File: trajectory.py

package info (click to toggle)
python-ase 3.24.0-1
  • links: PTS, VCS
  • area: main
  • in suites: trixie
  • size: 15,448 kB
  • sloc: python: 144,945; xml: 2,728; makefile: 113; javascript: 47
file content (506 lines) | stat: -rw-r--r-- 17,086 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
"""Trajectory"""
import contextlib
import io
import warnings
from typing import Tuple

import numpy as np

from ase import __version__
from ase.atoms import Atoms
from ase.calculators.calculator import PropertyNotImplementedError
from ase.calculators.singlepoint import SinglePointCalculator, all_properties
from ase.io.formats import is_compressed
from ase.io.jsonio import decode, encode
from ase.io.pickletrajectory import PickleTrajectory
from ase.parallel import world
from ase.utils import tokenize_version

__all__ = ['Trajectory', 'PickleTrajectory']


def Trajectory(filename, mode='r', atoms=None, properties=None, master=None,
               comm=world):
    """A Trajectory can be created in read, write or append mode.

    Parameters:

    filename: str
        The name of the file.  Traditionally ends in .traj.
    mode: str
        The mode.  'r' is read mode, the file should already exist, and
        no atoms argument should be specified.
        'w' is write mode.  The atoms argument specifies the Atoms
        object to be written to the file, if not given it must instead
        be given as an argument to the write() method.
        'a' is append mode.  It acts as write mode, except that
        data is appended to a preexisting file.
    atoms: Atoms object
        The Atoms object to be written in write or append mode.
    properties: list of str
        If specified, these calculator properties are saved in the
        trajectory.  If not specified, all supported quantities are
        saved.  Possible values: energy, forces, stress, dipole,
        charges, magmom and magmoms.
    master: bool
        Controls which process does the actual writing. The
        default is that process number 0 does this.  If this
        argument is given, processes where it is True will write.
    comm: Communicator object
        Communicator to handle parallel file reading and writing.

    The atoms, properties and master arguments are ignored in read mode.
    """
    if mode == 'r':
        return TrajectoryReader(filename)
    return TrajectoryWriter(filename, mode, atoms, properties, master=master,
                            comm=comm)


class TrajectoryWriter:
    """Writes Atoms objects to a .traj file."""

    def __init__(self, filename, mode='w', atoms=None, properties=None,
                 master=None, comm=world):
        """A Trajectory writer, in write or append mode.

        Parameters:

        filename: str
            The name of the file.  Traditionally ends in .traj.
        mode: str
            The mode.  'r' is read mode, the file should already exist, and
            no atoms argument should be specified.
            'w' is write mode.  The atoms argument specifies the Atoms
            object to be written to the file, if not given it must instead
            be given as an argument to the write() method.
            'a' is append mode.  It acts as write mode, except that
            data is appended to a preexisting file.
        atoms: Atoms object
            The Atoms object to be written in write or append mode.
        properties: list of str
            If specified, these calculator properties are saved in the
            trajectory.  If not specified, all supported quantities are
            saved.  Possible values: energy, forces, stress, dipole,
            charges, magmom and magmoms.
        master: bool
            Controls which process does the actual writing. The
            default is that process number 0 does this.  If this
            argument is given, processes where it is True will write.
        comm: MPI communicator
            MPI communicator for this trajectory writer, by default world.
            Passing a different communicator facilitates writing of
            different trajectories on different MPI ranks.
        """
        if master is None:
            master = comm.rank == 0

        self.filename = filename
        self.mode = mode
        self.atoms = atoms
        self.properties = properties
        self.master = master
        self.comm = comm

        self.description = {}
        self.header_data = None
        self.multiple_headers = False

        self._open(filename, mode)

    def __enter__(self):
        return self

    def __exit__(self, exc_type, exc_value, tb):
        self.close()

    def set_description(self, description):
        self.description.update(description)

    def _open(self, filename, mode):
        import ase.io.ulm as ulm
        if mode not in 'aw':
            raise ValueError('mode must be "w" or "a".')
        if self.master:
            self.backend = ulm.open(filename, mode, tag='ASE-Trajectory')
            if len(self.backend) > 0 and mode == 'a':
                with Trajectory(filename) as traj:
                    atoms = traj[0]
                self.header_data = get_header_data(atoms)
        else:
            self.backend = ulm.DummyWriter()

    def write(self, atoms=None, **kwargs):
        """Write the atoms to the file.

        If the atoms argument is not given, the atoms object specified
        when creating the trajectory object is used.

        Use keyword arguments to add extra properties::

            writer.write(atoms, energy=117, dipole=[0, 0, 1.0])
        """
        if atoms is None:
            atoms = self.atoms

        for image in atoms.iterimages():
            self._write_atoms(image, **kwargs)

    def _write_atoms(self, atoms, **kwargs):
        b = self.backend

        if self.header_data is None:
            b.write(version=1, ase_version=__version__)
            if self.description:
                b.write(description=self.description)
            # Atomic numbers and periodic boundary conditions are written
            # in the header in the beginning.
            #
            # If an image later on has other numbers/pbc, we write a new
            # header.  All subsequent images will then have their own header
            # whether or not their numbers/pbc change.
            self.header_data = get_header_data(atoms)
            write_header = True
        else:
            if not self.multiple_headers:
                header_data = get_header_data(atoms)
                self.multiple_headers = not headers_equal(self.header_data,
                                                          header_data)
            write_header = self.multiple_headers

        write_atoms(b, atoms, write_header=write_header)

        calc = atoms.calc

        if calc is None and len(kwargs) > 0:
            calc = SinglePointCalculator(atoms)

        if calc is not None:
            if not hasattr(calc, 'get_property'):
                calc = OldCalculatorWrapper(calc)
            c = b.child('calculator')
            c.write(name=calc.name)
            if hasattr(calc, 'todict'):
                c.write(parameters=calc.todict())
            for prop in all_properties:
                if prop in kwargs:
                    x = kwargs[prop]
                else:
                    if self.properties is not None:
                        if prop in self.properties:
                            x = calc.get_property(prop, atoms)
                        else:
                            x = None
                    else:
                        try:
                            x = calc.get_property(prop, atoms,
                                                  allow_calculation=False)
                        except (PropertyNotImplementedError, KeyError):
                            # KeyError is needed for Jacapo.
                            # XXX We can perhaps remove this.
                            x = None
                if x is not None:
                    if prop in ['stress', 'dipole']:
                        x = x.tolist()
                    c.write(prop, x)

        info = {}
        for key, value in atoms.info.items():
            try:
                encode(value)
            except TypeError:
                warnings.warn(f'Skipping "{key}" info.')
            else:
                info[key] = value
        if info:
            b.write(info=info)

        b.sync()

    def close(self):
        """Close the trajectory file."""
        self.backend.close()

    def __len__(self):
        return self.comm.sum_scalar(len(self.backend))


class TrajectoryReader:
    """Reads Atoms objects from a .traj file."""

    def __init__(self, filename):
        """A Trajectory in read mode.

        The filename traditionally ends in .traj.
        """
        self.filename = filename
        self.numbers = None
        self.pbc = None
        self.masses = None

        self._open(filename)

    def __enter__(self):
        return self

    def __exit__(self, exc_type, exc_value, tb):
        self.close()

    def _open(self, filename):
        import ase.io.ulm as ulm
        self.backend = ulm.open(filename, 'r')
        self._read_header()

    def _read_header(self):
        b = self.backend
        if b.get_tag() != 'ASE-Trajectory':
            raise OSError('This is not a trajectory file!')

        if len(b) > 0:
            self.pbc = b.pbc
            self.numbers = b.numbers
            self.masses = b.get('masses')
            self.constraints = b.get('constraints', '[]')
            self.description = b.get('description')
            self.version = b.version
            self.ase_version = b.get('ase_version')

    def close(self):
        """Close the trajectory file."""
        self.backend.close()

    def __getitem__(self, i=-1):
        if isinstance(i, slice):
            return SlicedTrajectory(self, i)
        b = self.backend[i]
        if 'numbers' in b:
            # numbers and other header info was written alongside the image:
            atoms = read_atoms(b, traj=self)
        else:
            # header info was not written because they are the same:
            atoms = read_atoms(b,
                               header=[self.pbc, self.numbers, self.masses,
                                       self.constraints],
                               traj=self)
        if 'calculator' in b:
            results = {}
            implemented_properties = []
            c = b.calculator
            for prop in all_properties:
                if prop in c:
                    results[prop] = c.get(prop)
                    implemented_properties.append(prop)
            calc = SinglePointCalculator(atoms, **results)
            calc.name = b.calculator.name
            calc.implemented_properties = implemented_properties

            if 'parameters' in c:
                calc.parameters.update(c.parameters)
            atoms.calc = calc

        return atoms

    def __len__(self):
        return len(self.backend)

    def __iter__(self):
        for i in range(len(self)):
            yield self[i]


class SlicedTrajectory:
    """Wrapper to return a slice from a trajectory without loading
    from disk. Initialize with a trajectory (in read mode) and the
    desired slice object."""

    def __init__(self, trajectory, sliced):
        self.trajectory = trajectory
        self.map = range(len(self.trajectory))[sliced]

    def __getitem__(self, i):
        if isinstance(i, slice):
            # Map directly to the original traj, not recursively.
            traj = SlicedTrajectory(self.trajectory, slice(0, None))
            traj.map = self.map[i]
            return traj
        return self.trajectory[self.map[i]]

    def __len__(self):
        return len(self.map)


def get_header_data(atoms):
    return {'pbc': atoms.pbc.copy(),
            'numbers': atoms.get_atomic_numbers(),
            'masses': atoms.get_masses() if atoms.has('masses') else None,
            'constraints': list(atoms.constraints)}


def headers_equal(headers1, headers2):
    assert len(headers1) == len(headers2)
    eq = True
    for key in headers1:
        eq &= np.array_equal(headers1[key], headers2[key])
    return eq


class VersionTooOldError(Exception):
    pass


def read_atoms(backend,
               header: Tuple = None,
               traj: TrajectoryReader = None,
               _try_except: bool = True) -> Atoms:
    from ase.constraints import dict2constraint

    if _try_except:
        try:
            return read_atoms(backend, header, traj, False)
        except Exception as ex:
            if (traj is not None and tokenize_version(__version__) <
                    tokenize_version(traj.ase_version)):
                msg = ('You are trying to read a trajectory file written '
                       f'by ASE-{traj.ase_version} from ASE-{__version__}. '
                       'It might help to update your ASE')
                raise VersionTooOldError(msg) from ex
            else:
                raise

    b = backend
    if header:
        pbc, numbers, masses, constraints = header
    else:
        pbc = b.pbc
        numbers = b.numbers
        masses = b.get('masses')
        constraints = b.get('constraints', '[]')

    atoms = Atoms(positions=b.positions,
                  numbers=numbers,
                  cell=b.cell,
                  masses=masses,
                  pbc=pbc,
                  info=b.get('info'),
                  constraint=[dict2constraint(d)
                              for d in decode(constraints)],
                  momenta=b.get('momenta'),
                  magmoms=b.get('magmoms'),
                  charges=b.get('charges'),
                  tags=b.get('tags'))
    return atoms


def write_atoms(backend, atoms, write_header=True):
    b = backend

    if write_header:
        b.write(pbc=atoms.pbc.tolist(),
                numbers=atoms.numbers)
        if atoms.constraints:
            if all(hasattr(c, 'todict') for c in atoms.constraints):
                b.write(constraints=encode(atoms.constraints))

        if atoms.has('masses'):
            b.write(masses=atoms.get_masses())

    b.write(positions=atoms.get_positions(),
            cell=atoms.get_cell().tolist())

    if atoms.has('tags'):
        b.write(tags=atoms.get_tags())
    if atoms.has('momenta'):
        b.write(momenta=atoms.get_momenta())
    if atoms.has('initial_magmoms'):
        b.write(magmoms=atoms.get_initial_magnetic_moments())
    if atoms.has('initial_charges'):
        b.write(charges=atoms.get_initial_charges())


def read_traj(fd, index):
    trj = TrajectoryReader(fd)
    for i in range(*index.indices(len(trj))):
        yield trj[i]


@contextlib.contextmanager
def defer_compression(fd):
    """Defer the file compression until all the configurations are read."""
    # We do this because the trajectory and compressed-file
    # internals do not play well together.
    # Be advised not to defer compression of very long trajectories
    # as they use a lot of memory.
    if is_compressed(fd):
        with io.BytesIO() as bytes_io:
            try:
                # write the uncompressed data to the buffer
                yield bytes_io
            finally:
                # write the buffered data to the compressed file
                bytes_io.seek(0)
                fd.write(bytes_io.read())
    else:
        yield fd


def write_traj(fd, images):
    """Write image(s) to trajectory."""
    if isinstance(images, Atoms):
        images = [images]
    with defer_compression(fd) as fd_uncompressed:
        trj = TrajectoryWriter(fd_uncompressed)
        for atoms in images:
            trj.write(atoms)


class OldCalculatorWrapper:
    def __init__(self, calc):
        self.calc = calc
        try:
            self.name = calc.name
        except AttributeError:
            self.name = calc.__class__.__name__.lower()

    def get_property(self, prop, atoms, allow_calculation=True):
        try:
            if (not allow_calculation and
                    self.calc.calculation_required(atoms, [prop])):
                return None
        except AttributeError:
            pass

        method = 'get_' + {'energy': 'potential_energy',
                           'magmom': 'magnetic_moment',
                           'magmoms': 'magnetic_moments',
                           'dipole': 'dipole_moment'}.get(prop, prop)
        try:
            result = getattr(self.calc, method)(atoms)
        except AttributeError:
            raise PropertyNotImplementedError
        return result


def convert(name):
    import os
    t = TrajectoryWriter(name + '.new')
    for atoms in PickleTrajectory(name, _warn=False):
        t.write(atoms)
    t.close()
    os.rename(name, name + '.old')
    os.rename(name + '.new', name)


def main():
    import optparse
    parser = optparse.OptionParser(usage='python -m ase.io.trajectory '
                                   'a1.traj [a2.traj ...]',
                                   description='Convert old trajectory '
                                   'file(s) to new format. '
                                   'The old file is kept as a1.traj.old.')
    _opts, args = parser.parse_args()
    for name in args:
        convert(name)


if __name__ == '__main__':
    main()