Python reader for FWH binary surface data

[riddhi1228]

Adds SU2.io.fwh.FWHData, a pure-Python reader for the FWH binary surface data format. The committed reference file TestCases/unsteady/square_cylinder/fwh_bin.dat had no corresponding Python reader in the repository. This PR closes that gap.

Binary format (decoded from reference file)

Header : 3 × uint32 little-endian
           [0] start_time_us  — simulation time tag at first timestep
                                (inferred as microseconds; TBC with maintainers)
           [1] n_timesteps    — number of timesteps in file
           [2] n_points       — number of FWH surface points
Data   : n_timesteps × n_points × float32 little-endian
           pressure[t, i]  — static pressure at point i, timestep t (Pa)

Verified against fwh_bin.dat: n_t=50, n_pts=200, pressure range 825–1104 Pa, temporal RMS max ≈74 Pa — physically consistent with Ma=0.1 RANS flow past a square cylinder.

API

from SU2.io.fwh import FWHData
fwh = FWHData.from_file("fwh_bin.dat", time_step=0.0015)
# fwh.pressure          shape (50, 200), float32, Pa
# fwh.time              shape (50,), float64, seconds
# fwh.pressure_perturbation()   p' = p - p_mean
# fwh.pressure_rms()            per-point RMS of p'
# fwh.to_dataframe()            pandas DataFrame

Testing

17 pytest tests in SU2_PY/tests/test_fwh_reader.py:

• Synthetic roundtrip (shape, values, time axis, edge cases)
• Error handling (file not found, truncated file)
• Regression guards against committed reference file

Notes

• Pure Python, no rebuild required
• Touches only SU2_PY/
• This reader is intended as the data-ingest layer for the GSoC 2026 SU2PY_FWH aeroacoustic post-processor implementing the Ffowcs Williams–Hawkings integral formulation

cc @pcarruscag

[pcarruscag]

Please setup your AI tool properly and respect the PR description template

[riddhi1228]

Sorry about that — I'll reopen using the proper PR description template. Thanks for the pointer.