Changelog#

0.5.0 - 2024-01-09#

Fixed

  • Update sklearn dependency to scikit-learn

0.4.6 - 2022-01-11#

Fixed

0.4.5 - 2022-01-09#

Added

  • Issue #32 - add tensor product method for combining graphs

  • Issue #27 - add option to plot spectrum of graph

  • Issue #31 - add graphs together

0.4.4 - 2022-01-07#

For brevity, the log below includes changes made in versions 0.4.1, 0.4.2, 0.4.3, and 0.4.4.

Added

  • Issue #25 - add methods for reading and writing edgelist files

  • Added Graph.adjacency_matrix() method.

  • Issue #29 - add spectrum methods for adjacency and laplacian matrices.

  • Issue #35 - add method to construct k nearest neighbour graph.

  • Issue #41 - construct graph using gaussian kernel function with threshold.

Fixed

  • Issue #39 - KNN graph construction should work with sparse data matrices.

0.4 - 2021-12-14#

Changed

  • Issue #3 - move members graph.num_edges and graph.num_vertices to new methods graph.number_of_edges and graph.number_of_vertices.

Added

  • Issue #3 - add total_volume method to sgtl.graph.Graph

  • Issue #2 - add _check_vert_num method to sgtl.graph.Graph

  • Issue #9 - add methods for converting to and from networkx graph objects.

  • Issue #6 - add cheeger cut algorithm

Fixed

  • Issue #10 - correct graph.weight calculation when edges have floating point weights or self-loops

  • Issue #3 - correct number of edges for weighted graphs

  • Issue #8 - return a meaningful error when computing the conductance or bipartiteness of the empty set

0.3.3 - 2021-11-12#

Fixed

  • Issue #4 - definition of normalised laplacian was incorrect

0.3.2 - 2021-11-11#

Fixed

  • The ssbm method will not allow non-float probabilities

  • The Graph.weight method will give the correct value when sets are equal

  • Prevent sbm method from generating duplicate edges

0.3 - 2021-11-11#

Changed

  • Changed the name of the sbm module to random.

Added

  • Generate standard graphs: complete, cycle, path, star.

  • Generate Erdos-Renyi graphs

Fixed

  • Spectral clustering checks that number of clusters and eigenvectors are positive integers.

0.2 - 2021-11-10#

Added

  • Changelog page to documentation

  • sbm module for generating graphs from the stochastic block model

0.1.1 - 2021-11-10#

Added

  • Graph class for representing graphs

  • clustering.spectral_clustering clustering method