AI Atlas Nexus Extensions

Scope

The purpose of this project is to provide a template for new open source ai-atlas-nexus extensions.

How to add a new AI Atlas Nexus extension.

Install cookiecutter using pipx package manager in your python environment.

pip install pipx
pipx install cookiecutter

Use a GitHub template to replicate extension project. Enter the relevant details of your extension. You can change these details later on.

pipx run cookiecutter gh:IBM/ai-atlas-nexus-extensions/extension-template

Once the extension project is generated, follow the instructions below.

1. Modify the pyproject.toml and make appropriate changes. Add your requirements to dependencies.

2. Add new files and directories to src/<EXTENSION_PACKAGE_NAME>/*

   mkdir src/<EXTENSION_PACKAGE_NAME>/<NEW_SUB_DIRECTORY>
   
   With an __init__.py in each new directory.

3. Integrate with the AI Atlas Nexus API

   • Your extension must use the AI Atlas Nexus API for core functionality.
   • Ensure proper error handling, and logging.

4. Your extension should only be executed via the run method of the class Extension in src/EXTENSION_PACKAGE_NAME/main.py.

5. Add and run tests, linting, formatting and static type checking:

   pytest --cov <YOUR_EXTENSION_NAME> tests # tests
   black --check --line-length 120 src # formatting
   pylint src # linting

6. Provide unit tests for core functionality in YOUR_EXTENSION_NAME/tests/ folder inside your extension directory.

7. Update the <YOUR_EXTENSION_NAME>/README.md inside your extension folder with:

   • Name and description
   • Usage instructions
   • License

Extension list

Name	Tags	Description
AI Atlas Nexus ARES Intgeration	AI robustness evaluation, AI risks, red-teaming	ARES Integration for AI Atlas Nexus allows you to run AI robustness evaluations on AI Systems derived from use cases.

License

AI Atlas Nexus Extensions is under Apache 2.0 license.

View the detailed LICENSE.

IBM ❤️ Open Source AI

AI Atlas Nexus has been brought to you by IBM.