Postgres Q&A

Postgres Q&A is a project that demonstrates how to use word embeddings and Postgres to build a chatbot. The chatbot is implemented using Vercel Edge Functions and the @neondatabase/serverless driver, and relies on OpenAI's GPT-3 API to generate responses.

Getting started

To get started with this project, you'll need to have:

• A Neon account, and project.
• an API key for the OpenAI GPT API, which you can obtain from https://openai.com/.

Once you have the prerequisites installed, follow these steps to get the project up and running:

Clone the repository:

git clone https://github.com/neondatabase/postgres-qa.git
cd ask-postgres

Prepare the data

Import embeddings to Neon

This section is derived from OpenAI’s cookbook example. You can use the Python code here to build a web crawler and extract the text you need to create embeddings. We’ve already created the text files based on https://www.postgresql.org/docs/, which you can find in data/text directory.

To get started, move to the data directory create a new environment and install the dependencies:

cd data
python -m venv env
source env/bin/activate
pip install -r requirements.txt

Import the schema to your database:

psql <database-url> -f database.sql

Let’s now add DATABASE_URL and OPENAI_API_KEY to our environment variables:

export DATABASE_URL=<YOUR_NEON_CONEECTION_STRING> OPENAI_API_KEY=<YOUR_OPENAI_API_KEY>

Run main.py to import the emebbeding to your Neon database:

python main.py

Relax and grab a cup of coffee as this section might take 10min to process!

Expcted result:

Saving to CSV...
Loading tokenizer...
Embedding text...
Connecting to database...
Done!

Install the project dependencies:

cd app
npm install

Create a .env file

touch .env.local

Set the following environment variables:

OPENAI_API_KEY= Your OpenAI API key.
DATABASE_URL= The connection URL for your Neon database.

Start the server:

npm run dev

Contributing

We welcome contributions to this project! If you find a bug, have a suggestion, or want to contribute code, please open an issue or pull request on the GitHub repository.

License

This project is licensed under the MIT License. See the LICENSE file for more information.