1818 load_json ,
1919 save_json ,
2020 create_file_if_missing ,
21- json_dumps_pretty ,
21+ json_dumps ,
2222 set_readonly ,
2323 is_readonly ,
2424 get_datetime ,
2525)
26+ from . import __version__
2627
2728lgr = logging .getLogger (__name__ )
2829
4041 ("Description" , "md5 hash of UIDs" )])),
4142])
4243
44+ #: JSON Key where we will embed our version in the newly produced .json files
45+ HEUDICONV_VERSION_JSON_KEY = 'HeudiconvVersion'
46+
47+
4348class BIDSError (Exception ):
4449 pass
4550
4651
52+ BIDS_VERSION = "1.4.1"
53+
54+
55+ def maybe_na (val ):
56+ """Return 'n/a' if non-None value represented as str is not empty
57+
58+ Primarily for the consistent use of lower case 'n/a' so 'N/A' and 'NA'
59+ are also treated as 'n/a'
60+ """
61+ if val is not None :
62+ val = str (val )
63+ val = val .strip ()
64+ return 'n/a' if (not val or val in ('N/A' , 'NA' )) else val
65+
66+
67+ def treat_age (age ):
68+ """Age might encounter 'Y' suffix or be a float"""
69+ age = str (age )
70+ if age .endswith ('M' ):
71+ age = age .rstrip ('M' )
72+ age = float (age ) / 12
73+ age = ('%.2f' if age != int (age ) else '%d' ) % age
74+ else :
75+ age = age .rstrip ('Y' )
76+ if age :
77+ # strip all leading 0s but allow to scan a newborn (age 0Y)
78+ age = '0' if not age .lstrip ('0' ) else age .lstrip ('0' )
79+ if age .startswith ('.' ):
80+ # we had float point value, let's prepend 0
81+ age = '0' + age
82+ return age
83+
84+
4785def populate_bids_templates (path , defaults = {}):
4886 """Premake BIDS text files with templates"""
4987
@@ -53,7 +91,7 @@ def populate_bids_templates(path, defaults={}):
5391 save_json (descriptor ,
5492 OrderedDict ([
5593 ('Name' , "TODO: name of the dataset" ),
56- ('BIDSVersion' , "1.0.1" ),
94+ ('BIDSVersion' , BIDS_VERSION ),
5795 ('License' , defaults .get ('License' ,
5896 "TODO: choose a license, e.g. PDDL "
5997 "(http://opendatacommons.org/licenses/pddl/)" )),
@@ -87,6 +125,9 @@ def populate_bids_templates(path, defaults={}):
87125 create_file_if_missing (op .join (path , 'README' ),
88126 "TODO: Provide description for the dataset -- basic details about the "
89127 "study, possibly pointing to pre-registration (if public or embargoed)" )
128+ create_file_if_missing (op .join (path , 'scans.json' ),
129+ json_dumps (SCANS_FILE_FIELDS , sort_keys = False )
130+ )
90131
91132 populate_aggregated_jsons (path )
92133
@@ -111,7 +152,8 @@ def populate_aggregated_jsons(path):
111152 # way too many -- let's just collect all which are the same!
112153 # FIELDS_TO_TRACK = {'RepetitionTime', 'FlipAngle', 'EchoTime',
113154 # 'Manufacturer', 'SliceTiming', ''}
114- for fpath in find_files ('.*_task-.*\_bold\.json' , topdir = path ,
155+ for fpath in find_files ('.*_task-.*\_bold\.json' ,
156+ topdir = glob (op .join (path , 'sub-*' )),
115157 exclude_vcs = True ,
116158 exclude = "/\.(datalad|heudiconv)/" ):
117159 #
@@ -120,7 +162,7 @@ def populate_aggregated_jsons(path):
120162 # TODO: if we are to fix it, then old ones (without _acq) should be
121163 # removed first
122164 task = re .sub ('.*_(task-[^_\.]*(_acq-[^_\.]*)?)_.*' , r'\1' , fpath )
123- json_ = load_json (fpath )
165+ json_ = load_json (fpath , retry = 100 )
124166 if task not in tasks :
125167 tasks [task ] = json_
126168 else :
@@ -172,10 +214,10 @@ def populate_aggregated_jsons(path):
172214 placeholders = {
173215 "TaskName" : ("TODO: full task name for %s" %
174216 task_acq .split ('_' )[0 ].split ('-' )[1 ]),
175- "CogAtlasID" : "TODO" ,
217+ "CogAtlasID" : "http://www.cognitiveatlas.org/task/id/ TODO" ,
176218 }
177219 if op .lexists (task_file ):
178- j = load_json (task_file )
220+ j = load_json (task_file , retry = 100 )
179221 # Retain possibly modified placeholder fields
180222 for f in placeholders :
181223 if f in j :
@@ -207,6 +249,10 @@ def tuneup_bids_json_files(json_files):
207249 # Let's hope no word 'Date' comes within a study name or smth like
208250 # that
209251 raise ValueError ("There must be no dates in .json sidecar" )
252+ # Those files should not have our version field already - should have been
253+ # freshly produced
254+ assert HEUDICONV_VERSION_JSON_KEY not in json_
255+ json_ [HEUDICONV_VERSION_JSON_KEY ] = str (__version__ )
210256 save_json (jsonfile , json_ )
211257
212258 # Load the beast
@@ -274,12 +320,13 @@ def add_participant_record(studydir, subject, age, sex):
274320 "control group)" )])),
275321 ]),
276322 sort_keys = False )
323+
277324 # Add a new participant
278325 with open (participants_tsv , 'a' ) as f :
279326 f .write (
280327 '\t ' .join (map (str , [participant_id ,
281- age . lstrip ( '0' ). rstrip ( 'Y' ) if age else 'N/A' ,
282- sex ,
328+ maybe_na ( treat_age ( age )) ,
329+ maybe_na ( sex ) ,
283330 'control' ])) + '\n ' )
284331
285332
@@ -369,11 +416,6 @@ def add_rows_to_scans_keys_file(fn, newrows):
369416 os .unlink (fn )
370417 else :
371418 fnames2info = newrows
372- # Populate _scans.json (an optional file to describe column names in
373- # _scans.tsv). This auto generation will make BIDS-validator happy.
374- scans_json = '.' .join (fn .split ('.' )[:- 1 ] + ['json' ])
375- if not op .lexists (scans_json ):
376- save_json (scans_json , SCANS_FILE_FIELDS , sort_keys = False )
377419
378420 header = SCANS_FILE_FIELDS
379421 # prepare all the data rows
@@ -404,10 +446,10 @@ def get_formatted_scans_key_row(dcm_fn):
404446 """
405447 dcm_data = dcm .read_file (dcm_fn , stop_before_pixels = True , force = True )
406448 # we need to store filenames and acquisition times
407- # parse date and time and get it into isoformat
449+ # parse date and time of start of run acquisition and get it into isoformat
408450 try :
409- date = dcm_data .ContentDate
410- time = dcm_data .ContentTime
451+ date = dcm_data .AcquisitionDate
452+ time = dcm_data .AcquisitionTime
411453 acq_time = get_datetime (date , time )
412454 except (AttributeError , ValueError ) as exc :
413455 lgr .warning ("Failed to get date/time for the content: %s" , str (exc ))
0 commit comments