This Streamlit application allows users to interact with their PostgreSQL database using natural language queries. Users can type questions in natural language, which the app translates into SQL queries executed against the database. The results are then displayed in the chat interface. Powered by Google Generative AI and LangChain's SQL agent tool

README.md

Project Title: Postgres Chatbot with Streamlit

Introduction

This project aims to develop a chatbot that can interact with a PostgreSQL database and answer queries in natural language. The chatbot will be able to handle queries related to an Orders table, which includes columns such as order_id, customer_email, tracking_number, shipping_service, tracking_URL, order_created_at, and order_shipped_at.

The chatbot will be designed to understand and respond to layman language questions, such as:

1. "Give me the status of order number 12222."
2. "What is the order ID of tracking number?"
3. "How many orders are shipped?"

A user interface will be created using Streamlit where users can input their PostgreSQL credentials and ask questions.

Setup Instructions

Follow these steps to set up the project on your local machine:

1. Clone the Repository Begin by cloning the repository to your local machine:

https://github.com/langchain-tech/postgres-chatbot.git
cd postgres-chatbot

2. Install Dependencies with Poetry It is recommended to use poetry to manage dependencies. Install the necessary packages and create a virtual environment:

poetry install

3. Activate the virtual environment: after installation the necessary packages activate the virtual env:

poetry shell

4. Set Up Environment Variables Create a .env file in the root directory of your project and add the required environment variables. For example:

GOOGLE_API_KEY=your_gemini_api_key

5. Start the Application

Run the application using Streamlit:

streamlit run app.py

How to use

Notes

• Questions will not be limited to one table, so it should be done to ask from database
• As long as entry is added into database, we should be able to answer questions for that new record