Changelog¶
0.4.6 - 2022-01-11¶
Fixed
Issue #46 - speed up weight calculation
0.4.5 - 2022-01-09¶
Added
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_edgesandgraph.num_verticesto new methodsgraph.number_of_edgesandgraph.number_of_vertices.
Added
Issue #3 - add
total_volumemethod tosgtl.graph.GraphIssue #2 - add
_check_vert_nummethod tosgtl.graph.GraphIssue #9 - add methods for converting to and from
networkxgraph objects.Issue #6 - add cheeger cut algorithm
Fixed
0.3.3 - 2021-11-12¶
Fixed
Issue #4 - definition of normalised laplacian was incorrect
0.3.2 - 2021-11-11¶
Fixed
The
ssbmmethod will not allow non-float probabilitiesThe
Graph.weightmethod will give the correct value when sets are equalPrevent
sbmmethod from generating duplicate edges
0.3 - 2021-11-11¶
Changed
Changed the name of the
sbmmodule torandom.
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
sbmmodule for generating graphs from the stochastic block model
0.1.1 - 2021-11-10¶
Added
Graphclass for representing graphsclustering.spectral_clusteringclustering method