def tune_geometry(graph, mol):
"""Fine tune a molecular geometry, starting from a (very) poor guess of
the initial geometry.
Do not expect all details to be in perfect condition. A subsequent
optimization with a more accurate level of theory is at least advisable.
Arguments:
graph -- The molecular graph of the system
mol -- The initial guess of the coordinates
"""
N = len(graph.numbers)
from molmod.minimizer import Minimizer, NewtonGLineSearch
ff = ToyFF(graph)
x_init = mol.coordinates.ravel()
# level 3 geometry optimization: bond lengths + pauli
ff.dm_reci = 0.2
ff.bond_quad = 1.0
minimizer = Minimizer(x_init, ff, NewtonGLineSearch, 1e-3, 1e-3, 2*N, 500, 50, do_gradient=True, verbose=False)
x_init = minimizer.x
# level 4 geometry optimization: bond lengths + bending angles + pauli
ff.bond_quad = 0.0
ff.bond_hyper = 1.0
ff.span_quad = 1.0
minimizer = Minimizer(x_init, ff, NewtonGLineSearch, 1e-6, 1e-6, 2*N, 500, 50, do_gradient=True, verbose=False)
x_init = minimizer.x
x_opt = x_init
mol = Molecule(graph.numbers, x_opt.reshape((N,3)))
return mol
class CGOptimizer(BaseOptimizer):
log_name = 'CGOPT'
def __init__(self, dof, state=None, hooks=None, counter0=0):
self.minimizer = Minimizer(
dof.x0,
self.fun,
ConjugateGradient(),
NewtonLineSearch(),
None,
None,
anagrad=True,
verbose=False,
)
BaseOptimizer.__init__(self, dof, state, hooks, counter0)
def initialize(self):
self.minimizer.initialize()
BaseOptimizer.initialize(self)
def propagate(self):
success = self.minimizer.propagate()
self.x = self.minimizer.x
if success == False:
if log.do_warning:
log.warn(
'Line search failed in optimizer. Aborting optimization. This is probably due to a dicontinuity in the energy or the forces. Check the truncation of the non-bonding interactions and the Ewald summation parameters.'
)
return True
return BaseOptimizer.propagate(self)
def __init__(self, dof, state=None, hooks=None, counter0=0):
self.minimizer = Minimizer(
dof.x0,
self.fun,
ConjugateGradient(),
NewtonLineSearch(),
None,
None,
anagrad=True,
verbose=False,
)
BaseOptimizer.__init__(self, dof, state, hooks, counter0)
def guess_geometry(graph, unitcell_active, unitcell, unitcell_reciproke):
"""Construct a molecular geometry based on a molecular graph.
This routine does not require initial coordinates and will give a very
rough picture of the initial geometry. Do not expect all details to be
in perfect condition. A subsequent optimization with a more accurate
level of theory is at least advisable.
Arguments:
graph -- The molecular graph of the system
"""
N = len(graph.numbers)
from molmod.minimizer import Minimizer, NewtonGLineSearch
ff = ToyFF(graph, unitcell_active, unitcell, unitcell_reciproke)
x_init = numpy.random.normal(0,1,N*3)
# level 1 geometry optimization: graph based
ff.dm_quad = 1.0
minimizer = Minimizer(x_init, ff, NewtonGLineSearch, 1e-10, 1e-8, 2*N, 500, 50, do_gradient=True, verbose=False)
x_init = minimizer.x
# level 2 geometry optimization: graph based + pauli repulsion
ff.dm_quad = 1.0
ff.dm_reci = 1.0
minimizer = Minimizer(x_init, ff, NewtonGLineSearch, 1e-10, 1e-8, 2*N, 500, 50, do_gradient=True, verbose=False)
x_init = minimizer.x
# Add a little noise to avoid saddle points
x_init += numpy.random.uniform(-0.01, 0.01, len(x_init))
# level 3 geometry optimization: bond lengths + pauli
ff.dm_quad = 0.0
ff.dm_reci = 0.2
ff.bond_quad = 1.0
minimizer = Minimizer(x_init, ff, NewtonGLineSearch, 1e-3, 1e-3, 2*N, 500, 50, do_gradient=True, verbose=False)
x_init = minimizer.x
# level 4 geometry optimization: bond lengths + bending angles + pauli
ff.bond_quad = 0.0
ff.bond_hyper = 1.0
ff.span_quad = 1.0
minimizer = Minimizer(x_init, ff, NewtonGLineSearch, 1e-6, 1e-6, 2*N, 500, 50, do_gradient=True, verbose=False)
x_init = minimizer.x
x_opt = x_init
mol = Molecule(graph.numbers, x_opt.reshape((N,3)))
return mol
class CGOptimizer(BaseOptimizer):
log_name = 'CGOPT'
def __init__(self, dof, state=None, hooks=None, counter0=0):
self.minimizer = Minimizer(
dof.x0, self.fun, ConjugateGradient(), NewtonLineSearch(), None,
None, anagrad=True, verbose=False,
)
BaseOptimizer.__init__(self, dof, state, hooks, counter0)
def initialize(self):
self.minimizer.initialize()
BaseOptimizer.initialize(self)
def propagate(self):
success = self.minimizer.propagate()
self.x = self.minimizer.x
if success == False:
if log.do_warning:
log.warn('Line search failed in optimizer. Aborting optimization. This is probably due to a dicontinuity in the energy or the forces. Check the truncation of the non-bonding interactions and the Ewald summation parameters.')
return True
return BaseOptimizer.propagate(self)
def __init__(self, dof, state=None, hooks=None, counter0=0):
self.minimizer = Minimizer(
dof.x0, self.fun, ConjugateGradient(), NewtonLineSearch(), None, None, anagrad=True, verbose=False
)
BaseOptimizer.__init__(self, dof, state, hooks, counter0)
def guess_geometry(graph, unit_cell=None, verbose=False):
"""Construct a molecular geometry based on a molecular graph.
This routine does not require initial coordinates and will give a very
rough picture of the initial geometry. Do not expect all details to be
in perfect condition. A subsequent optimization with a more accurate
level of theory is at least advisable.
Argument:
| ``graph`` -- The molecular graph of the system, see
:class:molmod.molecular_graphs.MolecularGraph
Optional argument:
| ``unit_cell`` -- periodic boundry conditions, see
:class:`molmod.unit_cells.UnitCell`
| ``verbose`` -- Show optimizer progress when True
"""
N = len(graph.numbers)
from molmod.minimizer import Minimizer, ConjugateGradient, \
NewtonLineSearch, ConvergenceCondition, StopLossCondition
search_direction = ConjugateGradient()
line_search = NewtonLineSearch()
convergence = ConvergenceCondition(grad_rms=1e-6, step_rms=1e-6)
stop_loss = StopLossCondition(max_iter=500, fun_margin=0.1)
ff = ToyFF(graph, unit_cell)
x_init = np.random.normal(0, 1, N * 3)
# level 1 geometry optimization: graph based
ff.dm_quad = 1.0
minimizer = Minimizer(x_init,
ff,
search_direction,
line_search,
convergence,
stop_loss,
anagrad=True,
verbose=verbose)
x_init = minimizer.x
# level 2 geometry optimization: graph based + pauli repulsion
ff.dm_quad = 1.0
ff.dm_reci = 1.0
minimizer = Minimizer(x_init,
ff,
search_direction,
line_search,
convergence,
stop_loss,
anagrad=True,
verbose=verbose)
x_init = minimizer.x
# Add a little noise to avoid saddle points
x_init += np.random.uniform(-0.01, 0.01, len(x_init))
# level 3 geometry optimization: bond lengths + pauli
ff.dm_quad = 0.0
ff.dm_reci = 0.2
ff.bond_quad = 1.0
minimizer = Minimizer(x_init,
ff,
search_direction,
line_search,
convergence,
stop_loss,
anagrad=True,
verbose=verbose)
x_init = minimizer.x
# level 4 geometry optimization: bond lengths + bending angles + pauli
ff.bond_quad = 0.0
ff.bond_hyper = 1.0
ff.span_quad = 1.0
minimizer = Minimizer(x_init,
ff,
search_direction,
line_search,
convergence,
stop_loss,
anagrad=True,
verbose=verbose)
x_init = minimizer.x
x_opt = x_init
mol = Molecule(graph.numbers, x_opt.reshape((N, 3)))
return mol
def tune_geometry(graph, mol, unit_cell=None, verbose=False):
"""Fine tune a molecular geometry, starting from a (very) poor guess of
the initial geometry.
Do not expect all details to be in perfect condition. A subsequent
optimization with a more accurate level of theory is at least advisable.
Arguments:
| ``graph`` -- The molecular graph of the system, see
:class:molmod.molecular_graphs.MolecularGraph
| ``mol`` -- A :class:molmod.molecules.Molecule class with the initial
guess of the coordinates
Optional argument:
| ``unit_cell`` -- periodic boundry conditions, see
:class:`molmod.unit_cells.UnitCell`
| ``verbose`` -- Show optimizer progress when True
"""
N = len(graph.numbers)
from molmod.minimizer import Minimizer, ConjugateGradient, \
NewtonLineSearch, ConvergenceCondition, StopLossCondition
search_direction = ConjugateGradient()
line_search = NewtonLineSearch()
convergence = ConvergenceCondition(grad_rms=1e-6, step_rms=1e-6)
stop_loss = StopLossCondition(max_iter=500, fun_margin=1.0)
ff = ToyFF(graph, unit_cell)
x_init = mol.coordinates.ravel()
# level 3 geometry optimization: bond lengths + pauli
ff.dm_reci = 0.2
ff.bond_quad = 1.0
minimizer = Minimizer(x_init,
ff,
search_direction,
line_search,
convergence,
stop_loss,
anagrad=True,
verbose=verbose)
x_init = minimizer.x
# level 4 geometry optimization: bond lengths + bending angles + pauli
ff.bond_quad = 0.0
ff.bond_hyper = 1.0
ff.span_quad = 1.0
minimizer = Minimizer(x_init,
ff,
search_direction,
line_search,
convergence,
stop_loss,
anagrad=True,
verbose=verbose)
x_init = minimizer.x
x_opt = x_init
mol = Molecule(graph.numbers, x_opt.reshape((N, 3)))
return mol