snake.core.sampling.factories

K-space trajectory factories.

Module Contents

Classes

VDSorder	Available ordering for variable density sampling.
VDSpdf	Available law for variable density sampling.
AngleRotation	Available rotation angle for density sampling.

Functions

validate_rng	Validate Random Number Generator.
get_kspace_slice_loc	Get slice index at a random position.
get_cartesian_mask	Get a cartesian mask for fMRI kspace data.
flip2center	Reorder a list by starting by a center_position and alternating left/right.
check_trajectory	Check if a trajectory is feasible or not.
vds_factory	Create a variable density sampling trajectory.
radial_factory	Create a radial sampling trajectory.
stack_spiral_factory	Generate a trajectory of stack of spiral.
rotate_trajectory	Incrementally rotate a trajectory.
stacked_epi2d	Generate a list of 2D epi plane, stacked.
stacked_epi_factory	Generate a VDS stack of fully sampled EPI trajectory.
evi_factory	Generate a Echo Volume Imaging trajectory.
trajectory_generator	Generate a trajectory.

Data

logger
SlicerType

API

snake.core.sampling.factories.logger = 'getLogger(...)'

snake.core.sampling.factories.SlicerType = None

snake.core.sampling.factories.validate_rng(rng: int | None | numpy.random.Generator = None) → numpy.random.Generator

Validate Random Number Generator.

class snake.core.sampling.factories.VDSorder

Bases: snake._meta.NoCaseEnum

Available ordering for variable density sampling.

CENTER_OUT = 'center-out'

RANDOM = 'random'

TOP_DOWN = 'top-down'

class snake.core.sampling.factories.VDSpdf

Bases: snake._meta.NoCaseEnum

Available law for variable density sampling.

GAUSSIAN = 'gaussian'

UNIFORM = 'uniform'

EQUISPACED = 'equispaced'

snake.core.sampling.factories.get_kspace_slice_loc(dim_size: int, center_prop: int | float, accel: int = 4, pdf: snake.core.sampling.factories.VDSpdf = VDSpdf.GAUSSIAN, rng: int | None | numpy.random.Generator = None, order: snake.core.sampling.factories.VDSorder = VDSorder.CENTER_OUT) → numpy.ndarray

Get slice index at a random position.

Parameters:

• dim_size (int) – Dimension size

• center_prop (float or int) – Proportion of center of kspace to continuouly sample

• accel (float) – Undersampling/Acceleration factor

• pdf (str, optional) – Probability density function for the remaining samples. “gaussian” (default) or “uniform”.

• rng (random state)

Returns:

np.ndarray

Return type:

array of size dim_size/accel.

snake.core.sampling.factories.get_cartesian_mask(shape: tuple[int, ...], n_frames: int, rng: int | None | numpy.random.Generator = None, constant: bool = False, center_prop: float | int = 0.3, accel: int = 4, accel_axis: int = 0, pdf: snake.core.sampling.factories.VDSpdf = VDSpdf.GAUSSIAN) → numpy.ndarray

Get a cartesian mask for fMRI kspace data.

Parameters:

• shape (tuple) – shape of fMRI volume.

• n_frames (int) – number of frames.

• rng (random state) – Random number generator or seed.

• constant (bool) – If True, the mask is constant across time.

• center_prop (float) – Proportion of center of kspace to continuouly sample

• accel (float) – Undersampling/Acceleration factor

• pdf (str, optional) – Probability density function for the remaining samples. “gaussian” (default) or “uniform”.

• rng

Returns:

np.ndarray

Return type:

random mask for an acquisition.

snake.core.sampling.factories.flip2center(mask_cols: collections.abc.Sequence[int], center_value: int) → numpy.ndarray

Reorder a list by starting by a center_position and alternating left/right.

Parameters:

• mask_cols (list or np.array) – List of columns to reorder.

• center_pos (int) – Position of the center column.

Returns:

np.array

Return type:

reordered columns.

snake.core.sampling.factories.check_trajectory(trajectory: numpy.typing.NDArray, osf: int, gmax: float, smax: float) → numpy.bool_

Check if a trajectory is feasible or not.

snake.core.sampling.factories.vds_factory(shape: tuple[int, ...], acs: float | int, accel: int, accel_axis: int, order: snake.core.sampling.factories.VDSorder = VDSorder.CENTER_OUT, pdf: snake.core.sampling.factories.VDSpdf = VDSpdf.GAUSSIAN, rng: int | None | numpy.random.Generator = None) → numpy.ndarray

Create a variable density sampling trajectory.

Parameters:

• shape – Shape of the kspace.

• acs – autocalibration line number (int) or proportion (float)

• direction – Direction of the sampling.

• TR – Time to acquire the k-space. Exclusive with base_TR.

• base_TR – Time to acquire a full volume in the base trajectory. Exclusive with TR.

• pdf – Probability density function of the sampling. “gaussian” or “uniform”

• rng – Random number generator or seed.

Returns:

Variable density sampling trajectory.

Return type:

KspaceTrajectory

snake.core.sampling.factories.radial_factory(shape: tuple[int, ...], n_shots: int, n_points: int, expansion: str | None = None, n_repeat: int = 0, **kwargs: collections.abc.Mapping[str, Any]) → numpy.ndarray

Create a radial sampling trajectory.

snake.core.sampling.factories.stack_spiral_factory(shape: tuple[int, ...], accelz: int, acsz: int | float, n_samples: int, nb_revolutions: int, in_out: bool = True, spiral: str = 'archimedes', n_shot_slices: int = 1, orderz: snake.core.sampling.factories.VDSorder = VDSorder.CENTER_OUT, pdfz: snake.core.sampling.factories.VDSpdf = VDSpdf.GAUSSIAN, rng: int | None | numpy.random.Generator = None, rotate_angle: AngleRotation | float = 0.0) → numpy.ndarray

Generate a trajectory of stack of spiral.

class snake.core.sampling.factories.AngleRotation

Bases: float, snake._meta.NoCaseEnum

Available rotation angle for density sampling.

Initialization

Initialize self. See help(type(self)) for accurate signature.

ZERO = 0

GOLDEN = 2.3999632297286535

GOLDEN_MRI = 1.941678793

snake.core.sampling.factories.rotate_trajectory(trajectories: collections.abc.Generator[numpy.ndarray, None, None], theta: snake.core.sampling.factories.AngleRotation | float = 0) → collections.abc.Generator[numpy.ndarray, None, None]

Incrementally rotate a trajectory.

Parameters:

trajectories – Trajectory to rotate.

snake.core.sampling.factories.stacked_epi2d(shape: tuple[int, int, int], freq_locs: numpy.typing.NDArray, phase_locs: numpy.typing.NDArray, slice_locs: numpy.typing.NDArray) → numpy.typing.NDArray

Generate a list of 2D epi plane, stacked.

snake.core.sampling.factories.stacked_epi_factory(shape: tuple[int, int, int], acsz: int | float, accelz: int, orderz: snake.core.sampling.factories.VDSorder = VDSorder.CENTER_OUT, pdfz: snake.core.sampling.factories.VDSpdf = VDSpdf.GAUSSIAN, rng: int | None | numpy.random.Generator = None) → numpy.ndarray

Generate a VDS stack of fully sampled EPI trajectory.

snake.core.sampling.factories.evi_factory(shape: tuple[int, int, int]) → numpy.ndarray

Generate a Echo Volume Imaging trajectory.

snake.core.sampling.factories.trajectory_generator(traj_factory: collections.abc.Callable[..., numpy.ndarray], shape: tuple[int, ...], **kwargs: Any) → collections.abc.Generator[numpy.ndarray, None, None]

Generate a trajectory.

Parameters:

• traj_factory – Trajectory factory function.

• n_batch – Number of shot to deliver at once.

• kwargs – Trajectory factory kwargs.

Yields:

np.ndarray – Kspace trajectory.