Pecanpy

Summary

New features
* Implemented fast and memory efificent first-order walk modes ``PreCompFirstOrder`` and ``FirstOrderUnweighted`` (when ``p = q = 1``, or even faster when the graph is unweighted)
* Added new ``gamma`` hyperparameter for _node2vec+_

Others
* Setup automated testing
* Refactor ``graph`` objects

What's Changed
* Refactor graph by RemyLau in https://github.com/krishnanlab/PecanPy/pull/43
* Refactor graph construction by RemyLau in https://github.com/krishnanlab/PecanPy/pull/50
* Test cli and remove deprecated options by RemyLau in https://github.com/krishnanlab/PecanPy/pull/51
* Refactor rwgraph by RemyLau in https://github.com/krishnanlab/PecanPy/pull/52
* First order walk by RemyLau in https://github.com/krishnanlab/PecanPy/pull/53
* Automated testing by RemyLau in https://github.com/krishnanlab/PecanPy/pull/56
* Fix windows by RemyLau in https://github.com/krishnanlab/PecanPy/pull/57
* Cleanup configs by RemyLau in https://github.com/krishnanlab/PecanPy/pull/58
* Gamma hyperparam by RemyLau in https://github.com/krishnanlab/PecanPy/pull/60


**Full Changelog**: https://github.com/krishnanlab/PecanPy/compare/v2.0.1...v2.0.2

New features
* Progress bar via [``numba-progress``](https://pypi.org/project/numba-progress/), enable by specifying ``--verbose``
* Read and write CSR sparse graph as ``.npz`` (with the convention of ``.csr.npz`` for csr graph and ``.dense.npz`` for dense graph)
* Convert weighted dense graph to unweighted dense graph

Minor changes
* Refactor ``Timer``
* Enforce BLACK formatting

- Node2vec+ implementation.
- Several fixes of node2vec with improved usage of PecanPy from Python script.

* Improved the documentation of the package and hosted on readthedoc
* Enforced coding style consistency through flake
* Fixed bug in setting in-out bias according to q
* Fixed gensim.word2vec keywords back compatibility for gensim 4.0.1