def test_add_subdataset(path, other):
subds = create(op.join(path, 'dir'), force=True)
ds = create(path, force=True)
ok_(subds.repo.dirty)
ok_(ds.repo.dirty)
assert_not_in('dir', ds.subdatasets(result_xfm='relpaths'))
# "add everything in subds to subds"
save(dataset=subds.path)
assert_repo_status(subds.path)
assert_not_in('dir', ds.subdatasets(result_xfm='relpaths'))
# but with a base directory we add the dataset subds as a subdataset
# to ds
res = ds.save(subds.path)
assert_in_results(res, action="add", path=subds.path, refds=ds.path)
assert_in('dir', ds.subdatasets(result_xfm='relpaths'))
# create another one
other = create(other)
# install into superdataset, but don't add
other_clone = install(source=other.path, path=op.join(ds.path, 'other'))
# little dance to get the revolution-type dataset
other_clone = Dataset(other_clone.path)
ok_(other_clone.is_installed)
assert_not_in('other', ds.subdatasets(result_xfm='relpaths'))
# now add, it should pick up the source URL
ds.save('other')
# and that is why, we can reobtain it from origin
ds.uninstall('other')
ok_(not other_clone.is_installed())
ds.get('other')
ok_(other_clone.is_installed())
def test_symlinked_relpath(path):
# initially ran into on OSX https://github.com/datalad/datalad/issues/2406
os.makedirs(op.join(path, "origin"))
dspath = op.join(path, "linked")
os.symlink('origin', dspath)
ds = Dataset(dspath).create()
create_tree(dspath, {
"mike1": 'mike1', # will be added from topdir
"later": "later", # later from within subdir
"d": {
"mike2": 'mike2', # to be added within subdir
}
})
# in the root of ds
with chpwd(dspath):
ds.repo.add("mike1", git=True)
ds.save(message="committing", path="./mike1")
# Let's also do in subdirectory as CWD, check that relative path
# given to a plain command (not dataset method) are treated as
# relative to CWD
with chpwd(op.join(dspath, 'd')):
save(dataset=ds.path,
message="committing",
path="mike2")
later = op.join(op.pardir, "later")
ds.repo.add(later, git=True)
save(dataset=ds.path, message="committing", path=later)
assert_repo_status(dspath)
def test_subsuperdataset_save(path):
# Verify that when invoked without recursion save does not
# cause querying of subdatasets of the subdataset
# see https://github.com/datalad/datalad/issues/4523
parent = Dataset(path).create()
# Create 3 levels of subdatasets so later to check operation
# with or without --dataset being specified
sub1 = parent.create('sub1')
sub2 = parent.create(sub1.pathobj / 'sub2')
sub3 = parent.create(sub2.pathobj / 'sub3')
assert_repo_status(path)
# now we will lobotomize that sub2 so git would fail if any query is performed.
rmtree(str(sub3.pathobj / '.git' / 'objects'))
# the call should proceed fine since neither should care about sub3
# default is no recursion
parent.save('sub1')
sub1.save('sub2')
assert_raises(CommandError, parent.save, 'sub1', recursive=True)
# and should fail if we request saving while in the parent directory
# but while not providing a dataset, since operation would run within
# pointed subdataset
with chpwd(sub1.path):
assert_raises(CommandError, save, 'sub2')
# but should not fail in the top level superdataset
with chpwd(parent.path):
save('sub1')
def test_subdataset_save(path):
parent = Dataset(path).create()
sub = parent.create('sub')
ok_clean_git(parent.path)
create_tree(parent.path, {"untracked": 'ignore', 'sub': {"new": "wanted"}})
sub.add('new')
# defined state: one untracked, modified (but clean in itself) subdataset
ok_clean_git(sub.path)
ok_clean_git(parent.path, untracked=['untracked'], index_modified=['sub'])
# `save sub` does not save the parent!!
with chpwd(parent.path):
assert_status('notneeded', save(path=sub.path))
ok_clean_git(parent.path, untracked=['untracked'], index_modified=['sub'])
# `save -d .` saves the state change in the subdataset, but leaves any untracked
# content alone
with chpwd(parent.path):
assert_status('ok', parent.save())
ok_clean_git(parent.path, untracked=['untracked'])
# get back to the original modified state and check that -S behaves in
# exactly the same way
create_tree(parent.path, {'sub': {"new2": "wanted2"}})
sub.add('new2')
ok_clean_git(parent.path, untracked=['untracked'], index_modified=['sub'])
with chpwd(parent.path):
assert_status(
# notneeded to save sub, but need to save parent
['ok', 'notneeded'],
# the key condition of this test is that no reference dataset is
# given!
save(path='sub', super_datasets=True))
# save super must not cause untracked content to be commited!
ok_clean_git(parent.path, untracked=['untracked'])
def test_symlinked_relpath(path):
# initially ran into on OSX https://github.com/datalad/datalad/issues/2406
os.makedirs(op.join(path, "origin"))
dspath = op.join(path, "linked")
os.symlink('origin', dspath)
ds = Dataset(dspath).create()
create_tree(
dspath,
{
"mike1": 'mike1', # will be added from topdir
"later": "later", # later from within subdir
"d": {
"mike2": 'mike2', # to be added within subdir
}
})
# in the root of ds
with chpwd(dspath):
ds.repo.add("mike1", git=True)
ds.save(message="committing", path="./mike1")
# Let's also do in subdirectory as CWD, check that relative path
# given to a plain command (not dataset method) are treated as
# relative to CWD
with chpwd(op.join(dspath, 'd')):
save(dataset=ds.path, message="committing", path="mike2")
later = op.join(op.pardir, "later")
ds.repo.add(later, git=True)
save(dataset=ds.path, message="committing", path=later)
assert_repo_status(dspath)
def test_gh2043p1(path):
# this tests documents the interim agreement on what should happen
# in the case documented in gh-2043
ds = Dataset(path).create(force=True)
ds.save('1')
assert_repo_status(ds.path, untracked=['2', '3'])
ds.unlock('1')
assert_repo_status(
ds.path,
# on windows we are in an unlocked branch by default, hence
# we would see no change
modified=[] if on_windows else ['1'],
untracked=['2', '3'])
# save(.) should recommit unlocked file, and not touch anything else
# this tests the second issue in #2043
with chpwd(path):
# only save modified bits
save(path='.', updated=True)
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
assert_repo_status(ds.path, untracked=['2', '3'])
with chpwd(path):
# but when a path is given, anything that matches this path
# untracked or not is added/saved
save(path='.')
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
assert_repo_status(ds.path)
def test_add_subdataset(path, other):
subds = create(op.join(path, 'dir'), force=True)
ds = create(path, force=True)
ok_(subds.repo.dirty)
ok_(ds.repo.dirty)
assert_not_in('dir', ds.subdatasets(result_xfm='relpaths'))
# "add everything in subds to subds"
save(dataset=subds.path)
assert_repo_status(subds.path)
assert_not_in('dir', ds.subdatasets(result_xfm='relpaths'))
# but with a base directory we add the dataset subds as a subdataset
# to ds
res = ds.save(subds.path)
assert_in_results(res, action="add", path=subds.path, refds=ds.path)
assert_in('dir', ds.subdatasets(result_xfm='relpaths'))
# create another one
other = create(other)
# install into superdataset, but don't add
other_clone = install(source=other.path, path=op.join(ds.path, 'other'))
# little dance to get the revolution-type dataset
other_clone = Dataset(other_clone.path)
ok_(other_clone.is_installed)
assert_not_in('other', ds.subdatasets(result_xfm='relpaths'))
# now add, it should pick up the source URL
ds.save('other')
# and that is why, we can reobtain it from origin
ds.uninstall('other')
ok_(not other_clone.is_installed())
ds.get('other')
ok_(other_clone.is_installed())
def test_gh2043p1(path):
# this tests documents the interim agreement on what should happen
# in the case documented in gh-2043
ds = Dataset(path).create(force=True)
ds.save('1')
assert_repo_status(ds.path, untracked=['2', '3'])
ds.unlock('1')
assert_repo_status(
ds.path,
# on windows we are in an unlocked branch by default, hence
# we would see no change
modified=[] if on_windows else ['1'],
untracked=['2', '3'])
# save(.) should recommit unlocked file, and not touch anything else
# this tests the second issue in #2043
with chpwd(path):
# only save modified bits
save(path='.', updated=True)
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
assert_repo_status(ds.path, untracked=['2', '3'])
with chpwd(path):
# but when a path is given, anything that matches this path
# untracked or not is added/saved
save(path='.')
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
assert_repo_status(ds.path)
def test_subsuperdataset_save(path):
# Verify that when invoked without recursion save does not
# cause querying of subdatasets of the subdataset
# see https://github.com/datalad/datalad/issues/4523
parent = Dataset(path).create()
# Create 3 levels of subdatasets so later to check operation
# with or without --dataset being specified
sub1 = parent.create('sub1')
sub2 = parent.create(sub1.pathobj / 'sub2')
sub3 = parent.create(sub2.pathobj / 'sub3')
assert_repo_status(path)
# now we will lobotomize that sub3 so git would fail if any query is performed.
(sub3.pathobj / '.git' / 'config').chmod(0o000)
try:
sub3.repo.call_git(['ls-files'], read_only=True)
raise SkipTest
except CommandError:
# desired outcome
pass
# the call should proceed fine since neither should care about sub3
# default is no recursion
parent.save('sub1')
sub1.save('sub2')
assert_raises(CommandError, parent.save, 'sub1', recursive=True)
# and should not fail in the top level superdataset
with chpwd(parent.path):
save('sub1')
# or in a subdataset above the problematic one
with chpwd(sub1.path):
save('sub2')
def test_relpath_add(path):
ds = Dataset(path).create(force=True)
with chpwd(op.join(path, 'dir')):
eq_(save('testindir')[0]['path'], op.join(ds.path, 'dir', 'testindir'))
# and now add all
save('..')
# auto-save enabled
assert_repo_status(ds.path)
def make_pretend_repo(dirname):
os.chdir(dirname)
datalad.create(str(dirname), force=True)
rev_log = datalad.save("add data", str(dirname))
(dirname / "useless.txt").write_text("who me.")
datalad.save("add superfluous change", str(dirname))
(dirname / "useless.txt").unlink()
datalad.save("make things better", str(dirname))
def test_path_arg_call(path):
ds = create(path)
for testfile in (ds.pathobj / 'abs.txt', ds.pathobj / 'rel.txt'):
testfile.write_text(u'123')
# we used to resolve relative paths against a dataset just given by
# a path, but we no longer do that
#save(dataset=ds.path, path=[testfile.name], to_git=True)
save(dataset=ds, path=[testfile.name], to_git=True)
def test_bf2043p2(path):
ds = Dataset(path).create(force=True)
ds.repo.add('staged')
assert_repo_status(ds.path, added=['staged'], untracked=['untracked'])
# save -u does not commit untracked content
# this tests the second issue in #2043
with chpwd(path):
save(updated=True)
assert_repo_status(ds.path, untracked=['untracked'])
def test_bf2043p2(path):
ds = Dataset(path).create(force=True)
ds.add('staged', save=False)
ok_clean_git(ds.path, head_modified=['staged'], untracked=['untracked'])
# plain save does not commit untracked content
# this tests the second issue in #2043
with chpwd(path):
save()
ok_clean_git(ds.path, untracked=['untracked'])
def test_bf2043p2(path):
ds = Dataset(path).create(force=True)
ds.repo.add('staged')
assert_repo_status(ds.path, added=['staged'], untracked=['untracked'])
# save -u does not commit untracked content
# this tests the second issue in #2043
with chpwd(path):
save(updated=True)
assert_repo_status(ds.path, untracked=['untracked'])
def test_relpath_add(path):
ds = Dataset(path).create(force=True)
with chpwd(op.join(path, 'dir')):
eq_(save('testindir')[0]['path'],
op.join(ds.path, 'dir', 'testindir'))
# and now add all
save('..')
# auto-save enabled
assert_repo_status(ds.path)
def test_bf2043p2(path):
ds = Dataset(path).create(force=True)
ds.add('staged', save=False)
ok_clean_git(ds.path, head_modified=['staged'], untracked=['untracked'])
# plain save does not commit untracked content
# this tests the second issue in #2043
with chpwd(path):
save()
ok_clean_git(ds.path, untracked=['untracked'])
def test_path_arg_call(path):
ds = create(path)
for testfile in (
ds.pathobj / 'abs.txt',
ds.pathobj / 'rel.txt'):
testfile.write_text(u'123')
# we used to resolve relative paths against a dataset just given by
# a path, but we no longer do that
#save(dataset=ds.path, path=[testfile.name], to_git=True)
save(dataset=ds, path=[testfile.name], to_git=True)
def test_save_directory(path):
# Sequence of save invocations on subdirectories.
ds = Dataset(path).create(force=True)
ds.save(path='sdir1')
ok_clean_git(ds.path, untracked=['sdir2/foo', 'sdir3/sdir/subsub/foo'])
# There is also difference from
with chpwd(path):
save(path='sdir2')
ok_clean_git(ds.path, untracked=['sdir3/sdir/subsub/foo'])
with chpwd(opj(path, 'sdir3')):
save(path='sdir')
ok_clean_git(ds.path)
def test_update_known_submodule(path):
def get_baseline(p):
ds = Dataset(p).create()
sub = create(str(ds.pathobj / 'sub'))
assert_repo_status(ds.path, untracked=['sub'])
return ds
# attempt one
ds = get_baseline(op.join(path, 'wo_ref'))
with chpwd(ds.path):
save(recursive=True)
assert_repo_status(ds.path)
# attempt two, same as above but call add via reference dataset
ds = get_baseline(op.join(path, 'w_ref'))
ds.save(recursive=True)
assert_repo_status(ds.path)
def test_subdataset_save(path):
parent = Dataset(path).create()
sub = parent.create('sub')
assert_repo_status(parent.path)
create_tree(parent.path, {"untracked": 'ignore', 'sub': {"new": "wanted"}})
sub.save('new')
# defined state: one untracked, modified (but clean in itself) subdataset
assert_repo_status(sub.path)
assert_repo_status(parent.path, untracked=['untracked'], modified=['sub'])
# `save sub` does not save the parent!!
with chpwd(parent.path):
assert_status('notneeded', save(dataset=sub.path))
assert_repo_status(parent.path, untracked=['untracked'], modified=['sub'])
# `save -u .` saves the state change in the subdataset,
# but leaves any untracked content alone
with chpwd(parent.path):
assert_status('ok', parent.save(updated=True))
assert_repo_status(parent.path, untracked=['untracked'])
# get back to the original modified state and check that -S behaves in
# exactly the same way
create_tree(parent.path, {'sub': {"new2": "wanted2"}})
sub.save('new2')
assert_repo_status(parent.path, untracked=['untracked'], modified=['sub'])
def test_update_known_submodule(path):
def get_baseline(p):
ds = Dataset(p).create()
sub = create(text_type(ds.pathobj / 'sub'))
assert_repo_status(ds.path, untracked=['sub'])
return ds
# attempt one
ds = get_baseline(op.join(path, 'wo_ref'))
with chpwd(ds.path):
save(recursive=True)
assert_repo_status(ds.path)
# attempt two, same as above but call add via reference dataset
ds = get_baseline(op.join(path, 'w_ref'))
ds.save(recursive=True)
assert_repo_status(ds.path)
def test_subdataset_save(path):
parent = Dataset(path).create()
sub = parent.create('sub')
assert_repo_status(parent.path)
create_tree(parent.path, {
"untracked": 'ignore',
'sub': {
"new": "wanted"}})
sub.save('new')
# defined state: one untracked, modified (but clean in itself) subdataset
assert_repo_status(sub.path)
assert_repo_status(parent.path, untracked=['untracked'], modified=['sub'])
# `save sub` does not save the parent!!
with chpwd(parent.path):
assert_status('notneeded', save(dataset=sub.path))
assert_repo_status(parent.path, untracked=['untracked'], modified=['sub'])
# `save -u .` saves the state change in the subdataset,
# but leaves any untracked content alone
with chpwd(parent.path):
assert_status('ok', parent.save(updated=True))
assert_repo_status(parent.path, untracked=['untracked'])
# get back to the original modified state and check that -S behaves in
# exactly the same way
create_tree(parent.path, {
'sub': {
"new2": "wanted2"}})
sub.save('new2')
assert_repo_status(parent.path, untracked=['untracked'], modified=['sub'])
def test_bf2541(path):
ds = create(path)
subds = ds.create('sub')
assert_repo_status(ds.path)
os.symlink('sub', op.join(ds.path, 'symlink'))
with chpwd(ds.path):
res = save(recursive=True)
assert_repo_status(ds.path)
def test_bf2541(path):
ds = create(path)
subds = ds.create('sub')
assert_repo_status(ds.path)
os.symlink('sub', op.join(ds.path, 'symlink'))
with chpwd(ds.path):
res = save(recursive=True)
assert_repo_status(ds.path)
def test_bf1886(path):
parent = Dataset(path).create()
parent.create('sub')
assert_repo_status(parent.path)
# create a symlink pointing down to the subdataset, and add it
os.symlink('sub', op.join(parent.path, 'down'))
parent.save('down')
assert_repo_status(parent.path)
# now symlink pointing up
os.makedirs(op.join(parent.path, 'subdir', 'subsubdir'))
os.symlink(op.join(op.pardir, 'sub'), op.join(parent.path, 'subdir', 'up'))
parent.save(op.join('subdir', 'up'))
# 'all' to avoid the empty dir being listed
assert_repo_status(parent.path, untracked_mode='all')
# now symlink pointing 2xup, as in #1886
os.symlink(
op.join(op.pardir, op.pardir, 'sub'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup'))
parent.save(op.join('subdir', 'subsubdir', 'upup'))
assert_repo_status(parent.path)
# simulatenously add a subds and a symlink pointing to it
# create subds, but don't register it
create(op.join(parent.path, 'sub2'))
os.symlink(
op.join(op.pardir, op.pardir, 'sub2'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup2'))
parent.save(['sub2', op.join('subdir', 'subsubdir', 'upup2')])
assert_repo_status(parent.path)
# full replication of #1886: the above but be in subdir of symlink
# with no reference dataset
create(op.join(parent.path, 'sub3'))
os.symlink(
op.join(op.pardir, op.pardir, 'sub3'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup3'))
# need to use absolute paths
with chpwd(op.join(parent.path, 'subdir', 'subsubdir')):
save([op.join(parent.path, 'sub3'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup3')])
assert_repo_status(parent.path)
def test_bf1886(path):
parent = Dataset(path).create()
parent.create('sub')
assert_repo_status(parent.path)
# create a symlink pointing down to the subdataset, and add it
os.symlink('sub', op.join(parent.path, 'down'))
parent.save('down')
assert_repo_status(parent.path)
# now symlink pointing up
os.makedirs(op.join(parent.path, 'subdir', 'subsubdir'))
os.symlink(op.join(op.pardir, 'sub'), op.join(parent.path, 'subdir', 'up'))
parent.save(op.join('subdir', 'up'))
# 'all' to avoid the empty dir being listed
assert_repo_status(parent.path, untracked_mode='all')
# now symlink pointing 2xup, as in #1886
os.symlink(
op.join(op.pardir, op.pardir, 'sub'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup'))
parent.save(op.join('subdir', 'subsubdir', 'upup'))
assert_repo_status(parent.path)
# simulatenously add a subds and a symlink pointing to it
# create subds, but don't register it
create(op.join(parent.path, 'sub2'))
os.symlink(
op.join(op.pardir, op.pardir, 'sub2'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup2'))
parent.save(['sub2', op.join('subdir', 'subsubdir', 'upup2')])
assert_repo_status(parent.path)
# full replication of #1886: the above but be in subdir of symlink
# with no reference dataset
create(op.join(parent.path, 'sub3'))
os.symlink(
op.join(op.pardir, op.pardir, 'sub3'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup3'))
# need to use absolute paths
with chpwd(op.join(parent.path, 'subdir', 'subsubdir')):
save([op.join(parent.path, 'sub3'),
op.join(parent.path, 'subdir', 'subsubdir', 'upup3')])
assert_repo_status(parent.path)
def test_subdataset_save(path):
parent = Dataset(path).create()
sub = parent.create('sub')
ok_clean_git(parent.path)
create_tree(parent.path, {
"untracked": 'ignore',
'sub': {
"new": "wanted"}})
sub.add('new')
# defined state: one untracked, modified (but clean in itself) subdataset
ok_clean_git(sub.path)
ok_clean_git(parent.path, untracked=['untracked'], index_modified=['sub'])
# `save sub` does not save the parent!!
with chpwd(parent.path):
assert_status('notneeded', save(path=sub.path))
ok_clean_git(parent.path, untracked=['untracked'], index_modified=['sub'])
# `save -d .` saves the state change in the subdataset, but leaves any untracked
# content alone
with chpwd(parent.path):
assert_status('ok', parent.save())
ok_clean_git(parent.path, untracked=['untracked'])
# get back to the original modified state and check that -S behaves in
# exactly the same way
create_tree(parent.path, {
'sub': {
"new2": "wanted2"}})
sub.add('new2')
ok_clean_git(parent.path, untracked=['untracked'], index_modified=['sub'])
with chpwd(parent.path):
assert_status(
# notneeded to save sub, but need to save parent
['ok', 'notneeded'],
# the key condition of this test is that no reference dataset is
# given!
save(path='sub', super_datasets=True))
# save super must not cause untracked content to be commited!
ok_clean_git(parent.path, untracked=['untracked'])
def test_gh2043p1(path):
# this tests documents the interim agreement on what should happen
# in the case documented in gh-2043
ds = Dataset(path).create(force=True)
ds.add('1')
ok_clean_git(ds.path, untracked=['2', '3'])
ds.unlock('1')
ok_clean_git(ds.path, index_modified=['1'], untracked=['2', '3'])
# save(.) should recommit unlocked file, and not touch anything else
# this tests the second issue in #2043
with chpwd(path):
# only save modified bits by default
save('.') # because the first arg is the dataset
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
skip_v6_or_later(method='pass')(ok_clean_git)(ds.path, untracked=['2', '3'])
with chpwd(path):
# but when a path is given, anything that matches this path
# untracked or not is added/saved
save(path='.')
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
skip_v6_or_later(method='pass')(ok_clean_git)(ds.path)
def test_gh2043p1(path):
# this tests documents the interim agreement on what should happen
# in the case documented in gh-2043
ds = Dataset(path).create(force=True)
ds.add('1')
ok_clean_git(ds.path, untracked=['2', '3'])
ds.unlock('1')
ok_clean_git(ds.path, index_modified=['1'], untracked=['2', '3'])
# save(.) should recommit unlocked file, and not touch anything else
# this tests the second issue in #2043
with chpwd(path):
# only save modified bits by default
save('.') # because the first arg is the dataset
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
skip_v6_or_later(method='pass')(ok_clean_git)(ds.path, untracked=['2', '3'])
with chpwd(path):
# but when a path is given, anything that matches this path
# untracked or not is added/saved
save(path='.')
# state of the file (unlocked/locked) is committed as well, and the
# test doesn't lock the file again
skip_v6_or_later(method='pass')(ok_clean_git)(ds.path)
def save_output_to_repo(config_obj):
base_comparison_dir_path = get_base_comparison_dir_path(config_obj)
update_msg = "Update data with test run on {d}".format(
d=dt.datetime.today().strftime("%Y-%m-%d"))
result = datalad.save(update_msg,
str(base_comparison_dir_path),
on_failure="stop")
sample_test_output = get_test_data_path() / "sample_test_output"
data_message = ("New sample output was saved to {sample_test_output} for "
"future comparisons. Consider publishing this new data to "
"the publicly accessible servers.. ")
print(data_message.format(**locals()))
def test_recursive_save(path):
ds = Dataset(path).create()
# nothing to save
assert_false(ds.save())
subds = ds.create('sub')
# subdataset presence already saved
ok_clean_git(ds.path)
subsubds = subds.create('subsub')
assert_equal(
ds.get_subdatasets(recursive=True, absolute=True, fulfilled=True),
[subsubds.path, subds.path])
newfile_name = opj(subsubds.path, 'test')
with open(newfile_name, 'w') as f:
f.write('some')
# saves the status change of the subdataset due to the subsubdataset addition
assert_equal(ds.save(all_changes=True), [ds])
# make the new file known to its dataset
# with #1141 this would be
#ds.add(newfile_name, save=False)
subsubds.add(newfile_name, save=False)
# but remains dirty because of the untracked file down below
assert ds.repo.dirty
# auto-add will save nothing deep down without recursive
assert_equal(ds.save(all_changes=True), [])
assert ds.repo.dirty
# with recursive pick up the change in subsubds
assert_equal(ds.save(all_changes=True, recursive=True), [subsubds, subds, ds])
# modify content in subsub and try saving
testfname = newfile_name
subsubds.unlock(testfname)
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
# the following should all do nothing
# no auto_add
assert_false(ds.save())
# no recursive
assert_false(ds.save(all_changes=True))
# an explicit target saves only the corresponding dataset
assert_equal(save(files=[testfname]), [subsubds])
# plain recursive without any files given will save the beast
assert_equal(ds.save(recursive=True), [subds, ds])
# there is nothing else to save
assert_false(ds.save(all_changes=True, recursive=True))
# one more time and check that all datasets in the hierarchy get updated
states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
testfname = opj('sub', 'subsub', 'saveme2')
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
assert_true(ds.save(all_changes=True, recursive=True))
newstates = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
for old, new in zip(states, newstates):
assert_not_equal(old, new)
# now let's check saving "upwards"
assert not subds.repo.dirty
create_tree(subds.path, {"testnew": 'smth', "testadded": "added"})
subds.repo.add("testadded")
indexed_files = subds.repo.get_indexed_files()
assert subds.repo.dirty
assert ds.repo.dirty
assert not subsubds.repo.dirty
create_tree(subsubds.path, {"testnew2": 'smth'})
assert subsubds.repo.dirty
# and indexed files didn't change
assert_equal(indexed_files, subds.repo.get_indexed_files())
ok_clean_git(subds.repo, untracked=['testnew'],
index_modified=['subsub'], head_modified=['testadded'])
subsubds.save(message="savingtestmessage", super_datasets=True,
all_changes=True)
ok_clean_git(subsubds.repo)
# but its super should have got only the subsub saved
# not the file we created
ok_clean_git(subds.repo, untracked=['testnew'], head_modified=['testadded'])
# check commits to have correct messages
# there are no more dedicated superdataset-save commits anymore, because
# superdatasets get saved as part of the processed hierarchy and can contain
# other parts in the commit (if so instructed)
assert_equal(next(subsubds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(next(subds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(next(ds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
def test_save(path):
ds = Dataset(path)
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("something")
ds.repo.add("new_file.tst", git=True)
ok_(ds.repo.dirty)
ds.save(message="add a new file")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("modify")
ok_(ds.repo.dirty)
ds.save(message="modified new_file.tst")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# save works without ds and files given in the PWD
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("rapunzel")
with chpwd(path):
save(message="love rapunzel")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# and also without `-a` when things are staged
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("exotic")
ds.repo.add("new_file.tst", git=True)
with chpwd(path):
save(message="love marsians")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
files = ['one.txt', 'two.txt']
for fn in files:
with open(op.join(path, fn), "w") as f:
f.write(fn)
ds.save([op.join(path, f) for f in files])
# superfluous call to save (alll saved it already), should not fail
# but report that nothing was saved
assert_status('notneeded', ds.save(message="set of new files"))
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# create subdataset
subds = ds.create('subds')
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# modify subds
with open(op.join(subds.path, "some_file.tst"), "w") as f:
f.write("something")
subds.save()
assert_repo_status(subds.path, annex=isinstance(subds.repo, AnnexRepo))
# ensure modified subds is committed
ds.save()
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# now introduce a change downstairs
subds.create('someotherds')
assert_repo_status(subds.path, annex=isinstance(subds.repo, AnnexRepo))
ok_(ds.repo.dirty)
# and save via subdataset path
ds.save('subds', version_tag='new_sub')
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
tags = ds.repo.get_tags()
ok_(len(tags) == 1)
eq_(tags[0], dict(hexsha=ds.repo.get_hexsha(), name='new_sub'))
# fails when retagged, like git does
res = ds.save(version_tag='new_sub', on_failure='ignore')
assert_status('error', res)
assert_result_count(
res, 1,
action='save', type='dataset', path=ds.path,
message=('cannot tag this version: %s',
"fatal: tag 'new_sub' already exists"))
def test_recursive_save(path):
ds = Dataset(path).create()
# nothing to save
assert_false(ds.save())
subds = ds.create('sub')
# subdataset presence already saved
ok_clean_git(ds.path)
subsubds = subds.create('subsub')
assert_equal(
ds.get_subdatasets(recursive=True, absolute=True, fulfilled=True),
[subsubds.path, subds.path])
newfile_name = opj(subsubds.path, 'test')
with open(newfile_name, 'w') as f:
f.write('some')
# saves the status change of the subdataset due to the subsubdataset addition
assert_equal(ds.save(all_changes=True), [ds])
# make the new file known to its dataset
# with #1141 this would be
#ds.add(newfile_name, save=False)
subsubds.add(newfile_name, save=False)
# but remains dirty because of the untracked file down below
assert ds.repo.dirty
# auto-add will save nothing deep down without recursive
assert_equal(ds.save(all_changes=True), [])
assert ds.repo.dirty
# with recursive pick up the change in subsubds
assert_equal(ds.save(all_changes=True, recursive=True),
[subsubds, subds, ds])
# modify content in subsub and try saving
testfname = newfile_name
subsubds.unlock(testfname)
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
# the following should all do nothing
# no auto_add
assert_false(ds.save())
# no recursive
assert_false(ds.save(all_changes=True))
# an explicit target saves only the corresponding dataset
assert_equal(save(files=[testfname]), [subsubds])
# plain recursive without any files given will save the beast
assert_equal(ds.save(recursive=True), [subds, ds])
# there is nothing else to save
assert_false(ds.save(all_changes=True, recursive=True))
# one more time and check that all datasets in the hierarchy get updated
states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
testfname = opj('sub', 'subsub', 'saveme2')
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
assert_true(ds.save(all_changes=True, recursive=True))
newstates = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
for old, new in zip(states, newstates):
assert_not_equal(old, new)
# now let's check saving "upwards"
assert not subds.repo.dirty
create_tree(subds.path, {"testnew": 'smth', "testadded": "added"})
subds.repo.add("testadded")
indexed_files = subds.repo.get_indexed_files()
assert subds.repo.dirty
assert ds.repo.dirty
assert not subsubds.repo.dirty
create_tree(subsubds.path, {"testnew2": 'smth'})
assert subsubds.repo.dirty
# and indexed files didn't change
assert_equal(indexed_files, subds.repo.get_indexed_files())
ok_clean_git(subds.repo,
untracked=['testnew'],
index_modified=['subsub'],
head_modified=['testadded'])
subsubds.save(message="savingtestmessage",
super_datasets=True,
all_changes=True)
ok_clean_git(subsubds.repo)
# but its super should have got only the subsub saved
# not the file we created
ok_clean_git(subds.repo,
untracked=['testnew'],
head_modified=['testadded'])
# check commits to have correct messages
# there are no more dedicated superdataset-save commits anymore, because
# superdatasets get saved as part of the processed hierarchy and can contain
# other parts in the commit (if so instructed)
assert_equal(
next(subsubds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(
next(subds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(
next(ds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
def test_recursive_save(path):
ds = Dataset(path).create()
# nothing to save
assert_status('notneeded', ds.save())
subds = ds.create('sub')
# subdataset presence already saved
ok_clean_git(ds.path)
subsubds = subds.create('subsub')
assert_equal(
ds.subdatasets(recursive=True, fulfilled=True, result_xfm='paths'),
[subds.path, subsubds.path])
newfile_name = opj(subsubds.path, 'test')
with open(newfile_name, 'w') as f:
f.write('some')
# saves the status change of the subdataset due to the subsubdataset addition
assert_result_values_equal(
ds.save(result_filter=is_ok_dataset),
'path',
[ds.path])
# make the new file known to its dataset
ds.add(newfile_name, save=False)
# but remains dirty because of the uncommited file down below
assert ds.repo.dirty
# auto-add will save nothing deep down without recursive
assert_status('notneeded', ds.save())
assert ds.repo.dirty
# with recursive pick up the change in subsubds
assert_result_values_equal(
ds.save(recursive=True, result_filter=is_ok_dataset),
'path',
[subsubds.path, subds.path, ds.path])
# at this point the entire tree is clean
ok_clean_git(ds.path)
states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
# now we save recursively, nothing should happen
res = ds.save(recursive=True)
# we do not get any report from a subdataset, because we detect at the
# very top that the entire tree is clean
assert_result_count(res, 1)
assert_result_count(res, 1, status='notneeded', action='save', path=ds.path)
# now we introduce new files all the way down
create_tree(subsubds.path, {"mike1": 'mike1'})
# because we cannot say from the top if there is anything to do down below,
# we have to traverse and we will get reports for all dataset, but there is
# nothing actually saved
res = ds.save(recursive=True)
assert_result_count(res, 3)
assert_status('notneeded', res)
subsubds_indexed = subsubds.repo.get_indexed_files()
assert_not_in('mike1', subsubds_indexed)
assert_equal(states, [d.repo.get_hexsha() for d in (ds, subds, subsubds)])
unlink(opj(subsubds.path, 'mike1'))
ok_clean_git(ds.path)
# modify content in subsub and try saving
testfname = newfile_name
subsubds.unlock(testfname)
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
# the following should all do nothing
# no auto_add
assert_status('notneeded', ds.save())
# no recursive
assert_status('notneeded', ds.save())
# an explicit target saves only the corresponding dataset
assert_result_values_equal(
save(path=[testfname]),
'path',
[subsubds.path])
# plain recursive without any files given will save the beast
assert_result_values_equal(
ds.save(recursive=True, result_filter=is_ok_dataset),
'path',
[subds.path, ds.path])
# there is nothing else to save
assert_status('notneeded', ds.save(recursive=True))
ok_clean_git(ds.path)
# one more time and check that all datasets in the hierarchy are not
# contaminated with untracked files
states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
testfname = opj('sub', 'subsub', 'saveme2')
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
assert_status('notneeded', ds.save(recursive=True))
newstates = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
for old, new in zip(states, newstates):
assert_equal(old, new)
assert ds.repo.dirty
unlink(opj(ds.path, testfname))
ok_clean_git(ds.path)
# now let's check saving "upwards"
create_tree(subds.path, {"testnew": 'smth', "testadded": "added"})
subds.repo.add("testadded")
indexed_files = subds.repo.get_indexed_files()
assert subds.repo.dirty
assert ds.repo.dirty
assert not subsubds.repo.dirty
create_tree(subsubds.path, {"testnew2": 'smth'})
assert subsubds.repo.dirty
# and indexed files didn't change
assert_equal(indexed_files, subds.repo.get_indexed_files())
ok_clean_git(subds.repo, untracked=['testnew'],
index_modified=['subsub'], head_modified=['testadded'])
old_states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
subsubds.save(message="savingtestmessage", super_datasets=True)
# this save actually didn't save anything in subsub (or anywhere),
# because there were only untracked bits pending
for old, new in zip(old_states, [d.repo.get_hexsha()
for d in (ds, subds, subsubds)]):
assert_equal(old, new)
# but now we are saving this untracked bit specifically
subsubds.save(message="savingtestmessage", path=['testnew2'],
super_datasets=True)
ok_clean_git(subsubds.repo)
# but its super should have got only the subsub saved
# not the file we created
ok_clean_git(subds.repo, untracked=['testnew'], head_modified=['testadded'])
# check commits to have correct messages
# there are no more dedicated superdataset-save commits anymore, because
# superdatasets get saved as part of the processed hierarchy and can contain
# other parts in the commit (if so instructed)
assert_equal(next(subsubds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(next(subds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(next(ds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
# and if we try to save while being within that subsubds path
subsubds.unlock('testnew2')
create_tree(subsubds.path, {"testnew2": 'smth2'})
# trying to replicate https://github.com/datalad/datalad/issues/1540
subsubds.save(message="saving new changes", all_updated=True) # no super
with chpwd(subds.path):
# no explicit dataset is provided by path is provided
save(path=['subsub'], message='saving sub', super_datasets=True)
# super should get it saved too
assert_equal(next(ds.repo.get_branch_commits('master')).message.rstrip(),
'saving sub')
def test_save(path):
ds = Dataset(path)
with open(opj(path, "new_file.tst"), "w") as f:
f.write("something")
ds.repo.add("new_file.tst", git=True)
ok_(ds.repo.dirty)
ds.save("add a new file")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
with open(opj(path, "new_file.tst"), "w") as f:
f.write("modify")
ok_(ds.repo.dirty)
ds.save("modified new_file.tst")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# save works without ds and files given in the PWD
with open(opj(path, "new_file.tst"), "w") as f:
f.write("rapunzel")
with chpwd(path):
save("love rapunzel")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# and also without `-a` when things are staged
with open(opj(path, "new_file.tst"), "w") as f:
f.write("exotic")
ds.repo.add("new_file.tst", git=True)
with chpwd(path):
save("love marsians")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
files = ['one.txt', 'two.txt']
for fn in files:
with open(opj(path, fn), "w") as f:
f.write(fn)
ds.add([opj(path, f) for f in files])
# superfluous call to save (add saved it already), should not fail
# but report that nothing was saved
assert_status('notneeded', ds.save("set of new files"))
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# create subdataset
subds = ds.create('subds')
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# modify subds
with open(opj(subds.path, "some_file.tst"), "w") as f:
f.write("something")
subds.add('.')
ok_clean_git(subds.path, annex=isinstance(subds.repo, AnnexRepo))
# Note/TODO: ok_clean_git is failing in direct mode, due to staged but
# uncommited .datalad (probably caused within create)
ok_(ds.repo.dirty)
# ensure modified subds is committed
ds.save()
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# now introduce a change downstairs
subds.create('someotherds')
ok_clean_git(subds.path, annex=isinstance(subds.repo, AnnexRepo))
ok_(ds.repo.dirty)
# and save via subdataset path
ds.save('subds')
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
def test_save(path):
ds = Dataset(path)
with open(opj(path, "new_file.tst"), "w") as f:
f.write("something")
ds.repo.add("new_file.tst", git=True)
ok_(ds.repo.dirty)
ds.save("add a new file")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
with open(opj(path, "new_file.tst"), "w") as f:
f.write("modify")
ok_(ds.repo.dirty)
ds.save("modified new_file.tst")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# save works without ds and files given in the PWD
with open(opj(path, "new_file.tst"), "w") as f:
f.write("rapunzel")
with chpwd(path):
save("love rapunzel")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# and also without `-a` when things are staged
with open(opj(path, "new_file.tst"), "w") as f:
f.write("exotic")
ds.repo.add("new_file.tst", git=True)
with chpwd(path):
save("love marsians")
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
files = ['one.txt', 'two.txt']
for fn in files:
with open(opj(path, fn), "w") as f:
f.write(fn)
ds.add([opj(path, f) for f in files])
# superfluous call to save (add saved it already), should not fail
# but report that nothing was saved
assert_status('notneeded', ds.save("set of new files"))
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# create subdataset
subds = ds.create('subds')
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# modify subds
with open(opj(subds.path, "some_file.tst"), "w") as f:
f.write("something")
subds.add('.')
ok_clean_git(subds.path, annex=isinstance(subds.repo, AnnexRepo))
# Note/TODO: ok_clean_git is failing in direct mode, due to staged but
# uncommited .datalad (probably caused within create)
ok_(ds.repo.dirty)
# ensure modified subds is committed
ds.save()
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
# now introduce a change downstairs
subds.create('someotherds')
ok_clean_git(subds.path, annex=isinstance(subds.repo, AnnexRepo))
ok_(ds.repo.dirty)
# and save via subdataset path
ds.save('subds')
ok_clean_git(path, annex=isinstance(ds.repo, AnnexRepo))
def add(
template_id,
osf_user,
osf_password,
osf_overwrite,
gh_user,
gh_token,
path,
nprocs,
):
"""Add a new template."""
from .io import run_command
from .utils import copy_template
import shutil
from datalad import api as dl
gh_password = getenv("GITHUB_PASSWORD")
if not gh_user or not gh_token:
raise click.BadParameter("Insufficient secrets to login into GitHub")
path = Path(path or f"tpl-{template_id}").absolute()
cwd = Path.cwd()
if not path.exists():
raise click.UsageError(f"<{path}> does not exist.")
metadata = {}
# Check metadata
if (path / "template_description.json").exists():
metadata = json.loads((path / "template_description.json").read_text())
metadata["Identifier"] = template_id
# Check license
license_path = path / "LICENSE"
if not license_path.exists():
license_path = path / "LICENCE"
if not license_path.exists():
license_path = path / "COPYING"
if not license_path.exists():
license_prompt = click.prompt(
text="""\
A LICENSE file MUST be distributed with the template. The TemplateFlow Manager can \
set a license (either CC0 or CC-BY) for you.""",
type=click.Choice(("CC0", "CC-BY", "Custom (abort)")),
default="Custom (abort)",
)
if license_prompt == "Custom (abort)":
raise click.UsageError(
"Cannot proceed without a valid license. Please write a LICENSE "
"file before uploading.")
license_path = Path(
pkgr_fn("tfmanager", f"data/{license_prompt}.LICENSE"))
metadata["License"] = license_prompt
# Check RRID
if not metadata.get("RRID"):
rrid = click.prompt(
text="Has a RRID (research resource ID) already been assigned?",
type=str,
default='') or None
if rrid:
metadata["RRID"] = rrid
# Check short description
if not metadata.get("Name", "").strip():
short_desc = click.prompt(
text="""\
The "Name" metadata is not found within the <template_description.json> file. \
Please provide a short description for this resource.""",
type=str,
)
if not short_desc:
raise click.UsageError(
"Cannot proceed without a short description.")
metadata["Name"] = short_desc
# Check authors
authors_prompt = [
a.strip() for a in metadata.get("Authors", []) if a.strip()
]
if not authors_prompt:
authors_prompt = [
n.strip() for n in click.prompt(
text="""\
The "Authors" metadata is not found within the <template_description.json> file. \
Please provide a list of authors separated by semicolon (;) in <Lastname Initial(s)> format.""",
type=str,
).split(";") if n
]
if not authors_prompt:
click.confirm("No authors were given, do you want to continue?",
abort=True)
metadata["Authors"] = authors_prompt
# Check references
refs_prompt = [
f"""\
{'https://doi.org/' if not a.strip().startswith('http') else ''}\
{a.replace("doi:", "").strip()}"""
for a in metadata.get("ReferencesAndLinks", []) if a.strip()
]
if not refs_prompt:
refs_prompt = [
n.replace('"', "").strip() for n in click.prompt(
text="""\
The "ReferencesAndLinks" metadata is not found within the <template_description.json> file. \
Please provide a list of links and publications within double-quotes \
(for example, "doi:10.1101/2021.02.10.430678") and separated by spaces (< >).""",
type=str,
).split(" ") if n
]
if not refs_prompt:
click.confirm("No authors were given, do you want to continue?",
abort=True)
metadata["ReferencesAndLinks"] = refs_prompt
with TemporaryDirectory() as tmpdir:
repodir = Path(tmpdir) / "templateflow"
# Clone root <user>/templateflow project - fork if necessary
click.echo(f"Preparing Pull-Request (wd={tmpdir}).")
clone = run_command(
f"git clone https://github.com/{gh_user}/templateflow.git "
"--branch tpl-intake --single-branch",
cwd=tmpdir,
capture_output=False,
)
if clone.returncode != 0:
run_command(
"hub clone templateflow/templateflow",
cwd=tmpdir,
capture_output=False,
env={
"GITHUB_USER": gh_user,
"GITHUB_PASSWORD": gh_password
},
)
run_command(
"hub fork --remote-name origin",
cwd=str(repodir),
capture_output=False,
env={
"GITHUB_USER": gh_user,
"GITHUB_PASSWORD": gh_password
},
)
else:
run_command(
"git remote add upstream https://github.com/templateflow/templateflow.git",
cwd=str(repodir),
capture_output=False,
)
chdir(repodir)
# Create datalad dataset
dl.create(
path=f"tpl-{template_id}",
cfg_proc="text2git",
initopts={"initial-branch": "main"},
description=metadata["Name"],
)
# Populate template
copy_template(
path=path,
dest=repodir / f"tpl-{template_id}",
)
# Copy license
shutil.copy(license_path, repodir / f"tpl-{template_id}" / "LICENSE")
# (Over)write template_description.json
(repodir / f"tpl-{template_id}" /
"template_description.json").write_text(json.dumps(metadata,
indent=2))
# Init/update CHANGELOG
changelog = repodir / f"tpl-{template_id}" / "CHANGES"
changes = [
f"""
## {datetime.date.today().ctime()} - TemplateFlow Manager Upload
Populated contents after NIfTI sanitizing by the TF Manager.
"""
]
if changelog.exists():
changes += [changelog.read_text()]
changelog.write_text("\n".join(changes))
# Init OSF sibling
rrid_str = f" (RRID: {metadata['RRID']})" if metadata.get(
"RRID") else ""
dl.create_sibling_osf(
title=f"TemplateFlow resource: <{template_id}>{rrid_str}",
name="osf",
dataset=f"./tpl-{template_id}",
public=True,
category="data",
description=metadata["Name"],
tags=["TemplateFlow dataset", template_id])
# Init GH sibling
dl.create_sibling_github(reponame=f"tpl-{template_id}",
dataset=str(repodir / f"tpl-{template_id}"),
github_login=gh_user,
publish_depends="osf-storage",
existing="replace",
access_protocol="ssh")
# Save added contents
dl.save(dataset=str(repodir / f"tpl-{template_id}"),
message="ADD: TemplateFlow Manager initialized contents")
# Push to siblings
dl.push(
dataset=str(repodir / f"tpl-{template_id}"),
to="github",
jobs=cpu_count(),
)
# Back home
chdir(cwd)
run_command(
"git fetch upstream tpl-intake",
cwd=str(repodir),
capture_output=False,
)
run_command(
f"git checkout -b pr/tpl-{template_id} upstream/tpl-intake",
cwd=str(repodir),
capture_output=False,
)
(repodir / f"{path.name}.toml").write_text(
toml.dumps({
"github": {
"user": gh_user
},
}))
run_command(
f"git add {path.name}.toml",
cwd=str(repodir),
capture_output=False,
)
run_command(
f"git commit -m 'add(tpl-{template_id}): create intake file'",
cwd=str(repodir),
capture_output=False,
)
run_command(
f"git push -u origin pr/tpl-{template_id}",
cwd=str(repodir),
capture_output=False,
env={
"GITHUB_USER": gh_user,
"GITHUB_TOKEN": gh_token
},
)
(repodir.parent / "message.md").write_text(f"""\
ADD: ``tpl-{template_id}``
## {metadata.get('Name', '<missing Name>')}
Identifier: {metadata.get('Identifier', '<missing Identifier>')}
Datalad: https://github.com/{gh_user}/tpl-{template_id}
### Authors
{', '.join(metadata['Authors'])}.
### License
{metadata.get('License', metadata.get('Licence', '<missing License>'))}
### Cohorts
{' '.join(('The dataset contains', str(len(metadata.get('cohort', []))), 'cohorts.'))
if metadata.get('cohort') else 'The dataset does not contain cohorts.'}
### References and links
{', '.join(metadata.get('ReferencesAndLinks', [])) or 'N/A'}
""")
run_command(
"hub pull-request -b templateflow:tpl-intake "
f"-h {gh_user}:pr/tpl-{template_id} "
f"-F {repodir.parent / 'message.md'}",
cwd=str(repodir),
capture_output=False,
env={
"GITHUB_USER": gh_user,
"GITHUB_TOKEN": gh_token
},
)
def test_recursive_save(path):
ds = Dataset(path).create()
# nothing to save
assert_status('notneeded', ds.save())
subds = ds.create('sub')
# subdataset presence already saved
ok_clean_git(ds.path)
subsubds = subds.create('subsub')
assert_equal(
ds.subdatasets(recursive=True, fulfilled=True, result_xfm='paths'),
[subds.path, subsubds.path])
newfile_name = opj(subsubds.path, 'test')
with open(newfile_name, 'w') as f:
f.write('some')
# saves the status change of the subdataset due to the subsubdataset addition
assert_result_values_equal(
ds.save(result_filter=is_ok_dataset),
'path',
[ds.path])
# make the new file known to its dataset
ds.add(newfile_name, save=False)
# but remains dirty because of the uncommited file down below
assert ds.repo.dirty
# auto-add will save nothing deep down without recursive
assert_status('notneeded', ds.save())
assert ds.repo.dirty
# with recursive pick up the change in subsubds
assert_result_values_equal(
ds.save(recursive=True, result_filter=is_ok_dataset),
'path',
[subsubds.path, subds.path, ds.path])
# at this point the entire tree is clean
ok_clean_git(ds.path)
states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
# now we save recursively, nothing should happen
res = ds.save(recursive=True)
# we do not get any report from a subdataset, because we detect at the
# very top that the entire tree is clean
assert_result_count(res, 1)
assert_result_count(res, 1, status='notneeded', action='save', path=ds.path)
# now we introduce new files all the way down
create_tree(subsubds.path, {"mike1": 'mike1'})
# because we cannot say from the top if there is anything to do down below,
# we have to traverse and we will get reports for all dataset, but there is
# nothing actually saved
res = ds.save(recursive=True)
assert_result_count(res, 3)
assert_status('notneeded', res)
subsubds_indexed = subsubds.repo.get_indexed_files()
assert_not_in('mike1', subsubds_indexed)
assert_equal(states, [d.repo.get_hexsha() for d in (ds, subds, subsubds)])
unlink(opj(subsubds.path, 'mike1'))
ok_clean_git(ds.path)
# modify content in subsub and try saving
testfname = newfile_name
subsubds.unlock(testfname)
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
# the following should all do nothing
# no auto_add
assert_status('notneeded', ds.save())
# no recursive
assert_status('notneeded', ds.save())
# an explicit target saves only the corresponding dataset
assert_result_values_equal(
save(path=[testfname]),
'path',
[subsubds.path])
# plain recursive without any files given will save the beast
assert_result_values_equal(
ds.save(recursive=True, result_filter=is_ok_dataset),
'path',
[subds.path, ds.path])
# there is nothing else to save
assert_status('notneeded', ds.save(recursive=True))
ok_clean_git(ds.path)
# one more time and check that all datasets in the hierarchy are not
# contaminated with untracked files
states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
testfname = opj('sub', 'subsub', 'saveme2')
with open(opj(ds.path, testfname), 'w') as f:
f.write('I am in here!')
assert_status('notneeded', ds.save(recursive=True))
newstates = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
for old, new in zip(states, newstates):
assert_equal(old, new)
assert ds.repo.dirty
unlink(opj(ds.path, testfname))
ok_clean_git(ds.path)
# now let's check saving "upwards"
create_tree(subds.path, {"testnew": 'smth', "testadded": "added"})
subds.repo.add("testadded")
indexed_files = subds.repo.get_indexed_files()
assert subds.repo.dirty
assert ds.repo.dirty
assert not subsubds.repo.dirty
create_tree(subsubds.path, {"testnew2": 'smth'})
assert subsubds.repo.dirty
# and indexed files didn't change
assert_equal(indexed_files, subds.repo.get_indexed_files())
ok_clean_git(subds.repo, untracked=['testnew'],
index_modified=['subsub'], head_modified=['testadded'])
old_states = [d.repo.get_hexsha() for d in (ds, subds, subsubds)]
subsubds.save(message="savingtestmessage", super_datasets=True)
# this save actually didn't save anything in subsub (or anywhere),
# because there were only untracked bits pending
for old, new in zip(old_states, [d.repo.get_hexsha()
for d in (ds, subds, subsubds)]):
assert_equal(old, new)
# but now we are saving this untracked bit specifically
subsubds.save(message="savingtestmessage", path=['testnew2'],
super_datasets=True)
ok_clean_git(subsubds.repo)
# but its super should have got only the subsub saved
# not the file we created
ok_clean_git(subds.repo, untracked=['testnew'], head_modified=['testadded'])
# check commits to have correct messages
# there are no more dedicated superdataset-save commits anymore, because
# superdatasets get saved as part of the processed hierarchy and can contain
# other parts in the commit (if so instructed)
assert_equal(next(subsubds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(next(subds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
assert_equal(next(ds.repo.get_branch_commits('master')).message.rstrip(),
'savingtestmessage')
# and if we try to save while being within that subsubds path
subsubds.unlock('testnew2')
create_tree(subsubds.path, {"testnew2": 'smth2'})
# trying to replicate https://github.com/datalad/datalad/issues/1540
subsubds.save(message="saving new changes", all_updated=True) # no super
with chpwd(subds.path):
# no explicit dataset is provided by path is provided
save(path=['subsub'], message='saving sub', super_datasets=True)
# super should get it saved too
assert_equal(next(ds.repo.get_branch_commits('master')).message.rstrip(),
'saving sub')
def test_save(path):
ds = Dataset(path)
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("something")
ds.repo.add("new_file.tst", git=True)
ok_(ds.repo.dirty)
ds.save(message="add a new file")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("modify")
ok_(ds.repo.dirty)
ds.save(message="modified new_file.tst")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# save works without ds and files given in the PWD
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("rapunzel")
with chpwd(path):
save(message="love rapunzel")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# and also without `-a` when things are staged
with open(op.join(path, "new_file.tst"), "w") as f:
f.write("exotic")
ds.repo.add("new_file.tst", git=True)
with chpwd(path):
save(message="love marsians")
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
files = ['one.txt', 'two.txt']
for fn in files:
with open(op.join(path, fn), "w") as f:
f.write(fn)
ds.save([op.join(path, f) for f in files])
# superfluous call to save (alll saved it already), should not fail
# but report that nothing was saved
assert_status('notneeded', ds.save(message="set of new files"))
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# create subdataset
subds = ds.create('subds')
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# modify subds
with open(op.join(subds.path, "some_file.tst"), "w") as f:
f.write("something")
subds.save()
assert_repo_status(subds.path, annex=isinstance(subds.repo, AnnexRepo))
# ensure modified subds is committed
ds.save()
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
# now introduce a change downstairs
subds.create('someotherds')
assert_repo_status(subds.path, annex=isinstance(subds.repo, AnnexRepo))
ok_(ds.repo.dirty)
# and save via subdataset path
ds.save('subds', version_tag='new_sub')
assert_repo_status(path, annex=isinstance(ds.repo, AnnexRepo))
tags = ds.repo.get_tags()
ok_(len(tags) == 1)
eq_(tags[0], dict(hexsha=ds.repo.get_hexsha(), name='new_sub'))
# fails when retagged, like git does
res = ds.save(version_tag='new_sub', on_failure='ignore')
assert_status('error', res)
assert_result_count(res,
1,
action='save',
type='dataset',
path=ds.path,
message=('cannot tag this version: %s',
"fatal: tag 'new_sub' already exists"))
def test_path_arg_call(path):
ds = create(path)
for testfile in (ds.pathobj / 'abs.txt', ds.pathobj / 'rel.txt'):
testfile.write_text(u'123')
save(dataset=ds.path, path=[testfile.name], to_git=True)
def _makeds(path, levels, ds=None, max_leading_dirs=2):
"""Create a hierarchy of datasets
Used recursively, with current invocation generating datasets for the
first level, and delegating sub-levels to recursive invocation
Parameters
----------
path : str
Path to the top directory under which dataset will be created.
If relative -- relative to current directory
levels : list of list
List of specifications for :func:`random.randint` call per each level.
ds : Dataset, optional
Super-dataset which would contain a new dataset (thus its path whould be
a parent of path. Note that ds needs to be installed.
max_leading_dirs : int, optional
Up to how many leading directories withing a dataset could lead to a
sub-dataset
Yields
------
str
Path to the generated dataset(s)
"""
# we apparently can't import api functionality within api
from datalad.api import save
# To simplify managing all the file paths etc
if not isabs(path):
path = abspath(path)
# make it a git (or annex??) repository... ok - let's do randomly one or another ;)
RepoClass = GitRepo if random.randint(0, 1) else AnnexRepo
lgr.info("Generating repo of class %s under %s", RepoClass, path)
repo = RepoClass(path, create=True)
# let's create some dummy file and add it to the beast
fn = opj(path, "file%d.dat" % random.randint(1, 1000))
with open(fn, 'w') as f:
f.write(fn)
repo.add(fn, git=True)
repo.commit(msg="Added %s" % fn)
yield path
if levels:
# make a dataset for that one since we want to add sub datasets
ds_ = Dataset(path)
# Process the levels
level, levels_ = levels[0], levels[1:]
nrepos = random.randint(*level) # how many subds to generate
for irepo in range(nrepos):
# we would like to have up to 2 leading dirs
subds_path = opj(*(['d%i' % i
for i in range(random.randint(0, max_leading_dirs+1))]
+ ['r%i' % irepo]))
subds_fpath = opj(path, subds_path)
# yield all under
for d in _makeds(subds_fpath, levels_, ds=ds_):
yield d
if ds:
assert ds.is_installed()
out = save(
path,
dataset=ds,
)
def process_files(files):
failed = []
for line in fileinput.input(files):
name, kind = parse_input(line)
if kind == "namespace":
try:
repos = list(get_namespace_repos(name))
except requests.HTTPError as exc:
lgr.warning(
"Failed to list repositories for %s (status %s). Skipping",
name, exc.response.status_code)
failed.append(name)
continue
else:
repos = [name]
target_architectures_re = re.compile(target_architectures)
target_tags_re = re.compile(target_tags)
for repo in repos:
lgr.info("Working on %s", repo)
try:
registry = RepoRegistry(repo)
#pprint(list(zip(sorted(_all_tags['latest'], key=lambda r: r['digest']), sorted(_all_tags['1.32.0'],
# key=lambda r: r['digest']))))
tag_images = dict(get_repo_tag_images(repo))
# 'latest' tag is special in docker, it is the default one
# which might typically point to some other release/version.
# If we find that it is the case, we do not create a dedicated "latest"
# image/datalad container -- we just add container entry pointing to that
# one. If there is no matching one -- we do get "latest"
latest_matching_tag = None
# NOTE: "master" is also often used to signal a moving target
# it might, or not, correspond to tagged release. I guess we are just
# doomed to breed those
if target_tags_re.match('latest'):
matching_tags = []
for tag, images in tag_images.items():
if tag == 'latest' or not target_tags_re.match(tag):
lgr.debug("Skipping tag %(tag)s")
continue
if images == tag_images['latest']:
matching_tags.append(tag)
if len(matching_tags) >= 1:
if len(matching_tags) > 1:
lgr.info(
"Multiple tags images match latest, taking the first: %s",
', '.join(matching_tags))
latest_matching_tag = matching_tags[0]
lgr.info("Taking %s as the one for 'latest'",
latest_matching_tag)
else:
# TODO: if there is no latest, we should at least establish the
# convenient one for each tag
pass
for tag, images in tag_images.items():
if tag == 'latest' and latest_matching_tag:
continue # skip since we will handle it
if not target_tags_re.match(tag):
lgr.debug("Skipping tag %(tag)s")
continue
multiarch = len({i['architecture'] for i in images}) > 1
for image in images:
architecture = image['architecture']
if not target_architectures_re.match(architecture):
lgr.debug("Skipping architecture %(architecture)s",
image)
continue
manifest = registry.get_manifest(image['digest'])
digest = manifest["config"]["digest"]
# yoh: if I got it right, it is actual image ID we see in docker images
assert digest.startswith("sha256:")
digest = digest[7:]
digest_short = digest[:12] # use short version in name
last_pushed = image.get('last_pushed')
if last_pushed:
assert last_pushed.endswith('Z')
# take only date
last_pushed = last_pushed[:10].replace('-', '')
assert len(last_pushed) == 8
cleaner_repo = repo
# this is how it looks on hub.docker.com URL
if repo.startswith('library/'):
cleaner_repo = "_/" + cleaner_repo[len('library/'
):]
image_name = f"{cleaner_repo}/{tag}/"
if multiarch:
image_name += f"{architecture}-"
if last_pushed:
# apparently not in all, e.g. no for repronim/neurodocker
# may be None for those built on the hub?
image_name += f"{last_pushed}-"
image_name += f"{digest_short}"
dl_container_name = clean_container_name(
str(image_name))
image_path = Path("images") / image_name
url = f"dhub://{repo}:{tag}@{image['digest']}"
save_paths = []
if image_path.exists():
lgr.info("%s already exists, skipping adding",
str(image_path))
else:
save_paths.append(
write_json(
Path(str(image_path) + '.manifest.json'),
manifest))
save_paths.append(
write_json(
Path(str(image_path) + '.image.json'),
image))
add_container(url, dl_container_name, image_path)
# TODO: either fix datalad-container for https://github.com/datalad/datalad-container/issues/98
# or here, since we have manifest, we can datalad download-url, and add-archive-content
# of the gzipped layers (but without untarring) - that should add datalad-archive
# urls to individual layers in the "saved" version
# TODO: make it in a single commit with add_container at least,
# or one commit for the whole repo sweep
save(
path=save_paths,
message=
f"Added manifest and image records for {dl_container_name}"
)
# TODO: ensure .datalad/config to have additional useful fields:
# architecture, os, and manually "updateurl" since not added for
# dhub:// ATM
if tag == latest_matching_tag and architecture == default_architecture:
# TODO remove section if exists, copy this one
lgr.warning(
"Tracking of 'latest' is not yet implemented")
except requests.HTTPError as exc:
lgr.warning(
"Failed processing %s. Skipping\n status %s for %s", repo,
exc.response.status_code, exc.response.url)
failed.append(name)
continue
return failed
