I-PI xyz file to an ase readable xyz file

[Shubhang Goswami]

Hello,

I have outputs generated from I-Pi saved as xyz files. Although these files are readable from the atomic simulation environment, but it doesnt read any of the cell dimension information.

The xyz file from I-py written as,

192
# CELL(abcABC): 7.63946 6.61313 9.30182 89.61648 90.34115 90.22212 Step: 0 Bead: 11 positions{angstrom} cell{angstrom}
H -8.98498e-03 3.73538e+00 -8.74215e-02
H -1.67969e-01 3.16369e+00 9.38250e+00
H 9.32450e-01 4.21329e+00 1.66229e+00 

and so on

if you read through ase.io.read() , the atom files obtained do not contain any of the #CELL dimension information.

Is there a nice way to convert the I-py output xyz format to an ase one (that can take in cell dimension information)?

[Mariana Rossi]

So, if you have ASE installed in your system, it is possible to set the i-PI output to extxyz format directly.

Check here:

https://ipi-code.org/i-pi/input-reference.html#trajectory

choose 'ase' as the mode.

Otherwise, if you want to convert something you already have, I think the best is to write a quick script. Maybe someone knows of some other way, but I just did it in the past by making my script write the comment line in the ASE format, and just echoing all other lines.

[Michele Ceriotti]

I should just add that I used many times `bin/i-pi-contract-trajectory` to do this, but it's not very well documented. might be useful and interesting to make a small script to convert trajectories, but I (and I think we) don't have the bandwidth right now, so if you feel like doing it, it'd be a very well appreciated contribution.

Michele

[Mariana Rossi]

Just in case this is still useful, here is a small code using ase:

import numpy as np
from ase.io import read, write
from sys import argv

# change here for your actual lattice vectors
a=11.554971
b=10.0069
c=65.000001

trajname=argv[1]

trajectory=read(trajname, index=':', format='xyz')

for i in trajectory:
    i.set_cell([a,b,c])
    i.set_pbc([True,True,True])
    i.wrap()

output=write('wrapped.xyz', trajectory, format='extxyz')

[Janos Daru]

Dear All,

I have a related question. 
Is there a way to dump bead-wise a single file containing coordinates, forces, cell parameters, and energies in ASE like manner?

Eg:
10
Lattice="10.0 0.0 0.0 0.0 10.0 0.0 0.0 0.0 10.0" Properties=species:S:1:pos:R:3:forces:R:3  energy=-314.2384855360319 pbc="TTT" 
C        0.14088468       0.71747048       0.58595283      -0.06500885       0.06629617      -0.11803488

...

many thanks in advance, 

Best,

Janos

[Michele Ceriotti]

Showt anwser is no, but you can do this easily as a post-processing step.

Assuming you are using the same output stride, something like

```

from ipi import read_output, read_trajectory

data, description = read_output("simulation.out")

traj = read_trajectory("simulation.pos_0.xyz")

force = read_trajectory("simulation.for_0.xyz")

for t, f, e in zip(traj, force, data["potential"]):

    t.arrays["force"] = f.positions

    t.info["energy"] = e

ase.write("combo.xyz", traj)

```

Should do to collate information from the standard output files

M