Source code for MDAnalysis.coordinates.XDR
# -*- Mode: python; tab-width: 4; indent-tabs-mode:nil; coding:utf-8 -*-
# vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
#
# MDAnalysis --- https://www.mdanalysis.org
# Copyright (c) 2006-2017 The MDAnalysis Development Team and contributors
# (see the file AUTHORS for the full list of names)
#
# Released under the GNU Public Licence, v2 or any higher version
#
# Please cite your use of MDAnalysis in published work:
#
# R. J. Gowers, M. Linke, J. Barnoud, T. J. E. Reddy, M. N. Melo, S. L. Seyler,
# D. L. Dotson, J. Domanski, S. Buchoux, I. M. Kenney, and O. Beckstein.
# MDAnalysis: A Python package for the rapid analysis of molecular dynamics
# simulations. In S. Benthall and S. Rostrup editors, Proceedings of the 15th
# Python in Science Conference, pages 102-109, Austin, TX, 2016. SciPy.
# doi: 10.25080/majora-629e541a-00e
#
# N. Michaud-Agrawal, E. J. Denning, T. B. Woolf, and O. Beckstein.
# MDAnalysis: A Toolkit for the Analysis of Molecular Dynamics Simulations.
# J. Comput. Chem. 32 (2011), 2319--2327, doi:10.1002/jcc.21787
#
"""\
XDR based trajectory files --- :mod:`MDAnalysis.coordinates.XDR`
================================================================
This module contains helper function and classes to read the XTC and TRR file
formats.
See Also
--------
MDAnalysis.coordinates.XTC: Read and write GROMACS XTC trajectory files.
MDAnalysis.coordinates.TRR: Read and write GROMACS TRR trajectory files.
MDAnalysis.lib.formats.libmdaxdr: Low level xdr format reader
"""
import errno
import numpy as np
from os.path import getctime, getsize, isfile, split, join
import warnings
import fasteners
from . import base
from ..lib.mdamath import triclinic_box
from ..lib.util import store_init_arguments
[docs]def offsets_filename(filename, ending='npz'):
"""Return offset or its lock filename for XDR files.
For this the filename is appended
with `_offsets.{ending}`.
Parameters
----------
filename : str
filename of trajectory
ending : str (optional)
fileending of offsets file
Returns
-------
offset_filename : str
"""
head, tail = split(filename)
return join(head, f'.{tail}_offsets.{ending}')
[docs]def read_numpy_offsets(filename):
"""read offsets into dictionary.
This assume offsets have been saved using numpy
Parameters
----------
filename : str
filename of offsets
Returns
-------
offsets : dict
dictionary of offsets information
"""
try:
return {k: v for k, v in np.load(filename).items()}
# `ValueError` is encountered when the offset file is corrupted.
except (ValueError, IOError):
warnings.warn(f"Failed to load offsets file {filename}\n")
return False
[docs]class XDRBaseReader(base.ReaderBase):
"""Base class for libmdaxdr file formats xtc and trr
This class handles integration of XDR based formats into MDAnalysis. The
XTC and TRR classes only implement `_write_next_frame` and
`_frame_to_ts`.
.. _offsets-label:
Notes
-----
XDR based readers store persistent offsets on disk. The offsets are used to
enable access to random frames efficiently. These offsets will be generated
automatically the first time the trajectory is opened. Generally offsets
are stored in hidden `*_offsets.npz` files. Afterwards opening the same
file again is fast. It sometimes can happen that the stored offsets get out
off sync with the trajectory they refer to. For this the offsets also store
the number of atoms, size of the file and last modification time. If any of
them change the offsets are recalculated. Writing of the offset file can
fail when the directory where the trajectory file resides is not writable
or if the disk is full. In this case a warning message will be shown but
the offsets will nevertheless be used during the lifetime of the trajectory
Reader. However, the next time the trajectory is opened, the offsets will
have to be rebuilt again.
.. versionchanged:: 1.0.0
XDR offsets read from trajectory if offsets file read-in fails
.. versionchanged:: 2.0.0
Add a InterProcessLock when generating offsets
.. versionchanged:: 2.4.0
Use a direct read into ts attributes
"""
@store_init_arguments
def __init__(self, filename, convert_units=True, sub=None,
refresh_offsets=False, **kwargs):
"""
Parameters
----------
filename : str
trajectory filename
convert_units : bool (optional)
convert units to MDAnalysis units
sub : array_like (optional)
`sub` is an array of indices to pick out the corresponding
coordinates and load only them; this requires that the topology
itself is that of the sub system.
refresh_offsets : bool (optional)
force refresh of offsets
**kwargs : dict
General reader arguments.
"""
super(XDRBaseReader, self).__init__(filename,
convert_units=convert_units,
**kwargs)
self._xdr = self._file(self.filename)
self._sub = sub
if self._sub is not None:
self.n_atoms = len(self._sub)
else:
self.n_atoms = self._xdr.n_atoms
if not refresh_offsets:
self._load_offsets()
else:
self._read_offsets(store=True)
frame = self._xdr.read()
try:
xdr_frame = self._xdr.read()
dt = xdr_frame.time - frame.time
self._xdr.seek(1)
except StopIteration:
dt = 0
self.ts = self._Timestep(self.n_atoms, **self._ts_kwargs)
self._frame = 0
self._frame_to_ts(frame, self.ts)
# these should only be initialized once
self.ts.dt = dt
self.ts.dimensions = triclinic_box(*frame.box)
if self.convert_units:
if self.ts.dimensions is not None:
self.convert_pos_from_native(self.ts.dimensions[:3])
[docs] @classmethod
def parse_n_atoms(cls, filename, **kwargs):
with cls._file(filename) as f:
n_atoms = f.n_atoms
return n_atoms
def _load_offsets(self):
"""load frame offsets from file, reread them from the trajectory if that
fails. To prevent the competition of generating the same offset file
from multiple processes, an `InterProcessLock` is used."""
fname = offsets_filename(self.filename)
lock_name = offsets_filename(self.filename,
ending='lock')
# check if the location of the lock is writable.
try:
with fasteners.InterProcessLock(lock_name) as filelock:
pass
except OSError as e:
if isinstance(e, PermissionError) or e.errno == errno.EROFS:
warnings.warn(f"Cannot write lock/offset file in same location as "
f"{self.filename}. Using slow offset calculation.")
self._read_offsets(store=True)
return
else:
raise
with fasteners.InterProcessLock(lock_name) as filelock:
if not isfile(fname):
self._read_offsets(store=True)
return
# if offsets file read correctly, data will be a dictionary of offsets
# if not, data will be False
# if False, offsets should be read from the trajectory
# this warning can be avoided by loading Universe like:
# u = mda.Universe(data.TPR, data.XTC, refresh_offsets=True)
# refer to Issue #1893
data = read_numpy_offsets(fname)
if not data:
warnings.warn(f"Reading offsets from {fname} failed, "
"reading offsets from trajectory instead.\n"
"Consider setting 'refresh_offsets=True' "
"when loading your Universe.")
self._read_offsets(store=True)
return
ctime_ok = size_ok = n_atoms_ok = False
try:
ctime_ok = getctime(self.filename) == data['ctime']
size_ok = getsize(self.filename) == data['size']
n_atoms_ok = self._xdr.n_atoms == data['n_atoms']
except KeyError:
# we tripped over some old offset formated file
pass
if not (ctime_ok and size_ok and n_atoms_ok):
warnings.warn("Reload offsets from trajectory\n "
"ctime or size or n_atoms did not match")
self._read_offsets(store=True)
else:
self._xdr.set_offsets(data['offsets'])
def _read_offsets(self, store=False):
"""read frame offsets from trajectory"""
fname = offsets_filename(self.filename)
offsets = self._xdr.offsets
if store:
ctime = getctime(self.filename)
size = getsize(self.filename)
try:
np.savez(fname,
offsets=offsets, size=size, ctime=ctime,
n_atoms=self._xdr.n_atoms)
except Exception as e:
warnings.warn(f"Couldn't save offsets because: {e}")
@property
def n_frames(self):
"""number of frames in trajectory"""
return len(self._xdr)
def _reopen(self):
"""reopen trajectory"""
self.ts.frame = 0
self._frame = -1
offsets = self._xdr.offsets.copy()
self._xdr.close()
self._xdr.open(self.filename.encode('utf-8'), 'r')
# only restore in case we actually had offsets
if len(offsets) != 0:
self._xdr.set_offsets(offsets)
def _read_frame(self, i):
"""read frame i"""
self._frame = i - 1
try:
self._xdr.seek(i)
timestep = self._read_next_timestep()
except IOError:
warnings.warn('seek failed, recalculating offsets and retrying')
offsets = self._xdr.calc_offsets()
self._xdr.set_offsets(offsets)
self._read_offsets(store=True)
self._xdr.seek(i)
timestep = self._read_next_timestep()
return timestep
[docs] def Writer(self, filename, n_atoms=None, **kwargs):
"""Return writer for trajectory format"""
if n_atoms is None:
n_atoms = self.n_atoms
return self._writer(filename, n_atoms=n_atoms, **kwargs)
[docs]class XDRBaseWriter(base.WriterBase):
"""Base class for libmdaxdr file formats xtc and trr"""
def __init__(self, filename, n_atoms, convert_units=True, **kwargs):
"""
Parameters
----------
filename : str
filename of trajectory
n_atoms : int
number of atoms to be written
convert_units : bool (optional)
convert from MDAnalysis units to format specific units
**kwargs : dict
General writer arguments
"""
self.filename = filename
self._convert_units = convert_units
self.n_atoms = n_atoms
self._xdr = self._file(self.filename, 'w')
def __del__(self):
self.close()