def test_to_atomistic2(self):
x0 = np.array(range(self.nrigid * 6), dtype=float)
x2 = x0.reshape([-1, 3])
for p in x2[self.nrigid:, :]:
p /= np.linalg.norm(p)
atomistic = self.topology.to_atomistic(x0).flatten()
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g)
rbpot = RBPotentialWrapper(self.topology, lj)
print rbpot.getEnergy(x0)
def test_to_atomistic2(self):
x0 = np.array(range(self.nrigid * 6), dtype=float)
x2 = x0.reshape([-1,3])
for p in x2[self.nrigid:,:]:
p /= np.linalg.norm(p)
atomistic = self.topology.to_atomistic(x0).flatten()
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g)
rbpot = RBPotentialWrapper(self.topology, lj)
print rbpot.getEnergy(x0)
def test_pot_wrapper(self):
from pele.angleaxis import _cpp_aa
from pele.potentials import LJ
rbpot_cpp = _cpp_aa.RBPotentialWrapper(self.topology, LJ())
rbpot = RBPotentialWrapper(self.topology, LJ())
self.assertAlmostEqual(rbpot_cpp.getEnergy(self.x0),
rbpot.getEnergy(self.x0), 4)
e1, grad1 = rbpot_cpp.getEnergyGradient(self.x0)
e2, grad2 = rbpot.getEnergyGradient(self.x0)
self.assertAlmostEqual(e1, e2, 4)
for g1, g2 in zip(grad1, grad2):
self.assertAlmostEqual(g1, g2, 3)
def test_pot_wrapper(self):
from pele.angleaxis import _cpp_aa
from pele.potentials import LJ
rbpot_cpp = _cpp_aa.RBPotentialWrapper(self.topology, LJ())
rbpot = RBPotentialWrapper(self.topology, LJ())
self.assertAlmostEqual(rbpot_cpp.getEnergy(self.x0),
rbpot.getEnergy(self.x0), 4)
e1, grad1 = rbpot_cpp.getEnergyGradient(self.x0)
e2, grad2 = rbpot.getEnergyGradient(self.x0)
self.assertAlmostEqual(e1, e2, 4)
for g1, g2 in zip(grad1, grad2):
self.assertAlmostEqual(g1, g2, 3)
def test_to_atomistic2(self):
x0 = np.array(range(self.nrigid * 6), dtype=float);
x2 = x0.reshape([-1,3])
for p in x2[self.nrigid:,:]:
p /= np.linalg.norm(p);
print x0
print "range to atomistic"
print x0
atomistic = self.topology.to_atomistic(x0).flatten()
print atomistic
print atomistic.size
print atomistic[14]
print atomistic[23]
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g);
print "transformed gradient"
print grb
print "rbpotential"
rbpot = RBPotentialWrapper(self.topology, lj);
print rbpot.getEnergy(x0);
class TestOTPExplicit(unittest.TestCase):
def setUp(self):
nrigid = 3
self.topology = RBTopology()
self.topology.add_sites([make_otp() for i in range(nrigid)])
self.topology.finalize_setup()
cartesian_potential = LJ()
self.pot = RBPotentialWrapper(self.topology, cartesian_potential)
self.x0 = _x03
self.x0 = np.array(self.x0)
self.e0 = -17.3387670023
assert nrigid * 6 == self.x0.size
self.x0atomistic = _x03_atomistic
self.nrigid = nrigid
def test_energy(self):
e = self.pot.getEnergy(self.x0)
self.assertAlmostEqual(e, self.e0, delta=1e-4)
def test_energy_gradient(self):
e = self.pot.getEnergy(self.x0)
gnum = self.pot.NumericalDerivative(self.x0)
e2, g = self.pot.getEnergyGradient(self.x0)
self.assertAlmostEqual(e, e2, delta=1e-4)
for i in range(g.size):
self.assertAlmostEqual(g[i], gnum[i], 2)
def test_to_atomistic(self):
xatom = self.topology.to_atomistic(self.x0).flatten()
for i in range(xatom.size):
self.assertAlmostEqual(xatom[i], self.x0atomistic[i], 2)
def test_site_to_atomistic(self):
rf = make_otp()
p = np.array([1., 2, 3])
p /= np.linalg.norm(p)
com = np.array([4., 5, 6])
print("otp to atomistic")
print(rf.to_atomistic(com, p))
print("otp transform grad")
g = np.array(list(range(9)), dtype=float).reshape([-1, 3])
print(g.reshape(-1))
print(rf.transform_grad(p, g))
def test_to_atomistic2(self):
x0 = np.array(list(range(self.nrigid * 6)), dtype=float)
x2 = x0.reshape([-1, 3])
for p in x2[self.nrigid:, :]:
p /= np.linalg.norm(p)
atomistic = self.topology.to_atomistic(x0).flatten()
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g)
rbpot = RBPotentialWrapper(self.topology, lj)
print(rbpot.getEnergy(x0))
class TestOTPExplicit(unittest.TestCase):
def make_otp(self):
"""this constructs a single OTP molecule"""
otp = RigidFragment()
otp.add_atom("O", np.array([0.0, -2. / 3 * np.sin(7. * pi / 24.),
0.0]), 1.)
otp.add_atom(
"O",
np.array([cos(7. * pi / 24.), 1. / 3. * sin(7. * pi / 24.), 0.0]),
1.)
otp.add_atom(
"O",
np.array([-cos(7. * pi / 24.), 1. / 3. * sin(7. * pi / 24), 0.0]),
1.)
otp.finalize_setup()
return otp
def setUp(self):
nrigid = 3
self.topology = RBTopology()
self.topology.add_sites([self.make_otp() for i in xrange(nrigid)])
self.topology.finalize_setup()
cartesian_potential = LJ()
self.pot = RBPotentialWrapper(self.topology, cartesian_potential)
self.x0 = _x03
self.x0 = np.array(self.x0)
self.e0 = -17.3387670023
assert nrigid * 6 == self.x0.size
self.x0atomistic = _x03_atomistic
self.nrigid = nrigid
def test_energy(self):
e = self.pot.getEnergy(self.x0)
self.assertAlmostEqual(e, self.e0, delta=1e-4)
def test_energy_gradient(self):
e = self.pot.getEnergy(self.x0)
gnum = self.pot.NumericalDerivative(self.x0)
e2, g = self.pot.getEnergyGradient(self.x0)
self.assertAlmostEqual(e, e2, delta=1e-4)
for i in xrange(g.size):
self.assertAlmostEqual(g[i], gnum[i], 2)
def test_to_atomistic(self):
xatom = self.topology.to_atomistic(self.x0).flatten()
for i in xrange(xatom.size):
self.assertAlmostEqual(xatom[i], self.x0atomistic[i], 2)
def test_site_to_atomistic(self):
rf = self.make_otp()
p = np.array([1., 2, 3])
p /= np.linalg.norm(p)
com = np.array([4., 5, 6])
print "otp to atomistic"
print rf.to_atomistic(com, p)
print "otp transform grad"
g = np.array(range(9), dtype=float).reshape([-1, 3])
print g.reshape(-1)
print rf.transform_grad(p, g)
def test_to_atomistic2(self):
x0 = np.array(range(self.nrigid * 6), dtype=float)
x2 = x0.reshape([-1, 3])
for p in x2[self.nrigid:, :]:
p /= np.linalg.norm(p)
atomistic = self.topology.to_atomistic(x0).flatten()
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g)
rbpot = RBPotentialWrapper(self.topology, lj)
print rbpot.getEnergy(x0)
class TestOTPExplicit(unittest.TestCase):
def setUp(self):
nrigid = 3
self.topology = RBTopology()
self.topology.add_sites([make_otp() for i in range(nrigid)])
self.topology.finalize_setup()
cartesian_potential = LJ()
self.pot = RBPotentialWrapper(self.topology, cartesian_potential)
self.x0 = _x03
self.x0 = np.array(self.x0)
self.e0 = -17.3387670023
assert nrigid * 6 == self.x0.size
self.x0atomistic = _x03_atomistic
self.nrigid = nrigid
def test_energy(self):
e = self.pot.getEnergy(self.x0)
self.assertAlmostEqual(e, self.e0, delta=1e-4)
def test_energy_gradient(self):
e = self.pot.getEnergy(self.x0)
gnum = self.pot.NumericalDerivative(self.x0)
e2, g = self.pot.getEnergyGradient(self.x0)
self.assertAlmostEqual(e, e2, delta=1e-4)
for i in range(g.size):
self.assertAlmostEqual(g[i], gnum[i], 2)
def test_to_atomistic(self):
xatom = self.topology.to_atomistic(self.x0).flatten()
for i in range(xatom.size):
self.assertAlmostEqual(xatom[i], self.x0atomistic[i], 2)
def test_site_to_atomistic(self):
rf = make_otp()
p = np.array([1., 2, 3])
p /= np.linalg.norm(p)
com = np.array([4., 5, 6])
print("otp to atomistic")
print(rf.to_atomistic(com, p))
print("otp transform grad")
g = np.array(list(range(9)), dtype=float).reshape([-1,3])
print(g.reshape(-1))
print(rf.transform_grad(p, g))
def test_to_atomistic2(self):
x0 = np.array(list(range(self.nrigid * 6)), dtype=float)
x2 = x0.reshape([-1,3])
for p in x2[self.nrigid:,:]:
p /= np.linalg.norm(p)
atomistic = self.topology.to_atomistic(x0).flatten()
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g)
rbpot = RBPotentialWrapper(self.topology, lj)
print(rbpot.getEnergy(x0))
class TestOTP(unittest.TestCase):
def make_otp(self):
"""this constructs a single OTP molecule"""
otp = RigidFragment()
otp.add_atom("O", np.array([0.0, -2./3 * np.sin( 7.*pi/24.), 0.0]), 1.)
otp.add_atom("O", np.array([cos( 7.*pi/24.), 1./3. * sin( 7.* pi/24.), 0.0]), 1.)
otp.add_atom("O", np.array([-cos( 7.* pi/24.), 1./3. * sin( 7.*pi/24), 0.0]), 1.)
otp.finalize_setup()
print "otp"
print otp.atom_positions
return otp
def setUp(self):
nrigid = 3
self.topology = RBTopology()
self.topology.add_sites([self.make_otp() for i in xrange(nrigid)])
cartesian_potential = LJ()
self.pot = RBPotentialWrapper(self.topology, cartesian_potential)
self.x0 = [2.550757898788, 2.591553038507, 3.696836364193,
2.623281513163, 3.415794212648, 3.310786279789,
1.791383852327, 2.264321752809, 4.306217333671,
0.761945654023, -0.805817782109, 1.166981882601,
0.442065301864, -2.747066418223, -1.784325262714,
-1.520905562598, 0.403670860200, -0.729768985400, ]
self.x0 = np.array(self.x0)
self.e0 = -17.3387670023
assert nrigid * 6 == self.x0.size
self.x0atomistic = [ 3.064051819556, 2.474533745459, 3.646107658946,
2.412011983074, 2.941152759499, 4.243695098053,
2.176209893734, 2.358972610563, 3.200706335581,
2.786627589565, 3.211876105193, 2.850924310983,
1.962626909252, 3.436918873216, 3.370903763850,
3.120590040673, 3.598587659535, 3.710530764535,
1.697360211099, 2.317229950712, 4.823998989452,
2.283487958310, 1.840698306602, 4.168734267290,
1.393303387573, 2.635037001113, 3.925918744272, ]
self.nrigid = nrigid
def test_energy(self):
e = self.pot.getEnergy(self.x0)
self.assertAlmostEqual(e, self.e0, delta=1e-4)
def test_energy_gradient(self):
e = self.pot.getEnergy(self.x0)
gnum = self.pot.NumericalDerivative(self.x0)
e2, g = self.pot.getEnergyGradient(self.x0)
self.assertAlmostEqual(e, e2, delta=1e-4)
for i in xrange(g.size):
self.assertAlmostEqual(g[i], gnum[i], 2)
def test_to_atomistic(self):
xatom = self.topology.to_atomistic(self.x0).flatten()
for i in xrange(xatom.size):
self.assertAlmostEqual(xatom[i], self.x0atomistic[i], 2)
def test_site_to_atomistic(self):
rf = self.make_otp()
p = np.array([1., 2, 3])
p /= np.linalg.norm(p)
com = np.array([4., 5, 6])
print "otp to atomistic"
print rf.to_atomistic(com, p)
print "otp transform grad"
g = np.array(range(9), dtype=float).reshape([-1,3])
print g.reshape(-1)
print rf.transform_grad(p, g)
def test_to_atomistic2(self):
x0 = np.array(range(self.nrigid * 6), dtype=float);
x2 = x0.reshape([-1,3])
for p in x2[self.nrigid:,:]:
p /= np.linalg.norm(p);
print x0
print "range to atomistic"
print x0
atomistic = self.topology.to_atomistic(x0).flatten()
print atomistic
print atomistic.size
print atomistic[14]
print atomistic[23]
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g);
print "transformed gradient"
print grb
print "rbpotential"
rbpot = RBPotentialWrapper(self.topology, lj);
print rbpot.getEnergy(x0);
class TestOTPExplicit(unittest.TestCase):
def make_otp(self):
"""this constructs a single OTP molecule"""
otp = RigidFragment()
otp.add_atom("O", np.array([0.0, -2./3 * np.sin( 7.*pi/24.), 0.0]), 1.)
otp.add_atom("O", np.array([cos( 7.*pi/24.), 1./3. * sin( 7.* pi/24.), 0.0]), 1.)
otp.add_atom("O", np.array([-cos( 7.* pi/24.), 1./3. * sin( 7.*pi/24), 0.0]), 1.)
otp.finalize_setup()
return otp
def setUp(self):
nrigid = 3
self.topology = RBTopology()
self.topology.add_sites([self.make_otp() for i in xrange(nrigid)])
self.topology.finalize_setup()
cartesian_potential = LJ()
self.pot = RBPotentialWrapper(self.topology, cartesian_potential)
self.x0 = _x03
self.x0 = np.array(self.x0)
self.e0 = -17.3387670023
assert nrigid * 6 == self.x0.size
self.x0atomistic = _x03_atomistic
self.nrigid = nrigid
def test_energy(self):
e = self.pot.getEnergy(self.x0)
self.assertAlmostEqual(e, self.e0, delta=1e-4)
def test_energy_gradient(self):
e = self.pot.getEnergy(self.x0)
gnum = self.pot.NumericalDerivative(self.x0)
e2, g = self.pot.getEnergyGradient(self.x0)
self.assertAlmostEqual(e, e2, delta=1e-4)
for i in xrange(g.size):
self.assertAlmostEqual(g[i], gnum[i], 2)
def test_to_atomistic(self):
xatom = self.topology.to_atomistic(self.x0).flatten()
for i in xrange(xatom.size):
self.assertAlmostEqual(xatom[i], self.x0atomistic[i], 2)
def test_site_to_atomistic(self):
rf = self.make_otp()
p = np.array([1., 2, 3])
p /= np.linalg.norm(p)
com = np.array([4., 5, 6])
print "otp to atomistic"
print rf.to_atomistic(com, p)
print "otp transform grad"
g = np.array(range(9), dtype=float).reshape([-1,3])
print g.reshape(-1)
print rf.transform_grad(p, g)
def test_to_atomistic2(self):
x0 = np.array(range(self.nrigid * 6), dtype=float)
x2 = x0.reshape([-1,3])
for p in x2[self.nrigid:,:]:
p /= np.linalg.norm(p)
atomistic = self.topology.to_atomistic(x0).flatten()
from pele.potentials import LJ
lj = LJ()
e, g = lj.getEnergyGradient(atomistic.reshape(-1))
grb = self.topology.transform_gradient(x0, g)
rbpot = RBPotentialWrapper(self.topology, lj)
print rbpot.getEnergy(x0)