def fetch(self):
"""Download and extract the dataset."""
home = self.home()
lock = lockfile.FileLock(home)
if lock.is_locked():
log.warn('%s is locked, waiting for release' % home)
with lock:
# -- download pair labels
for fname, sha1 in PAIRS_FILENAMES:
url = path.join(PAIRS_BASE_URL, fname)
basename = path.basename(url)
filename = path.join(home, basename)
if not path.exists(filename):
if not path.exists(home):
os.makedirs(home)
download(url, filename, sha1=sha1)
# -- download and extract images
url = self.URL
sha1 = self.SHA1
output_dirname = self.home('images')
if not path.exists(output_dirname):
os.makedirs(output_dirname)
# -- various disruptions might cause this to fail
# but if any process gets as far as writing the completion
# marker, then it should be all good.
done_marker = os.path.join(output_dirname, 'completion_marker')
if not path.exists(done_marker):
download_and_extract(url, output_dirname, sha1=sha1)
open(done_marker, 'w').close()
def fetch(self, download_if_missing=True):
"""Download and extract the dataset."""
home = self.home()
if not download_if_missing:
raise IOError("'%s' exists!" % home)
lock = lockfile.FileLock(home)
if lock.is_locked():
log.warn('%s is locked, waiting for release' % home)
with lock:
for fkey, (fname, sha1) in self.FILES.iteritems():
url = path.join(BASE_URL, fname)
basename = path.basename(url)
archive_filename = self.home(basename)
marker = self.home(basename + '.marker')
if ('extra' not in url) or self.need_extra:
if not path.exists(marker):
if not download_if_missing:
return
if not path.exists(home):
os.makedirs(home)
download(url, archive_filename, sha1=sha1)
open(marker, 'w').close()
def fetch(self, download_if_missing=True):
"""Download and extract the dataset."""
home = self.home()
if not os.path.exists(home):
if download_if_missing:
raise NotImplementedError()
else:
raise IOError("'%s' does not exists!" % home)
for filename, url in urls.items():
download(url, self.home(filename), md5=md5s[filename])
return # XXX REST IS CUT AND PASTE FROM ELSEWHERE
for fkey, (fname, sha1) in self.FILES.iteritems():
url = path.join(BASE_URL, fname)
basename = path.basename(url)
archive_filename = path.join(home, basename)
if not path.exists(archive_filename):
if not download_if_missing:
return
if not path.exists(home):
os.makedirs(home)
download(url, archive_filename, sha1=sha1)
def fetch(self, download_if_missing=True):
"""Download and extract the dataset."""
home = self.home()
if not os.path.exists(home):
if download_if_missing:
raise NotImplementedError()
else:
raise IOError("'%s' does not exists!" % home)
for filename, url in urls.items():
download(url, self.home(filename), md5=md5s[filename])
return # XXX REST IS CUT AND PASTE FROM ELSEWHERE
for fkey, (fname, sha1) in self.FILES.iteritems():
url = path.join(BASE_URL, fname)
basename = path.basename(url)
archive_filename = path.join(home, basename)
if not path.exists(archive_filename):
if not download_if_missing:
return
if not path.exists(home):
os.makedirs(home)
download(url, archive_filename, sha1=sha1)
def fetch(self, download_if_missing=True):
"""Download and extract the dataset."""
home = self.home()
if not download_if_missing:
raise IOError("'%s' exists!" % home)
lock = lockfile.FileLock(home)
if lock.is_locked():
log.warn('%s is locked, waiting for release' % home)
with lock:
for fkey, (fname, sha1) in self.FILES.iteritems():
url = path.join(BASE_URL, fname)
basename = path.basename(url)
archive_filename = self.home(basename)
marker = self.home(basename + '.marker')
if ('extra' not in url) or self.need_extra:
if not path.exists(marker):
if not download_if_missing:
return
if not path.exists(home):
os.makedirs(home)
download(url, archive_filename, sha1=sha1)
open(marker, 'w').close()
def fetch(self, download_if_missing=True):
"""Download and extract the dataset."""
home = self.home()
if not download_if_missing:
raise IOError("'%s' exists!" % home)
for fkey, (fname, sha1) in self.FILES.iteritems():
url = path.join(BASE_URL, fname)
basename = path.basename(url)
archive_filename = path.join(home, basename)
if not path.exists(archive_filename):
if not download_if_missing:
return
if not path.exists(home):
os.makedirs(home)
download(url, archive_filename, sha1=sha1)
def fetch(self, download_if_missing=True):
"""Download and extract the dataset."""
home = self.home()
if not download_if_missing:
raise IOError("'%s' exists!" % home)
for fkey, (fname, sha1) in self.FILES.iteritems():
url = path.join(BASE_URL, fname)
basename = path.basename(url)
archive_filename = path.join(home, basename)
if not path.exists(archive_filename):
if not download_if_missing:
return
if not path.exists(home):
os.makedirs(home)
download(url, archive_filename, sha1=sha1)
def fetch(self, download_if_missing=True):
home = self.home()
if not download_if_missing:
raise IOError("'%s' exists!" % home)
# download archive
url = self.URL
sha1 = self.SHA1
basename = os.path.basename(url)
archive_filename = os.path.join(home, basename)
if not os.path.exists(archive_filename):
if not download_if_missing:
return
if not os.path.exists(home):
os.makedirs(home)
download(url, archive_filename, sha1=sha1)
# extract it
if not os.path.exists(self.home(self.SUBDIR)):
extract(archive_filename, home, sha1=sha1, verbose=True)
def fetch(self, download_if_missing=True):
if not download_if_missing:
return
if not os.path.exists(self.home()):
os.makedirs(self.home())
def checkmd5md5():
md5sums = open(self.home('md5sums'), 'rb').read()
md5md5 = hashlib.md5(md5sums).hexdigest()
if md5md5 != 'da55092603cb2628e91e759aec79f654':
print 'Re-downloading corrupt md5sums file'
download(self.BASE_URL + 'md5sums', self.home('md5sums'))
try:
checkmd5md5()
except IOError:
download(self.BASE_URL + 'md5sums', self.home('md5sums'))
checkmd5md5()
meta = self._get_meta()
for ii, item in enumerate(meta):
if self.n_item_limit is None:
required = True
else:
required = ii < self.n_item_limit
try:
data = open(self.home(item['basename']), 'rb').read()
if hashlib.md5(data).hexdigest() != item['md5']:
# -- ignore 'required' flag for incorrect files
print 'Re-downloading incorrect file', item['basename']
download(self.BASE_URL + item['basename'],
self.home(item['basename']),
md5=item['md5'])
# TODO: catch ctrl-C, check md5,
# and remove partial download
except IOError:
if required:
download(self.BASE_URL + item['basename'],
self.home(item['basename']),
md5=item['md5'])
def fetch(self, download_if_missing=True):
if not download_if_missing:
return
if not os.path.exists(self.home()):
os.makedirs(self.home())
def checkmd5md5():
md5sums = open(self.home('md5sums'), 'rb').read()
md5md5 = hashlib.md5(md5sums).hexdigest()
if md5md5 != 'da55092603cb2628e91e759aec79f654':
print 'Re-downloading corrupt md5sums file'
download(self.BASE_URL + 'md5sums', self.home('md5sums'))
try:
checkmd5md5()
except IOError:
download(self.BASE_URL + 'md5sums', self.home('md5sums'))
checkmd5md5()
meta = self._get_meta()
for ii, item in enumerate(meta):
if self.n_item_limit is None:
required = True
else:
required = ii < self.n_item_limit
try:
data = open(self.home(item['basename']), 'rb').read()
if hashlib.md5(data).hexdigest() != item['md5']:
# -- ignore 'required' flag for incorrect files
print 'Re-downloading incorrect file', item['basename']
download(self.BASE_URL + item['basename'],
self.home(item['basename']),
md5=item['md5'])
# TODO: catch ctrl-C, check md5,
# and remove partial download
except IOError:
if required:
download(self.BASE_URL + item['basename'],
self.home(item['basename']),
md5=item['md5'])
def checkmd5md5():
md5sums = open(self.home('md5sums'), 'rb').read()
md5md5 = hashlib.md5(md5sums).hexdigest()
if md5md5 != 'da55092603cb2628e91e759aec79f654':
print 'Re-downloading corrupt md5sums file'
download(self.BASE_URL + 'md5sums', self.home('md5sums'))
def checkmd5md5():
md5sums = open(self.home('md5sums'), 'rb').read()
md5md5 = hashlib.md5(md5sums).hexdigest()
if md5md5 != 'da55092603cb2628e91e759aec79f654':
print 'Re-downloading corrupt md5sums file'
download(self.BASE_URL + 'md5sums', self.home('md5sums'))