0.0.2-alpha release

ATTENTION This software is in alpha. APIs are liable to change, along with the set of supported features. Please do not use this software in production at this time.

Features in this release, and planned features for the future (excerpt from project readme)

Features

Current

Python 3.X

• read/write access to the file formats:
  • mtx
  • bxf
  • valcol
  • mat
• matrix operations
  • conversion of matrix to/from scipy.sparse and numpy dense matrix formats
  • append (COO) element
  • retrieve (COO) element by index
  • retrieve indicies which match a COO element
  • remove COO element by index
  • retrieve matrix values by (row, col) coordinate pair
  • modify matrix values by (row, col) coordinate pair
  • remove elements from the matrix whose value is zero
  • replace matrix contents with a scipy.sparse matrix
  • check if any elements are stored in the upper/lower triangle of the matrix
  • check if the matrix is symmetric
  • convert the matrix to symmetric/asymmetric
  • sort matrix elements to row-major
  • transpose the matrix
• HercExplorer interface
  • wrappers for all of the above
  • paint values by (row, col) pair, in a rectangle pattern, or a diagonal pattern
  • retrieve values by (row, col) pair, in a rectangular region, by row, or by col
  • display graphical (via matplotlib) and pseudo-graphical overviews of the matrix
  • modify matrix dimensions

C

• read matrices stored in BXF format

BXF

• store symmetric and asymmetric matrices in COO format
• store comments about matrix in REMARKS field

Planned

Python 3.X

• read/write support for harwell-boeing format
• read/write support for serialized python objects
• matrix computation operations (eg. computing matrix eigenvector)

C

• read matricies stored in valcol format
• read+write support for BXF and valcol
• matrix transpose operation
• COO-to-CSR and CSR-to-COO operations
• sort matrix elements to row major

BXF

• error checking

0.0.1 ALPHA release

Overview

This is the first release of the HeRC Matrix Tools. It represents a number of exciting new changes, updates and features. See below for a more extensive list of changes and features. Keep in mind, this is not a production ready release by any stretch of the imagination, but it represents another step in that direction.

Major Milestones

• consistent use of Doxygen compatible comments throughout Python code (C++ code is next)
• drastically improved interface, with an updated help system, and plugin support
• support for a new file format, valcol (used internally within the HeRC lab, but other may find use for it thanks to it's simplicity)

Features supported as of this release

• Python 3.X libraries provided IO for the following file formats: valcol, bxf, mat, and mtx
• Full scipy interoperability for libHercMatrix
• Most HercExplorer commands now wrap functions from various libraries, so you can now automate things easily without making HercExplorer process commands directly
• Wide variety of matrix operations and manipulation available via libraries and HercExplorer (transpose, resize, modify values individually, paint values, paint values diagonally)