrivAerNet   (39 TB, Globus)
  https://github.com/Mohamedelrefaie/DrivAerNet

Usage:
    python fetch_drivaernet.py
"""
from __future__ import annotations

import sys
import urllib.request
from pathlib import Path

DATA_DIR = Path(__file__).parent / "DrivAerNet"

# Official author-hosted Dropbox links (dl=1 forces a direct download).
FILES = {
    "DrivAerNetPlusPlus_Cd_8k.csv":
        "https://www.dropbox.com/scl/fi/2rtchqnpmzy90uwa9wwny/"
        "DrivAerNetPlusPlus_Cd_8k_Updated.csv"
        "?rlkey=vjnjurtxfuqr40zqgupnks8sn&dl=1",
    "DrivAerNetPlusPlus_Areas.csv":
        "https://www.dropbox.com/scl/fi/b7fenj0wmhzqx64bj82t1/"
        "DrivAerNetPlusPlus_CarDesign_Areas.csv"
        "?rlkey=usbunuupxwmx6g49r9r7dh8zk&dl=1",
}


def fetch():
    DATA_DIR.mkdir(parents=True, exist_ok=True)
    for name, url in FILES.items():
        dest = DATA_DIR / name
        if dest.exists() and dest.stat().st_size > 0:
            print(f"[fetch] {name} already present ({dest.stat().st_size:,} B)")
            continue
        print(f"[fetch] downloading {name} ...")
        req = urllib.request.Request(url, headers={"User-Agent": "Mozilla/5.0"})
        with urllib.request.urlopen(req, timeout=60) as r, open(dest, "wb") as f:
            f.write(r.read())
        n = dest.stat().st_size
        print(f"[fetch]   -> {dest}  ({n:,} B)")
        if n < 1000:
            print(f"[fetch]   WARNING: file looks too small; check the URL.")
    # quick sanity report
    try:
        import drivaernet_data as DN
        s = DN.dataset_stats()
        if s.get("available"):
            print(f"[fetch] OK — {s['n']} real designs, "
                  f"Cd mean={s['cd_mean']:.3f} "
                  f"[{s['cd_min']:.3f}, {s['cd_max']:.3f}]")
    except Exception as e:
        print(f"[fetch] note: stats check skipped ({e})")


if __name__ == "__main__":
    try:
        fetch()
    except Exception as e:
        print(f"[fetch] FAILED: {e}", file=sys.stderr)
        sys.exit(1)