-----
* docs: update docstring for compress/decompress + reference
* test: add size restriction
* test: add reconstitution test
* chore: add Dockerfile for manylinux distribution
* Revert "perf: ~1/3 faster connected components algorithm"
* refactor: remove redundant array inititalizations
* perf: ~1/3 faster connected components algorithm
* fix: used the right implementation of compresso (from experiments)
* chore: update gitignore
* refactor: setup such that it is packageable
* Fixed installation instructions
* Updated requirements.txt to work with conda
* README: fixed formatting
* Removed wrong instructions
* README: added datasets
* Updated README
* Removed old compression location
* requirements.txt: moved to conda environment
* experiments: updated all files to work with new framework
* .gitignore: ignore temp cython files
* .gitignore: ignore the results folder
* compression: removed failed methods
* compression: removed old files
* README: fixed formatting
* README: updated instructions
* \_LEGACY: removed old outdated experiments
* Mention paper explicitly
* Minimize
* Update results
* Update name
* Add link to paper
* Update name
* Update name
* Add plotting script
* Fix failure when using lowercase names
* Add missing no-compression class
* Sanity assertion runs only once
* Added experiment runner. Fixed a bug. Updated README
* Set to root path
* Update to latest banner
* Add logo
* Update references visually
* Update figure composition to save space
* Bockwurst encoding and experiments
* Initial commit