SQL Chat Assistant 🤖

The SQL Chat Assistant is a Streamlit-based web application that allows users to interact with a SQL database using natural language. It uses a Language Learning Model (LLM) to generate SQL queries from user questions, executes them on a SQLite database, and provides the results in a user-friendly format. Checkout the demo using following URL= https://sql-chat-aiassistant-omkarharer.streamlit.app/

How It Works

User Input: The user enters a natural language question (e.g., "Who is the manager of the Sales department?").

Query Generation: The app uses an LLM (e.g., Groq or Google Generative AI) to generate a SQL query from the question.

Query Execution: The generated SQL query is executed on a SQLite database.

Result Formatting: The app formats the query results into a natural language response and displays it to the user.

       +-------------------+
       |   User Interface  |   <-- User Input (Natural Language)
       +-------------------+
               |
               v
       +---------------------------+
       |   LLM (ChatGroq/Gemini)   |   <-- Converts Query to SQL
       +---------------------------+
               |
               v
       +--------------------------+
       |  SQL Query Generation   |   <-- Formats SQL Query
       +--------------------------+
               |
               v
       +-------------------+
       |   SQL Database    |   <-- Executes SQL Query
       +-------------------+
               |
               v
       +--------------------------+
       |   Response Generation    |   <-- Converts Data to Natural Language
       +--------------------------+
               |
               v
       +-------------------+
       |   User Interface  |   <-- Displays Answer
       +-------------------+

🚀 Steps to Run the SQL Chat Assistant Locally

1️⃣ Clone the GitHub Repository

Open your terminal or command prompt and run:

git clone <repo-url>
cd <your-repo-folder>

2️⃣ Create and Activate a Virtual Environment

Using Conda:

conda create -n sqlbot python=3.12 -y
conda activate sqlbot

Or using virtualenv (alternative to Conda):

python -m venv sqlbot
source sqlbot/bin/activate  # On macOS/Linux
sqlbot\Scripts\activate     # On Windows

3️⃣ Install Required Dependencies

pip install -r requirements.txt

4️⃣ Get API Key for LLM

• Obtain your API key from Google AI Studio using the link below:
  • 🔗 Get Your Google AI Studio API Key
    Create a .env file in the project directory and add your API key:

GOOGLE_API_KEY=your_google_api_key_here

5️⃣ Run the Streamlit App

streamlit run app.py

🚀Deployment to Streamlit Community Cloud

Push your code to GitHub.

Deploy the app to Streamlit Community Cloud.

Add your API key to Streamlit’s Secrets:

GOOGLE_API_KEY=your_google_api_key_here

Share the public URL (e.g., https://sql-chat-aiassistant-omkarharer.streamlit.app/) with others.

Known Limitations

Database Dependency

• The app currently works only with SQLite databases.
• Support for other databases (e.g., PostgreSQL, MySQL) can be added in future versions.

LLM Limitations

• The accuracy of generated SQL queries depends on the LLM’s capabilities.
• Complex queries may not always generate correct SQL statements.

Error Handling

• The app handles common errors (e.g., missing tables, invalid queries).
• Some edge cases may not be fully covered.

Scalability

• Designed for small to medium-sized datasets.
• For large datasets, consider using a cloud-based database.

---

Suggestions for Improvement

🚀 Few-Shot Learning for Complex Queries

• Implement few-shot learning to improve handling of complex SQL queries.
• Provide examples of complex queries and their corresponding SQL statements to guide the model.

🎤 Audio-to-Text & Text-to-Audio

• Support for voice input (audio-to-text) and voice output (text-to-audio) to enhance user interaction.

🗄️ Support for Multiple Databases

• Extend support to PostgreSQL, MySQL, MongoDB, and other database systems.

🔐 User Authentication

• Implement authentication to restrict access and enhance security.

🖥️ UI Improvements

• Add a history of previous queries and results for a better user experience.