Too strict requirements when parsing valid NWChem input

[francescalb]

Python version

python3.14

Pymatgen version

2025.10.7

Operating system version

Ubutnu 24.04 LTS (on WSL)

Current behavior

There are valid geometry inputs for NWChem that are not accepted. I am not sure whether to report this as bug or feature request.

Example of valig geometry directive that works:

geometry start units angstrom
symmetry Cs
 B   ...
 P   ...
 ...
end

Example that fails:

geometry start units angstrom
 B   ...
 P   ...
 ...
symmetry Cs
end

Expected Behavior

Expected behaviour is that the two inputs above read the same and have the same output when using
NwChem.from_file().

Minimal example

# requires pip install deepdiff

from deepdiff import diff

i = NwInput.from_file('test.nw')
i2 = NwInput.from_file('test2.nw') # currently fails here

diff = DeepDiff(dict1, dict2)
assert not diff

Relevant files to reproduce this bug

start test
title test

geometry start  units angstrom
symmetry Cs
 B                    -1.39033432     0.00006055     0.00000000
 P                     0.55743699     0.00026317     0.00000000
 H                     1.21288219    -1.24261287     0.00000000
 H                     1.21371191     0.62124034     1.07614375
 H                     1.21371191     0.62124034    -1.07614375
 H                    -1.67029855    -0.58705888    -1.01673217
 H                    -1.67029855    -0.58705888     1.01673217
 H                    -1.67111928     1.17392623     0.00000000
end

set geometry start

basis
 * library cc-pvqz
end

charge 0

start test
title test

geometry start  units angstrom
 B                    -1.39033432     0.00006055     0.00000000
 P                     0.55743699     0.00026317     0.00000000
 H                     1.21288219    -1.24261287     0.00000000
 H                     1.21371191     0.62124034     1.07614375
 H                     1.21371191     0.62124034    -1.07614375
 H                    -1.67029855    -0.58705888    -1.01673217
 H                    -1.67029855    -0.58705888     1.01673217
 H                    -1.67111928     1.17392623     0.00000000
symmetry Cs
end

set geometry start

basis
 * library cc-pvqz
end

charge 0

[DanielYang59]

should have been fixed by #4567 ?