Datasets:
File size: 4,988 Bytes
e579e7e | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 | """Round-trip verification of the TRIP50 release artefacts.
Three checks, run in order:
1. MANIFEST.sha256 matches every payload file's actual digest.
2. Recomputed per-method MAE table from species.parquet reproduces the
published trip50_dft_mae.csv to within 0.005 kcal/mol per cell.
3. Extended-XYZ round-trip: every frame's species_id, atomic_numbers, and
positions agree with species.parquet to within 1e-7 Å (extxyz default
8-digit precision).
Run from the release root or the examples/ directory.
"""
from __future__ import annotations
import hashlib
import sys
from pathlib import Path
import numpy as np
import pandas as pd
from ase.io import read
ROOT = Path(__file__).resolve().parent.parent
DATA = ROOT / "data"
PRODUCTION = Path("/Users/rpaton/TRIP50/production")
sys.path.insert(0, str(ROOT / "scripts"))
from trip50_core import reaction_quantities # noqa: E402
POSITION_TOL = 1e-7
MAE_TOL = 0.005 # kcal/mol; published table has 3-dp rounding
def check_manifest() -> int:
manifest = DATA / "MANIFEST.sha256"
failed = 0
for line in manifest.read_text().strip().splitlines():
digest, name = line.split(" ", 1)
actual = hashlib.sha256((DATA / name).read_bytes()).hexdigest()
ok = actual == digest
print(f" {'ok ' if ok else 'BAD'} {name}")
if not ok:
failed += 1
return failed
def _build_series(species: pd.DataFrame, slug_to_canonical: dict[str, str]) -> dict[str, pd.Series]:
"""For every method slug, return a Series indexed by species_id (post-alias
canonical names) with energies in Hartree."""
out = {}
for slug, _display in slug_to_canonical.items():
out[slug] = species.set_index("species_id")[slug]
return out
def check_mae_table() -> int:
species = pd.read_parquet(DATA / "species.parquet")
methods = pd.read_json(DATA / "methods.json")["energy_columns"]
slug_to_display = {m["slug"]: m["display_name"] for m in methods}
aliases_df = pd.read_parquet(DATA / "aliases.parquet")
alias_map = dict(zip(aliases_df["alias"], aliases_df["canonical"]))
sp_indexed = species.set_index("species_id")
dlpno = sp_indexed["energy_dlpno_ccsd_t"]
ref_rxn, ref_fwd, ref_rev = reaction_quantities(dlpno, dlpno, alias_map)
pub_path = PRODUCTION / "trip50_dft_mae.csv"
if not pub_path.exists():
print(f" SKIP: {pub_path} not present (offline verification only)")
return 0
pub = pd.read_csv(pub_path).set_index("Theory")
failed = 0
for slug, display in slug_to_display.items():
if display == "DLPNO-CCSD(T)" or display not in pub.index:
continue
col = sp_indexed[slug]
t_rxn, t_fwd, t_rev = reaction_quantities(col, dlpno, alias_map)
keys = sorted(set(ref_rxn) & set(t_rxn))
mae_rxn = sum(abs(ref_rxn[k] - t_rxn[k]) for k in keys) / len(keys)
mae_fwd = sum(abs(ref_fwd[k] - t_fwd[k]) for k in keys) / len(keys)
mae_rev = sum(abs(ref_rev[k] - t_rev[k]) for k in keys) / len(keys)
diffs = [
abs(round(mae_rxn, 3) - pub.loc[display, "ΔE_rxn (P-R)"]),
abs(round(mae_fwd, 3) - pub.loc[display, "ΔE_fwd (TS-R)"]),
abs(round(mae_rev, 3) - pub.loc[display, "ΔE_rev (TS-P)"]),
]
ok = max(diffs) < MAE_TOL
if not ok:
failed += 1
print(f" BAD {display}: max abs diff = {max(diffs):.4f} kcal/mol")
print(f" {'ok ' if failed == 0 else 'BAD'} MAE table reproduces "
f"trip50_dft_mae.csv (max tol {MAE_TOL})")
return failed
def check_extxyz_roundtrip() -> int:
frames = read(DATA / "trip50.extxyz", index=":")
species = pd.read_parquet(DATA / "species.parquet").set_index("species_id")
failed = 0
for atoms in frames:
sid = atoms.info.get("species_id")
if sid is None or sid not in species.index:
print(f" BAD frame missing/unknown species_id: {sid!r}")
failed += 1
continue
row = species.loc[sid]
if atoms.numbers.tolist() != list(row["atomic_numbers"]):
print(f" BAD {sid}: atomic_numbers mismatch")
failed += 1
continue
pos = np.asarray(row["positions"].tolist())
max_dev = float(np.abs(atoms.positions - pos).max())
if max_dev > POSITION_TOL:
print(f" BAD {sid}: positions max dev {max_dev:.2e} Å > {POSITION_TOL}")
failed += 1
print(f" {'ok ' if failed == 0 else 'BAD'} extxyz round-trip "
f"({len(frames)} frames, tol {POSITION_TOL} Å)")
return failed
def main() -> int:
print("MANIFEST")
f1 = check_manifest()
print("\nMAE table")
f2 = check_mae_table()
print("\nExtxyz round-trip")
f3 = check_extxyz_roundtrip()
total = f1 + f2 + f3
print(f"\n{'PASS' if total == 0 else 'FAIL'} ({total} failure(s))")
return total
if __name__ == "__main__":
sys.exit(main())
|