Sparse Matrix Manipulation in Libigl

Here I’ve compiled a list of functions to manipulate Eigen::SparseMatrix in Libigl, due to the lack of documentation.

Up-to-date with commit 1f5a0c1.

Creation

cat.h

Concatenate two SparseMatrix.

diag.h

Extract the diagonal entries of a SparseMatrix into a Vector and vice-versa. This is superceded by

VectorXd V = X.diagonal() and
SparseVector<double> V = X.diagonal().sparseView()
SparseMatrix<double> X = V.asDiagonal().sparseView()

invert_diag.h

Invert the diagonal entries of a SparseMatrix.

repdiag.h

Repeat a SparseMatrix on the diagonal n times to form a new SparseMatrix.

repmat.h

Repeat a SparseMatrix along the columns and rows to form a r by c SparseMatrix.

sparse.h

Create a SparseMatrix from indices and values. It’s a simple wrapper to the Eigen::SparseMatrix::setFromTriplets method.

sparse_cached.h

Faster version of sparse using cached data.

speye.h

Create a sparse identity matrix.

Predicate

is_sparse.h

Determine if a matrix is SparseMatrx.

is_symmetric.h

Determine if a SparseMatrix is symmetric.

is_diag.h

Determine if a SparseMatrix is diagonal.

Coefficient-wise Operation

for_each

Apply a function for each non-zero element. Could be replaced with Eigen::SparseMatrix::unaryExpr.

Note that Eigen::SparseMatrix also provides a lot of coefficient-wise operations to use.

Reduction and Visitor

all.h

Return true if all coefficients in a row/column of a SparseMatrix are true.

any.h

Return true if any coefficient in a row/column of a SparseMatrix is true.

count.h

Return the number of coefficients that are true in each row/column in a given SparseMatrix.

find.h

Get the values and indices of non-zero elements in a SparseMatrix. It’s a convinient wrapper to:

SparseMatrix<double> mat(rows,cols);
for (int k=0; k<mat.outerSize(); ++k)
  for (SparseMatrix<double>::InnerIterator it(mat,k); it; ++it)
  {
    it.value();
    it.row();   // row index
    it.col();   // col index (here it is equal to k)
    it.index(); // inner index, here it is equal to it.row()
  }