def read_dtrajs_from_pattern(patterns, logger=getLogger()):
"""
Parameters
----------
patterns : single pattern or list of patterns
eg. '*.txt' or ['/foo/*/bar/*.txt', '*.txt']
Returns
-------
list of discrete trajectories : list of numpy arrays, dtype=int
"""
dtrajs = []
filenames = paths_from_patterns(patterns)
if filenames == []:
raise ValueError('no match to given pattern')
for dt in filenames:
# skip directories
if os.path.isdir(dt):
continue
logger.info('reading discrete trajectory: %s' % dt)
try:
dtrajs.append(read_discrete_trajectory(dt))
except Exception as e:
logger.error(
'Exception occurred during reading of %s:\n%s' % (dt, e))
raise
return dtrajs
def __init__(self, chain, chunksize=100, param_stride=1):
r"""Data processing pipeline.
Parameters
----------
chain : list of transformers like objects
the order in the list defines the direction of data flow.
chunksize : int, optional
how many frames shall be processed at once.
param_stride : int, optional
omit every n'th data point
"""
self._chain = []
self.chunksize = chunksize
self.param_stride = param_stride
self.chunksize = chunksize
# add given elements in chain
for e in chain:
self.add_element(e)
self._parametrized = False
name = "%s[%s]" % (self.__class__.__name__, hex(id(self)))
self._logger = getLogger(name)
def __init__(self, topfile):
self.topologyfile = topfile
self.topology = (mdtraj.load(topfile)).topology
self.active_features = []
self._dim = 0
self._logger = getLogger("%s[%s]" %
(self.__class__.__name__, hex(id(self))))
def __init__(self, topfile):
self.topologyfile = topfile
self.topology = (mdtraj.load(topfile)).topology
self.active_features = []
self._dim = 0
self._logger = getLogger("%s[%s]" %
(self.__class__.__name__, hex(id(self))))
def setUpClass(cls):
cls.logger = getLogger(cls.__class__.__name__)
d = np.arange(3 * 100).reshape((100, 3))
d2 = np.arange(300, 900).reshape((200, 3))
d_1d = np.random.random(100)
cls.dir = tempfile.mkdtemp(prefix='pyemma_npyreader')
cls.f1 = tempfile.mktemp(suffix='.npy', dir=cls.dir)
cls.f2 = tempfile.mktemp(suffix='.npy', dir=cls.dir)
cls.f3 = tempfile.mktemp(suffix='.npz', dir=cls.dir)
cls.f4 = tempfile.mktemp(suffix='.npy', dir=cls.dir)
# 2d
np.save(cls.f1, d)
np.save(cls.f4, d2)
# 1d
np.save(cls.f2, d_1d)
np.savez(cls.f3, d, d)
cls.files2d = [cls.f1, cls.f4] #cls.f3]
cls.files1d = [cls.f2]
cls.d = d
cls.d_1d = d_1d
cls.npy_files = [f for f in cls.files2d if f.endswith('.npy')]
cls.npz = cls.f3
return cls
def setUpClass(cls):
cls.logger = getLogger(cls.__class__.__name__)
d = np.arange(3 * 100).reshape((100, 3))
d_1d = np.random.random(100)
cls.dir = tempfile.mkdtemp(prefix='pyemma_npyreader')
cls.f1 = tempfile.mktemp(suffix='.npy', dir=cls.dir)
cls.f2 = tempfile.mktemp(suffix='.npy', dir=cls.dir)
cls.f3 = tempfile.mktemp(suffix='.npz', dir=cls.dir)
# 2d
np.save(cls.f1, d)
# 1d
np.save(cls.f2, d_1d)
np.savez(cls.f3, d, d)
cls.files2d = [cls.f1, cls.f3]
cls.files1d = [cls.f2]
cls.d = d
cls.d_1d = d_1d
cls.npy_files = [f for f in cls.files2d if f.endswith('.npy')]
cls.npz = cls.f3
return cls
def __init__(self, chain, chunksize=100, param_stride=1):
"""
TODO:chunksize should be estimated from memory requirements (max memory usage)
"""
self._chain = []
self.chunksize = chunksize
self.param_stride = param_stride
# add given elements in chain
for e in chain:
self.add_element(e)
self._parametrized = False
name = "%s[%s]" % (self.__class__.__name__, hex(id(self)))
self._logger = getLogger(name)
def __init__(self, chain, chunksize=100, param_stride=1):
"""
TODO:chunksize should be estimated from memory requirements (max memory usage)
"""
self._chain = []
self.chunksize = chunksize
self.param_stride = param_stride
# add given elements in chain
for e in chain:
self.add_element(e)
self._parametrized = False
name = "%s[%s]" % (self.__class__.__name__, hex(id(self)))
self._logger = getLogger(name)
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
from __future__ import absolute_import
import mdtraj as md
import numpy as np
from pyemma.util.log import getLogger
from pyemma.coordinates.data.util.reader_utils import copy_traj_attributes as _copy_traj_attributes, \
preallocate_empty_trajectory as _preallocate_empty_trajectory, enforce_top as _enforce_top
from mdtraj.core.trajectory import Trajectory
__all__ = ['frames_from_file']
log = getLogger(__name__)
def frames_from_files(files, top, frames, chunksize=1000, stride=1, verbose=False, copy_not_join=None):
from pyemma.coordinates import source
# Enforce topology to be a md.Topology object
top = _enforce_top(top)
reader = source(files, top=top)
stride = int(stride)
if stride != 1:
frames[:, 1] *= int(stride)
if verbose:
log.info('A stride value of = %u was parsed, '
'interpreting "indexes" accordingly.' % stride)
import unittest
import os
import numpy as np
import tempfile
from pyemma.util.log import getLogger
import pyemma.coordinates as coor
import pyemma.util.types as types
logger = getLogger('TestCluster')
class TestCluster(unittest.TestCase):
@classmethod
def setUpClass(cls):
super(TestCluster, cls).setUpClass()
cls.dtraj_dir = tempfile.mkdtemp()
# generate Gaussian mixture
means = [np.array([-3,0]),
np.array([-1,1]),
np.array([0,0]),
np.array([1,-1]),
np.array([4,2])]
widths = [np.array([0.1,0.1]),
np.array([0.1,0.1]),
np.array([0.1,0.1]),
np.array([0.1,0.1]),
np.array([0.1,0.1])]
#
# You should have received a copy of the GNU Lesser General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
from __future__ import absolute_import
import unittest
import os
import numpy as np
from pyemma.coordinates.data import MDFeaturizer
from pyemma.util.log import getLogger
import pyemma.coordinates.api as api
import pyemma.util.types as types
import pkg_resources
logger = getLogger('pyemma.' + 'TestReaderUtils')
class TestSource(unittest.TestCase):
def setUp(self):
path = pkg_resources.resource_filename('pyemma.coordinates.tests',
'data') + os.path.sep
self.pdb_file = os.path.join(path, 'bpti_ca.pdb')
self.traj_files = [
os.path.join(path, 'bpti_001-033.xtc'),
os.path.join(path, 'bpti_067-100.xtc')
]
def tearDown(self):
pass
'''
from __future__ import print_function
from __future__ import absolute_import
import unittest
import os
import tempfile
import numpy as np
import mdtraj
from pyemma.coordinates import api
from pyemma.coordinates.data.feature_reader import FeatureReader
from pyemma.util.log import getLogger
from six.moves import range
log = getLogger('pyemma.' + 'TestFeatureReaderAndTICA')
class TestFeatureReaderAndTICA(unittest.TestCase):
@classmethod
def setUpClass(cls):
cls.dim = 9 # dimension (must be divisible by 3)
N = 50000 # length of single trajectory # 500000
N_trajs = 10 # number of trajectories
cls.w = 2.0 * np.pi * 1000.0 / N # have 1000 cycles in each trajectory
# get random amplitudes and phases
cls.A = np.random.randn(cls.dim)
cls.phi = np.random.random_sample((cls.dim, )) * np.pi * 2.0
mean = np.random.randn(cls.dim)
__author__ = 'noe'
from pyemma.util.log import getLogger
import numpy as np
log = getLogger('Transformer')
__all__ = ['Transformer']
class Transformer(object):
"""
Parameters
----------
chunksize : int (optional)
the chunksize used to batch process underlying data
lag : int (optional)
if you want to process time lagged data, set this to a value > 0.
"""
def __init__(self, chunksize=100, lag=0):
self.chunksize = chunksize
self._lag = lag
self._in_memory = False
self._dataproducer = None
@property
def data_producer(self):
#
# You should have received a copy of the GNU Lesser General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
from __future__ import absolute_import
import unittest
import os
import numpy as np
import tempfile
from pyemma.util.log import getLogger
import pyemma.coordinates as coor
import pyemma.util.types as types
from six.moves import range
logger = getLogger('TestCluster')
class TestCluster(unittest.TestCase):
@classmethod
def setUpClass(cls):
super(TestCluster, cls).setUpClass()
cls.dtraj_dir = tempfile.mkdtemp()
# generate Gaussian mixture
means = [
np.array([-3, 0]),
np.array([-1, 1]),
np.array([0, 0]),
np.array([1, -1]),
np.array([4, 2])
'''
Test feature reader and Tica by checking the properties of the ICs.
cov(ic_i,ic_j) = delta_ij and cov(ic_i,ic_j,tau) = lambda_i delta_ij
@author: Fabian Paul
'''
import unittest
import os
import tempfile
import numpy as np
import mdtraj
from pyemma.coordinates.api import tica, _TICA as TICA
from pyemma.coordinates.data.feature_reader import FeatureReader
from pyemma.util.log import getLogger
log = getLogger('TestFeatureReaderAndTICAProjection')
def random_invertible(n, eps=0.01):
'generate real random invertible matrix'
m = np.random.randn(n, n)
u, s, v = np.linalg.svd(m)
s = np.maximum(s, eps)
return u.dot(np.diag(s)).dot(v)
from nose.plugins.attrib import attr
@attr(slow=True)
class TestFeatureReaderAndTICAProjection(unittest.TestCase):
@classmethod
def setUpClass(cls):
c = super(TestFeatureReaderAndTICAProjection, cls).setUpClass()
'''
Created on 18.10.2013
@author: marscher
'''
import numpy as np
from decimal import Decimal # for wrapping java.math.BigDecimal
from pyemma.util.log import getLogger
from pyemma.util.pystallone import ndarray_to_stallone_array, stallone, JArray, JInt, JDouble
__all__ = ['PathwayDecomposition']
log = getLogger()
# TODO: test
class PathwayDecomposition(object):
def __init__(self, F, Q, A, B):
"""
Parameters
----------
F : The net fluxes matrix
ndarray(dtype=float, shape=(n,n))
Q : The committor vector
ndarray(dtype=float, shape=(n)
A : set of representatives (indices defining set A in F)
ndarray(dtype=int)
B : set of representatives
ndarray(dtype=int)
@author: marscher
'''
import mdtraj
import os
import tempfile
import unittest
from pyemma.coordinates import api
from pyemma.coordinates.data.feature_reader import FeatureReader
from pyemma.util.log import getLogger
import pkg_resources
import numpy as np
from pyemma.coordinates.api import feature_reader, discretizer, tica
log = getLogger('TestFeatureReader')
class TestFeatureReader(unittest.TestCase):
@classmethod
def setUpClass(cls):
c = super(TestFeatureReader, cls).setUpClass()
# create a fake trajectory which has 3 atoms and coordinates are just a range
# over all frames.
cls.trajfile = tempfile.mktemp('.xtc')
cls.n_frames = 1000
cls.xyz = np.random.random(cls.n_frames * 3 * 3).reshape((cls.n_frames, 3, 3))
log.debug("shape traj: %s" % str(cls.xyz.shape))
cls.topfile = pkg_resources.resource_filename(
'pyemma.coordinates.tests.test_featurereader', 'data/test.pdb')
'''
Created on 04.02.2015
@author: marscher
'''
import unittest
from pyemma.coordinates.data.data_in_memory import DataInMemory
from pyemma.util.log import getLogger
import numpy as np
import tempfile
import os
logger = getLogger('TestDataInMemory')
class TestDataInMemory(unittest.TestCase):
@classmethod
def setUpClass(cls):
d = np.random.random((100, 3))
d_1d = np.random.random(100)
f1 = tempfile.mktemp()
f2 = tempfile.mktemp(suffix='.npy')
f3 = tempfile.mktemp()
f4 = tempfile.mktemp(suffix='.npy')
npz = tempfile.mktemp(suffix='.npz')
'''
Created on 02.02.2015
@author: marscher
'''
import unittest
import numpy as np
from pyemma.coordinates import pca
from pyemma.util.log import getLogger
import pyemma.util.types as types
logger = getLogger('TestTICA')
class TestPCAExtensive(unittest.TestCase):
@classmethod
def setUpClass(cls):
import pyemma.msm.generation as msmgen
# generate HMM with two Gaussians
cls.P = np.array([[0.99, 0.01], [0.01, 0.99]])
cls.T = 10000
means = [np.array([-1, 1]), np.array([1, -1])]
widths = [np.array([0.3, 2]), np.array([0.3, 2])]
# continuous trajectory
cls.X = np.zeros((cls.T, 2))
# hidden trajectory
dtraj = msmgen.generate_traj(cls.P, cls.T)
for t in range(cls.T):
from pyemma.coordinates.data.featurizer import MDFeaturizer as _MDFeaturizer
from pyemma.coordinates.data.feature_reader import FeatureReader as _FeatureReader
from pyemma.coordinates.data.data_in_memory import DataInMemory as _DataInMemory
from pyemma.coordinates.data.util.reader_utils import create_file_reader as _create_file_reader
from pyemma.coordinates.data.frames_from_file import frames_from_file as _frames_from_file
# transforms
from pyemma.coordinates.transform.transformer import Transformer as _Transformer
from pyemma.coordinates.transform.pca import PCA as _PCA
from pyemma.coordinates.transform.tica import TICA as _TICA
# clustering
from pyemma.coordinates.clustering.kmeans import KmeansClustering as _KmeansClustering
from pyemma.coordinates.clustering.uniform_time import UniformTimeClustering as _UniformTimeClustering
from pyemma.coordinates.clustering.regspace import RegularSpaceClustering as _RegularSpaceClustering
from pyemma.coordinates.clustering.assign import AssignCenters as _AssignCenters
logger = getLogger('coordinates.api')
__author__ = "Frank Noe, Martin Scherer"
__copyright__ = "Copyright 2015, Computational Molecular Biology Group, FU-Berlin"
__credits__ = ["Benjamin Trendelkamp-Schroer", "Martin Scherer", "Frank Noe"]
__license__ = "FreeBSD"
__version__ = "2.0.0"
__maintainer__ = "Martin Scherer"
__email__ = "m.scherer AT fu-berlin DOT de"
__all__ = ['featurizer', # IO
'load',
'source',
'pipeline',
'discretizer',
'save_traj',
'''
Created on 18.02.2015
@author: marscher
'''
from pyemma.coordinates.transform.transformer import Transformer
from pyemma.util.log import getLogger
import numpy as np
log = getLogger('Clustering')
class AbstractClustering(Transformer):
"""
provides a common interface for cluster algorithms.
"""
def __init__(self):
super(AbstractClustering, self).__init__()
self.clustercenters = None
self.dtrajs = []
def map(self, x):
"""get closest index of point in :attr:`clustercenters` to x."""
d = self.data_producer.distances(x, self.clustercenters)
return np.argmin(d)
def save_dtrajs(self,
trajfiles=None,
prefix='',
output_format='ascii',
extension='.dtraj'):
'''
Created on 22.01.2015
@author: marscher
'''
from pyemma.util.log import getLogger
import numpy as np
from pyemma.coordinates.transform.transformer import Transformer
log = getLogger('WriterCSV')
__all__ = ['WriterCSV']
class WriterCSV(Transformer):
'''
shall write to csv files
'''
def __init__(self, filename):
'''
Constructor
'''
super(WriterCSV, self).__init__()
# filename should be obtained from source trajectory filename,
# eg suffix it to given filename
self.filename = filename
self.last_frame = False
'''
Created on 02.02.2015
@author: marscher
'''
import unittest
import numpy as np
from pyemma.coordinates import pca
from pyemma.util.log import getLogger
import pyemma.util.types as types
logger = getLogger('TestTICA')
class TestPCAExtensive(unittest.TestCase):
@classmethod
def setUpClass(cls):
import pyemma.msm.generation as msmgen
# generate HMM with two Gaussians
cls.P = np.array([[0.99, 0.01],
[0.01, 0.99]])
cls.T = 10000
means = [np.array([-1,1]), np.array([1,-1])]
widths = [np.array([0.3,2]),np.array([0.3,2])]
# continuous trajectory
cls.X = np.zeros((cls.T, 2))
'''
Created on 19.01.2015
@author: marscher
'''
from .transformer import Transformer
from pyemma.util.linalg import eig_corr
from pyemma.util.log import getLogger
from pyemma.util.annotators import doc_inherit
import numpy as np
log = getLogger('TICA')
__all__ = ['TICA']
class TICA(Transformer):
r"""
Time-lagged independent component analysis (TICA)
Given a sequence of multivariate data :math:`X_t`, computes the mean-free
covariance and time-lagged covariance matrix:
.. math::
C_0 &= (X_t - \mu)^T (X_t - \mu) \\
C_{\tau} &= (X_t - \mu)^T (X_t+\tau - \mu)
and solves the eigenvalue problem
.. math:: C_{\tau} r_i = C_0 \lambda_i r_i
'''
from __future__ import absolute_import
import mdtraj
import tempfile
import unittest
from pyemma.coordinates import api
from pyemma.coordinates.data.feature_reader import FeatureReader
from pyemma.util.log import getLogger
import pkg_resources
import numpy as np
from pyemma.coordinates.api import discretizer, tica, source
from six.moves import range
log = getLogger('TestFeatureReader')
def create_traj(top, format='.xtc', dir=None):
trajfile = tempfile.mktemp(suffix=format, dir=dir)
n_frames = np.random.randint(500, 1500)
log.debug("create traj with %i frames" % n_frames)
xyz = np.arange(n_frames * 3 * 3).reshape((n_frames, 3, 3))
t = mdtraj.load(top)
t.xyz = xyz
t.unitcell_vectors = np.array(n_frames*[[0,0,1], [0,1,0], [1,0,0]]).reshape(n_frames, 3,3)
t.time = np.arange(n_frames)
t.save(trajfile)
return trajfile, xyz, n_frames
#!/usr/bin/env python
# encoding: utf-8
"""
"""
import argparse
import sys
import os
from pyemma.util.log import getLogger
from pyemma.msm.generation import generate_traj
import pyemma
log = getLogger('mm_generate')
def handleArgs():
parser = argparse.ArgumentParser()
parser.add_argument('-T', type=str, help='path to transition matrix.')
parser.add_argument('-o',
'--output',
dest='output',
required=True,
help='output filename of trajectory.')
parser.add_argument('-dt', type=int, default=1)
parser.add_argument('-steps', type=int)
parser.add_argument('-start_state', type=int)
args = parser.parse_args()
return args
from __future__ import absolute_import
import pyemma
from six.moves import range
from six.moves import zip
'''
Created on 04.02.2015
@author: marscher
'''
import unittest
import numpy as np
from pyemma.coordinates.data.data_in_memory import DataInMemory
from pyemma.util.log import getLogger
logger = getLogger('pyemma.' + 'TestDataInMemory')
class TestDataInMemory(unittest.TestCase):
@classmethod
def setUpClass(cls):
d = np.random.random((100, 3))
d_1d = np.random.random(100)
cls.d = d
cls.d_1d = d_1d
return cls
def test_skip(self):
for skip in [0, 3, 13]:
r1 = DataInMemory(self.d)
from __future__ import absolute_import
import unittest
import os
import tempfile
from pyemma.coordinates.data import MDFeaturizer
from pyemma.util.log import getLogger
import pyemma.coordinates.api as api
import numpy as np
from pyemma.coordinates.data.numpy_filereader import NumPyFileReader
from pyemma.coordinates.data.py_csv_reader import PyCSVReader as CSVReader
import shutil
logger = getLogger('TestReaderUtils')
class TestApiSourceFileReader(unittest.TestCase):
@classmethod
def setUpClass(cls):
data_np = np.random.random((100, 3))
data_raw = np.arange(300 * 4).reshape(300, 4)
cls.dir = tempfile.mkdtemp("test-api-src")
cls.npy = tempfile.mktemp(suffix='.npy', dir=cls.dir)
cls.npz = tempfile.mktemp(suffix='.npz', dir=cls.dir)
cls.dat = tempfile.mktemp(suffix='.dat', dir=cls.dir)
def create_logger(self):
# creates a logger based on the the attribe "name" of self
self._logger_instance = getLogger(self.name)
r = weakref.ref(self, _cleanup_logger(self))
_refs[self.name] = r
return self._logger_instance
def __init__(self, counts, mu=None, reversible=False, Tinit=None):
"""
Sets the count matrix used for sampling. Assumes that the prior
(if desired) is included.
Parameters
----------
counts : ndarray (n, n)
the posterior count matrix
mu : ndarray (n)
optional stationary distribution, if given, the sampled transition matrix
will have this this stat dist.
reversible : boolean
should sample a reversible transition matrix.
Tinit : ndarray(n, n)
optional start point for sampling algorithm.
Example
-------
>>> C = np.array([[5, 2], [1,10]])
>>> sampler = ITransitionMatrixSampler(C)
>>> T = sampler.sample(10**6)
>>> print T
"""
if issparse(counts):
counts = counts.toarray()
# the interface in stallone takes counts as doubles
counts = counts.astype(np.float64)
try:
C = ndarray_to_stallone_array(counts)
jpackage = stallone.mc.sampling
# convert types to java
if Tinit is not None:
Tinit = ndarray_to_stallone_array(Tinit)
if mu is not None:
mu = ndarray_to_stallone_array(mu)
if reversible:
if mu: # fixed pi
if Tinit:
self.sampler = jpackage.TransitionMatrixSamplerRevFixPi(
C, Tinit, mu)
else:
self.sampler = jpackage.TransitionMatrixSamplerRevFixPi(
C, mu)
else: # sample reversible matrix, with arbitrary pi
if Tinit:
self.sampler = jpackage.TransitionMatrixSamplerRev(
C, Tinit)
else:
self.sampler = jpackage.TransitionMatrixSamplerRev(C)
else: # sample non rev
if Tinit:
self.sampler = jpackage.TransitionMatrixSamplerNonrev(
C, Tinit)
else:
self.sampler = jpackage.TransitionMatrixSamplerNonrev(C)
except JavaException as je:
log = getLogger()
log.exception("Error during creation of tmatrix sampling wrapper:"
" stack\n%s" % je.stacktrace())
raise
#
# You should have received a copy of the GNU Lesser General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
from __future__ import absolute_import
import os
import unittest
from pyemma.util.files import TemporaryDirectory
from pyemma.util.log import getLogger
from six.moves import range
import numpy as np
import pyemma.coordinates as coor
import pyemma.util.types as types
logger = getLogger('pyemma.' + 'TestCluster')
class TestClusterAssign(unittest.TestCase):
@classmethod
def setUpClass(cls):
super(TestClusterAssign, cls).setUpClass()
# generate Gaussian mixture
means = [
np.array([-3, 0]),
np.array([-1, 1]),
np.array([0, 0]),
np.array([1, -1]),
np.array([4, 2])
]
from __future__ import absolute_import
import unittest
import os
import pkg_resources
import numpy as np
from pyemma.coordinates import api
from pyemma.coordinates.data.data_in_memory import DataInMemory
from pyemma.coordinates import source, tica
from pyemma.util.contexts import numpy_random_seed
from pyemma.util.log import getLogger
import pyemma.util.types as types
from six.moves import range
logger = getLogger('pyemma.' + 'TestTICA')
def mycorrcoef(X, Y, lag):
X = X.astype(np.float64)
Y = Y.astype(np.float64)
mean_X = 0.5 * (np.mean(X[lag:], axis=0) + np.mean(X[0:-lag], axis=0))
mean_Y = 0.5 * (np.mean(Y[lag:], axis=0) + np.mean(Y[0:-lag], axis=0))
cov = (
(X[0:-lag] - mean_X).T.dot(Y[0:-lag] - mean_Y) +
(X[lag:] - mean_X).T.dot(Y[lag:] - mean_Y)) / (2 *
(X.shape[0] - lag) - 1)
autocov_X = (
(X[0:-lag] - mean_X).T.dot(X[0:-lag] - mean_X) +
(X[lag:] - mean_X).T.dot(X[lag:] - mean_X)) / (2 *
import unittest
import os
import tempfile
import numpy as np
from pyemma.util.log import getLogger
import pyemma.coordinates as coor
import pyemma.util.types as types
logger = getLogger('TestReaderUtils')
class TestCluster(unittest.TestCase):
@classmethod
def setUpClass(cls):
super(TestCluster, cls).setUpClass()
cls.dtraj_dir = tempfile.mkdtemp()
# generate Gaussian mixture
means = [np.array([-3,0]),
np.array([-1,1]),
np.array([0,0]),
np.array([1,-1]),
np.array([4,2])]
widths = [np.array([0.3,2]),
np.array([0.3,2]),
np.array([0.3,2]),
np.array([0.3,2]),
#!/usr/bin/env python
'''
Created on 17.02.2014
@author: marscher
'''
import argparse
import sys
from pyemma.msm.estimation.api import count_matrix, largest_connected_set
from pyemma.msm.io.api import write_matrix
from pyemma.util.files import read_dtrajs_from_pattern
from pyemma.util.log import getLogger
log = getLogger()
def handleArgs():
parser = argparse.ArgumentParser()
parser.add_argument('-i',
dest='discTraj',
required=True,
nargs='+',
help='list of discrete trajectories')
parser.add_argument('-o',
dest='output',
help='output filename of largest connected set')
parser.add_argument(
'-lag',
dest='lag',
help='lag time for which connectivity should be calculated for',
def __create_logger(self):
name = "%s[%s]" % (self.__class__.__name__, hex(id(self)))
self._logger = getLogger(name)
__author__ = 'noe'
import numpy as np
from .transformer import Transformer
from pyemma.util.log import getLogger
from pyemma.util.annotators import doc_inherit
log = getLogger('PCA')
__all__ = ['PCA']
class PCA(Transformer):
r"""Principal component analysis.
Given a sequence of multivariate data :math:`X_t`,
computes the mean-free covariance matrix.
.. math:: C = (X - \mu)^T (X - \mu)
and solves the eigenvalue problem
.. math:: C r_i = \sigma_i r_i,
where :math:`r_i` are the principal components and :math:`\sigma_i` are
their respective variances.
When used as a dimension reduction method, the input data is projected onto
the dominant principal components.
Parameters
@author: marscher
'''
import numpy as np
import warnings
from mdtraj.utils.validation import cast_indices
from mdtraj.core.trajectory import load, Trajectory, _parse_topology
from mdtraj.formats.hdf5 import HDF5TrajectoryFile
from mdtraj.utils.unit import in_units_of
from mdtraj.formats.lh5 import LH5TrajectoryFile
from mdtraj.formats import DCDTrajectoryFile
from mdtraj.formats import XTCTrajectoryFile
from pyemma.util.log import getLogger
log = getLogger('patches')
def iterload(filename, chunk=100, **kwargs):
"""An iterator over a trajectory from one or more files on disk, in fragments
This may be more memory efficient than loading an entire trajectory at
once
Parameters
----------
filename : str
Path to the trajectory file on disk
chunk : int
Number of frames to load at once from disk per iteration. If 0, load all.
__author__ = 'noe'
from pyemma.util.log import getLogger
from pyemma.coordinates.clustering.interface import AbstractClustering
import numpy as np
log = getLogger('UniformTimeClustering')
__all__ = ['UniformTimeClustering']
class UniformTimeClustering(AbstractClustering):
"""
Uniform time clustering
Parameters
----------
k : int
"""
def __init__(self, k=2):
super(UniformTimeClustering, self).__init__()
self.k = k
def describe(self):
return "[Uniform time clustering, k = %i]" % self.k
def dimension(self):
return 1
def get_memory_per_frame(self):
"""
import pyemma
from six.moves import range
from six.moves import zip
'''
Created on 04.02.2015
@author: marscher
'''
import unittest
import numpy as np
from pyemma.coordinates.data.data_in_memory import DataInMemory
from pyemma.coordinates.transform.transformer import TransformerIteratorContext
from pyemma.util.log import getLogger
logger = getLogger('TestDataInMemory')
class TestDataInMemory(unittest.TestCase):
@classmethod
def setUpClass(cls):
d = np.random.random((100, 3))
d_1d = np.random.random(100)
cls.d = d
cls.d_1d = d_1d
return cls
def testWrongArguments(self):
with self.assertRaises(ValueError):
reader = DataInMemory("foo")
'''
Test feature reader and Tica by checking the properties of the ICs.
cov(ic_i,ic_j) = delta_ij and cov(ic_i,ic_j,tau) = lambda_i delta_ij
@author: Fabian Paul
'''
import unittest
import os
import tempfile
import numpy as np
import mdtraj
from pyemma.coordinates.api import tica, _TICA as TICA
from pyemma.coordinates.data.feature_reader import FeatureReader
from pyemma.util.log import getLogger
log = getLogger('TestFeatureReaderAndTICAProjection')
def random_invertible(n, eps=0.01):
'generate real random invertible matrix'
m = np.random.randn(n, n)
u, s, v = np.linalg.svd(m)
s = np.maximum(s, eps)
return u.dot(np.diag(s)).dot(v)
from nose.plugins.attrib import attr
@attr(slow=True)
class TestFeatureReaderAndTICAProjection(unittest.TestCase):
@classmethod
'''
Created on 22.01.2015
@author: marscher
'''
import numpy as np
from sklearn.cluster import MiniBatchKMeans
from pyemma.util.log import getLogger
from pyemma.util.annotators import doc_inherit
from pyemma.coordinates.clustering.interface import AbstractClustering
log = getLogger('KmeansClustering')
__all__ = ['KmeansClustering']
class KmeansClustering(AbstractClustering):
r"""
Kmeans clustering
Parameters
----------
n_clusters : int
amount of cluster centers
max_iter : int
how many iterations per chunk?
"""
def __init__(self, n_clusters, max_iter=1000):
super(KmeansClustering, self).__init__()
def tmatrix_sampler(C, reversible=False, mu=None, T0=None):
r"""Generate transition matrix sampler object.
Parameters
----------
C : (M, M) ndarray or scipy.sparse matrix
Count matrix
reversible : bool
If true sample from the ensemble of transition matrices
restricted to those obeying a detailed balance condition,
else draw from the whole ensemble of stochastic matrices.
mu : array_like
The stationary distribution of the transition matrix samples.
T0 : ndarray, shape=(n, n) or scipy.sparse matrix
Starting point of the MC chain of the sampling algorithm.
Has to obey the required constraints.
Returns
-------
sampler : A :py:class:dense.ITransitionMatrixSampler object.
Notes
-----
The transition matrix sampler generates transition matrices from
the posterior distribution. The posterior distribution is given as
a product of Dirichlet distributions
.. math:: \mathbb{P}(T|C) \propto \prod_{i=1}^{M} \left( \prod_{j=1}^{M} p_{ij}^{c_{ij}} \right)
The method can generate samples from the posterior under the follwing two constraints
**Reversible sampling**
Using a MCMC sampler outlined in .. [1] it is ensured that samples
from the posterior are reversible, i.e. there is a probability
vector :math:`(\mu_i)` such that :math:`\mu_i t_{ij} = \mu_j
t_{ji}` holds for all :math:`i,j`.
**Reversible sampling with fixed stationary vector**
Using a MCMC sampler outlined in .. [2] it is ensured that samples
from the posterior fulfill detailed balance with respect to a given
probability vector :math:`(\mu_i)`.
References
----------
.. [1] Noe, F. 2008. Probability distributions of molecular observables
computed from Markov state models. J Chem Phys 128: 244103.
.. [2] Trendelkamp-Schroer, B and F Noe. 2013. Efficient Bayesian estimation
of Markov model transition matrices with given stationary distribution.
J Chem Phys 138: 164113.
"""
if issparse(C):
_showSparseConversionWarning()
C=C.toarray()
from pyemma.util.pystallone import JavaException
try:
return ITransitionMatrixSampler(C, mu, reversible, Tinit=T0)
except JavaException as je:
log = getLogger()
log.exception("Error during tmatrix sampling")
raise
'''
Test feature reader and Tica with a set of cosine time series.
@author: Fabian Paul
'''
import unittest
import os
import tempfile
import numpy as np
import mdtraj
from pyemma.coordinates import api
from pyemma.coordinates.data.feature_reader import FeatureReader
from pyemma.util.log import getLogger
log = getLogger('TestFeatureReaderAndTICA')
from nose.plugins.attrib import attr
@attr(slow=True)
class TestFeatureReaderAndTICA(unittest.TestCase):
@classmethod
def setUpClass(cls):
c = super(TestFeatureReaderAndTICA, cls).setUpClass()
cls.dim = 99 # dimension (must be divisible by 3)
N = 50000 # length of single trajectory # 500000
N_trajs = 10 # number of trajectories
cls.w = 2.0*np.pi*1000.0/N # have 1000 cycles in each trajectory
# get random amplitudes and phases
