def get_ergodic_set(dtrajs, tau):
m = MaximumLikelihoodMSM(lag=tau, connectivity='largest', reversible=True)
m.fit(dtrajs)
erg_dtrajs = [x for x in m.dtrajs_active if not -1 in x]
m = MaximumLikelihoodMSM(lag=tau, connectivity='largest', reversible=True)
m.fit(erg_dtrajs)
assert m.active_count_fraction == 1.0, 'Active count fraction not 1'
return erg_dtrajs
def test_score_cv(self):
self._score_cv(MaximumLikelihoodMSM(lag=10, reversible=True))
self._score_cv(MaximumLikelihoodMSM(lag=10, reversible=True, statdist_constraint=self.statdist))
self._score_cv(MaximumLikelihoodMSM(lag=10, reversible=False))
self._score_cv(MaximumLikelihoodMSM(lag=10, reversible=True, sparse=True))
self._score_cv(MaximumLikelihoodMSM(lag=10, reversible=True, statdist_constraint=self.statdist, sparse=True))
self._score_cv(MaximumLikelihoodMSM(lag=10, reversible=False, sparse=True))
def setUpClass(cls):
P = np.array([[0.5, .25, .25, 0.],
[0., .25, .5, .25],
[.25, .25, .5, 0],
[.25, .25, .25, .25],
])
# bogus its object
lags = [1, 2, 3, 5, 10]
dtraj = generate_traj(P, 1000)
estimator = MaximumLikelihoodMSM(dt_traj='10 ps')
cls.its = ImpliedTimescales(estimator=estimator)
cls.its.estimate(dtraj, lags=lags)
cls.refs = cls.its.timescales[-1]
return cls
def setUpClass(cls):
P = np.array([
[0.5, .25, .25, 0.],
[0., .25, .5, .25],
[.25, .25, .5, 0],
[.25, .25, .25, .25],
])
# bogus its object
lags = [1, 2, 3, 5, 10]
cls.dtraj = MarkovStateModel(P).simulate(1000)
cls.estimator = MaximumLikelihoodMSM(dt_traj='10 ps')
cls.its = ImpliedTimescales(estimator=cls.estimator, n_jobs=1)
cls.its.estimate(cls.dtraj, lags=lags)
cls.refs = cls.its.timescales[-1]
return cls
from pyemma.msm import MaximumLikelihoodMSM
from pyemma.coordinates.util import DtrajReshape
from msmbuilder.cluster import NDGrid
from sklearn.pipeline import Pipeline
import pickle
from glob import glob
import numpy as np
# traj_paths = glob('data/000.5pc/*.npy')
# X = [np.load(traj_path) for traj_path in traj_paths]
xmin, xmax = -1.2, 1.2
tau = 1
model = Pipeline([('cluster',NDGrid(min=xmin, max=xmax, n_bins_per_feature=200)),
('msm', MaximumLikelihoodMSM(lag=1, score_method='vamp1'))])
pickle.dump(model, open('model_lag1.pickl', 'wb'))
results = {
'tau': [],
'k': [],
'bic': [],
'aic': [],
'icl': [],
'entropy': [],
'n_obs': [],
'dofs': []
}
print('tau = ', tau)
for k in [ks[-1]]:
print('\tk = ', k)
m = MaximumLikelihoodMSM(lag=tau,
connectivity='largest',
reversible=True)
m.fit(erg_dtrajs)
assert m.active_count_fraction == 1.0, 'Active count fraction not 1.0'
print('\tFitting HMM')
hmm = estimate_hidden_markov_model(dtrajs=erg_dtrajs,
nstates=int(k),
lag=tau,
stationary=False,
reversible=True,
connectivity='largest')
#hmm = MaximumLikelihoodHMSM(nstates=int(k), lag=tau, stationary=False, reversible=True, connectivity='largest', msm_init=m)
#hmm.fit(erg_dtrajs[:2])