cr.sparse.lop.Operator

class cr.sparse.lop.Operator(times: Callable[[jax.Array], jax.Array], trans: Callable[[jax.Array], jax.Array], shape: Tuple[int, int], linear: bool = True, jit_safe: bool = True, matrix_safe: bool = True, real: bool = True)

Represents a finite linear operator \(T : A -> B\) where \(A\) and \(B\) are finite vector spaces.

Parameters

• times – A function implementing \(T(x)\)

• trans – A function implementing \(T^H (x)\)

• m – The dimension of the destination vector space \(B\)

• n – The dimension of the source vector space \(A\)

• linear – Indicates if the operator is linear or not

• jit_safe – Indicates if the operator can be safely JIT compiled

• matrix_safe – Indicates if the operator can accept a matrix of vectors

• real – Indicates if a linear operator is real i.e. has a matrix representation of real numbers

Note

While most of the operators in the library are linear operators, some are not. Prominent examples include operators like real part operator, imaginary part operator. These operators are provided for convenience.

Most operators in this collection are real.

Attributes

input_ndim	Returns the number of dimensions of input to the operator
input_shape	Returns the shape of input to the operator as a tuple
input_size	Returns the size of input to the operator
jit_safe	Indicates if the times and trans functions can be safely jit compiled
linear	Indicates if the operator is linear or not
matrix_safe	Indicates if the operator can accept a matrix of vectors
output_ndim	Returns the number of dimensions of output of the operator
output_shape	Returns the shape of output of the operator as a tuple
output_size	Returns the size of output of the operator
real	Indicates if a linear operator is real i.e. has a matrix representation of real numbers.
shape	Dimension of the linear operator (m, n)
times	A linear function mapping from A to B
trans	Corresponding adjoint linear function mapping from B to A