Skip to main content

A hub of AI/ML models for open source drug discovery and global health

Project description

Welcome to the Ersilia Model Hub!

Donate Contributor Covenant License: MIT

documentation EOSI PyPI version fury.io Python 3.7 Code style: black Github All Releases Gitpod Ready-to-Code

logo

Table of Contents:

  1. Project Description
  2. Getting started
  3. Contribute
  4. Roadmap
  5. License and citation
  6. About us

Project Description

The Ersilia Model Hub is the main project of the Ersilia Open Source Initiative. The aim is to provide a platform for a user-friendly deployment of AI/ML models, where scientists can browse through the assets, identify those relevant to their research and obtain predictions without the need of coding expertise. Currently, most of the tools developed, even when published and fully open-sourced, remain unusable by a large majority of the scientific community who does not have the necessary expertise. This gap becomes even larger in Low and Middle Income Country institutions where access to bioinformatic facilities or data science experts is scarce. With this project, we hope to democratize access to this expertise and support research into neglected and infectious diseases.

The models embedded in the hub include both models published in the literature (with appropriate third party acknowledgment) and models developed by the Ersilia team or contributors. All assets are open source, but please do check the Licenses of each model before using them.

Getting started

Follow the installation instructions from the Ersilia Book.

Once Ersilia is installed, you can browse models in the Ersilia Model Hub.

Select one model. For example chemprop-antibiotic. You can fetch your model with the Ersilia CLI:

ersilia fetch chemprop-antibiotic

Then, serve your model:

ersilia serve chemprop-antibiotic

And run the prediction API:

ersilia api -i my_molecules.csv -o my_predictions.csv

Finally, close the service when you are done.

ersilia close

Please see the Ersilia Book for more examples and detailes explanations.

Contribute

The Ersilia Model Hub is developed and maintained by a small team of Ersilia employees and volunteers, and any contribution is highly valued! There are several ways in which you can contribute to the project:

  • If you are a developer, check the issues and help us improve the tool
  • If you have developed a model and would like to include it in the Hub to increase its visibility and usability, please contact us or open an issue. We are currently working on an automated contribution tool to facilitate the process.
  • If you are a scientist with a cool dataset, also contact us or open an issue as we might be interested in developing a model based on the data!
  • If there is a third-party model you have identified and would like to see it in the Hub, open an issue with the relevant information and we will get back to you as soon as possible.

The Ersilia Open Source Initiative is adhered to the Contributor Covenant guidelines.

Roadmap

We are working to grow the Hub organically and responding to our users needs. Here a detail of the next features to come, stay tuned!

  1. Deployment for Windows System (expected: February 2022)
  2. Automated third-party model contributions (expected: March 2022)
  3. Possibility to run lite models online (expected: May 2022)

License and citation

This repository is open-sourced under an MIT License. Please cite us if you use it.

About us

The Ersilia Open Source Initiative is a Non Profit Organization incorporated with the Charity Commission for England and Wales (number 1192266). Our mission is to reduce the imbalance in biomedical research productivity between countries by supporting research in underfunded settings.

You can support us via Open Collective.

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

ersilia-0.0.10.tar.gz (1.8 MB view hashes)

Uploaded Source

Built Distribution

ersilia-0.0.10-py3-none-any.whl (1.8 MB view hashes)

Uploaded Python 3

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