oakx-grape

🌳 🍇 Grape wrapper for OAK 🌳 🍇

ALPHA

Usage

Macbook users with M1 processor need to do a few extra steps as follows:

• Download Anaconda.
• conda create --name oakx-grape-env python=3.9
• conda activate oakx-grape-env
• pip install poetry
• poetry install

The steps below are common to everyone.

pip install oakx-grape
poetry run runoak -i grape:sqlite:obo:pato relationships --direction both shape

Install NVM + NPM

These instructions help setup nvm and npm on one's system.

Install GraphViz and OboGraphViz

• brew install graphviz
• npm install -g obographviz

How it works

This plugin implements a grape wrapper. The wrapper in fact wraps two adapters:

1. An adaptor to ensmallen/grape, for performing performance-intensive graph operations
2. An OAK adapter for handling everything else, including lookup by labels, search, predicate filtering, etc

There are two choices of selector:

1. grape:kgobo:{go,pato,uberon,...}
2. grape:OAK-SELECTOR

with the first pattern, the grape graph is loaded from kgobo, and the oak adapter is loaded from semantic sql

with the second, you can most existing existing OAK adapters:

• sqlite/semsql
• obo
• rdf/owl

Note you CANNOT use a backend like ubegraph or bioportal that relies on API calls

The idea is we will be able to run a notebook like this:

https://github.com/INCATools/ontology-access-kit/blob/main/notebooks/Monarch/PhenIO-Tutorial.ipynb

With the semsim part handled by OAK

Acknowledgements

This cookiecutter project was developed from the oakx-plugin-cookiecutter template and will be kept up-to-date using cruft.