Releases: pnnl/HyperNetX
Releases · pnnl/HyperNetX
HNX v2.2.0p
This is release for the HyperNetX paper to be reviewed and published under the Journal of Open Source Software (JOSS).
The main difference between v2.2.0 and v2.2.0p is that this release contains documents, images, and scripts related to the HyperNetX paper. There are no source code changes.
Full Changelog: v2.2.0...v2.2.0p
HNX v2.2.0
Improvements
- Update drawing submodule
- Reorganize tutorials
- Update documentation, README
- Update dependencies
- Update CI/CD scripts, Makefile
- Updated tests
Fixes
- Fixes minor bugs in Hypergraph constructor and helper method
What's Changed
New Contributors
Full Changelog: v2.1.4...v2.2.0
HNX v.2.1.4
Fixes
Full Changelog: v2.1.3...v2.1.4
HNX v2.1.3
HNX v2.1.2
HNX v.2.1.1
HNX v2.1.0
Improvements, Enhancements
- Refactor hypergraph_modularity; add tutorial; thank you @ftheberge
- Add conductance, tests; thank you @NicolBuratti @uninicol
Full Changelog: v2.0.5...v2.1.0
Modularity module updates
Unchanged functions:
- dict2part(D)
- part2dict(A)
- linear(d, c)
- majority(d, c)
- strict(d, c)
- two_section(HG)
- kumar(HG, delta=0.01, verbose=False)
No longer required
- precompute_attributes(H)
- _compute_partition_probas(HG, A)
- _degree_tax(HG, Pr, wdc)
- _edge_contribution(HG, A, wdc)
- _delta_ec(HG, P, v, a, b, wdc)
- _bin_ppmf(d, c, p)
- _delta_dt(HG, P, v, a, b, wdc)
New version
- modularity(HG, A, wdc=linear)
- last_step(HG, L, wdc=linear, delta=0.01, verbose=False)
New (hidden) functions
- _last_step_unweighted
- _last_step_weighted
HNX v2.0.5
Fixes
- Updates docstrings and documentation examples for
Hypergraph.collapse_nodes
andHypergraph.collapse_nodes_and_edges
methods - Enables deprecation warnings for Hypergraph class to be displayed on REPL, Jupyter Notebooks
Full Changelog: v2.0.4...v2.0.5
HNX v2.0.4
Fixes
- Fixes errors in Jupyter where H.dataframe and H.dual().dataframe had common references.
Full Changelog: v2.0.3...v2.0.4