semsimian

Semsimian is a package to provide fast semantic similarity calculations for ontologies. It is a Rust library with a Python interface.

This includes implementation of Jaccard and Resnik similarity of terms in an ontology, as well as a method to calculate the similarity of two sets of terms (so-called termset similarity). Other methods will be added in the future.

Semsimian is currently integrated into OAK and the Monarch app to provide fast semantic similarity calculations.

Installation

Set up your virtual environment of choice.
cd semsimian (home directory of this project)
pip install maturin
maturin develop
python

Python 3.9.16 (main, Jan 11 2023, 10:02:19) 
[Clang 14.0.6 ] :: Anaconda, Inc. on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> from semsimian import Semsimian
>>> s = Semsimian([('banana', 'is_a', 'fruit'), ('cherry', 'is_a', 'fruit')])
>>> s.jaccard_similarity('banana', 'cherry')

This should yield a value of 1.0.

Releases

As of version 0.1.14, the semsimian source is released on GitHub, with a corresponding set of Python wheels released to Pypi.

To trigger a new set of builds, first update the version number in Cargo.toml, then create a new release.

Wheels are prepared for the following environments and architectures:

OS	Architectures	Python Versions
Linux	x86_64, x86_64-unknown-linux-musl, aarch64-unknown-linux-gnu, aarch64-unknown-linux-musl	3.7, 3.8, 3.9, 3.10, 3.11
MacOS	x86_64, universal2	3.7, 3.8, 3.9, 3.10, 3.11
Windows	x86_64	3.7, 3.8, 3.9, 3.10, 3.11

Troubleshooting

Building for Mac ARM M1 architectures

If a import semsimian results in a ImportError warning about incompatible architecture, try the following:

Install conda. This guide may be helpful.
Set up a virtual environment with conda so that your Python build is aligned with your processor architecture (in this case, ARM). Try something like:

$ conda create -n myenv python=3.9
...setup happens...
$ conda activate myenv

and then proceed as above.

Code Coverage via Docker

Build a docker image:

docker build -t my-rust-app .

Run your tests inside a Docker container and generate coverage:

docker run -v "$(pwd)":/usr/src/app -t my-rust-app bash -c "CARGO_INCREMENTAL=0 RUSTFLAGS='-Zprofile -Ccodegen-units=1 -Cinline-threshold=0 -Coverflow-checks=off -Zpanic_abort_tests -Cpanic=abort' cargo test && grcov . -s . --binary-path ./target/debug/ -t html --branch --ignore-not-existing -o ./target/debug/coverage/"

Get Coverage report from: