Implements 2D based matrix visualisation.
#include <TMatrixVis.hh>
Inheritance diagram for T2DMatrixVis:
Public Member Functions | |
| T2DMatrixVis () | |
| construct matrix visualisation object | |
| T2DMatrixVis & | border (const bool b) |
| turn on/off printing of matrix block border (default: on) | |
| T2DMatrixVis & | structure (const bool b) |
| turn on/off printing of structure of matrix (default: on) | |
| T2DMatrixVis & | indexset (const bool b) |
| turn on/off printing of indexsets (default: off) | |
| T2DMatrixVis & | id (const bool b) |
| turn on/off printing of id (default: off) | |
| T2DMatrixVis & | nonempty (const bool b) |
| turn on/off printing of only nonempty blocks (default: off) | |
| T2DMatrixVis & | only_local (const bool b) |
| turn on/off printing of local/non-local matrix blocks (default: off) | |
| T2DMatrixVis & | entries (const bool b) |
| turn on/off printing of matrix coefficients (default: off) | |
| T2DMatrixVis & | pattern (const bool b) |
| turn on/off printing of sparsity pattern (default: off) | |
| T2DMatrixVis & | sparse_pattern (const bool b) |
| turn on/off printing of sparsity pattern of sparse matrices only (default: off) | |
| T2DMatrixVis & | svd (const bool b) |
| turn on/off printing of SVD of matrix-blocks (default: off) | |
| T2DMatrixVis & | rank_col (const bool b, const std::string &cmap="jet") |
| turn on/off colouring matrix blocks according to rank (default: off) | |
| T2DMatrixVis & | max_size_ratio (const double r) |
| T2DMatrixVis & | svd_ref (const double ref_max, const double ref_min=Limits::epsilon< double >()) |
| virtual void | print (const TMatrix *A, const std::string &filename) const |
| print matrix A to file filename | |
Member Function Documentation
| T2DMatrixVis& max_size_ratio | ( | const double | r | ) |
set maximal allowed ratio of block size compared to largest block size (default: 1000); if ratio is exceeded, block will not be printed to limit file size for very large matrices
| T2DMatrixVis& svd_ref | ( | const double | ref_max, |
| const double | ref_min = Limits::epsilon< double >() |
||
| ) |
set value of largest/smallest singular value to compare with for each block;
- if ref_max/ref_min ≤ 0, the largest/smallest local singular values per block are used
- good choice for ref_max is ∥A∥₂
