This is a custom streamlit chat for Azure PromptFlow starter 'Bring Your Own Data Chat QnA'. It adapts the streamlit open-ai-chat component to work with the interface of the example.
Following env variables are needed to be configured:
- URL=https://MYDEPLOYMENT.REGION.inference.ml.azure.com/score
- KEY=APIKEYFORDEPLOYMENT
- MODEL=MODELNAME
docker run -d -t -i -e URL='https://MYDEPLOYMENT' -e KEY='MYAPIKEY' -e MODEL='my-model-name' -p 8501:8501 ghcr.io/piizei/promptflow-custom-ui:latest
After that, you can open browser to http://localhost:8501
and you should see the chat ui.
You can add a .env
file with these values.
To run with streamlit, first do pip install (and preferably start virtual env) pip install -r requirements.txt
and then run streamlit run app.py