"""Generate publication-quality data acquisition methods section from BIDS dataset."""
import json
import os.path as op
from collections import Counter
from bids.reports import parsing, utils
[docs]
class BIDSReport:
"""Generate publication-quality data acquisition section from BIDS dataset.
Parameters
----------
layout : :obj:`bids.layout.BIDSLayout`
Layout object for a BIDS dataset.
config : :obj:`str` or :obj:`dict`, optional
Configuration info for methods generation. Can be a path to a file
(str), a dictionary, or None. If None, loads and uses default
configuration information.
Keys in the dictionary include:
'dir': a dictionary for converting encoding direction strings
(e.g., j-) to descriptions (e.g., anterior to
posterior)
'seq': a dictionary of sequence abbreviations (e.g., EP) and
corresponding names (e.g., echo planar)
'seqvar': a dictionary of sequence variant abbreviations
(e.g., SP) and corresponding names (e.g., spoiled)
Warning
-------
pybids' automatic report generation is experimental and currently under
active development, and as such should be used with caution.
Please remember to verify any generated report before putting it to use.
Additionally, only MRI datatypes (func, anat, fmap, and dwi) are currently
supported.
"""
[docs]
def __init__(self, layout, config=None):
self.layout = layout
if config is None:
config = op.join(
op.dirname(op.abspath(__file__)),
"config",
"converters.json",
)
if isinstance(config, str):
with open(config) as fobj:
config = json.load(fobj)
if not isinstance(config, dict):
raise ValueError(
"Input config must be None, dict, or path to "
"json file containing dict."
)
self.config = config
def generate_from_files(self, files):
r"""Generate a methods section from a list of files.
Parameters
----------
files : list of BIDSImageFile objects
List of files from which to generate methods description.
Returns
-------
counter : :obj:`collections.Counter`
A dictionary of unique descriptions across subjects in the file list,
along with the number of times each pattern occurred. In cases
where all subjects underwent the same protocol, the most common
pattern is most likely the most complete. In cases where the
file list contains multiple protocols, each pattern will need to be
inspected manually.
Examples
--------
>>> from os.path import join
>>> from bids.layout import BIDSLayout
>>> from bids.reports import BIDSReport
>>> from bids.tests import get_test_data_path
>>> layout = BIDSLayout(join(get_test_data_path(), 'synthetic'))
>>> report = BIDSReport(layout)
>>> files = layout.get(session='01', extension=['.nii.gz', '.nii'])
>>> counter = report.generate_from_files(files)
Number of patterns detected: 1
Remember to double-check everything and to replace <deg> with a degree symbol.
>>> counter.most_common()[0][0] # doctest: +ELLIPSIS
'In session 01, MR data were...'
"""
descriptions = []
subjects = sorted(list(set([f.get_entities().get("subject") for f in files])))
sessions = sorted(list(set([f.get_entities().get("session") for f in files])))
for sub in subjects:
subject_files = [f for f in files if f.get_entities().get("subject") == sub]
description_list = []
for ses in sessions:
data_files = [
f for f in subject_files if f.get_entities().get("session") == ses
]
if data_files:
ses_description = parsing.parse_files(
self.layout,
data_files,
sub,
self.config,
)
ses_description[0] = "In session {0}, ".format(ses) + ses_description[0]
description_list += ses_description
metadata = self.layout.get_metadata(data_files[0].path)
else:
raise Exception("No imaging files for subject {0}".format(sub))
# Assume all data were converted the same way and use the last nifti
# file's json for conversion information.
if "metadata" not in vars():
raise Exception(
"No valid jsons found. Cannot generate final paragraph."
)
description = "\n\t".join(description_list)
description += "\n\n{0}".format(parsing.final_paragraph(metadata))
descriptions.append(description)
counter = Counter(descriptions)
print("Number of patterns detected: {0}".format(len(counter.keys())))
print(utils.reminder())
return counter
def generate(self, **kwargs):
r"""Generate the methods section.
Parameters
----------
kwargs : dict
Keyword arguments passed to BIDSLayout to select subsets of the
dataset.
Returns
-------
counter : :obj:`collections.Counter`
A dictionary of unique descriptions across subjects in the dataset,
along with the number of times each pattern occurred. In cases
where all subjects underwent the same protocol, the most common
pattern is most likely the most complete. In cases where the
dataset contains multiple protocols, each pattern will need to be
inspected manually.
Examples
--------
>>> from os.path import join
>>> from bids.layout import BIDSLayout
>>> from bids.reports import BIDSReport
>>> from bids.tests import get_test_data_path
>>> layout = BIDSLayout(join(get_test_data_path(), 'synthetic'))
>>> report = BIDSReport(layout)
>>> counter = report.generate(session='01')
Number of patterns detected: 1
Remember to double-check everything and to replace <deg> with a degree symbol.
>>> counter.most_common()[0][0] # doctest: +ELLIPSIS
'In session 01, MR data were...'
"""
descriptions = []
subjects = self.layout.get_subjects(**kwargs)
kwargs = {k: v for k, v in kwargs.items() if k != "subject"}
for sub in subjects:
descriptions.append(self._report_subject(subject=sub, **kwargs))
counter = Counter(descriptions)
print("Number of patterns detected: {0}".format(len(counter.keys())))
print(utils.reminder())
return counter
def _report_subject(self, subject, **kwargs):
"""Write a report for a single subject.
Parameters
----------
subject : :obj:`str`
Subject ID.
Attributes
----------
layout : :obj:`bids.layout.BIDSLayout`
Layout object for a BIDS dataset.
config : :obj:`dict`
Configuration info for methods generation.
Returns
-------
description : :obj:`str`
A publication-ready report of the dataset's data acquisition
information. Each scan type is given its own paragraph.
"""
description_list = []
# Remove session from kwargs if provided, else set session as all available
sessions = kwargs.pop(
"session", self.layout.get_sessions(subject=subject, **kwargs)
)
if not sessions:
sessions = [None]
elif not isinstance(sessions, list):
sessions = [sessions]
for ses in sessions:
data_files = self.layout.get(
subject=subject,
extension=[".nii", ".nii.gz"],
**kwargs,
)
if data_files:
ses_description = parsing.parse_files(
self.layout,
data_files,
subject,
self.config,
)
ses_description[0] = "In session {0}, ".format(ses) + ses_description[0]
description_list += ses_description
metadata = self.layout.get_metadata(data_files[0].path)
else:
raise Exception("No imaging files for subject {0}".format(subject))
# Assume all data were converted the same way and use the first nifti
# file's json for conversion information.
description = "\n\t".join(description_list)
description += "\n\n{0}".format(parsing.final_paragraph(metadata))
return description