def test_store_and_retrieve(self):
x = int("0b01010100100100100100100010010100100100101001001010101010", 2)
mask = BitMask(x)
mask.set_kth_bit(11)
mask.set_kth_bit(22)
mask.set_kth_bit(33)
mask.set_kth_bit(44)
mask.set_kth_bit(55)
mask.set_kth_bit(66)
mask.set_kth_bit(77)
mask.set_kth_bit(88)
mask.set_kth_bit(99)
somewhere = mask.to_hex()
somewhere_else = mask.to_bin()
mask_somewhere = BitMask.from_hex(somewhere)
mask_somewhere_else = BitMask.from_bin(somewhere_else)
self.assertEqual(mask.x, mask_somewhere.x)
self.assertEqual(mask.x, mask_somewhere_else.x)
def create_tag_from_filenames(self,
fnames_new_tag: List[str],
new_tag_name: str,
parent_tag_id: str = None) -> TagData:
"""Creates a new tag from a list of filenames.
Args:
fnames_new_tag:
A list of filenames to be included in the new tag.
new_tag_name:
The name of the new tag.
parent_tag_id:
The tag defining where to sample from, default: None resolves to the initial-tag.
Returns:
The newly created tag.
Raises:
RuntimeError
"""
# make sure the tag name does not exist yet
tags = self.get_all_tags()
if new_tag_name in [tag.name for tag in tags]:
raise RuntimeError(
f'There already exists a tag with tag_name {new_tag_name}.')
if len(tags) == 0:
raise RuntimeError('There exists no initial-tag for this dataset.')
# fallback to initial tag if no parent tag is provided
if parent_tag_id is None:
parent_tag_id = next(tag.id for tag in tags
if tag.name == 'initial-tag')
# get list of filenames from tag
fnames_server = self.get_filenames()
tot_size = len(fnames_server)
# create new bitmask for the new tag
bitmask = BitMask(0)
fnames_new_tag = set(fnames_new_tag)
for i, fname in enumerate(fnames_server):
if fname in fnames_new_tag:
bitmask.set_kth_bit(i)
# quick sanity check
num_selected_samples = len(bitmask.to_indices())
if num_selected_samples != len(fnames_new_tag):
raise RuntimeError(
f'An error occured when creating the new subset! '
f'Out of the {len(fnames_new_tag)} filenames you provided '
f'to create a new tag, only {num_selected_samples} have been '
f'found on the server. '
f'Make sure you use the correct filenames. '
f'Valid filename example from the dataset: {fnames_server[0]}')
# create new tag
tag_data_dict = {
'name': new_tag_name,
'prevTagId': parent_tag_id,
'bitMaskData': bitmask.to_hex(),
'totSize': tot_size
}
new_tag = self._tags_api.create_tag_by_dataset_id(
tag_data_dict, self.dataset_id)
return new_tag
