Skip to content
Compare
Choose a tag to compare
@jcpassy jcpassy released this 27 Jan 20:25
Merge branch 'develop'

* develop: (24 commits)
  Added build status tag.
  Added example and jupyter notebook. Tidying up everything…
  Rewrote the Python extension quite a bit. Can now reproduce the results from the C++ tests.
  Fixed the Python tests and the files used for the benchmark.
  Fixed benchmark runner. Additional: * added COPYRIGHT * updated README * minor fixes
  Documentations: * updated Doxygen doc * added Sphinx doc
  Removed legacy code + some cython cleaning.
  Cleaning Windows compilation by adding headers hack for debug config. Additional: * reduced tests duration by limiting tests to 100 iterations * removed legacy tests
  Added benchmark functionality from Caterina. Additional: * cmake updated * minor cleaning and updating the README
  Reorganized cmake and repository for making the code cross-platform and setting the CI. This includes: * builds passing on Win10, OSX, and Linux * python extension, including the infrastructure to build wheels * cleaning of the cmake
  Added a python wrapper to the algorithm.
  Added assortative case. Additional: * many tests including some for the algorithm types * added template for initializing w * major refactoring * printing out likelihood * minor review comments
  Implemented case when affinity is specified in a file. Additional: * added test for undirected graphs * cleaning + bug fixes in the solver and the application * a bit more consistency + comments as suggested by Ivan
  Improved interface, writing files is now part of the application. Additional: * random generator is now a template. Added CLI for the user to specify the seed * added tensor tests. * labels must be extracted to write the output files * writing files is now part of the application. * implementation of undirected graphs;
  Some general improvements: * added template types (vertex_t, weight_t…) * reorganized tests inside fixtures * added report class * added termination reason * initialization is now a template of the solver * added report output file * API simplified * more tests * documentation updated
  Removing unnecessary Makefile
  Added a lot of unit tests and ctests. Additional: * refactored of the network logic * comments from review * moved headers to separate folder. * added methods writing outputs into files
  Awesome documentation and consistent versioning across the project. Additional: * number of groups is now a required argument. * fixed some warnings during compilation
  Some additional changes: * reverted back legacy code * added timer * removed dependency on boost program options - not easy to handle required arguments
  First implementation finished. Reproduces the results for an non-assortative, undirected network with the adjacency data file.
  ...