1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
|
# fmt: off
import argparse
import traceback
from math import pi
from time import time
from typing import Union
import numpy as np
import ase.db
import ase.optimize
from ase import Atoms
from ase.calculators.emt import EMT
from ase.io import Trajectory
all_optimizers = ase.optimize.__all__ + ['PreconLBFGS', 'PreconFIRE',
'SciPyFminCG', 'SciPyFminBFGS']
all_optimizers.remove('QuasiNewton')
all_optimizers.remove('RestartError')
def get_optimizer(name):
# types: (str) -> ase.optimize.Optimizer
if name.startswith('Precon'):
import ase.optimize.precon as precon
return getattr(precon, name)
if name.startswith('SciPy'):
import ase.optimize.sciopt as sciopt
return getattr(sciopt, name)
return getattr(ase.optimize, name)
class Wrapper:
"""Atoms-object wrapper that can count number of moves."""
def __init__(
self,
atoms: Atoms,
gridspacing: float = 0.2,
eggbox: float = 0.0,
) -> None:
self.t0 = time()
self.texcl = 0.0
self.nsteps = 0
self.atoms = atoms
self.ready = False
self.pos: Union[np.ndarray, None] = None
self.eggbox = eggbox
self.x = None
if eggbox:
# Find small unit cell for grid-points
h = []
for axis in atoms.get_cell(complete=True):
L = np.linalg.norm(axis)
n = int(L / gridspacing)
h.append(axis / n)
self.x = np.linalg.inv(h)
def get_potential_energy(self, force_consistent=False):
t1 = time()
e = self.atoms.get_potential_energy(force_consistent)
if self.eggbox:
# Add egg-box error:
s = np.dot(self.atoms.positions, self.x)
e += np.cos(2 * pi * s).sum() * self.eggbox / 6
t2 = time()
self.texcl += t2 - t1
if not self.ready:
self.nsteps += 1
self.ready = True
return e
def get_forces(self):
t1 = time()
f = self.atoms.get_forces()
if self.eggbox:
# Add egg-box error:
s = np.dot(self.atoms.positions, self.x)
f += np.dot(np.sin(2 * pi * s),
self.x.T) * (2 * pi * self.eggbox / 6)
t2 = time()
self.texcl += t2 - t1
if not self.ready:
self.nsteps += 1
self.ready = True
return f
def set_positions(self, pos):
if self.pos is not None and abs(pos - self.pos).max() > 1e-15:
self.ready = False
if self.nsteps == 200:
raise RuntimeError('Did not converge!')
self.pos = pos
self.atoms.set_positions(pos)
def get_positions(self):
return self.atoms.get_positions()
def get_calculator(self):
return self.atoms.calc
def __len__(self):
return len(self.atoms)
def __ase_optimizable__(self):
from ase.optimize.optimize import OptimizableAtoms
return OptimizableAtoms(self)
def run_test(atoms, optimizer, tag, fmax=0.02, eggbox=0.0):
"""Optimize atoms with optimizer."""
wrapper = Wrapper(atoms, eggbox=eggbox)
relax = optimizer(wrapper, logfile=tag + '.log')
relax.attach(Trajectory(tag + '.traj', 'w', atoms=atoms))
tincl = -time()
error = ''
try:
relax.run(fmax=fmax, steps=10000000)
except Exception as x:
wrapper.nsteps = float('inf')
error = f'{x.__class__.__name__}: {x}'
tb = traceback.format_exc()
with open(tag + '.err', 'w') as fd:
fd.write(f'{error}\n{tb}\n')
tincl += time()
return error, wrapper.nsteps, wrapper.texcl, tincl
def test_optimizer(systems, optimizer, calculator, prefix='', db=None,
eggbox=0.0):
"""Test optimizer on systems."""
for name, atoms in systems:
if db is not None:
optname = optimizer.__name__
id = db.reserve(optimizer=optname, name=name)
if id is None:
continue
atoms = atoms.copy()
tag = f'{prefix}{optname}-{name}'
atoms.calc = calculator(txt=tag + '.txt')
error, nsteps, texcl, tincl = run_test(atoms, optimizer, tag,
eggbox=eggbox)
if db is not None:
db.write(atoms,
id=id,
optimizer=optname,
name=name,
error=error,
n=nsteps,
t=texcl,
T=tincl,
eggbox=eggbox)
def main():
parser = argparse.ArgumentParser(
description='Test ASE optimizers')
parser.add_argument('systems', help='File containing test systems.')
parser.add_argument('optimizer', nargs='*',
help='Optimizer name(s). Choose from: {}. '
.format(', '.join(all_optimizers)) +
'Default is all optimizers.')
parser.add_argument('-e', '--egg-box', type=float, default=0.0,
help='Fake egg-box error in eV.')
args = parser.parse_args()
systems = [(row.name, row.toatoms())
for row in ase.db.connect(args.systems).select()]
db = ase.db.connect('results.db')
if not args.optimizer:
args.optimizer = all_optimizers
for opt in args.optimizer:
print(opt)
optimizer = get_optimizer(opt)
test_optimizer(systems, optimizer, EMT, db=db, eggbox=args.egg_box)
if __name__ == '__main__':
main()
|
