Streamlit Chatbot for Kishiva Internship

Tech-Stack: Google Gemini API, Langchain, Streamlit

This project is a Streamlit application that integrates with Google Generative AI to create a conversational chatbot. It includes user authentication, chat history management, and context-based responses.

Features

• User Authentication: Register and log in to use the chatbot.
• Chat Context: Maintains context of the conversation for more accurate responses.
• Chat History: Save and display chat history for each user.
• Clear History: Option to clear the chat history.
• Logout: Option to log out of the application.

Requirements

• Python
• Streamlit
• Langchain
• Google Generative AI
• SQLite (used for local database)
• python-dotenv (for managing environment variables)

The app will be accessible at http://localhost:8501 in your web browser.

Project Structure

• app.py: Main application file where the Streamlit app is defined.
• database.py: Contains database functions for user authentication and chat history.
• .env: Environment variables file (not included in version control).
• requirements.txt: List of required Python packages.

Code Explanation

database.py:

• This file contains many functions to add, delete and validate users.
• It uses sqlite3 module of python.
• This file is responsible for all the database features, We just have to call these functions in our main file.
• Further explanation can be found in the comments.

main.py:

• This is the main python file of our project.
• We have used streamlit for building the front-end using python.
• The entry-point of the project is the login page.
• If login is successful, then the chat page gets launched.
• Users can see their own chat history.
• This file contains code for Langchain, Google-Genai and Streamlit.

.env:

• This file has my API key for Google's Gemini API.
• I have erased the key from github for privacy and confidential nature of the API keys.
• Anyone can get their own API key from Google's AI studio for free.

Instructions to Use

• Clone this repository on your device.

• Get your own API key from https://aistudio.google.com/app/apikey

• Paste your API key in ".env" file between the quotes.

• You have to install python virtual environment using this command on terminal:

    pip install virtualenv
  

• From here on, all the commands are to run on terminal of the project directory.

• Activate the virtual environment inside the directory using this command on terminal:

    .\venv\Scripts\activate
  

• Install python-dotenv on this virtual environment:

    pip install python-dotenv
  

• Install other dependencies of the project using this command on terminal:

    pip install -r requirements.txt
  

• Run the streamlit app using this command on terminal at your project directory:

    streamlit run main.py
  

Bonus points activity:

There are at least 2 ways deploy our app to cloud.

Deploy on Popular cloud services like AWS, Azure and GCP

• To deploy production applications, It is advised to use these cloud services which provides better options of reliablity and scalability.
• I have learned how to deploy our project using AWS.

Steps to deploy on AWS:

• Set up a free-tier account on aws.

• Setup an EC2 instance.

• Clone this git repository.

• Install every dependency mentioned above.

• run the app using:

    streamlit run main.py
  

• You will get the IP address where the application is served.

• To make it persistent, that means to make this app available all the times even when you close the aws window in your browser: use nohup command.

Deploy feature of Streamlit:

• We can use community version for free.
• We can get a url of our project as well.
• I have deployed our app using this feature and you can use it directly here: https://kishiva-chatbot.streamlit.app/
• This method is not recommended for production grade applications.

Note

Here's the link to demo video: https://drive.google.com/file/d/1ufGY7ETL74q4iWDaCJsGxQyijZn_QugR/view?usp=drive_link