Document Chatbot 📄

A Conversational AI chatbot that allows users to ask questions based on the content of a PDF document. It leverages FAISS for vector storage, Google Generative AI (Flash Gemini 2.0) for embeddings and responses, and Gradio for an interactive chat interface.

Features ✨

• Conversational AI with Flash Gemini 2.0: Uses Google's Gemini 2.0 Flash model to generate accurate and context-aware responses.
• Automatic PDF Processing: Extracts text from PDFs and converts them into searchable chunks.
• Efficient Text Search: Uses FAISS for fast and efficient document retrieval.
• Interactive Chat Interface: Built with Gradio for an easy-to-use web-based chatbot.
• Customizable & Expandable: Modify document sources and enhance AI capabilities as needed.

Video Demo 📽️

• Click to watch the demo on YouTube! >> Video Demo's Link

How It Works 🏗️

1. Load PDF Document: Extracts text from the provided "BDO-Malaysia-Budget-2025-Highlights.pdf" file.
2. Create Vector Store: The extracted text is split into chunks and embedded using Google Generative AI (models/embedding-001), then stored in FAISS.
3. User Query Processing: When a user asks a question, the system searches for relevant document chunks using FAISS.
4. AI-Generated Responses: The chatbot, powered by Gemini 2.0 Flash, provides intelligent responses based on the retrieved document context.

Installation & Setup 🚀

1. Ensure you have Python installed, then install dependencies:

pip install -r requirements.txt

2. add your Google API Key in .env file.

GOOGLE_API_KEY=your_google_api_key_here

3. The system uses "BDO-Malaysia-Budget-2025-Highlights.pdf" as the default document to demonstrate its capabilities. You can replace it with any other document. To use a different document instead of "BDO-Malaysia-Budget-2025-Highlights.pdf", modify the DEFAULT_DOCUMENT_PATHS in main.py:

DEFAULT_DOCUMENT_PATHS = ["Your_document_name.pdf"]

4. Running the Chatbot in your local host.

python main.py

The chatbot will launch in your web browser via Gradio. http://127.0.0.1:7860

Project Structure 🛠️

📂 Chat with Doc 
├── 📄 main.py                # Main script for running the chatbot
├── 📄 requirements.txt       # Required dependencies
├── 📄 .env                   # Stores API keys (not shared in repo)
├── 📂 faiss_index/           # Directory where FAISS index is stored
└── 📄 README.md              # Project documentation

Chatbot Interface (Gradio) 💬

The chatbot runs on Gradio, providing a simple and interactive web-based interface. It is suitable for rapid prototyping, allowing users to showcase and test the system using a straightforward UI.

Interface Features 🎨

1. User-Friendly Chatbot:

• Users can input questions and receive AI-generated responses.
• The interface allows easy scrolling and interaction.
• Users can undo and edit their prompt.
• If the chatbot’s answer does not meet expectations, users can ask the chatbot to regenerate the response with one click for a refined answer.

2. Predefined Example Questions:

• The chatbot includes predefined example questions to help users quickly understand how to interact with the system. These questions demonstrate the chatbot's ability to extract and summarize information from the document.
• Example questions in this system are "What is the content of the document?", "Summarize the tax personal update."
• These predefined examples serve as guidelines for users, showing how they can ask customized questions to retrieve specific information from the document.

---

Feel free to reach out with suggestions or improvements for the project! If you like or are using this project, please consider giving it a star⭐. Thanks! (●'◡'●)