Case 2: Batch-varying constant @ variable using interleaved structure. Each batch element uses a DIFFERENT constant matrix.

Batch-varying parametric constant @ variable: interleaved matrix structure. Each batch element uses a different slice of the paramet

2D constant @ ND variable: use Kronecker structure I_n ⊗ C ⊗ I_batch. This is the most common case: a 2D constant matrix multiplies

Case 3: 2D constant @ ND variable using Kronecker structure. The SAME constant matrix C is applied to all batch elements. Us

Parametric constant @ variable: expand each parameter slice with Kronecker. When the constant depends on Parameters, we expand each

Case 1: Parametric constant @ variable. The constant is a cp.Parameter - values unknown at canonicalization. Each parameter

Collect atom types that need conic canonicalization under quad_obj=True. Mirrors the Dcp2Cone.canonicalize_tree walk: atoms in the affine head

Pad Q to num_vars x num_vars for cuOpt SOC variable permutation.

.. index: default :refguide: something, else, and more

func_name : Descriptive text continued text another_func_name : Descriptive text func_name1, func_name2, :meth:`f

Grab signature (if given) and summary

Promote bound entries to canonical form. None → 0-d array (-inf/inf), scalars → 0-d arrays, sparse → COO arrays, dense arrays kept as

Frobenius norm of (sign * A - L[p, :] @ L[p, :].T), used to validate QDLDL's factorization when zero-pivot columns were dropped.

Reshape non-parametric constant data from column to matrix format. For a column vector of shape (p * m * k, 1), reshape to (p * m, k

Reshape parametric constant data from column to matrix format. For parametric data, entries may be duplicated by broadcast operation

Validate and resolve a solver argument to a concrete Solver instance. Parameters ---------- solver : str, Solver, or None The sol

Batch-varying variable @ constant: use interleaved matrix structure. When the constant has batch dimensions (e.g., X(B,m,k) @ C(B,k,

ND rmul Case 2: Batch-varying rmul using interleaved structure. Each batch element uses a DIFFERENT constant matrix. Exampl

Batch-varying variable @ parametric constant: interleaved matrix structure. Each batch element uses a different slice of the paramet

ND variable @ 2D constant: use Kronecker structure C.T ⊗ I_{batch*m}. This is the most common case: a 2D constant matrix multiplies

ND rmul Case 3: ND variable @ 2D constant using Kronecker structure. The SAME constant matrix C is applied to all batch elements.

Parametric variable @ constant: expand each parameter slice with Kronecker. When the constant depends on Parameters, we expand each

ND rmul Case 1: Variable @ parametric constant. The constant is a cp.Parameter - values unknown at canonicalization. Each pa

Fast path for select_rows when there are no duplicate rows. Uses a reverse mapping approach: for each old row, compute which new row

General path for select_rows that handles duplicate rows. This method is used when broadcasting causes the same source row to be

Set model solve parameters.

Generate a random initial point for DNLP problems. A variable is initialized randomly if: 1. 'sample_bounds' is set for that variable. 2.

Skip tests that exceed the XPRESS Community license size limit.

Solve and return a solution if one exists.

Solve a problem using multiple solvers. Arguments --------- solvers : list of (str, dict) tuples or strings. The

\internal */

Key sparse Constant values without converting them to object arrays.

Given a stacked lhs with the following entries: [[a11, a12], [a21, a22], ... Apply the Kronecker product wi

Generate a member listing, autosummary:: table where possible, and a table where not.

Sum a single operand along specified axes. This function handles the case where there is only one input operand. It sums the operand along ax

Determines whether the current atom is implemented in C++. This method should be overridden in derived atom classes that are not impl

Helper function for partial transpose. Parameters ---------- expr : :class:`~cvxpy.expressions.expression.Expression` The 2D expr

Helper function for partial trace. Parameters ---------- expr : :class:`~cvxpy.expressions.expression.Expression` The 2D expressi

Given v, which is a stacked matrix of shape (p * n, m), transpose each slice of v, returning a stacked matrix of shape (p * m, n).

Update the positions and provisional input_sets of ``results`` based on performing the contraction result ``best``. Remove any involving the t

Return True if this class uses Atom.name without override.

Validate the input arguments for einsum operation. This function checks that the number of expressions matches the number of input subscript

Validate the sparsity pattern for a leaf node. Parameters: indices: List or tuple of indices indicating the positions of non

Validate a single sparse bound entry. Checks that the sparse bound has matching shape and sparsity pattern. Raises ------