Download code/ezflow_v3/cfd/normalize_external.py from BlidReview/steady-rans-surrogates: direct link, hf CLI and curl.
- Browser
- Download file 876 Bytes
-
https://huggingface.co/BlidReview/steady-rans-surrogates/resolve/main/code/ezflow_v3/cfd/normalize_external.py
- Command line
-
hf download hf://BlidReview/steady-rans-surrogates/code/ezflow_v3/cfd/normalize_external.py
-
curl -L -o normalize_external.py https://huggingface.co/BlidReview/steady-rans-surrogates/resolve/main/code/ezflow_v3/cfd/normalize_external.py
876 Bytes
| """Normalize an external STL (AhmedML / DrivAerML / NASA-CRM) to the EZFlow | |
| convention: max-extent L=1, centred at origin, and compute the frontal (y-z | |
| projected) reference area for Cd. Reuses the SAME normalizer the training data | |
| used (family_to_stl._normalize_and_measure), so the zero-shot geometries enter | |
| the model on exactly the same scale as training. | |
| Usage: | |
| python normalize_external.py <stl_in> <name> <out_dir> | |
| Prints one line: NORM <name> Aref=<aref> -> <out_stl> | |
| """ | |
| import sys, os | |
| sys.path.insert(0, r"C:\dev\EZFlow") | |
| from ezflow_v3.cfd.family_to_stl import _normalize_and_measure | |
| stl_in, name, out_dir = sys.argv[1], sys.argv[2], sys.argv[3] | |
| os.makedirs(out_dir, exist_ok=True) | |
| out_stl = os.path.join(out_dir, f"{name}.stl") | |
| lref, aref = _normalize_and_measure(stl_in, out_stl) | |
| print(f"NORM {name} Lref={lref:.3f} Aref={aref:.6f} -> {out_stl}", flush=True) | |