Skip to main content

'ANOVA supercharged' supercharged

Project description

Peroxymanova

This project is essentially an implementation of PERMANOVA (wiki link, doi:10.1111/j.1442-9993.2001.01070.pp.x) in Rust.

PERMANOVA (Permutational Multivariate Analysis of Variance) is a method for comparing groups of mathematical objects, requiring only a dissimilarity matrix between them, as opposed to having a notion of an average, like the one used in classical ANOVA. This is incredibly useful, since it is massively easier to define a dissimilarity than a mean: there is no obvious "average graph", "average neural network" or an "average RL policy", but with a little bit of hand waving one can define distances, or dissimilarities between a pair of such entities.

This package aims to provide quality-of-life bells and whistles that turn this incredible method into something useful day to day. It implements the following workflow:

  1. Accept a Collection of some things T, a Callable that can compare two of those T, returning a float, and a Collection of labels that indicate to which group a given thing T belongs to
  2. Efficiently run the Callable[[T,T], float] for every possible pair of objects in the Collection and build a dissimilarity matrix
  3. Given the dissimilarity matrix and a Collection of group-indicating labels, run the PERMANOVA algorithm to get a test statistic and a p-value for the null hypothesis of the groups being all and the same. This step requires a lot of permutations to get the p-value so run it blazingly fast in Rust

Strategic roadmap:

  • Actually make it multivariate since a single p-value for a single "difference" is kind of just one-way permutational ANOVA.
  • Make a fancy parallelization backend interface for computing pairwise distances. Maybe there could be a backend='ray' that would actually search for a full ray cluster?
  • Since we dream of ray, should we get a cluster for rust side as well? :)

Development

Releasing

Since were using pyproject.toml, the source of truth for the version is the file, and the version tag is based on the project.version attribute. It means bumping the version necessitates a separate commit. At least there is a way to ensure the version tag consistency with the version attribute in the file. Use:

git config --local include.path ../.gitconfig

to add the repos' .gitconfig with an ugly but nifty "git bump" alias

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

peroxymanova-0.3.1.tar.gz (20.8 kB view hashes)

Uploaded Source

Built Distributions

peroxymanova-0.3.1-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded PyPy manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded PyPy manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-pp310-pypy310_pp73-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded PyPy manylinux: glibc 2.5+ i686

peroxymanova-0.3.1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded PyPy manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded PyPy manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-pp39-pypy39_pp73-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded PyPy manylinux: glibc 2.5+ i686

peroxymanova-0.3.1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded PyPy manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-pp38-pypy38_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded PyPy manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-pp38-pypy38_pp73-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded PyPy manylinux: glibc 2.5+ i686

peroxymanova-0.3.1-cp312-none-win_amd64.whl (164.5 kB view hashes)

Uploaded CPython 3.12 Windows x86-64

peroxymanova-0.3.1-cp312-none-win32.whl (155.8 kB view hashes)

Uploaded CPython 3.12 Windows x86

peroxymanova-0.3.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded CPython 3.12 manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded CPython 3.12 manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded CPython 3.12 manylinux: glibc 2.5+ i686

peroxymanova-0.3.1-cp312-cp312-macosx_11_0_arm64.whl (278.7 kB view hashes)

Uploaded CPython 3.12 macOS 11.0+ ARM64

peroxymanova-0.3.1-cp312-cp312-macosx_10_12_x86_64.whl (287.2 kB view hashes)

Uploaded CPython 3.12 macOS 10.12+ x86-64

peroxymanova-0.3.1-cp311-none-win_amd64.whl (166.0 kB view hashes)

Uploaded CPython 3.11 Windows x86-64

peroxymanova-0.3.1-cp311-none-win32.whl (157.9 kB view hashes)

Uploaded CPython 3.11 Windows x86

peroxymanova-0.3.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded CPython 3.11 manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded CPython 3.11 manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded CPython 3.11 manylinux: glibc 2.5+ i686

peroxymanova-0.3.1-cp311-cp311-macosx_11_0_arm64.whl (280.9 kB view hashes)

Uploaded CPython 3.11 macOS 11.0+ ARM64

peroxymanova-0.3.1-cp311-cp311-macosx_10_12_x86_64.whl (291.0 kB view hashes)

Uploaded CPython 3.11 macOS 10.12+ x86-64

peroxymanova-0.3.1-cp310-none-win_amd64.whl (166.0 kB view hashes)

Uploaded CPython 3.10 Windows x86-64

peroxymanova-0.3.1-cp310-none-win32.whl (158.1 kB view hashes)

Uploaded CPython 3.10 Windows x86

peroxymanova-0.3.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded CPython 3.10 manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded CPython 3.10 manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded CPython 3.10 manylinux: glibc 2.5+ i686

peroxymanova-0.3.1-cp310-cp310-macosx_11_0_arm64.whl (281.2 kB view hashes)

Uploaded CPython 3.10 macOS 11.0+ ARM64

peroxymanova-0.3.1-cp310-cp310-macosx_10_12_x86_64.whl (290.9 kB view hashes)

Uploaded CPython 3.10 macOS 10.12+ x86-64

peroxymanova-0.3.1-cp39-none-win_amd64.whl (166.2 kB view hashes)

Uploaded CPython 3.9 Windows x86-64

peroxymanova-0.3.1-cp39-none-win32.whl (158.4 kB view hashes)

Uploaded CPython 3.9 Windows x86

peroxymanova-0.3.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded CPython 3.9 manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded CPython 3.9 manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded CPython 3.9 manylinux: glibc 2.5+ i686

peroxymanova-0.3.1-cp38-none-win_amd64.whl (166.1 kB view hashes)

Uploaded CPython 3.8 Windows x86-64

peroxymanova-0.3.1-cp38-none-win32.whl (159.4 kB view hashes)

Uploaded CPython 3.8 Windows x86

peroxymanova-0.3.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view hashes)

Uploaded CPython 3.8 manylinux: glibc 2.17+ x86-64

peroxymanova-0.3.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view hashes)

Uploaded CPython 3.8 manylinux: glibc 2.17+ ARM64

peroxymanova-0.3.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.whl (1.1 MB view hashes)

Uploaded CPython 3.8 manylinux: glibc 2.5+ i686

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page