def test_get_unnamed_file_in_archive():
with hold_file_root(get_artemis_data_path('file_getter_tests'),
delete_after=True,
delete_before=True):
path = get_file_in_archive(
url=
'https://drive.google.com/uc?export=download&id=0B4IfiNtPKeSATWZXWjEyd1FsRG8',
subpath='testzip/test_file.txt')
with open(path) as f:
txt = f.read()
assert txt == 'blah blah blah'
def load_ilsvrc_video(identifier,
size=None,
resize_mode='scale_crop',
cut_edges=True,
cut_edges_thresh=5):
"""
Load a file from the ILSVRC Dataset. The first time this is run, it will download an 8GB file, so be patient.
Note: If you are using the same videos repeatedly, and applying resizing, you may want to call this function as:
memoize_to_disk(load_ilsvrc_video)(identifier, size, ...)
This will save you time on future runs.
:param identifier: The file-name of the video, not including the extension. Eg: 'ILSVRC2015_train_00249001'
:param size:
:param cut_edges:
:param cut_edges_thresh:
:return:
"""
print('Downloading ILSVER2015... this may take a while...')
archive_folder_path = get_archive(
relative_path='data/ILSVRC2015',
url=
'http://vision.cs.unc.edu/ilsvrc2015/ILSVRC2015_VID_snippets_final.tar.gz'
)
print('Done.')
subpath = \
'ILSVRC2015/Data/VID/snippets/test' if 'test' in identifier else \
'ILSVRC2015/Data/VID/snippets/val' if 'val' in identifier else \
'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0001/' if os.path.exists(os.path.join(archive_folder_path, 'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0001/', identifier + '.mp4')) else \
'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0002/' if os.path.exists(os.path.join(archive_folder_path, 'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0002/', identifier + '.mp4')) else \
'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0003/' if os.path.exists(os.path.join(archive_folder_path, 'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0003/', identifier + '.mp4')) else \
bad_value(identifier, 'Could not find identifier: {}'.format(identifier, ))
print('Loading %s' % (identifier, ))
full_path = get_file_in_archive(
relative_path='data/ILSVRC2015',
subpath=os.path.join(subpath, identifier + '.mp4'),
url=
'http://vision.cs.unc.edu/ilsvrc2015/ILSVRC2015_VID_snippets_final.tar.gz'
)
video = smart_load_video(full_path,
size=size,
cut_edges=cut_edges,
resize_mode=resize_mode,
cut_edges_thresh=cut_edges_thresh)
print('Done.')
return video
def test_unpack_zip():
with hold_file_root(get_artemis_data_path('file_getter_tests'), delete_after=True, delete_before=True):
assert not os.path.exists(get_artemis_data_path('tests/test_tar_zip'))
for _ in xrange(2): # (Second time to check caching)
local_file = get_file_in_archive(
relative_path= 'tests/test_zip_zip',
url = 'https://drive.google.com/uc?export=download&id=0B4IfiNtPKeSATWZXWjEyd1FsRG8',
subpath = 'testzip/test_file.txt'
)
with open(local_file) as f:
txt = f.read()
assert txt == 'blah blah blah'
def test_unpack_zip():
with hold_file_root(get_artemis_data_path('file_getter_tests'),
delete_after=True,
delete_before=True):
assert not os.path.exists(get_artemis_data_path('tests/test_tar_zip'))
for _ in xrange(2): # (Second time to check caching)
local_file = get_file_in_archive(
relative_path='tests/test_zip_zip',
url=
'https://drive.google.com/uc?export=download&id=0B4IfiNtPKeSATWZXWjEyd1FsRG8',
subpath='testzip/test_file.txt')
with open(local_file) as f:
txt = f.read()
assert txt == 'blah blah blah'
def test_unpack_tar_gz():
if os.path.exists(get_local_path('tests/test_tar_zip')):
shutil.rmtree(get_local_path('tests/test_tar_zip'))
if os.path.exists(get_local_path('tests/test_tar_zip.tar.gz')):
os.remove(get_local_path('tests/test_tar_zip.tar.gz'))
for _ in xrange(2): # (Second time to check caching)
local_file = get_file_in_archive(
relative_path= 'tests/test_tar_zip',
url = 'https://drive.google.com/uc?export=download&id=0B4IfiNtPKeSAbmp6VEVJdjdSSlE',
subpath = 'testzip/test_file.txt'
)
with open(local_file) as f:
txt = f.read()
assert txt == 'blah blah blah'
def test_unpack_zip():
if os.path.exists(get_artemis_data_path('tests/test_zip_zip')):
shutil.rmtree(get_artemis_data_path('tests/test_zip_zip'))
if os.path.exists(get_artemis_data_path('tests/test_zip_zip.zip')):
os.remove(get_artemis_data_path('tests/test_zip_zip.zip'))
for _ in xrange(2): # (Second time to check caching)
local_file = get_file_in_archive(
relative_path= 'tests/test_zip_zip',
url = 'https://drive.google.com/uc?export=download&id=0B4IfiNtPKeSATWZXWjEyd1FsRG8',
subpath = 'testzip/test_file.txt'
)
with open(local_file) as f:
txt = f.read()
assert txt == 'blah blah blah'
def load_ilsvrc_video(identifier, size = None, resize_mode='scale_crop', cut_edges=True, cut_edges_thresh=5):
"""
Load a file from the ILSVRC Dataset. The first time this is run, it will download an 8GB file, so be patient.
Note: If you are using the same videos repeatedly, and applying resizing, you may want to call this function as:
memoize_to_disk(load_ilsvrc_video)(identifier, size, ...)
This will save you time on future runs.
:param identifier: The file-name of the video, not including the extension. Eg: 'ILSVRC2015_train_00249001'
:param size:
:param cut_edges:
:param cut_edges_thresh:
:return:
"""
print ('Downloading ILSVER2015... this may take a while...')
archive_folder_path = get_archive(
relative_path='data/ILSVRC2015',
url='http://vision.cs.unc.edu/ilsvrc2015/ILSVRC2015_VID_snippets_final.tar.gz'
)
print ('Done.')
subpath = \
'ILSVRC2015/Data/VID/snippets/test' if 'test' in identifier else \
'ILSVRC2015/Data/VID/snippets/val' if 'val' in identifier else \
'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0001/' if os.path.exists(os.path.join(archive_folder_path, 'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0001/', identifier + '.mp4')) else \
'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0002/' if os.path.exists(os.path.join(archive_folder_path, 'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0002/', identifier + '.mp4')) else \
'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0003/' if os.path.exists(os.path.join(archive_folder_path, 'ILSVRC2015/Data/VID/snippets/train/ILSVRC2015_VID_train_0003/', identifier + '.mp4')) else \
bad_value(identifier, 'Could not find identifier: {}'.format(identifier, ))
print('Loading %s' % (identifier, ))
full_path = get_file_in_archive(
relative_path='data/ILSVRC2015',
subpath=os.path.join(subpath, identifier+'.mp4'),
url='http://vision.cs.unc.edu/ilsvrc2015/ILSVRC2015_VID_snippets_final.tar.gz'
)
video = smart_load_video(full_path, size=size, cut_edges=cut_edges, resize_mode=resize_mode, cut_edges_thresh=cut_edges_thresh)
print('Done.')
return video
def test_get_unnamed_file_in_archive():
with hold_file_root(get_artemis_data_path('file_getter_tests'), delete_after=True, delete_before=True):
path = get_file_in_archive(url='https://drive.google.com/uc?export=download&id=0B4IfiNtPKeSATWZXWjEyd1FsRG8', subpath='testzip/test_file.txt')
with open(path) as f:
txt = f.read()
assert txt == 'blah blah blah'