def decompress_file(archive, dir_):
"""Decompress `archive` into a directory `dir_`
Parameters
----------
archive: str
dir_: str
"""
with swallow_outputs() as cmo:
archive = ensure_bytes(archive)
dir_ = ensure_bytes(dir_)
patoolib.util.check_existing_filename(archive)
patoolib.util.check_existing_filename(dir_, onlyfiles=False)
# Call protected one to avoid the checks on existence on unixified path
outdir = unixify_path(dir_)
# should be supplied in PY3 to avoid b''
outdir = ensure_unicode(outdir)
archive = ensure_unicode(archive)
format_compression = patoolib.get_archive_format(archive)
if format_compression == ('gzip', None):
# Yarik fell into the trap of being lazy and not providing proper
# support for .gz .xz etc "stream archivers" formats in handling
# of archives. ATM out support for .gz relies on behavior of 7z while
# extracting them and respecting possibly present .gz filename
# header field.
# See more https://github.com/datalad/datalad/pull/3176#issuecomment-466819861
# TODO: provide proper handling of all those archives without
# relying on any filename been stored in the header
program = patoolib.find_archive_program(format_compression[0],
'extract')
if basename(program) != '7z':
raise MissingExternalDependency(
"cmd:7z",
msg="(Not) Funny enough but ATM we need p7zip installation "
"to handle .gz files extraction 'correctly'")
patoolib._extract_archive(unixify_path(archive),
outdir=outdir,
verbosity=100)
if cmo.out:
lgr.debug("patool gave stdout:\n%s", cmo.out)
if cmo.err:
lgr.debug("patool gave stderr:\n%s", cmo.err)
# Note: (ben) Experienced issue, where extracted tarball
# lacked execution bit of directories, leading to not being
# able to delete them while having write permission.
# Can't imagine a situation, where we would want to fail on
# that kind of mess. So, to be sure set it.
if not on_windows:
os.chmod(dir_, os.stat(dir_).st_mode | os.path.stat.S_IEXEC)
for root, dirs, files in os.walk(dir_, followlinks=False):
for d in dirs:
subdir = opj(root, d)
os.chmod(subdir,
os.stat(subdir).st_mode | os.path.stat.S_IEXEC)
def check_7z():
try:
patoolib.find_archive_program('7z', 'extract')
except patoolib.util.PatoolError as e:
warning = ('This step requires a working installation of the 7zip program. <br>' +
'Please install 7zip on your system before continuing.<br>')
if sys.platform == 'win32':
warning += 'You can download it from <a href="http://www.7zip.org">www.7zip.org</a>.'
elif sys.platform == 'darwin':
warning += ('You can install 7zip with <a href="https://brew.sh">homebrew</a>.<br>' +
'brew install p7zip')
elif sys.platform.startswith('linux'):
warning += ("Use your linux distribtion's package manager for the installation.<br>" +
'eg: sudo apt install p7zip-full')
else:
warning += 'More infos here: <a href="http://www.7zip.org">www.7zip.org</a>'
return warning
return ''
def decompress_file(archive, dir_, leading_directories='strip'):
"""Decompress `archive` into a directory `dir_`
Parameters
----------
archive: str
dir_: str
leading_directories: {'strip', None}
If `strip`, and archive contains a single leading directory under which
all content is stored, all the content will be moved one directory up
and that leading directory will be removed.
"""
if not exists(dir_):
lgr.debug("Creating directory %s to extract archive into" % dir_)
os.makedirs(dir_)
with swallow_outputs() as cmo:
archive = assure_bytes(archive)
dir_ = assure_bytes(dir_)
patoolib.util.check_existing_filename(archive)
patoolib.util.check_existing_filename(dir_, onlyfiles=False)
# Call protected one to avoid the checks on existence on unixified path
outdir = unixify_path(dir_)
if not PY2:
# should be supplied in PY3 to avoid b''
outdir = assure_unicode(outdir)
archive = assure_unicode(archive)
format_compression = patoolib.get_archive_format(archive)
if format_compression == ('gzip', None):
# Yarik fell into the trap of being lazy and not providing proper
# support for .gz .xz etc "stream archivers" formats in handling
# of archives. ATM out support for .gz relies on behavior of 7z while
# extracting them and respecting possibly present .gz filename
# header field.
# See more https://github.com/datalad/datalad/pull/3176#issuecomment-466819861
# TODO: provide proper handling of all those archives without
# relying on any filename been stored in the header
program = patoolib.find_archive_program(format_compression[0],
'extract')
if basename(program) != '7z':
raise MissingExternalDependency(
"cmd:7z",
msg="(Not) Funny enough but ATM we need p7zip installation "
"to handle .gz files extraction 'correctly'")
patoolib._extract_archive(unixify_path(archive),
outdir=outdir,
verbosity=100)
if cmo.out:
lgr.debug("patool gave stdout:\n%s" % cmo.out)
if cmo.err:
lgr.debug("patool gave stderr:\n%s" % cmo.err)
# Note: (ben) Experienced issue, where extracted tarball
# lacked execution bit of directories, leading to not being
# able to delete them while having write permission.
# Can't imagine a situation, where we would want to fail on
# that kind of mess. So, to be sure set it.
if not on_windows:
os.chmod(dir_, os.stat(dir_).st_mode | os.path.stat.S_IEXEC)
for root, dirs, files in os.walk(dir_, followlinks=False):
for d in dirs:
subdir = opj(root, d)
os.chmod(subdir,
os.stat(subdir).st_mode | os.path.stat.S_IEXEC)
if leading_directories == 'strip':
_, dirs, files = next(os.walk(dir_))
if not len(files) and len(dirs) == 1:
# move all the content under dirs[0] up 1 level
widow_dir = opj(dir_, dirs[0])
lgr.debug("Moving content within %s upstairs" % widow_dir)
subdir, subdirs_, files_ = next(os.walk(opj(dir_, dirs[0])))
for f in subdirs_ + files_:
os.rename(opj(subdir, f), opj(dir_, f))
rmdir(widow_dir)
elif leading_directories is None:
pass # really do nothing
else:
raise NotImplementedError("Not supported %s" % leading_directories)
def decompress_file(archive, dir_, leading_directories='strip'):
"""Decompress `archive` into a directory `dir_`
Parameters
----------
archive: str
dir_: str
leading_directories: {'strip', None}
If `strip`, and archive contains a single leading directory under which
all content is stored, all the content will be moved one directory up
and that leading directory will be removed.
"""
if not exists(dir_):
lgr.debug("Creating directory %s to extract archive into" % dir_)
os.makedirs(dir_)
with swallow_outputs() as cmo:
archive = assure_bytes(archive)
dir_ = assure_bytes(dir_)
patoolib.util.check_existing_filename(archive)
patoolib.util.check_existing_filename(dir_, onlyfiles=False)
# Call protected one to avoid the checks on existence on unixified path
outdir = unixify_path(dir_)
if not PY2:
# should be supplied in PY3 to avoid b''
outdir = assure_unicode(outdir)
archive = assure_unicode(archive)
format_compression = patoolib.get_archive_format(archive)
if format_compression == ('gzip', None):
# Yarik fell into the trap of being lazy and not providing proper
# support for .gz .xz etc "stream archivers" formats in handling
# of archives. ATM out support for .gz relies on behavior of 7z while
# extracting them and respecting possibly present .gz filename
# header field.
# See more https://github.com/datalad/datalad/pull/3176#issuecomment-466819861
# TODO: provide proper handling of all those archives without
# relying on any filename been stored in the header
program = patoolib.find_archive_program(
format_compression[0], 'extract')
if basename(program) != '7z':
raise MissingExternalDependency(
"cmd:7z",
msg="(Not) Funny enough but ATM we need p7zip installation "
"to handle .gz files extraction 'correctly'"
)
patoolib._extract_archive(unixify_path(archive),
outdir=outdir,
verbosity=100)
if cmo.out:
lgr.debug("patool gave stdout:\n%s" % cmo.out)
if cmo.err:
lgr.debug("patool gave stderr:\n%s" % cmo.err)
# Note: (ben) Experienced issue, where extracted tarball
# lacked execution bit of directories, leading to not being
# able to delete them while having write permission.
# Can't imagine a situation, where we would want to fail on
# that kind of mess. So, to be sure set it.
if not on_windows:
os.chmod(dir_,
os.stat(dir_).st_mode |
os.path.stat.S_IEXEC)
for root, dirs, files in os.walk(dir_, followlinks=False):
for d in dirs:
subdir = opj(root, d)
os.chmod(subdir,
os.stat(subdir).st_mode |
os.path.stat.S_IEXEC)
if leading_directories == 'strip':
_, dirs, files = next(os.walk(dir_))
if not len(files) and len(dirs) == 1:
# move all the content under dirs[0] up 1 level
widow_dir = opj(dir_, dirs[0])
lgr.debug("Moving content within %s upstairs" % widow_dir)
subdir, subdirs_, files_ = next(os.walk(opj(dir_, dirs[0])))
for f in subdirs_ + files_:
os.rename(opj(subdir, f), opj(dir_, f))
rmdir(widow_dir)
elif leading_directories is None:
pass # really do nothing
else:
raise NotImplementedError("Not supported %s" % leading_directories)
