spacetransformer.core.validation
core.validation
Validation utilities for SpaceTransformer parameters.
This module provides validation functions for common SpaceTransformer parameters including points, point sets, transformation matrices, spaces, and image data. These functions ensure data integrity and provide clear error messages.
Design Philosophy: Provides a set of core validation functions that can be reused across the library to ensure consistent error handling. Each function focuses on validating a specific data type, with helpful error messages to guide users in correcting their inputs.
Example: Validating inputs in a function:
>>> from spacetransformer.core.validation import validate_pointset, validate_space
>>> def warp_points(points, source, target):
... points_np = validate_pointset(points)
... source = validate_space(source)
... target = validate_space(target)
... # Proceed with validated inputsFunctions
| Name | Description |
|---|---|
| validate_bbox | Validate a bounding box for space operations. |
| validate_image_data | Validate image data array. |
| validate_orientation | Validate a single 3D point. |
| validate_orthonormal_basis | Validate that three orientation vectors form an orthonormal basis. |
| validate_point | Validate a single 3D point. |
| validate_pointset | Validate a set of 3D points. |
| validate_shape | Validate image shape parameters. |
| validate_space | Validate a Space object. |
| validate_spacing | Validate voxel spacing parameters. |
| validate_transform_matrix | Validate a 4x4 transformation matrix. |
validate_bbox
core.validation.validate_bbox(bbox, check_int=False)Validate a bounding box for space operations.
Args: bbox: Input bounding box to validate, should have shape (3, 2) check_int: Whether to require integer values in the bbox name: Parameter name for error messages
Returns: np.ndarray: Validated bbox as a numpy array with shape (3, 2)
Raises: ValidationError: If bbox is invalid
Example: >>> validate_bbox([[10, 90], [20, 80], [5, 45]]) array([[10, 90], [20, 80], [5, 45]]) >>> validate_bbox([[10.5, 90.2], [20.3, 80.1], [5.0, 45.9]], check_int=False) array([[10.5, 90.2], [20.3, 80.1], [5. , 45.9]])
validate_image_data
core.validation.validate_image_data(image, expected_ndim=None, name='image')Validate image data array.
Args: image: Input image data to validate expected_ndim: Expected number of dimensions (None to skip check) name: Parameter name for error messages
Returns: np.ndarray: The validated image data
Raises: ValidationError: If the image data is invalid
Example: >>> import numpy as np >>> image = np.zeros((100, 100, 50)) >>> validate_image_data(image, expected_ndim=3).shape (100, 100, 50) >>> validate_image_data(image, expected_ndim=2) # Will raise ValidationError
validate_orientation
core.validation.validate_orientation(orientation, name='orientation')Validate a single 3D point.
Args: orientation: Input orientation to validate name: Parameter name for error messages
Returns: np.ndarray: Validated point as a numpy array
Raises: ValidationError: If point is invalid
Example: >>> validate_orientation([1, 2, 3]) array([1., 2., 3.]) >>> validate_orientation([1, 2]) # Will raise ValidationError
validate_orthonormal_basis
core.validation.validate_orthonormal_basis(
x_orientation,
y_orientation,
z_orientation,
)Validate that three orientation vectors form an orthonormal basis.
This function checks if the given vectors form a proper orthonormal right-handed basis by examining properties of the rotation matrix they form.
Args: x_orientation: X-axis orientation vector (unit vector) y_orientation: Y-axis orientation vector (unit vector) z_orientation: Z-axis orientation vector (unit vector)
Raises: ValidationError: If vectors don’t form an orthonormal right-handed basis
validate_point
core.validation.validate_point(point, name='point')Validate a single 3D point.
Args: point: Input point to validate name: Parameter name for error messages
Returns: np.ndarray: Validated point as a numpy array
Raises: ValidationError: If point is invalid
Example: >>> validate_point([1, 2, 3]) array([1., 2., 3.]) >>> validate_point([1, 2]) # Will raise ValidationError
validate_pointset
core.validation.validate_pointset(points, name='points', allow_empty=False)Validate a set of 3D points.
Args: points: Input points to validate, should have shape (N, 3) name: Parameter name for error messages allow_empty: Whether to allow empty point sets
Returns: np.ndarray: Validated points as a numpy array with shape (N, 3)
Raises: ValidationError: If points are invalid
Example: >>> validate_pointset([[1, 2, 3], [4, 5, 6]]) array([[1., 2., 3.], [4., 5., 6.]]) >>> validate_pointset([1, 2, 3]) # Single point, will be reshaped array([[1., 2., 3.]])
validate_shape
core.validation.validate_shape(shape)Validate image shape parameters.
Args: shape: Image dimensions as (width, height, depth) or similar sequence
Raises: ValidationError: If shape is invalid
Example: >>> validate_shape((100, 100, 50)) # No exception raised >>> validate_shape([512, 512, 200]) # Lists are also accepted >>> validate_shape((100, 100)) # Will raise ValidationError
validate_space
core.validation.validate_space(space, name='space')Validate a Space object.
Args: space: Input Space object to validate name: Parameter name for error messages
Returns: Space: The validated Space object
Raises: ValidationError: If the input is not a valid Space object
Example: >>> from spacetransformer.core import Space >>> space = Space(shape=(100, 100, 50)) >>> validate_space(space) is space True >>> validate_space(“not a space”) # Will raise ValidationError
validate_spacing
core.validation.validate_spacing(spacing)Validate voxel spacing parameters.
Args: spacing: Voxel spacing as (x_spacing, y_spacing, z_spacing) in mm
Raises: ValidationError: If spacing is invalid
Example: >>> validate_spacing((1.0, 1.0, 2.0)) # No exception raised >>> validate_spacing([0.5, 0.5, 1.0]) # Lists are also accepted >>> validate_spacing((1.0, 1.0)) # Will raise ValidationError
validate_transform_matrix
core.validation.validate_transform_matrix(
matrix,
name='transform_matrix',
check_invertible=True,
)Validate a 4x4 transformation matrix.
Args: matrix: Input matrix to validate name: Parameter name for error messages check_invertible: Whether to check if the matrix is invertible
Returns: np.ndarray: Validated transformation matrix
Raises: ValidationError: If matrix shape is invalid NumericalError: If matrix is not invertible (when check_invertible=True)
Example: >>> validate_transform_matrix(np.eye(4)) array([[1., 0., 0., 0.], [0., 1., 0., 0.], [0., 0., 1., 0.], [0., 0., 0., 1.]]) >>> validate_transform_matrix(np.eye(3)) # Will raise ValidationError