https://wiki.fysik.dtu.dk/ase/ase/filters.html#the-frechetcellfilter-class

The initial implementation of ExpCellFilter gave inconsistent gradients for cell variables (matrix log of the deformation tensor). If you would like to keep the previous behavior, please use ExpCellFilter.

In reply to @shivansh_tiwari:matrix.org
Hi everyone, I need some help in ase. I have created a 2d lattice and trying to apply Lennard jones potential. but it is giving me nan values. Is lennard algo is only for 3d lattices? Please anyone clarify

from ase.constraints import FixedLine
c = FixedLine(
indices=[atom.index for atom in atoms if atom.symbol == 'Cu'],
direction=[1, 0, 0],
)
atoms.set_constraint(c)

When I run this code in my latop

c = FixedLine(
^^^^^^^^^^
TypeError: FixedLine.init() got an unexpected keyword argument 'indices'

works for me

>>> from ase.build import bulk
>>> atoms = bulk('Cu') * (2,2,2)
>>> from ase.constraints import FixedLine
>>> c = FixedLine(
... indices=[atom.index for atom in atoms if atom.symbol == 'Cu'],
... direction=[1, 0, 0],
... )
>>> c
FixedLine(indices=[0 1 2 3 4 5 6 7], [1.0, 0.0, 0.0])
>>> import ase
>>> ase.__version__
'3.23.0b1'

In reply to @mikaelkuisma1:matrix.org

works for me

>>> from ase.build import bulk
>>> atoms = bulk('Cu') * (2,2,2)
>>> from ase.constraints import FixedLine
>>> c = FixedLine(
... indices=[atom.index for atom in atoms if atom.symbol == 'Cu'],
... direction=[1, 0, 0],
... )
>>> c
FixedLine(indices=[0 1 2 3 4 5 6 7], [1.0, 0.0, 0.0])
>>> import ase
>>> ase.__version__
'3.23.0b1'

print( ase.version)
3.22.1

This is my source code:

class FixedLine(FixConstraintSingle):
"""Constrain an atom index a to move on a given line only.

The line is defined by its vector *direction*."""

def __init__(self, a, direction):
    self.a = a
    self.dir = np.asarray(direction) / sqrt(np.dot(direction, direction))

def get_removed_dof(self, atoms):
    return 2

def adjust_positions(self, atoms, newpositions):
    step = newpositions[self.a] - atoms.positions[self.a]
    x = np.dot(step, self.dir)
    newpositions[self.a] = atoms.positions[self.a] + x * self.dir

def adjust_forces(self, atoms, forces):
    forces[self.a] = self.dir * np.dot(forces[self.a], self.dir)

def __repr__(self):
    return 'FixedLine(%d, %s)' % (self.a, self.dir.tolist())

def todict(self):
    return {'name': 'FixedLine',
            'kwargs': {'a': self.a, 'direction': self.dir.tolist()}}

So i change me code to this

c = FixedPlane([atom.index for atom in atoms if atom.symbol == 'Ba'], direction=[0, 0, 1],)
It does not have the indices

Ask Hjorth Larsen: Andrew Rosen

Thanks for taking the time to review the merge request. There is one comment that I did not understand, would you mind elaborating a bit more?

Ask Hjorth Larsen: Andrew Rosen

Thanks for taking the time to review the merge request. There is one comment that I did not understand, would you mind elaborating a bit more? (See image above)

    opt = FIRE2(a,
                dt=0.1,
                maxstep=0.2,
                dtmax=1.0,
                Nmin=20,
                finc=1.1,
                fdec=0.5,
                astart=0.25,
                fa=0.99)
    opt.run(fmax=0.001)
    e1 = a.get_potential_energy()
    n1 = opt.nsteps

    a = system_setup()

    reset_history = []

    def callback(a, r, e, e_last):
        reset_history.append([e - e_last])

    opt = FIRE2(a,
                dt=0.1,
                maxstep=0.2,
                dtmax=1.0,
                Nmin=20,
                finc=1.1,
                fdec=0.5,
                astart=0.25,
                fa=0.99,
                abc=True,
                position_reset_callback=callback)

Here (for example) there are two calls to FIRE2. Many of the parameters are the same. What about:

def new_fire2():

    opt = FIRE2(a,
                dt=0.1,
                maxstep=0.2,
                dtmax=1.0,
                Nmin=20,
                finc=1.1,
                fdec=0.5,
                astart=0.25,
                fa=0.99)
    opt.run(fmax=0.001)
    e1 = a.get_potential_energy()
    n1 = opt.nsteps

    a = system_setup()

    reset_history = []

    def callback(a, r, e, e_last):
        reset_history.append([e - e_last])

    opt = FIRE2(a,
                dt=0.1,
                maxstep=0.2,
                dtmax=1.0,
                Nmin=20,
                finc=1.1,
                fdec=0.5,
                astart=0.25,
                fa=0.99,
                abc=True,
                position_reset_callback=callback)

Here (for example) there are two calls to FIRE2. Many of the parameters are the same. What about:

def new_fire2(arg1, arg2):
    return FIRE2(<all the things that are the same go here>)

    opt = FIRE2(a,
                dt=0.1,
                maxstep=0.2,
                dtmax=1.0,
                Nmin=20,
                finc=1.1,
                fdec=0.5,
                astart=0.25,
                fa=0.99)
    opt.run(fmax=0.001)
    e1 = a.get_potential_energy()
    n1 = opt.nsteps

    a = system_setup()

    reset_history = []

    def callback(a, r, e, e_last):
        reset_history.append([e - e_last])

    opt = FIRE2(a,
                dt=0.1,
                maxstep=0.2,
                dtmax=1.0,
                Nmin=20,
                finc=1.1,
                fdec=0.5,
                astart=0.25,
                fa=0.99,
                abc=True,
                position_reset_callback=callback)

Here (for example) there are two calls to FIRE2. Many of the parameters are the same. What about:

def new_fire2(arg1, arg2):
    return FIRE2(arg1=arg1, arg2=arg2, <all the things that are the same go here>)