BILBY-PRO / construction /tests /test_normalisation_pipeline.py
EdeBruyn's picture
Initial upload for camera-ready version of NeurIPS paper.
23a728e verified
Raw History Blame Contribute Delete
2.24 kB
# Copyright 2026 Emile de Bruyn
# SPDX-License-Identifier: Apache-2.0
from pathlib import Path
import pandas as pd
from motifbench_prosite.normalisation import NormalisationConfig, run_normalisation
from test_normalisation_pdb import _write_tiny_pdb
def test_run_normalisation_roundtrip(tmp_path: Path):
output_root = tmp_path / "motifbench_candidates"
output_root.mkdir(parents=True, exist_ok=True)
motif_pdb_dir = output_root / "motif_pdbs"
motif_pdb_dir.mkdir()
benchmark_id = "01_FOO"
pdb_path = motif_pdb_dir / f"{benchmark_id}.pdb"
_write_tiny_pdb(pdb_path, chain_id="L", n_res=3)
meta_df = pd.DataFrame(
[
{
"benchmark_pdb_id": benchmark_id,
"motif_segments": "L24-26",
"redesignable_segments": "L24-25",
"not_motif_segments": "L23;L27",
}
]
)
meta_path = output_root / "motif_cases_metadata.csv"
meta_df.to_csv(meta_path, index=False)
test_df = pd.DataFrame(
[
{
"pdb_id": benchmark_id,
"motif_residues": "L24-26",
"redesign_idcs": "L24-25",
"length": 75,
"group": 4,
"not_motif": "L23;L27",
}
]
)
test_path = output_root / "test_cases.csv"
test_df.to_csv(test_path, index=False)
contig_df = pd.DataFrame(
[
{
"problem": benchmark_id,
"length": 75,
"contig": "0-10;L24-26;0-10",
}
]
)
contig_path = output_root / "contig_specifications.csv"
contig_df.to_csv(contig_path, index=False)
cfg = NormalisationConfig(output_root=output_root, backup_originals=False)
new_meta, new_test, new_contig = run_normalisation(cfg)
assert new_meta.loc[0, "motif_segments"] == "A1-3"
assert new_meta.loc[0, "redesignable_segments"] == "A1-2"
assert new_meta.loc[0, "not_motif_segments"] == ""
assert new_test.loc[0, "motif_residues"] == "A1-3"
assert new_test.loc[0, "redesign_idcs"] == "A1-2"
assert new_test.loc[0, "not_motif"] == ""
assert new_contig.loc[0, "contig"] == "0-10;A1-3;0-10"