[ase-users] [ase-developers] Adjusting atom.momentum?

[Ask Hjorth Larsen]

Hmm, I thought it was the back-reference which had changed.

Anyway this is actually fixable: If the user accesses a quantity and
modifies it, the Atom can edit the properties of the Atoms object
including creating the array if it is not there.  (I rarely use Atom
objects on their own though)

In fact the Atom does not need to store any data unless it is popped
from the Atoms, in which case it needs a copy.

But I guess there are more important things to do....

Best regards
Ask

2016-10-24 13:03 GMT+02:00 Jens Jørgen Mortensen <jensj at fysik.dtu.dk>:
> On 10/24/2016 12:44 PM, Ask Hjorth Larsen wrote:
>>
>>
>> Hello
>>
>> It can access the array and modify it, knowing its own index in the atoms
>> object.
>>
>
> Problem is that the Atoms object doesn't have a momenta array.  So,
> "atoms[2].momentum" will create a fresh array with "np.zeros(3)" :-(
>
>     https://gitlab.com/ase/ase/issues/3
>
> Jens Jørgen
>
>> Or is there something more to it?
>>
>> Actually I guess it is still tricky to handle the case where someone sets
>> the momenta after creating the Atom
>>
>> Best regards
>> Ask
>>
>>
>> El 24 oct. 2016 7:43, "Jens Jørgen Mortensen" <jensj at fysik.dtu.dk
>> <mailto:jensj at fysik.dtu.dk>> escribió:
>>
>>     On 10/23/2016 01:27 AM, Ask Hjorth Larsen wrote:
>>
>>         Hello
>>
>>         An Atom now has a reference to the Atoms from which it was
>>         extracted.
>>
>>         Therefore this is trivial to fix now.  It only requires the
>>         will to do so.
>>
>>
>>     What has happened that makes it trivial to fix now and how would
>>     you do it?
>>
>>     Jens Jørgen
>>
>>
>>         Best regards
>>         Ask
>>
>>         2015-10-19 9:37 GMT+02:00 Jens Jørgen Mortensen
>>         <jensj at fysik.dtu.dk <mailto:jensj at fysik.dtu.dk>>:
>>
>>
>>             On 10/16/2015 09:46 PM, Peterson, Andrew wrote:
>>
>>             In writing the NVE test for the Hookean constraint, I
>>             noticed that modifying
>>             the momentum attribute of an atom was not permanent. E.g.,
>>
>>             In [1]: from ase.structure import molecule
>>             In [2]: atoms = molecule('CH4')
>>             In [3]: atoms[2].momentum[1] += 8.
>>             In [4]: atoms.get_momenta()
>>             Out[4]:
>>             array([[ 0.,  0.,  0.],
>>                     [ 0.,  0.,  0.],
>>                     [ 0.,  0.,  0.],
>>                     [ 0.,  0.,  0.],
>>                     [ 0.,  0.,  0.]])
>>
>>             I know I can use the set_momentum function, but had
>>             expected this to behave
>>             the same as adjusting a position --- the analogous script
>>             with positions
>>             does result in changing the position array:
>>
>>             In [1]: from ase.structure import molecule
>>             In [2]: atoms = molecule('CH4')
>>             In [3]: atoms[2].position[1] += 8.
>>             In [4]: atoms.get_positions()
>>             Out[4]:
>>             array([[ 0.      ,  0.      ,  0.      ],
>>                     [ 0.629118,  0.629118,  0.629118],
>>                     [-0.629118,  7.370882,  0.629118],
>>                     [ 0.629118, -0.629118, -0.629118],
>>                     [-0.629118,  0.629118, -0.629118]])
>>
>>             Is this how it is supposed to work, or should I file an issue?
>>
>>
>>             I guess you should file an issue.  I really don't know how
>>             to fix this.
>>             Problem is that the momenta are zero even if you did not
>>             explicitly set them
>>             and we don't want to allocate an ndarray for them before
>>             they become
>>             non-zero.  The positions are always backed by an ndarray
>>             in the Atoms object
>>             ...
>>
>>             The best I can come of with is to make the ndarray you get
>>             from
>>             atoms[2].momentum not writeable and document this weird
>>             behaviour.
>>
>>             If the Atoms object already has momenta then it will work
>>             fine:
>>
>>                  atoms[2].momentum = (0, 0, 0)
>>                  atoms[2].momentum[1] += 8
>>
>>             This will actually also work:
>>
>>                  atoms[2].momentum += (0, 8, 0)
>>
>>             Jens Jørgen
>>
>>             thanks,
>>             Andy
>>
>>
>>             --
>>             Andrew Peterson
>>             Assistant Professor
>>             Brown University School of Engineering
>>             Barus & Holley 247
>>             184 Hope Street
>>             Providence, RI 02912
>>             (401) 863-2153 <tel:%28401%29%20863-2153>
>>             http://brown.edu/go/catalyst
>>
>>
>>             _______________________________________________
>>             ase-developers mailing list
>>             ase-developers at listserv.fysik.dtu.dk
>>             <mailto:ase-developers at listserv.fysik.dtu.dk>
>>             https://listserv.fysik.dtu.dk/mailman/listinfo/ase-developers
>>
>> <https://listserv.fysik.dtu.dk/mailman/listinfo/ase-developers>
>>
>>
>>
>>             _______________________________________________
>>             ase-developers mailing list
>>             ase-developers at listserv.fysik.dtu.dk
>>             <mailto:ase-developers at listserv.fysik.dtu.dk>
>>             https://listserv.fysik.dtu.dk/mailman/listinfo/ase-developers
>>
>> <https://listserv.fysik.dtu.dk/mailman/listinfo/ase-developers>
>>
>>
>