The purpose of this repo is to quickly and easily get you a streamlit app deployed to datarobot. The purpose of the app is to provide an interface for folks to interact with llm blueprints that have been deployed from the datarobot llm playground to datarobot mlops / console.
Before pushing the app to DataRobot, give it a spin locally via
streamlit run app/app.py
Recommended -> create a virtual environment and pip install -r ./app/requirements.txt
as well as streamlit.
You will also need the following environment variables set:
DATAROBOT_API_TOKEN
DATAROBOT_ENDPOINT
install dr-apps
run the following
drapps create --path ./app --base-env 6542cd582a9d3d51bf4ac71e 'Connect to any LLM Proxied by DR'
This will create an app in datarobot nammed Connect to any LLM Proxied by DR
.
If you don't already have DATAROBOT_API_TOKEN
and DATAROBOT_ENDPOINT
set and environment variables, either set them, or pass them to the drapps create
command via -t
and -E
arguments as follows
drapps create -t <datarobot-api-token> -E <datarobot-endpoint> --path ./app --base-env 6542cd582a9d3d51bf4ac71e 'Connect to any LLM Proxied by DR'