DAC Faculty Outreach Initiative

A program that scrapes faculty profiles from UVA websites, embeds them with OpenAI's embedding model, then serves a Flask web app that allows for non-exact keyword search of these profiles.

Setup

Populating the database

Set API keys in rc-DACFOI.env:

OpenAI API

OPENAI_API_KEY="your-api-key"
DATABASE_URL="sqlite:///database.db"

Azure OpenAI

OPENAI_API_KEY="your-api-key"
OPENAI_BASE_URL="https://YOUR-RESOURCE-NAME.openai.azure.com/openai/v1"
DATABASE_URL="sqlite:///database.db"

Note that your deployment of the Azure OpenAI embedding model should be named text-embedding-ada-002.

Then set database population options in backend/core/populate_config.py, then run:

python -m backend.core.populate

Running the server

flask --app backend/app.py run --debug