Fix Wannier90 k-point export path for symmetry-reconstructed SCF MOs (#5233)

Co-authored-by: Thomas D. Kuehne <tkuehne@cp2k.org>
This commit is contained in:
Dynamics of Condensed Matter 2026-05-17 22:20:32 +02:00 committed by GitHub
parent 5311037518
commit 358cae4de3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 273 additions and 5 deletions

View file

@ -2812,11 +2812,10 @@ CONTAINS
CALL keyword_release(keyword)
CALL keyword_create(keyword, __LOCATION__, name="USE_BLOCH_PHASES", &
description="Write use_bloch_phases=.true. to the generated Wannier90 input. "// &
"This lets Wannier90 initialize a complete exported band subspace without an "// &
"external .amn projection file. It is only valid when WANNIER_FUNCTIONS matches "// &
"the number of exported bands; disentangled Wannier90 calculations still need "// &
"explicit projections.", &
description="Apply the CP2K Bloch-phase gauge to complete exported band subspaces "// &
"and write an explicit identity .amn projection file for Wannier90. It is only "// &
"valid when WANNIER_FUNCTIONS matches the number of exported bands; disentangled "// &
"Wannier90 calculations still need explicit projections.", &
usage="USE_BLOCH_PHASES T", default_l_val=.FALSE.)
CALL section_add_keyword(print_key, keyword)
CALL keyword_release(keyword)

View file

@ -484,6 +484,7 @@ CONTAINS
END DO
WRITE (iunit, "(A,/)") "end unit_cell_cart"
WRITE (iunit, "(/,A)") "begin atoms_cart"
WRITE (iunit, "(A)") "bohr"
DO i = 1, num_atoms
WRITE (iunit, "(A,3F15.10)") atom_symbols(i), atoms_cart(1:3, i)
END DO

View file

@ -48,6 +48,7 @@ run_test mypy --strict ./tools/minimax_tools/minimax_to_fortran_source.py
run_test mypy --strict ./tools/dashboard/generate_dashboard.py
run_test mypy --strict ./tools/regtesting/optimize_test_dirs.py
run_test mypy --strict ./tools/regtesting/compare_wannier90_mmn.py
run_test mypy --strict ./tools/regtesting/validate_wannier90_export.py
run_test mypy --strict ./tools/precommit/precommit.py
run_test mypy --strict ./tools/precommit/check_file_properties.py
run_test mypy --strict ./tools/precommit/format_makefile.py
@ -61,6 +62,22 @@ run_test mypy --strict ./docs/generate_input_reference.py
run_test mypy --strict ./docs/fix_github_links.py
run_test mypy --strict ./tools/vibronic_spec/
# Test Wannier90 .wout parsing without requiring an external wannier90.x binary.
W90_TEST_DIR=$(mktemp -d)
cat > "${W90_TEST_DIR}/test.wout" << EOF
Final State
WF centre and spread 1 ( 1.000000, 2.000000, 3.000000 ) 1.25000000
Sum of centres and spreads ( 1.000000, 2.000000, 3.000000 ) 1.25000000
Spreads (Bohr^2) Omega I = 1.250000000
================ Omega D = 0.000000000
Omega OD = 0.000000000
Final Spread (Bohr^2) Omega Total = 1.250000000
All done: wannier90 exiting
EOF
run_test ./tools/regtesting/validate_wannier90_export.py run test \
--workdir "${W90_TEST_DIR}" --require-all-done --max-omega-total 2.0
# TODO: Find a way to test generate_dashboard.py without git repository.
#
# # Test generate_dashboard.py. Running it twice to also execute its caching.

View file

@ -0,0 +1,251 @@
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""Run and validate CP2K-generated Wannier90 exports.
CP2K regtests can validate the generated ``.mmn`` and ``.eig`` files internally,
but the external Wannier90 executable has a few additional failure modes. This
helper is intentionally small: it can run ``wannier90.x`` for one seed, parse the
resulting ``.wout`` file, and compare final spreads/centres between two runs.
"""
from __future__ import annotations
import argparse
import re
import subprocess
import sys
from dataclasses import dataclass
from pathlib import Path
from typing import Sequence
WF_RE = re.compile(
r"WF centre and spread\s+(\d+)\s+\(\s*"
r"([+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[Ee][+-]?\d+)?),\s*"
r"([+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[Ee][+-]?\d+)?),\s*"
r"([+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[Ee][+-]?\d+)?)\s*\)\s*"
r"([+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[Ee][+-]?\d+)?)"
)
OMEGA_RE = re.compile(
r"Omega\s+(I|D|OD|Total)\s*=\s*" r"([+-]?(?:\d+(?:\.\d*)?|\.\d+)(?:[Ee][+-]?\d+)?)"
)
ERROR_RE = re.compile(r"(^|\s)(ERROR:|Error in|Exiting\.\.\.\.\.\.\.)")
@dataclass(frozen=True)
class WannierFunction:
index: int
center: tuple[float, float, float]
spread: float
@dataclass(frozen=True)
class WoutSummary:
seed: str
path: Path
functions: tuple[WannierFunction, ...]
omega_i: float
omega_d: float
omega_od: float
omega_total: float
all_done: bool
def parse_wout(seed: str, workdir: Path) -> WoutSummary:
path = workdir / f"{seed}.wout"
text = path.read_text(encoding="utf-8", errors="replace")
if ERROR_RE.search(text):
raise ValueError(f"{path}: Wannier90 reported an error")
final_pos = text.rfind("Final State")
if final_pos < 0:
raise ValueError(f"{path}: missing final Wannier90 state")
final_text = text[final_pos:]
functions: list[WannierFunction] = []
for match in WF_RE.finditer(final_text):
functions.append(
WannierFunction(
index=int(match.group(1)),
center=(
float(match.group(2)),
float(match.group(3)),
float(match.group(4)),
),
spread=float(match.group(5)),
)
)
if not functions:
raise ValueError(f"{path}: missing final Wannier-function centres")
omegas: dict[str, float] = {}
for match in OMEGA_RE.finditer(final_text):
omegas[match.group(1)] = float(match.group(2))
required = {"I", "D", "OD", "Total"}
missing = required - omegas.keys()
if missing:
raise ValueError(
f"{path}: missing omega field(s): {', '.join(sorted(missing))}"
)
return WoutSummary(
seed=seed,
path=path,
functions=tuple(functions),
omega_i=omegas["I"],
omega_d=omegas["D"],
omega_od=omegas["OD"],
omega_total=omegas["Total"],
all_done="All done: wannier90 exiting" in text,
)
def run_wannier90(executable: str, seed: str, workdir: Path, timeout: float) -> None:
result = subprocess.run(
[executable, seed],
cwd=workdir,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
text=True,
encoding="utf-8",
errors="replace",
timeout=timeout,
check=False,
)
(workdir / f"{seed}.wannier90.stdout").write_text(result.stdout, encoding="utf-8")
if result.returncode != 0:
raise RuntimeError(
f"{executable} {seed} failed with exit code {result.returncode}"
)
def print_summary(summary: WoutSummary) -> None:
done = "yes" if summary.all_done else "no"
print(f"seed = {summary.seed}")
print(f"wout = {summary.path}")
print(f"all_done = {done}")
print(f"num_wann = {len(summary.functions)}")
print(f"omega_i = {summary.omega_i:.12e}")
print(f"omega_d = {summary.omega_d:.12e}")
print(f"omega_od = {summary.omega_od:.12e}")
print(f"omega_total = {summary.omega_total:.12e}")
for function in summary.functions:
cx, cy, cz = function.center
print(
f"wf {function.index:5d}: center = "
f"({cx:.12e}, {cy:.12e}, {cz:.12e}), "
f"spread = {function.spread:.12e}"
)
def max_abs_delta(left: Sequence[float], right: Sequence[float]) -> float:
if len(left) != len(right):
raise ValueError(f"length mismatch: {len(left)} != {len(right)}")
return max((abs(x - y) for x, y in zip(left, right)), default=0.0)
def compare_summaries(
left: WoutSummary, right: WoutSummary
) -> tuple[float, float, float]:
if len(left.functions) != len(right.functions):
raise ValueError(
f"number of Wannier functions differs: "
f"{len(left.functions)} != {len(right.functions)}"
)
left_spreads = [function.spread for function in left.functions]
right_spreads = [function.spread for function in right.functions]
spread_delta = max_abs_delta(left_spreads, right_spreads)
left_centers = [value for function in left.functions for value in function.center]
right_centers = [value for function in right.functions for value in function.center]
center_delta = max_abs_delta(left_centers, right_centers)
omega_delta = abs(left.omega_total - right.omega_total)
return omega_delta, spread_delta, center_delta
def command_run(args: argparse.Namespace) -> int:
workdir = Path(args.workdir)
if args.run:
run_wannier90(args.executable, args.seed, workdir, args.timeout)
summary = parse_wout(args.seed, workdir)
print_summary(summary)
if args.require_all_done and not summary.all_done:
print("Wannier90 did not report normal completion.", file=sys.stderr)
return 1
if args.max_omega_total is not None and summary.omega_total > args.max_omega_total:
print(
f"Omega total {summary.omega_total:.12e} exceeds "
f"{args.max_omega_total:.12e}.",
file=sys.stderr,
)
return 1
return 0
def command_compare(args: argparse.Namespace) -> int:
left = parse_wout(args.left_seed, Path(args.left_workdir))
right = parse_wout(args.right_seed, Path(args.right_workdir))
omega_delta, spread_delta, center_delta = compare_summaries(left, right)
print(f"omega_total_delta = {omega_delta:.12e}")
print(f"spread_delta = {spread_delta:.12e}")
print(f"center_delta = {center_delta:.12e}")
failed = False
if args.omega_total_tol is not None and omega_delta > args.omega_total_tol:
failed = True
if args.spread_tol is not None and spread_delta > args.spread_tol:
failed = True
if args.center_tol is not None and center_delta > args.center_tol:
failed = True
return 1 if failed else 0
def create_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(description=__doc__)
subparsers = parser.add_subparsers(dest="command", required=True)
run_parser = subparsers.add_parser(
"run", help="run and/or parse one Wannier90 seed"
)
run_parser.add_argument("seed", help="Wannier90 seed name")
run_parser.add_argument(
"--workdir", default=".", help="directory containing the seed files"
)
run_parser.add_argument(
"--run", action="store_true", help="run wannier90.x before parsing"
)
run_parser.add_argument(
"--executable", default="wannier90.x", help="Wannier90 executable"
)
run_parser.add_argument(
"--timeout", type=float, default=300.0, help="execution timeout in seconds"
)
run_parser.add_argument("--max-omega-total", type=float, default=None)
run_parser.add_argument("--require-all-done", action="store_true")
run_parser.set_defaults(func=command_run)
compare_parser = subparsers.add_parser(
"compare", help="compare two parsed .wout files"
)
compare_parser.add_argument("left_seed")
compare_parser.add_argument("right_seed")
compare_parser.add_argument("--left-workdir", default=".")
compare_parser.add_argument("--right-workdir", default=".")
compare_parser.add_argument("--omega-total-tol", type=float, default=None)
compare_parser.add_argument("--spread-tol", type=float, default=None)
compare_parser.add_argument("--center-tol", type=float, default=None)
compare_parser.set_defaults(func=command_compare)
return parser
def main(argv: Sequence[str] | None = None) -> int:
parser = create_parser()
args = parser.parse_args(argv)
try:
return int(args.func(args))
except (OSError, RuntimeError, subprocess.TimeoutExpired, ValueError) as exc:
print(f"ERROR: {exc}", file=sys.stderr)
return 1
if __name__ == "__main__":
sys.exit(main())