def test_fitghmm():
with tempdir():
RawPositionsFeaturizer().save('featurizer.pickl')
shell('hmsm fit-ghmm --featurizer featurizer.pickl --n-init 10 '
' --n-states 4 --dir %s --ext h5 --top %s' % (
DATADIR, os.path.join(DATADIR, 'Trajectory0.h5')))
shell('hmsm inspect -i hmms.jsonlines --details')
shell('hmsm sample-ghmm --no-match-vars -i hmms.jsonlines --lag-time 1 --n-state 4 '
'--featurizer featurizer.pickl --dir %s --ext h5 --top %s' % (
DATADIR, os.path.join(DATADIR, 'Trajectory0.h5')))
shell('hmsm means-ghmm -i hmms.jsonlines --lag-time 1 --n-state 4 '
'--featurizer featurizer.pickl --dir %s --ext h5 --top %s' % (
DATADIR, os.path.join(DATADIR, 'Trajectory0.h5')))
shell('hmsm structures means.csv --ext pdb --prefix means --top %s' % os.path.join(DATADIR, 'Trajectory0.h5'))
samples_csv = pd.read_csv('samples.csv', skiprows=1)
means_csv = pd.read_csv('means.csv', skiprows=1)
model = next(iterobjects('hmms.jsonlines'))
means_pdb = md.load(glob('means-*.pdb'))
means = np.array(sorted(model['means'], key=lambda e: e[0]))
print('true\n', HMM.means_)
print('learned\n', means)
eq(HMM.means_, means, decimal=0)
means_pdb_xyz = np.array(sorted(means_pdb.xyz.reshape(4, 3), key=lambda e: e[0]))
eq(means_pdb_xyz, np.array(sorted(model['means'], key=lambda e:e[0])), decimal=0)
def test_fitghmm():
with tempdir():
RawPositionsFeaturizer(n_features=3).save('featurizer.pickl')
shell('hmsm fit-ghmm --featurizer featurizer.pickl '
' --n-states 4 --dir %s --ext h5 --top %s' % (
DATADIR, os.path.join(DATADIR, 'Trajectory0.h5')))
shell('hmsm inspect -i hmms.jsonlines --details')
shell('hmsm sample-ghmm --no-match-vars -i hmms.jsonlines --lag-time 1 --n-state 4 '
'--featurizer featurizer.pickl --dir %s --ext h5 --top %s' % (
DATADIR, os.path.join(DATADIR, 'Trajectory0.h5')))
shell('hmsm means-ghmm -i hmms.jsonlines --lag-time 1 --n-state 4 '
'--featurizer featurizer.pickl --dir %s --ext h5 --top %s' % (
DATADIR, os.path.join(DATADIR, 'Trajectory0.h5')))
shell('hmsm structures means.csv --ext pdb --prefix means --top %s' % os.path.join(DATADIR, 'Trajectory0.h5'))
samples_csv = pd.read_csv('samples.csv', skiprows=1)
means_csv = pd.read_csv('means.csv', skiprows=1)
model = next(iterobjects('hmms.jsonlines'))
means_pdb = md.load(glob('means-*.pdb'))
means = np.array(sorted(model['means'], key=lambda e: e[0]))
eq(HMM.means_, means, decimal=0)
print(samples_csv)
print(means_csv)
means_pdb_xyz = means_pdb.xyz.reshape(4, 3)
eq(means_pdb_xyz, np.array(model['means']), decimal=0)
def __init__(self, args):
if os.path.exists(args.out):
self.error('IOError: file exists: %s' % args.out)
matches = [o for o in iterobjects(args.filename)
if o['n_states'] == args.n_states]
if len(matches) == 0:
self.error('No model with n_states=%d in %s.'
% (args.n_states, args.filename))
self.args = args
self.model_dict = matches[0]
self.out = args.out
self.topology = md.load(args.top)
self.filenames = glob.glob(
os.path.join(os.path.expanduser(args.dir), '*.%s' % args.ext))
self.featurizer = mixtape.featurizer.load(args.featurizer)
self.stride = stride
if len(self.filenames) == 0:
self.error('No files matched.')
def __init__(self, args):
if os.path.exists(args.out):
self.error('IOError: file exists: %s' % args.out)
matches = [o for o in iterobjects(args.filename)
if o['n_states'] == args.n_states
and o['train_lag_time'] == args.lag_time]
if len(matches) == 0:
self.error('No model with n_states=%d, train_lag_time=%d in %s.' % (
args.n_states, args.lag_time, args.filename))
self.args = args
self.model = matches[0]
self.out = args.out
self.topology = md.load(args.top)
self.filenames = glob.glob(os.path.join(os.path.expanduser(args.dir), '*.%s' % args.ext))
self.featurizer = mixtape.featurizer.load(args.featurizer)
self.match_vars = args.match_vars
if len(self.filenames) == 0:
self.error('No files matched.')
import numpy as np import pandas as pd import mdtraj as md from mixtape.utils import iterobjects timestep = 1.0 n_states = 2 input_filename = "HMM/4000pairs-lag10.jsonlines" models = list(iterobjects(input_filename)) models = [model for model in models if model["n_states"] == n_states] lagtimes = np.array([model["train_lag_time"] for model in models]) * timestep timescales = np.array([model["timescales"] for model in models]) * timestep plot(lagtimes, timescales[:,0], 'o')
import numpy as np import pandas as pd import mdtraj as md from mixtape.utils import iterobjects, assign import mixtape.ghmm, mixtape.featurizer import sklearn.hmm import os name = "tica" json_filename = "./%s.jsonlines" % name feature_filename = "./%s.pkl" % name featurizer = mixtape.featurizer.load(feature_filename) models = list(iterobjects(json_filename)) df = pd.DataFrame(models) x = df.ix[0] T = np.array(x["transmat"]) p = np.array(x["populations"]) n_states = len(p) model = mixtape.ghmm.GaussianFusionHMM(n_states, featurizer.n_features) model.means_ = x["means"] model.vars_ = x["vars"] model.transmat_ = x["transmat"] model.populations_ = x["populations"] means = model.means_
def __init__(self, args):
self.args = args
self.models = list(iterobjects(self.args.filename))
self.df = pd.DataFrame(self.models)
self.details = args.details
def start(self):
exclude = set(
[
"means",
"vars",
"kappas",
"train_logprobs",
"transmat",
"split",
"test_logprob",
"n_test_observations",
"fusion_prior",
"test_lag_time",
"cross_validation_fold",
"train_time",
"cross_validation_nfolds",
]
)
models = [
{k: v for k, v in list(m.items()) if k not in exclude} for f in self.args.filename for m in iterobjects(f)
]
explode_key_in_listofdicts("timescales", models)
explode_key_in_listofdicts("populations", models)
self.models = pd.DataFrame(models)
print("Columns:\n %s" % "\n ".join(wrap(", ".join(self.models.columns))))
print('Saving models as CSV file to "%s"' % self.args.out_csv)
self.models.to_csv(self.args.out_csv, index=False)
if self.args.x is not None and self.args.y is not None:
self.plot()
def __init__(self, args):
self.args = args
self.models = list(iterobjects(self.args.filename))
self.df = pd.DataFrame(self.models)
self.details = args.details
import shutil
import numpy as np
import pandas as pd
import mdtraj as md
from mixtape.utils import iterobjects
import mixtape.ghmm
import mixtape.featurizer
import os
name = "atomindices"
json_filename = "./%s.jsonlines" % name
feature_filename = "./%s.pkl" % name
models = list(iterobjects(json_filename))
df = pd.DataFrame(models)
x = df.ix[0]
T = np.array(x["transmat"])
p = np.array(x["populations"])
featurizer = mixtape.featurizer.load(feature_filename)
model = mixtape.ghmm.GaussianFusionHMM(3, featurizer.n_features)
model.means_ = x["means"]
model.vars_ = x["vars"]
model.transmat_ = x["transmat"]
model.populations_ = x["populations"]
trj0 = md.load("./system.subset.pdb")
atom_indices = np.loadtxt("./AtomIndices.dat", "int")