Update README.md

[pee8379]

'pert_box' dose not seems to work as percentage.

[AnguseZhang]

Thanks for your contribution! Though, I do think pert_box serves as changing the percentage, instead of an absolute length for box length. (See

dpgen/dpgen/data/tools/create_random_disturb.py

Line 93 in 0767dce

def gen_random_emat(etmax, diag=0):

). How does it behave in your circumstance?

[pee8379]

Right. I misunderstood, only saw emat = emat + np.eye(3) and not
cell = np.dot(cell0, gen_random_emat(etmax, diag))
part.

I originally had a interest this part, because some of my POSCARs generated by init_bulk has highly deformed cell.
does
"pert_box": 1
means 1% of changing in box? or should it be 0.01 to get 1% of change?

Thanks.

[HuangJiameng]

I see what you mean, maybe "ratio" is more appropriate than "percentage"
see 'pert_cmd = 'python3 ' + pert_cmd + ' -etmax %f -ofmt %s %s %d %f > /dev/null' %(pert_box, fp_style, poscar_name, pert_numb, pert_atom)' in https://github.com/deepmodeling/dpgen/blob/0767dce7cad29367edb2e4a55fd0d8724dbda642/dpgen/data/gen.py