This is a reverse chatbot that asks the users questions about data integration practices and then gives advice based on a body of knowledge. This version operates a bit like an agent which tries to gather enough information to be able to give advice. So it may ask an unspecified number of questions.
You should clone the project and then initialize the UI module with these commands:
git submodule init
git submodule update
We suggest to use Conda to manage the virtual environment and then install poetry.
conda activate base
conda remove -n data_wellness_agent --all
conda create -n data_wellness_agent python=3.12
conda activate data_wellness_agent
pip install poetry
Change the directory to the root folder of the project.
poetry install
Before you run, you should compile the UI with this command from the root folder of the project. Please note that this requires the installation of Yarn and node 18.18.0 or later.
poetry run build-ui
The UI connects the websocket to port 8085 natively. If the Python server runs on a different port, please change the port accordingly.
Before you run for the first time, you will need to create the Postgres database. Execute this command to start psql (make sure Postgres is installed)
On Linux you can install Postgres like this:
sudo -u postgres psql
On Windows you will need to download the distribution from here: https://www.postgresql.org/download/
Then create the database:
CREATE DATABASE data_wellness_companion
WITH
OWNER = postgres
ENCODING = 'UTF8'
CONNECTION LIMIT = -1
IS_TEMPLATE = False;
and add the initial tables with some data after exiting psql:
Run on the console:
sudo -u postgres psql data_wellness_companion
And run the script sql/db_setup.sql
The knowledge base is based on a directory (see environment variable RAW_TEXT_FOLDER). The folder needs to have *.txt documents in it.
The server fails if there are no documents in this folder.
You will have to have a .env file. To get you started you can copy the .env.local to .env and then change the configuration parameters accordingly.
You will need to have an OpenAI API key.
Make the Postgres DB is available.
python ./data_questionnaire_agent/server/questionnaire_server_main.py
You can then check the UI on http://localhost:8085/index.html
On Windows you can use the .\start.ps1 script to start the server.
You can run tests like this:
pytest