def test_uninstall_without_super(path):
# a parent dataset with a proper subdataset, and another dataset that
# is just placed underneath the parent, but not an actual subdataset
parent = Dataset(path).create()
sub = parent.create('sub')
ok_clean_git(parent.path)
nosub = create(opj(parent.path, 'nosub'))
ok_clean_git(nosub.path)
subreport = parent.subdatasets()
assert_result_count(subreport, 1, path=sub.path)
assert_result_count(subreport, 0, path=nosub.path)
# it should be possible to uninstall the proper subdataset, even without
# explicitly calling the uninstall methods of the parent -- things should
# be figured out by datalad
uninstall(sub.path)
assert not sub.is_installed()
# no present subdatasets anymore
subreport = parent.subdatasets()
assert_result_count(subreport, 1)
assert_result_count(subreport, 1, path=sub.path, state='absent')
assert_result_count(subreport, 0, path=nosub.path)
# but we should fail on an attempt to uninstall the non-subdataset
res = uninstall(nosub.path, on_failure='ignore')
assert_result_count(
res, 1, path=nosub.path, status='error',
message="will not uninstall top-level dataset (consider `remove` command)")
def test_safetynet(path):
ds = Dataset(path).create()
os.makedirs(opj(ds.path, 'deep', 'down'))
for p in (ds.path, opj(ds.path, 'deep'), opj(ds.path, 'deep', 'down')):
with chpwd(p):
# will never remove PWD, or anything outside the dataset
for target in (ds.path, os.curdir, os.pardir,
opj(os.pardir, os.pardir)):
assert_status(('error', 'impossible'),
uninstall(path=target, on_failure='ignore'))
sub = ds.create('sub')
subsub = sub.create('subsub')
for p in (sub.path, subsub.path):
with chpwd(p):
assert_status('error', uninstall(on_failure='ignore'))
def test_safetynet(path):
ds = Dataset(path).create()
os.makedirs(opj(ds.path, 'deep', 'down'))
for p in (ds.path, opj(ds.path, 'deep'), opj(ds.path, 'deep', 'down')):
with chpwd(p):
# will never remove PWD, or anything outside the dataset
for target in (ds.path, os.curdir, os.pardir, opj(os.pardir, os.pardir)):
assert_status(
('error', 'impossible'),
uninstall(path=target, on_failure='ignore'))
def test_uninstall_invalid(path):
assert_raises(InsufficientArgumentsError, uninstall)
# makes no sense to call uninstall, but ask it to do nothing
assert_raises(ValueError,
uninstall,
remove_handles=False,
remove_data=False)
ds = Dataset(path)
# TODO make these two cases uniform
if hasattr(ds.repo, 'drop'):
assert_raises(Exception, uninstall, dataset=ds, path='not_existent')
else:
eq_(uninstall(dataset=ds, path='not_existent'), [])
def test_uninstall_subdataset(src, dst):
ds = install(dst, source=src, recursive=True)
ok_(ds.is_installed())
known_subdss = ds.subdatasets(result_xfm='datasets')
for subds in ds.subdatasets(result_xfm='datasets'):
ok_(subds.is_installed())
repo = subds.repo
annexed_files = repo.get_content_annexinfo(init=None)
repo.get([str(f) for f in annexed_files])
# drop data of subds:
res = ds.drop(path=subds.path, result_xfm='paths')
ok_(all(str(f) in res for f in annexed_files))
ainfo = repo.get_content_annexinfo(paths=annexed_files,
eval_availability=True)
ok_(all(not st["has_content"] for st in ainfo.values()))
# subdataset is still known
assert_in(subds.path, ds.subdatasets(result_xfm='paths'))
eq_(ds.subdatasets(result_xfm='datasets'), known_subdss)
for subds in ds.subdatasets(result_xfm='datasets'):
# uninstall subds itself:
# simulate a cmdline invocation pointing to the subdataset
# with a relative path from outside the superdataset to catch
# https://github.com/datalad/datalad/issues/4001
pwd = Path(dst).parent
with chpwd(str(pwd)):
res = uninstall(
dataset=ds.path,
path=str(subds.pathobj.relative_to(pwd)),
result_xfm='datasets',
)
eq_(res[0], subds)
ok_(not subds.is_installed())
# just a deinit must not remove the subdataset registration
eq_(ds.subdatasets(result_xfm='datasets'), known_subdss)
# mountpoint of subdataset should still be there
ok_(exists(subds.path))
def main(argv):
parser = ArgumentParser(
description='This program will copy augmented BIDS sidecar files back'
'to the original BIDS data and run bidsmri2nidm on it')
parser.add_argument(
'-datalad_dir',
dest='datalad_dir',
required=True,
help="Path to directory containing "
"datalad datasets to copy BIDS sidecar files to (e.g. /datasets.datalad.org/openfmri)"
)
parser.add_argument(
'-new_sidecar_dir',
dest='sidecar_dir',
required=True,
help="Path to directory containing "
"sub-directories for each BIDS dataset with augmented BIDS sidecar files"
)
parser.add_argument(
'-nidm_dir',
dest='nidm_dir',
required=False,
help="If path is included then "
"NIDM files will be put in this location under dataset name otherwise they will be"
"added to the datalad dataset directly.")
parser.add_argument('-logfile',
dest='logfile',
required=True,
help="Logfile to track progress and errors")
args = parser.parse_args()
# open log file
logger = logging.getLogger('bidsidecar_to_datset')
hdlr = logging.FileHandler(args.logfile)
formatter = logging.Formatter('%(asctime)s %(levelname)s %(message)s')
hdlr.setFormatter(formatter)
logger.addHandler(hdlr)
logger.setLevel(logging.INFO)
# step 1 check directories exist
if not isdir(args.datalad_dir):
logger.error("Datalad directory not found: %s" % args.datalad_dir)
exit(1)
if not isdir(args.sidecar_dir):
logger.error("BIDS augmented sidecar directory not found: %s" %
args.sidecar_dir)
exit(1)
# set working directory to args.datalad_dir
#os.chdir(args.datalad_dir)
# step 2 loop through all datasets in args.sidecar_dir
bids_datasets = [
x for x in os.listdir(args.sidecar_dir)
if isdir(join(args.sidecar_dir, x))
]
# for each dataset
for ds in bids_datasets:
# search recursively in ds looking for json files. Path should be same as found within the dataset in
# datalad so it'll be a simple copy.
json_files = glob2.glob(join(args.sidecar_dir, ds, '**', "*.json"))
# check of we have json files before we download dataset from datalad
if len(json_files) == 0:
continue
# download datalad dataset and install
#cmd = ["datalad","get", "-r", join(args.datalad_dir,ds)]
# replacing with datalad api
#cmd = ["datalad", "get", "-r", ds]
try:
dl.get(path=join(args.datalad_dir, ds), recursive=True)
logger.info("Running datalad get command on dataset: %s" %
join(args.datalad_dir, ds))
#ret = subprocess.run(cmd, stdout=subprocess.PIPE, text=True)
except:
logger.error("Datalad returned error: %s for dataset %s." %
(sys.exc_info()[0], ds))
continue
# now copy each of the json_files into the datalad dataset
for file in json_files:
# changing copy to use copyfile from shutil
#cmd = ["cp",join(args.sidecar_dir,ds,file),join(args.datalad_dir,ds)]
if not isdir(join(args.datalad_dir, ds)):
os.mkdir(join(args.datalad_dir, ds))
logger.info(
"Copying file: source=%s, dest=%s" %
(join(args.sidecar_dir, ds, file), join(args.datalad_dir, ds)))
copyfile(join(args.sidecar_dir, ds, file),
join(args.datalad_dir, ds, file))
#ret = subprocess.run(cmd, stdout=subprocess.PIPE, text=True)
# make sure it's there
if not isfile(join(args.datalad_dir, ds, file)):
logger.error(
"ERROR: copy of file %s to %s didn't complete successfully!" %
(join(args.sidecar_dir, ds, file), join(args.datalad_dir, ds)))
# now run bidsmri2nidm
if args.nidm_dir is not None:
cmd = [
"bidsmri2nidm", "-d",
join(args.datalad_dir, ds), "-o",
join(args.nidm_dir, ds, "nidm.ttl"), "-bidsignore",
"-no_concepts"
]
else:
cmd = [
"bidsmri2nidm", "-d",
join(args.datalad_dir, ds), "-o",
join(args.datalad_dir, ds, "nidm.ttl"), "-bidsignore",
"-no_concepts"
]
logger.info("Running command: %s" % cmd)
ret = subprocess.run(cmd, stdout=subprocess.PIPE, text=True)
# now remove the datalad dataset to save space
# replacing with datalad api call
#cmd = ["datalad", "uninstall", "-r", join(args.datalad_dir, ds)]
dl.uninstall(path=join(args.datalad_dir, ds), recursive=True)
logger.info("Running datalad uninstall command on dataset: %s" %
join(args.datalad_dir, ds))
