def setup(self):
self.dyn = CalvinistDynamics([-0.1, 0.1, 0.3, 0.5, 0.7,
-0.1, 0.2, 0.4, 0.6, 0.8,
])
self.dyn.initialized = True
SampleMover.engine = self.dyn
op = CV_Function("myid", f=lambda snap :
snap.coordinates[0][0])
stateA = CVRangeVolume(op, -100, 0.0)
stateB = CVRangeVolume(op, 0.65, 100)
volX = CVRangeVolume(op, -100, 0.25)
volY = CVRangeVolume(op, -100, 0.40)
self.ens1 = paths.TISEnsemble(stateA, stateB, volX, op)
self.ens2 = paths.TISEnsemble(stateA, stateB, volY, op)
init_traj1 = make_1d_traj(
coordinates=[-0.1, 0.1, 0.2, 0.3, 0.24, 0.15, 0.06, -0.07],
velocities=[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
)
init_traj2 = make_1d_traj(
coordinates=[-0.1, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7],
velocities=[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
)
self.samp1 = Sample(trajectory=init_traj1, replica=0,
ensemble=self.ens1)
self.samp2 = Sample(trajectory=init_traj2, replica=1,
ensemble=self.ens2)
self.shooter = ForwardShootMover(selector=UniformSelector(),
ensemble=self.ens2)
self.pathrev = PathReversalMover(ensemble=self.ens1)
ens_dict = {self.ens1 : self.pathrev, self.ens2 : self.shooter}
# self.mover = EnsembleDictionaryMover(ens_dict)
self.mover = RandomAllowedChoiceMover([self.shooter, self.pathrev])
def setup(self):
self.l1 = LengthEnsemble(1)
self.l2 = LengthEnsemble(2)
self.l3 = LengthEnsemble(3)
self.s1 = Sample(replica=1, ensemble=self.l2)
self.s2 = Sample(replica=2, ensemble=self.l1)
self.s3 = Sample(replica=3, ensemble=self.l1)
self.s4 = Sample(replica=2, ensemble=self.l3)
self.sset = SampleSet([self.s1, self.s2, self.s3, self.s4])
def setup(self):
self.ensA = LengthEnsemble(1)
self.ensB = LengthEnsemble(2)
traj0A = Trajectory([0.5])
traj1A = Trajectory([1.0])
traj2B = Trajectory([0.5, 0.75])
traj2B_ = Trajectory([0.8, 0.9])
self.s0A = Sample(replica=0, trajectory=traj0A, ensemble=self.ensA)
self.s1A = Sample(replica=1, trajectory=traj1A, ensemble=self.ensA)
self.s2B = Sample(replica=2, trajectory=traj2B, ensemble=self.ensB)
self.s2B_ = Sample(replica=2, trajectory=traj2B_, ensemble=self.ensB)
self.testset = SampleSet([self.s0A, self.s1A, self.s2B])
def test_setitem_replica(self):
ensC = LengthEnsemble(3)
traj3C = Trajectory([-0.5, -0.25, 0.1])
s3C = Sample(replica=3, trajectory=traj3C, ensemble=ensC)
self.testset[3] = s3C
self.testset.consistency_check()
assert_equal(len(self.testset), 4)
def test_extension_fails(self):
innermost_bad_extension = [-0.25, 0.1, 0.5, 0.1, -0.25]
traj_bad_extension = make_1d_traj(innermost_bad_extension, [1.0]*5)
samp_bad_extension = Sample(
replica=0,
trajectory=traj_bad_extension,
ensemble=self.innermost
)
assert_equal(self.innermost(traj_bad_extension), True)
gs = SampleSet([self.minus_sample, samp_bad_extension])
change = self.mover.move(gs)
assert_equal(change.accepted, False) # whole minus has failed
# Minus : Filter :ChooseFB : CondSeq
sub = change.subchange.subchange.subchange
assert_equal(len(sub.trials), 2)
assert_equal(len(change.trials), 0) # no trials survive filtering
assert_subchanges_set_accepted(sub, [True, False])
# first two work and the extention fails
# this only happens due to length
assert_equal(
len(sub[-1].trials[0].trajectory),
len(traj_bad_extension)+self.dyn.n_frames_max-1
)
def test_successful_move(self):
init_innermost = make_1d_traj(self.list_innermost, [1.0]*5)
init_sample = Sample(
replica=0,
trajectory=init_innermost,
ensemble=self.innermost
)
gs = SampleSet([init_sample, self.minus_sample])
extend_forward = self.list_innermost + [0.12, 0.32, -0.131]
extend_backward = [-0.13, 0.13, 0.33] + self.list_innermost
assert_equal(self.minus(make_1d_traj(extend_forward)), True)
assert_equal(self.minus(make_1d_traj(extend_backward)), True)
seg_dir = {}
for i in range(100):
change = self.mover.move(gs)
samples = change.results
sub_samples = change.subchange.subchange.results
assert_equal(len(samples), 2)
assert_equal(len(sub_samples), 4)
s_inner = [s for s in sub_samples if s.ensemble==self.innermost]
s_minus = [s for s in sub_samples if s.ensemble==self.minus]
s_sub = [s for s in sub_samples if s.ensemble==self.minus._segment_ensemble]
assert_equal(len(s_inner), 1)
assert_equal(len(s_minus), 1)
assert_equal(len(s_sub), 2)
for c in change:
assert_equal(c.accepted, True)
assert_equal(change.canonical.mover, self.mover)
key = ""
s_inner0_xvals = [s.coordinates[0,0] for s in s_inner[0].trajectory]
if items_equal(s_inner0_xvals, self.first_segment):
key += "1"
elif items_equal(s_inner0_xvals, self.second_segment):
key += "2"
else:
print "s_inner0_xvals:", s_inner0_xvals
raise RuntimeError("Chosen segment neither first nor last!")
# final sample s_minus is accepted
s_minus_xvals = [s.coordinates[0,0] for s in s_minus[-1].trajectory]
if items_equal(s_minus_xvals, extend_forward):
key += "f"
elif items_equal(s_minus_xvals, extend_backward):
key += "b"
else:
print "s_minus_xvals:", s_minus_xvals
raise RuntimeError("Unexpected minus extension result!")
try:
seg_dir[key] += 1
except KeyError:
seg_dir[key] = 1
assert_equal(len(seg_dir.keys()), 4)
def add_empty_to_cache(self, idx, trajectory_idx, replica_idx, bias):
obj = Sample(
trajectory=self.storage.trajectories[int(trajectory_idx)],
replica=int(replica_idx),
bias=float(bias)
)
obj.idx[self.storage] = idx
obj._origin = self.storage
del obj.details
del obj.ensemble
del obj.mover
del obj.parent
self.cache[idx] = obj
return obj
def test_A_A_path(self):
trajA_A = make_1d_traj(coordinates=[-0.3, 0.1, -0.4])
sampA_A = Sample(trajectory=trajA_A,
ensemble=self.tis,
replica=0)
gs_A_A = SampleSet([sampA_A])
change = self.move.move(gs_A_A)
assert_equal(change.accepted, False)
def test_AB_path(self):
trajAXB = make_1d_traj(coordinates=[-0.2, 0.75, 1.8])
sampAXB = Sample(trajectory=trajAXB,
ensemble=self.tis,
replica=0)
gs_AXB = SampleSet([sampAXB])
change = self.move.move(gs_AXB)
assert_equal(change.accepted, False)
def setup(self):
traj = Trajectory([-0.5, 0.7, 1.1])
op = CallIdentity()
volA = CVRangeVolume(op, -100, 0.0)
volB = CVRangeVolume(op, 1.0, 100)
volX = CVRangeVolume(op, -100, 0.25)
tis = paths.TISEnsemble(volA, volB, volX)
tps = ef.A2BEnsemble(volA, volB)
len3 = LengthEnsemble(3)
len2 = LengthEnsemble(2)
self.hop_to_tis = RandomAllowedChoiceMover(
map(lambda ens : EnsembleHopMover(*ens),
[[tis, tis],
[tps, tis],
[len3, tis],
[len2, tis]]
)
)
self.hop_to_tps = RandomAllowedChoiceMover(
map(lambda ens : EnsembleHopMover(*ens),
[[tis, tps],
[tps, tps],
[len3, tps],
[len2, tps]]
)
)
self.hop_to_len3 = RandomAllowedChoiceMover(
map(lambda ens : EnsembleHopMover(*ens),
[[tis, len3],
[tps, len3],
[len3, len3],
[len2, len3]]
)
)
self.hop_to_len2 = RandomAllowedChoiceMover(
map(lambda ens : EnsembleHopMover(*ens),
[[tis, len2],
[tps, len2],
[len3, len2],
[len2, len2]]
)
)
self.init_sample = Sample(trajectory=traj,
ensemble=len3,
replica=0)
self.tis = tis
self.tps = tps
self.len3 = len3
self.len2 = len2
self.everything_accepted_movers = [
self.hop_to_tis, self.hop_to_len3, self.hop_to_tps
]
self.first_rejected_movers = [
self.hop_to_len2, self.hop_to_len3, self.hop_to_tps
]
self.last_rejected_movers = [
self.hop_to_tis, self.hop_to_tps, self.hop_to_len2
]
def test_BA_path(self):
trajBXA = make_1d_traj(coordinates=[1.2, 0.7, -0.25])
sampBXA = Sample(trajectory=trajBXA,
ensemble=self.tis,
replica=0)
gs_BXA = SampleSet([sampBXA])
change = self.move.move(gs_BXA)
# print [[v.coordinates[0] for v in t.trajectory] for t in change.trials]
assert_equal(change.accepted, True)
def setup(self):
op = CV_Function("myid", f=lambda snap :
snap.coordinates[0][0])
state1 = CVRangeVolume(op, -100, 0.0)
state2 = CVRangeVolume(op, 1, 100)
volA = CVRangeVolume(op, -100, 0.25)
volB = CVRangeVolume(op, -100, 0.50)
self.tisA = paths.TISEnsemble(state1, state2, volA)
self.tisB = paths.TISEnsemble(state1, state2, volB)
self.traj0 = make_1d_traj([-0.1, 0.2, 0.3, 0.1, -0.2])
self.traj1 = make_1d_traj([-0.1, 0.1, 0.4, 0.6, 0.3, 0.2, -0.15])
self.traj2 = make_1d_traj([-0.1, 0.2, 0.3, 0.7, 0.6, 0.4, 0.1, -0.15])
self.sampA0 = Sample(replica=0, trajectory=self.traj0, ensemble=self.tisA)
self.sampB1 = Sample(replica=1, trajectory=self.traj1, ensemble=self.tisB)
self.sampA2 = Sample(replica=2, trajectory=self.traj2, ensemble=self.tisA)
self.gs_B1A2 = SampleSet([self.sampB1, self.sampA2])
self.gs_A0B1 = SampleSet([self.sampA0, self.sampB1])
def setup(self):
self.mytraj = make_1d_traj(coordinates=[-0.5, 0.1, 0.2, 0.3, 0.5],
velocities=[1.0, 1.0, 1.0, 1.0, 1.0])
self.dyn = CalvinistDynamics(
[-0.5, -0.4, -0.3, -0.2, -0.1, 0.1, 0.2, 0.3, 0.4, 0.5])
self.dyn.initialized = True
self.ens = LengthEnsemble(5)
self.gs = SampleSet(
Sample(replica=0, trajectory=self.mytraj, ensemble=self.ens))
def test_AXA_path(self):
trajAXA = make_1d_traj(coordinates=[-0.1, 0.75, -0.6],
velocities=[0.1, 0.05, -0.05])
assert_equal(self.tis(trajAXA), True)
sampAXA = Sample(trajectory=trajAXA,
ensemble=self.tis,
replica=0)
gs_AXA = SampleSet([sampAXA])
change = self.move.move(gs_AXA)
assert_equal(change.accepted, True)
def _load(self, idx):
obj = Sample(
trajectory=self.vars['trajectory'][idx],
replica=self.vars['replica'][idx],
ensemble=self.vars['ensemble'][idx],
parent=self.vars['parent'][idx],
details=self.vars['details'][idx],
bias=self.vars['bias'][idx],
mover=self.vars['mover'][idx]
)
return obj
def setup(self):
op = CV_Function("myid", f=lambda snap :
snap.coordinates[0][0])
volA = CVRangeVolume(op, -100, 0.0)
volB = CVRangeVolume(op, 1.0, 100)
volX = CVRangeVolume(op, -100, 0.25)
self.dyn = CalvinistDynamics([
# successful move: (backward extension then forward)
-0.13, 0.13, 0.33, -0.11, -0.12, 0.12, 0.32, -0.131,
# never leaves state:
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.25,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
# goes to other state:
1.16, 1.26, 1.16, -0.16, 1.16, 1.26, 1.16
])
SampleMover.engine = self.dyn
self.dyn.initialized = True
self.innermost = paths.TISEnsemble(volA, volB, volX)
self.minus = paths.MinusInterfaceEnsemble(volA, volX)
self.mover = MinusMover(
minus_ensemble=self.minus,
innermost_ensembles=self.innermost
)
self.first_segment = [-0.1, 0.1, 0.3, 0.1, -0.15]
self.list_innermost = [-0.11, 0.11, 0.31, 0.11, -0.12]
self.second_segment = [-0.25, 0.2, 0.4, 0.2, -0.2]
init_minus = make_1d_traj(
coordinates=self.first_segment + [-0.35] + self.second_segment,
velocities=[1.0]*11
)
self.minus_sample = Sample(
replica=-1,
trajectory=init_minus,
ensemble=self.minus
)
def test_repex_fails_minus_crosses_to_state(self):
minus_crosses_to_state = make_1d_traj(
[-0.11, 0.5, 1.8, 0.6, -0.12, 0.7, 1.7, 0.4, -0.13]
)
badminus_sample = Sample(
replica=-1,
trajectory=minus_crosses_to_state,
ensemble=self.minus
)
init_sample = Sample(
replica=0,
trajectory=make_1d_traj(self.list_innermost, [1.0]*5),
ensemble=self.innermost
)
gs = SampleSet([badminus_sample, init_sample])
assert_equal(self.minus(minus_crosses_to_state), True)
change = self.mover.move(gs)
sub = change.subchange.subchange
assert_equal(len(sub.trials), 3) # stop after failed repex
assert_equal(len(change.trials), 1)
assert_subchanges_set_accepted(sub, [True, False, False])
def test_first_hop_fails(self):
crossing_traj = make_1d_traj([-0.11, 0.11, 0.31, 1.01], [1.0]*4)
crossing_samp = Sample(replica=0, trajectory=crossing_traj,
ensemble=self.innermost)
gs = SampleSet([crossing_samp])
gs.sanity_check()
change = self.mover.move(gs)
assert_equal(change.accepted, False)
assert_equal(len(change.results), 0)
sub_trials = change.subchange.subchange.subchange.trials
assert_equal(len(sub_trials), 1)
assert_equal(sub_trials[0].trajectory, crossing_traj)
assert_equal(sub_trials[0].ensemble, self.minus._segment_ensemble)
def test_repex_fails_innermost_crosses_state(self):
innermost_crosses_to_state = make_1d_traj([-0.11, 0.5, 1.8])
samp_crosses_to_state = Sample(
replica=0,
trajectory=innermost_crosses_to_state,
ensemble=self.innermost
)
gs = SampleSet([samp_crosses_to_state, self.minus_sample])
change = self.mover.move(gs)
assert_equal(len(change.trials), 1) # stop after failed repex
sub = change.subchange.subchange
assert_equal(self.innermost(innermost_crosses_to_state), True)
assert_equal(len(sub.trials), 3) # stop after failed repex
assert_subchanges_set_accepted(sub, [True, False, False])
def test_repex_fails_other_ensemble(self):
innermost_other_ensemble = make_1d_traj([-0.11, 0.1, -0.12])
samp_other_ensemble = Sample(
replica=0,
trajectory=innermost_other_ensemble,
ensemble=self.innermost
)
gs = SampleSet([samp_other_ensemble, self.minus_sample])
change = self.mover.move(gs)
assert_equal(len(change.trials), 1)
sub = change.subchange.subchange
assert_equal(self.innermost(innermost_other_ensemble), False)
assert_equal(sub[0].accepted, True)
assert_equal(sub[1].accepted, False)
assert_equal(len(sub.trials), 3) # stop after failed repex
def _add_empty_to_cache(self, idx, trajectory_idx, replica_idx, bias,
ensemble_idx, parent_idx, details_idx, mover_idx):
storage = self.storage
obj = Sample(
trajectory=storage.trajectories[int(trajectory_idx)],
replica=int(replica_idx),
bias=float(bias),
ensemble=storage.ensembles[int(ensemble_idx)],
mover=storage.pathmovers[int(mover_idx)],
parent=LoaderProxy(storage.samples, int(parent_idx)),
details=LoaderProxy(storage.details, int(details_idx))
)
self.index[obj] = idx
self.cache[idx] = obj
return obj
def setup(self):
self.dyn = CalvinistDynamics([-0.1, 0.1, 0.3, 0.5, 0.7,
-0.1, 0.2, 0.4, 0.6, 0.8,
])
SampleMover.engine = self.dyn
op = CV_Function("myid", f=lambda snap :
snap.coordinates[0][0])
stateA = CVRangeVolume(op, -100, 0.0)
stateB = CVRangeVolume(op, 0.65, 100)
self.tps = ef.A2BEnsemble(stateA, stateB)
init_traj = make_1d_traj(
coordinates=[-0.1, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7],
velocities=[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
)
self.init_samp = SampleSet([Sample(
trajectory=init_traj,
replica=0,
ensemble=self.tps
)])
def setup(self):
traj = Trajectory([-0.5, 0.7, 1.1])
op = CallIdentity()
volA = CVRangeVolume(op, -100, 0.0)
volB = CVRangeVolume(op, 1.0, 100)
volX = CVRangeVolume(op, -100, 0.25)
self.tis = paths.TISEnsemble(volA, volB, volX)
self.tps = ef.A2BEnsemble(volA, volB)
self.len3 = LengthEnsemble(3)
self.init_samp = SampleSet([Sample(trajectory=traj,
ensemble=self.len3,
replica=0)])
self.hop_to_tis = EnsembleHopMover(
ensemble=self.len3,
target_ensemble=self.tis
)
self.hop_to_tps = EnsembleHopMover(
ensemble=self.len3,
target_ensemble=self.tps
)
self.mover = RandomChoiceMover([self.hop_to_tis, self.hop_to_tps])
def setup(self):
op = CV_Function("myid", fcn=lambda snap :
snap.coordinates[0][0])
volA = CVRangeVolume(op, -100, 0.0)
volB = CVRangeVolume(op, 1.0, 100)
volX = CVRangeVolume(op, -100, 0.25)
self.dyn = CalvinistDynamics([
# successful move: (backward extension then forward)
-0.13, 0.13, 0.33, -0.11, -0.12, 0.12, 0.32, -0.131,
# never leaves state:
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.25,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
# goes to other state:
1.16, 1.26, 1.16, -0.16, 1.16, 1.26, 1.16
])
SampleGeneratingMover.engine = self.dyn
self.dyn.initialized = True
self.innermost = paths.TISEnsemble(volA, volB, volX)
self.minus = paths.MinusInterfaceEnsemble(volA, volX)
self.mover = MinusMover(
minus_ensemble=self.minus,
innermost_ensembles=self.innermost
)
self.first_segment = [-0.1, 0.1, 0.3, 0.1, -0.15]
self.list_innermost = [-0.11, 0.11, 0.31, 0.11, -0.12]
self.second_segment = [-0.25, 0.2, 0.4, 0.2, -0.2]
init_minus = make_1d_traj(
coordinates=self.first_segment + [-0.35] + self.second_segment,
velocities=[1.0]*11
)
self.minus_sample = Sample(
replica=-1,
trajectory=init_minus,
ensemble=self.minus
)
def setup(self):
op = CallIdentity()
vol = paths.CVRangeVolume(op, -0.5, 0.5)
inX = paths.AllInXEnsemble(vol)
outX = paths.AllOutXEnsemble(vol)
self.ensemble = paths.SequentialEnsemble([
inX, outX, inX, outX, inX, outX, inX
])
self.subensemble = paths.SequentialEnsemble([
paths.SingleFrameEnsemble(inX),
outX,
paths.SingleFrameEnsemble(inX)
])
self.traj_with_3_subtrajs = Trajectory(
[0.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0, 2.0, 0.0]
)
self.subtraj0 = Trajectory([0.0, 1.0, 1.0, 0.0])
self.subtraj1 = Trajectory([0.0, 1.0, 0.0])
self.subtraj2 = Trajectory([0.0, 2.0, 0.0])
self.gs = SampleSet(Sample(
replica=0,
ensemble=self.ensemble,
trajectory=self.traj_with_3_subtrajs
))
class testMinusMover(object):
def setup(self):
op = CV_Function("myid", f=lambda snap :
snap.coordinates[0][0])
volA = CVRangeVolume(op, -100, 0.0)
volB = CVRangeVolume(op, 1.0, 100)
volX = CVRangeVolume(op, -100, 0.25)
self.dyn = CalvinistDynamics([
# successful move: (backward extension then forward)
-0.13, 0.13, 0.33, -0.11, -0.12, 0.12, 0.32, -0.131,
# never leaves state:
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.25,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
# goes to other state:
1.16, 1.26, 1.16, -0.16, 1.16, 1.26, 1.16
])
SampleMover.engine = self.dyn
self.dyn.initialized = True
self.innermost = paths.TISEnsemble(volA, volB, volX)
self.minus = paths.MinusInterfaceEnsemble(volA, volX)
self.mover = MinusMover(
minus_ensemble=self.minus,
innermost_ensembles=self.innermost
)
self.first_segment = [-0.1, 0.1, 0.3, 0.1, -0.15]
self.list_innermost = [-0.11, 0.11, 0.31, 0.11, -0.12]
self.second_segment = [-0.25, 0.2, 0.4, 0.2, -0.2]
init_minus = make_1d_traj(
coordinates=self.first_segment + [-0.35] + self.second_segment,
velocities=[1.0]*11
)
self.minus_sample = Sample(
replica=-1,
trajectory=init_minus,
ensemble=self.minus
)
def test_is_ensemble_change_mover(self):
assert_equal(self.mover.is_ensemble_change_mover, True)
def test_is_canonical(self):
assert_equal(self.mover.is_canonical, True)
def test_setup_sanity(self):
# sanity checks to make sure that what we set up makes sense
assert_equal(self.minus_sample.ensemble(self.minus_sample.trajectory),
True)
first_subtraj = FirstSubtrajectorySelectMover(
ensemble=self.minus,
sub_ensemble=self.minus._segment_ensemble
)
change = first_subtraj.move(SampleSet(self.minus_sample))
samples = change.results
assert_equal(samples[0].ensemble(samples[0].trajectory), True)
final_subtraj = FinalSubtrajectorySelectMover(
ensemble=self.minus,
sub_ensemble=self.minus._segment_ensemble
)
change = final_subtraj.move(SampleSet(self.minus_sample))
samples = change.results
assert_equal(samples[0].ensemble(samples[0].trajectory), True)
assert_equal(samples[0].ensemble, self.minus._segment_ensemble)
def test_successful_move(self):
init_innermost = make_1d_traj(self.list_innermost, [1.0]*5)
init_sample = Sample(
replica=0,
trajectory=init_innermost,
ensemble=self.innermost
)
gs = SampleSet([init_sample, self.minus_sample])
extend_forward = self.list_innermost + [0.12, 0.32, -0.131]
extend_backward = [-0.13, 0.13, 0.33] + self.list_innermost
assert_equal(self.minus(make_1d_traj(extend_forward)), True)
assert_equal(self.minus(make_1d_traj(extend_backward)), True)
seg_dir = {}
for i in range(100):
change = self.mover.move(gs)
samples = change.results
sub_samples = change.subchange.subchange.results
assert_equal(len(samples), 2)
assert_equal(len(sub_samples), 4)
s_inner = [s for s in sub_samples if s.ensemble==self.innermost]
s_minus = [s for s in sub_samples if s.ensemble==self.minus]
s_sub = [s for s in sub_samples if s.ensemble==self.minus._segment_ensemble]
assert_equal(len(s_inner), 1)
assert_equal(len(s_minus), 1)
assert_equal(len(s_sub), 2)
for c in change:
assert_equal(c.accepted, True)
assert_equal(change.canonical.mover, self.mover)
key = ""
s_inner0_xvals = [s.coordinates[0,0] for s in s_inner[0].trajectory]
if items_equal(s_inner0_xvals, self.first_segment):
key += "1"
elif items_equal(s_inner0_xvals, self.second_segment):
key += "2"
else:
print "s_inner0_xvals:", s_inner0_xvals
raise RuntimeError("Chosen segment neither first nor last!")
# final sample s_minus is accepted
s_minus_xvals = [s.coordinates[0,0] for s in s_minus[-1].trajectory]
if items_equal(s_minus_xvals, extend_forward):
key += "f"
elif items_equal(s_minus_xvals, extend_backward):
key += "b"
else:
print "s_minus_xvals:", s_minus_xvals
raise RuntimeError("Unexpected minus extension result!")
try:
seg_dir[key] += 1
except KeyError:
seg_dir[key] = 1
assert_equal(len(seg_dir.keys()), 4)
def test_repex_fails_other_ensemble(self):
innermost_other_ensemble = make_1d_traj([-0.11, 0.1, -0.12])
samp_other_ensemble = Sample(
replica=0,
trajectory=innermost_other_ensemble,
ensemble=self.innermost
)
gs = SampleSet([samp_other_ensemble, self.minus_sample])
change = self.mover.move(gs)
assert_equal(len(change.trials), 1)
sub = change.subchange.subchange
assert_equal(self.innermost(innermost_other_ensemble), False)
assert_equal(sub[0].accepted, True)
assert_equal(sub[1].accepted, False)
assert_equal(len(sub.trials), 3) # stop after failed repex
# only one sample which is not a segment
def test_repex_fails_innermost_crosses_state(self):
innermost_crosses_to_state = make_1d_traj([-0.11, 0.5, 1.8])
samp_crosses_to_state = Sample(
replica=0,
trajectory=innermost_crosses_to_state,
ensemble=self.innermost
)
gs = SampleSet([samp_crosses_to_state, self.minus_sample])
change = self.mover.move(gs)
assert_equal(len(change.trials), 1) # stop after failed repex
sub = change.subchange.subchange
assert_equal(self.innermost(innermost_crosses_to_state), True)
assert_equal(len(sub.trials), 3) # stop after failed repex
assert_subchanges_set_accepted(sub, [True, False, False])
def test_repex_fails_minus_crosses_to_state(self):
minus_crosses_to_state = make_1d_traj(
[-0.11, 0.5, 1.8, 0.6, -0.12, 0.7, 1.7, 0.4, -0.13]
)
badminus_sample = Sample(
replica=-1,
trajectory=minus_crosses_to_state,
ensemble=self.minus
)
init_sample = Sample(
replica=0,
trajectory=make_1d_traj(self.list_innermost, [1.0]*5),
ensemble=self.innermost
)
gs = SampleSet([badminus_sample, init_sample])
assert_equal(self.minus(minus_crosses_to_state), True)
change = self.mover.move(gs)
sub = change.subchange.subchange
assert_equal(len(sub.trials), 3) # stop after failed repex
assert_equal(len(change.trials), 1)
assert_subchanges_set_accepted(sub, [True, False, False])
def test_extension_fails(self):
innermost_bad_extension = [-0.25, 0.1, 0.5, 0.1, -0.25]
traj_bad_extension = make_1d_traj(innermost_bad_extension, [1.0]*5)
samp_bad_extension = Sample(
replica=0,
trajectory=traj_bad_extension,
ensemble=self.innermost
)
assert_equal(self.innermost(traj_bad_extension), True)
gs = SampleSet([self.minus_sample, samp_bad_extension])
change = self.mover.move(gs)
assert_equal(change.accepted, False) # whole minus has failed
sub = change.subchange.subchange
assert_equal(len(sub.trials), 4)
# after filtering there are only 2 trials
assert_equal(len(change.trials), 2)
assert_subchanges_set_accepted(sub, [True] * 2 + [False])
# first two work and the extention fails
# this only happens due to length
assert_equal(
len(sub[-1][0].trials[0].trajectory),
len(traj_bad_extension)+self.dyn.n_frames_max-1
)
class testMinusMover(object):
def setup(self):
op = CV_Function("myid", fcn=lambda snap :
snap.coordinates[0][0])
volA = CVRangeVolume(op, -100, 0.0)
volB = CVRangeVolume(op, 1.0, 100)
volX = CVRangeVolume(op, -100, 0.25)
self.dyn = CalvinistDynamics([
# successful move: (backward extension then forward)
-0.13, 0.13, 0.33, -0.11, -0.12, 0.12, 0.32, -0.131,
# never leaves state:
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.25,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
-0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15, -0.15,
# goes to other state:
1.16, 1.26, 1.16, -0.16, 1.16, 1.26, 1.16
])
SampleGeneratingMover.engine = self.dyn
self.dyn.initialized = True
self.innermost = paths.TISEnsemble(volA, volB, volX)
self.minus = paths.MinusInterfaceEnsemble(volA, volX)
self.mover = MinusMover(
minus_ensemble=self.minus,
innermost_ensembles=self.innermost
)
self.first_segment = [-0.1, 0.1, 0.3, 0.1, -0.15]
self.list_innermost = [-0.11, 0.11, 0.31, 0.11, -0.12]
self.second_segment = [-0.25, 0.2, 0.4, 0.2, -0.2]
init_minus = make_1d_traj(
coordinates=self.first_segment + [-0.35] + self.second_segment,
velocities=[1.0]*11
)
self.minus_sample = Sample(
replica=-1,
trajectory=init_minus,
ensemble=self.minus
)
def test_is_ensemble_change_mover(self):
assert_equal(self.mover.is_ensemble_change_mover, True)
def test_is_canonical(self):
assert_equal(self.mover.is_canonical, True)
def test_setup_sanity(self):
# sanity checks to make sure that what we set up makes sense
assert_equal(self.minus_sample.ensemble(self.minus_sample.trajectory),
True)
first_subtraj = FirstSubtrajectorySelectMover(
ensemble=self.minus,
sub_ensemble=self.minus._segment_ensemble
)
change = first_subtraj.move(SampleSet(self.minus_sample))
samples = change.results
assert_equal(samples[0].ensemble(samples[0].trajectory), True)
final_subtraj = FinalSubtrajectorySelectMover(
ensemble=self.minus,
sub_ensemble=self.minus._segment_ensemble
)
change = final_subtraj.move(SampleSet(self.minus_sample))
samples = change.results
assert_equal(samples[0].ensemble(samples[0].trajectory), True)
assert_equal(samples[0].ensemble, self.minus._segment_ensemble)
def test_successful_move(self):
init_innermost = make_1d_traj(self.list_innermost, [1.0]*5)
init_sample = Sample(
replica=0,
trajectory=init_innermost,
ensemble=self.innermost
)
gs = SampleSet([init_sample, self.minus_sample])
extend_forward = self.list_innermost + [0.12, 0.32, -0.131]
extend_backward = [-0.13, 0.13, 0.33] + self.list_innermost
assert_equal(self.minus(make_1d_traj(extend_forward)), True)
assert_equal(self.minus(make_1d_traj(extend_backward)), True)
seg_dir = {}
for i in range(100):
change = self.mover.move(gs)
samples = change.results
sub_samples = change.subchange.subchange.results
assert_equal(len(samples), 2)
assert_equal(len(sub_samples), 4)
s_inner = [s for s in sub_samples if s.ensemble==self.innermost]
s_minus = [s for s in sub_samples if s.ensemble==self.minus]
s_sub = [s for s in sub_samples if s.ensemble==self.minus._segment_ensemble]
assert_equal(len(s_inner), 1)
assert_equal(len(s_minus), 1)
assert_equal(len(s_sub), 2)
for c in change:
assert_equal(c.accepted, True)
assert_equal(change.canonical.mover, self.mover)
key = ""
s_inner0_xvals = [s.coordinates[0,0] for s in s_inner[0].trajectory]
if items_equal(s_inner0_xvals, self.first_segment):
key += "1"
elif items_equal(s_inner0_xvals, self.second_segment):
key += "2"
else:
print "s_inner0_xvals:", s_inner0_xvals
raise RuntimeError("Chosen segment neither first nor last!")
# final sample s_minus is accepted
s_minus_xvals = [s.coordinates[0,0] for s in s_minus[-1].trajectory]
if items_equal(s_minus_xvals, extend_forward):
key += "f"
elif items_equal(s_minus_xvals, extend_backward):
key += "b"
else:
print "s_minus_xvals:", s_minus_xvals
raise RuntimeError("Unexpected minus extension result!")
try:
seg_dir[key] += 1
except KeyError:
seg_dir[key] = 1
assert_equal(len(seg_dir.keys()), 4)
def test_repex_fails_other_ensemble(self):
innermost_other_ensemble = make_1d_traj([-0.11, 0.1, -0.12])
samp_other_ensemble = Sample(
replica=0,
trajectory=innermost_other_ensemble,
ensemble=self.innermost
)
gs = SampleSet([samp_other_ensemble, self.minus_sample])
change = self.mover.move(gs)
assert_equal(len(change.trials), 1)
sub = change.subchange.subchange
assert_equal(self.innermost(innermost_other_ensemble), False)
assert_equal(sub[0].accepted, True)
assert_equal(sub[1].accepted, False)
assert_equal(len(sub.trials), 3) # stop after failed repex
# only one sample which is not a segment
def test_repex_fails_innermost_crosses_state(self):
innermost_crosses_to_state = make_1d_traj([-0.11, 0.5, 1.8])
samp_crosses_to_state = Sample(
replica=0,
trajectory=innermost_crosses_to_state,
ensemble=self.innermost
)
gs = SampleSet([samp_crosses_to_state, self.minus_sample])
change = self.mover.move(gs)
assert_equal(len(change.trials), 1) # stop after failed repex
sub = change.subchange.subchange
assert_equal(self.innermost(innermost_crosses_to_state), True)
assert_equal(len(sub.trials), 3) # stop after failed repex
assert_subchanges_set_accepted(sub, [True, False, False])
def test_repex_fails_minus_crosses_to_state(self):
minus_crosses_to_state = make_1d_traj(
[-0.11, 0.5, 1.8, 0.6, -0.12, 0.7, 1.7, 0.4, -0.13]
)
badminus_sample = Sample(
replica=-1,
trajectory=minus_crosses_to_state,
ensemble=self.minus
)
init_sample = Sample(
replica=0,
trajectory=make_1d_traj(self.list_innermost, [1.0]*5),
ensemble=self.innermost
)
gs = SampleSet([badminus_sample, init_sample])
assert_equal(self.minus(minus_crosses_to_state), True)
change = self.mover.move(gs)
sub = change.subchange.subchange
assert_equal(len(sub.trials), 3) # stop after failed repex
assert_equal(len(change.trials), 1)
assert_subchanges_set_accepted(sub, [True, False, False])
def test_extension_fails(self):
innermost_bad_extension = [-0.25, 0.1, 0.5, 0.1, -0.25]
traj_bad_extension = make_1d_traj(innermost_bad_extension, [1.0]*5)
samp_bad_extension = Sample(
replica=0,
trajectory=traj_bad_extension,
ensemble=self.innermost
)
assert_equal(self.innermost(traj_bad_extension), True)
gs = SampleSet([self.minus_sample, samp_bad_extension])
change = self.mover.move(gs)
assert_equal(change.accepted, False) # whole minus has failed
sub = change.subchange.subchange
assert_equal(len(sub.trials), 4)
# after filtering there are only 2 trials
assert_equal(len(change.trials), 2)
assert_subchanges_set_accepted(sub, [True] * 2 + [False])
# first two work and the extention fails
# this only happens due to length
assert_equal(
len(sub[-1][0].trials[0].trajectory),
len(traj_bad_extension)+self.dyn.n_frames_max-1
)
def test_sanity_insane(self):
traj0A = self.s0A.trajectory
ensB = self.s2B.ensemble
bad_samp = Sample(replica=0, trajectory=traj0A, ensemble=ensB)
testset = SampleSet([bad_samp])
testset.sanity_check()
