# Copyright 2026 Emile de Bruyn # SPDX-License-Identifier: Apache-2.0 import csv import importlib.util import os import subprocess import sys import tempfile import unittest from pathlib import Path ROOT = Path(__file__).resolve().parents[1] sys.path.insert(0, str(ROOT / "scripts")) from prepare_motifbench import adapt_pdb from validate_release import residues, validate class ReleaseTests(unittest.TestCase): def test_reference(self): validate() def test_corrected_contig_preserves_motif_and_target(self): with (ROOT / "contig_specifications.csv").open(newline="") as handle: row = next(r for r in csv.DictReader(handle) if r["problem"] == "1A22_PS01352_c1") self.assertEqual(row["length"], "75") self.assertEqual(row["contig"], "0-0;A1-5;16-20;B1-3;21-25;C1-30;0-0") self.assertEqual(row["contig_original"], "0-0;B353-357;16-20;B374-376;21-25;B402-431;0-0") def test_adapter_preserves_backbone_and_masks_redesign(self): case = "1AHD_PS00027_c1" with (ROOT / "test_cases.csv").open(newline="") as handle: row = next(r for r in csv.DictReader(handle) if r["pdb_id"] == case) redesign = residues(row["redesign_idcs"]) with tempfile.TemporaryDirectory() as temp: path = Path(temp) / "adapted.pdb" adapt_pdb(ROOT / "motif_pdbs" / f"{case}.pdb", path, redesign) for line in path.read_text().splitlines(): if line.startswith("ATOM "): key = (line[21], int(line[22:26])) if key in redesign: self.assertEqual(line[17:20], "UNK") self.assertIn(line[12:16].strip(), {"N", "CA", "C", "O"}) else: self.assertNotEqual(line[17:20], "UNK") def test_reconstruction_rejects_unpinned_prosite(self): with tempfile.TemporaryDirectory() as temp: temp = Path(temp) fake = temp / "prosite.dat" fake.write_text("not the pinned PROSITE release\n") config = temp / "config.yaml" config.write_text(f"input:\n prosite_dat: {fake}\n") result = subprocess.run( [sys.executable, str(ROOT / "scripts/reconstruct.py"), "--config", str(config)], text=True, capture_output=True, ) self.assertNotEqual(result.returncode, 0) self.assertIn("wrong PROSITE input", result.stderr) @unittest.skipUnless(os.environ.get("MOTIFBENCH_WORK"), "set MOTIFBENCH_WORK for upstream smoke test") def test_upstream_motif_info_and_segment_checker(self): work = Path(os.environ["MOTIFBENCH_WORK"]) with (ROOT / "test_cases.csv").open(newline="") as handle: cases = list(csv.DictReader(handle)) module_spec = importlib.util.spec_from_file_location( "upstream_motif_info", work / "scripts/write_motifInfo_from_scaffoldInfo.py" ) upstream = importlib.util.module_from_spec(module_spec) module_spec.loader.exec_module(upstream) for index, row in enumerate(cases, 1): name = f"{index:02d}_{row['pdb_id']}" chains = upstream.read_pdb_residues(work / "motif_pdbs" / f"{name}.pdb") observed = {(chain, number) for chain, entries in chains.items() for number, name in entries.items() if name == "UNK"} self.assertEqual(observed, residues(row["redesign_idcs"]), row["pdb_id"]) with tempfile.TemporaryDirectory() as temp: info = Path(temp) / "scaffold_info.csv" output = Path(temp) / "motif_info.csv" indices = ( 0, next(i for i, r in enumerate(cases) if ";" in r["motif_residues"]), next(i for i, r in enumerate(cases) if r["pdb_id"] == "1A22_PS01352_c1"), ) for index in indices: row = cases[index] name = f"{index + 1:02d}_{row['pdb_id']}" chains = [segment[0] for segment in row["motif_residues"].split(";")] with info.open("w", newline="") as handle: writer = csv.writer(handle) writer.writerow(["sample_num", "motif_placements"]) placement = "0/A/16/B/21/C/0" if row["pdb_id"] == "1A22_PS01352_c1" else "/".join(["0"] + chains + ["0"]) writer.writerow([0, placement]) subprocess.run([sys.executable, str(work / "scripts/write_motifInfo_from_scaffoldInfo.py"), str(info), str(work / "motif_pdbs" / f"{name}.pdb"), str(output)], check=True) with output.open(newline="") as handle: result = next(csv.DictReader(handle)) self.assertEqual(result["pdb_name"], name) self.assertEqual(result["segment_order"], ";".join(chains)) self.assertEqual(residues(result["redesign_positions"]), residues(row["redesign_idcs"])) subprocess.run([sys.executable, str(work / "scripts/check_segment_validity.py"), "-i", str(work / "test_cases.csv"), "-t", str(output)], check=True) if row["pdb_id"] == "1A22_PS01352_c1": with (work / "example/contig_specifications.csv").open(newline="") as handle: spec = next(r for r in csv.DictReader(handle) if r["problem"] == name) self.assertEqual(spec["contig"], "0-0;A1-5;16-20;B1-3;21-25;C1-30;0-0") @unittest.skipUnless(os.environ.get("MOTIFBENCH_WORK"), "set MOTIFBENCH_WORK for upstream smoke test") def test_upstream_summary_is_ordered_before_scoring(self): work = Path(os.environ["MOTIFBENCH_WORK"]) with (ROOT / "test_cases.csv").open(newline="") as handle: cases = list(csv.DictReader(handle)) first = f"01_{cases[0]['pdb_id']}" last = f"100_{cases[99]['pdb_id']}" with tempfile.TemporaryDirectory() as temp: result = Path(temp) cluster = result / first / "esm_successful_backbones" / "diversity_cluster.tsv" cluster.parent.mkdir(parents=True) cluster.write_text(f"{first}_0.pdb {first}_0.pdb\n") with (result / "summary_by_problem.csv").open("w", newline="") as handle: writer = csv.writer(handle) writer.writerow(["Problem", "Num_Solutions", "Novelty", "Success_Rate", "Num_Evaluated"]) writer.writerow([last, 0, 0, 0, 100]) writer.writerow([first, 1, 0, 1, 100]) subprocess.run([sys.executable, str(ROOT / "scripts/summarize_motifbench.py"), str(work), str(result)], check=True) with (result / "summary_by_problem_ordered.csv").open(newline="") as handle: scored = list(csv.DictReader(handle)) self.assertEqual([r["Problem"] for r in scored], [first, last]) self.assertGreater(float(scored[0]["overall_score"]), 0) self.assertEqual(float(scored[1]["overall_score"]), 0) if __name__ == "__main__": unittest.main()