bids.py 6.62 KB
Newer Older
Paul McCarthy's avatar
Paul McCarthy committed
1
2
3
4
5
6
7
8
9
10
11
#!/usr/bin/env python
#
# bids.py - Simple BIDS metadata reader.
#
# Author: Paul McCarthy <pauldmccarthy@gmail.com>
#
"""This module provides a few functions for working with BIDS data sets.

.. autosummary::
   :nosignatures:

12
   BIDSFile
Paul McCarthy's avatar
Paul McCarthy committed
13
14
15
16
17
   isBIDSDir
   inBIDSDir
   isBIDSFile
   loadMetadata

18
19
20
All of the other functions in this module should not be considered part of the
public API.

Paul McCarthy's avatar
Paul McCarthy committed
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

.. note::  The `pybids <https://bids-standard.github.io/pybids/>`_ library is
           a more suitable choice if you are after a more robust and featured
           interface for working with BIDS datasets.
"""


import os.path   as op
import itertools as it
import              re
import              glob
import              json

import fsl.utils.memoize as memoize
import fsl.utils.path    as fslpath


class BIDSFile(object):
    """The ``BIDSFile`` class parses and stores the entities and suffix contained
    in a BIDS file. See the :func:`parseFilename` function.

    The :meth:`match` method can be used to compare two ``BIDSFile`` instances.

    The following attributes are available on a ``BIDSFile`` instance:

     - ``filename``: Absolute path to the file
     - ``entities``: Dict of ``key : value`` pairs, the entities that are
       present in the file name (e.g. ``{'sub' : '01}``)
     - ``suffix``: File suffix (e.g. ``T1w``, ``bold``, etc.)
    """


    def __init__(self, filename):
        """Create a ``BIDSFile``. """
        entities, suffix = parseFilename(filename)
        self.filename    = op.abspath(filename)
        self.entities    = entities
        self.suffix      = suffix


61
62
63
64
65
66
67
68
69
70
71
    def __str__(self):
        """Return a strimg representation of this ``BIDSFile``. """
        return 'BIDSFile({})'.format(self.filename)


    def __repr__(self):
        """Return a strimg representation of this ``BIDSFile``. """
        return str(self)


    def match(self, other, suffix=True):
Paul McCarthy's avatar
Paul McCarthy committed
72
73
        """Compare this ``BIDSFile`` to ``other``.

74
75
76
77
78
79
80
81
        :arg other:  ``BIDSFile`` to compare

        :arg suffix: Defaults to ``True``. If ``False``, the comparison
                     is made solely on the entity values.

        :returns:    ``True`` if ``self.suffix == other.suffix`` (unless
                     ``suffix`` is ``False``) and if all of the entities in
                     ``other`` are present in ``self``, ``False`` otherwise.
Paul McCarthy's avatar
Paul McCarthy committed
82
83
        """

84
        suffix   = (not suffix) or (self.suffix == other.suffix)
Paul McCarthy's avatar
Paul McCarthy committed
85
86
87
        entities = True

        for key, value in other.entities.items():
88
            entities = entities and (self.entities.get(key, None) == value)
Paul McCarthy's avatar
Paul McCarthy committed
89
90
91
92
93

        return suffix and entities


def parseFilename(filename):
94
95
96
97
    """Parses a BIDS-like file name. The file name must consist of zero or more
    "entities" (alpha-numeric ``name-value`` pairs), a "suffix", all separated
    by underscores, and a regular file extension. For example, the following
    file::
Paul McCarthy's avatar
Paul McCarthy committed
98
99
100

        sub-01_ses-01_task-stim_bold.nii.gz

101
102
103
104
105
    has suffix ``bold``, entities ``sub=01``, ``ses=01`` and ``task=stim``, and
    extension ``.nii.gz``.

    .. note:: This function assumes that no period (``.``) characters occur in
              the body of a BIDS filename.
Paul McCarthy's avatar
Paul McCarthy committed
106
107

    :returns: A tuple containing:
108
109
               - A dict containing the entities
               - The suffix
Paul McCarthy's avatar
Paul McCarthy committed
110
111
    """

112
113
114
115
    if not isBIDSFile(filename, strict=False):
        raise ValueError('Does not look like a BIDS '
                         'file: {}'.format(filename))

Paul McCarthy's avatar
Paul McCarthy committed
116
117
118
    suffix   = None
    entities = []
    filename = op.basename(filename)
119
    filename = fslpath.removeExt(filename, firstDot=True)
Paul McCarthy's avatar
Paul McCarthy committed
120
121
122
123
124
    parts    = filename.split('_')

    for part in parts[:-1]:
        entities.append(part.split('-'))

125
    suffix   = parts[-1]
Paul McCarthy's avatar
Paul McCarthy committed
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
    entities = dict(entities)

    return entities, suffix


def isBIDSDir(dirname):
    """Returns ``True`` if ``dirname`` is the root directory of a BIDS dataset.
    """
    return op.exists(op.join(dirname, 'dataset_description.json'))


def inBIDSDir(filename):
    """Returns ``True`` if ``filename`` looks like it is within a BIDS dataset
    directory, ``False`` otherwise.
    """

    dirname = op.abspath(op.dirname(filename))
    inBIDS  = False

    while True:

        if isBIDSDir(dirname):
            inBIDS = True
            break

        prevdir = dirname
        dirname = op.dirname(dirname)

        # at filesystem root
        if prevdir == dirname:
            break

    return inBIDS


161
def isBIDSFile(filename, strict=True):
Paul McCarthy's avatar
Paul McCarthy committed
162
    """Returns ``True`` if ``filename`` looks like a BIDS image or JSON file.
163
164
165
166

    :arg filename: Name of file to check
    :arg strict:   If ``True`` (the default), the file must be within a BIDS
                   dataset directory, as defined by :func:`inBIDSDir`.
Paul McCarthy's avatar
Paul McCarthy committed
167
168
    """

169
    name    = op.basename(filename)
170
    pattern = r'([a-z0-9]+-[a-z0-9]+_)*([a-z0-9])+\.(.+)'
171
    flags   = re.ASCII | re.IGNORECASE
172
    match   = re.fullmatch(pattern, name, flags) is not None
Paul McCarthy's avatar
Paul McCarthy committed
173

174
    return ((not strict) or inBIDSDir(filename)) and match
Paul McCarthy's avatar
Paul McCarthy committed
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


@memoize.memoize
def loadMetadataFile(filename):
    """Load ``filename`` (assumed to be JSON), returning its contents. """
    with open(filename, 'rt') as f:
        return json.load(f)


def loadMetadata(filename):
    """Load all of the metadata associated with ``filename``.

    :arg filename: Path to a data file in a BIDS dataset.
    :returns:      A dict containing all of the metadata associated with
                   ``filename``
    """

    filename  = op.realpath(op.abspath(filename))
    bfile     = BIDSFile(filename)
    dirname   = op.dirname(filename)
    prevdir   = filename
    metafiles = []
    metadata  = {}

    # Walk up the directory tree until
    # we hit the BIDS dataset root, or
    # the filesystem root
    while True:

        # Gather all json files in this
        # directory with matching entities
        # and suffix, sorted alphabetically
207
208
209
        # and reversed, so that earlier
        # ones take precedence
        files = reversed(sorted(glob.glob(op.join(dirname, '*.json'))))
Paul McCarthy's avatar
Paul McCarthy committed
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
        files = [BIDSFile(f) for f in files if isBIDSFile(f)]
        files = [f.filename  for f in files if bfile.match(f)]

        # build a list of all files
        metafiles.append(files)

        # move to the next dir up
        prevdir = dirname
        dirname = op.dirname(dirname)

        # stop when we hit the dataset or filesystem root
        if isBIDSDir(prevdir) or dirname == prevdir:
            break

    # Load in each json file, from
    # shallowest to deepest, so entries
    # in deeper files take precedence
    # over shallower ones.
    for f in it.chain(*reversed(metafiles)):

        # assuming here that every file contains a dict
        metadata.update(loadMetadataFile(f))

    return metadata