cryosparc.row

cryosparc.row#

Classes:

R

Type variable for a Row subclass.

Row(cols, idx)

Provides row-by-row access to a dataset.

Spool(items[, rng])

List-like dataset row accessor class with support for splitting and randomizing based on row fields
class cryosparc.row.R#

Type variable for a Row subclass.

alias of TypeVar(‘R’, bound=Row)

class cryosparc.row.Row(cols: Dict[str, Column], idx: int)#

Provides row-by-row access to a dataset. Do not initialize directly. See dataset module.

Methods:

to_list([exclude_uid])

Convert into a list of native python types in the same order as the declared fields.
to_list(exclude_uid=False)#

Convert into a list of native python types in the same order as the declared fields.

class cryosparc.row.Spool(items: Iterable[R], rng: Generator | None = None)#

List-like dataset row accessor class with support for splitting and randomizing based on row fields

Parameters:

  • items (Iterable[R]) – List of rows

  • rng (Generator, optional) – Numpy random number generator. Uses numpy.random.default_rng() if not specified. Defaults to None.

Methods:

get_random_subset(num)

Randomly selected subset of the given size, without replacement.

make_batches([num])

Get a list of lists, each one a consecutive list of num images.

set_default_random(rng)

Reset the default random number generator for all Spools

set_random(rng)

Reset the random number generator for this Spool.

setup_spooling([random])

Determine the iteration order for the spool() method.

split(num[, random, prefix])

Return two Spools with the elements of this spool.

split_by(field)

Split into a dictionary of lists, where each key is a possible value for the given field and each value is a list if elements that have that value.

split_by_splits([prefix])

Return two Spools divided by the value of the {prefix}/split field.

split_from_field(field[, vals])

Split into two lists based on the given possible vals of the given field.

split_half_in_order(prefix[, random])

Split into two spools of approximately equal size.

split_into_quarter(num)

Randomly assign the elements of this Spool to two new Spools.

spool(num[, peek])

Get a list consisting of num randomly-selected elements.
get_random_subset(num: int)#

Randomly selected subset of the given size, without replacement.

Parameters:

num (int) – number of elements to select from the Spool

Returns:

selected elements

Return type:

list[R]

make_batches(num: int = 200)#

Get a list of lists, each one a consecutive list of num images.

Parameters:

num (int, optional) – Size of each batch. Defaults to 200.

Returns:

Split batches

Return type:

list[list[R]]

classmethod set_default_random(rng: Generator)#

Reset the default random number generator for all Spools

Parameters:

rng (Generator) – Numpy random generator.

set_random(rng: Generator)#

Reset the random number generator for this Spool.

Parameters:

rng (Generator) – Numpy random generator.

setup_spooling(random=True)#

Determine the iteration order for the spool() method.

Parameters:

random (bool, optional) – Randomize spooling order. Defaults to True.

split(num: int, random: bool = True, prefix: str | None = None)#

Return two Spools with the elements of this spool. The first Spool contains num elements. The second contains len(self) - num elements.

Parameters:

  • num (int) – Number of elements in first split.

  • random (bool, optional) – If True, add elements to each split in random order. Defaults to True.

  • prefix (str, optional) – If specified, set each {prefix}/split field to 0 in the first split and 1 in the second split. Defaults to None.

Returns:

the two split lists

Return type:

tuple[Spool, Spool]

split_by(field: str) → Dict[Any, List[R]]#

Split into a dictionary of lists, where each key is a possible value for the given field and each value is a list if elements that have that value.

Parameters:

field (str) – dataset field to split on

Returns:

dict of split element lists

Return type:

dict[any, list[R]]

split_by_splits(prefix: str = 'alignments')#

Return two Spools divided by the value of the {prefix}/split field.

Parameters:

prefix (str, optional) – Field to prefix to use to determine which element goes into which split list. Defaults to “alignments”.

Returns:

the two split lists

Return type:

tuple[Spool, Spool]

split_from_field(field: str, vals: Tuple[Any, Any] = (0, 1))#

Split into two lists based on the given possible vals of the given field.

Parameters:

  • field (str) – Split field to test for split value.

  • vals (tuple[any, any], optional) – Possible split field values. Defaults to (0, 1).

Returns:

the two split lists

Return type:

tuple[Spool, Spool]

split_half_in_order(prefix: str, random: bool = True)#

Split into two spools of approximately equal size. Elements are added to each split in stable order.

Parameters:

  • prefix (str) – Set each element’s {prefix}/split field to 0 or 1 depending on which split it’s added to.

  • random (bool, optional) – If True, randomly assign each element to a split. Defaults to True.

Returns:

the two split lists

Return type:

tuple[Spool, Spool]

split_into_quarter(num: int)#

Randomly assign the elements of this Spool to two new Spools. The first Spool contains num elements. The second contains len(self) - num elements.

Parameters:

num (int) – Number of elements in the first split.

Returns:

the two split lists

Return type:

tuple[Spool, Spool]

spool(num: int, peek: bool = False)#

Get a list consisting of num randomly-selected elements. Advance the spool. If peek is true, don’t advance the spool, just return the first num elements.

Parameters:

  • num (int) – Number of elements to get from the spool.

  • peek (bool, optional) – If True, does not advance internal spool index. Will return the same elements on the next spool() call. Defaults to False.

Returns:

list of selected spool elements

Return type:

list[R]