nvalchemi.csp.get_space_group_operations#
- nvalchemi.csp.get_space_group_operations(space_group, *, dtype=torch.float32, device=None)[source]#
Return standard fractional operations for one space group.
- Parameters:
space_group (int) – International space-group number in
[1, 230].dtype (torch.dtype, default=torch.float32) – Floating dtype for the returned operation tensor.
device (torch.device or str, optional) – Device for the returned tensor. The bundled table itself remains on CPU.
- Returns:
operations – Fractional rotations in
[..., :3]and translations in[..., 3]. For row-vector fractional coordinates,f_out = f @ R.T + t.- Return type:
torch.Tensor, shape
[S, 3, 4]- Raises:
TypeError – If
space_groupis not a non-boolean integer ordtypeis not a floating Torch dtype.ValueError – If the International number is outside
[1, 230].
Examples
>>> operations = get_space_group_operations(2) # P-1 space group. >>> print(operations) # Identity, then inversion. tensor([[[ 1., 0., 0., 0.], [ 0., 1., 0., 0.], [ 0., 0., 1., 0.]], [[-1., 0., 0., 0.], [ 0., -1., 0., 0.], [ 0., 0., -1., 0.]]])