The goal of this project is to simplify the developer experience when interacting with Onyx GenAI Services. This project provides wrappers around the the underlying APIs provided by the service.
To install with pip
, run:
pip install onyxgenai
To install with conda
, run:
conda install -c conda-forge onyxgenai
-
Create a Conda Store Environment with all dependencies listed in the requirements.txt
-
Start your JupyterLab Server
-
Create a new Jupyter Notebook
-
Install the SDK as indicated above
-
Add the onyxgenai client imports to your project
For more in depth examples, see notebooks section of this repo.
The Embedding Client provides access to the Onyx GenAI Embedding Service. The client provides access to functionality such as:
- Generating Text and Image Embeddings and Vector Storage
- Retrieving Vector Store Collections
- Vector Database Search
The Model Client provides access to the Onyx GenAI Model Store Service. The client provides access to functionality such as:
- Retrieving Model Info
- Retrieving Active Model Deployment Info
- Deploying and Deleting Model Deployments
- Performing Text and Image Prediction and Embedding
- Generating Text Completions from an LLM
- To run unit tests, run the following:
pytest
- To run code quality checks, run the following:
ruff check .