ManaGenie

Overview

This project is a chatbot implemented in Python using technique such as RAG (Retrieval Augmented Generation) & various libraries such as LangChain, LlamaCPP, Faiss, Transformers, PyPDF2, and Streamlit. The chatbot provides two main features:

1. Upload New PDFs: Users can upload PDF documents, which the chatbot then processes and stores for later use.
2. Chat with Bot: Users can engage in conversation with the chatbot, utilizing either the uploaded PDFs or a combination of uploaded and pre-existing data.

Additionally, the chatbot employs LangChain to prompt for the next dialogue in the conversation, ensuring coherence and context continuity.

Features

• PDF Upload: Enables users to upload PDF documents.
• Chat Interface: Provides a user-friendly chat interface powered by Streamlit.
• Integration with Transformers: Utilizes Transformer-based models for natural language processing tasks.
• History Saving: Utilizes LangChain to save conversation history and prompt for the next dialogue.

Requirements

• Python 3.x
• RAG
• LangChain
• LlamaCPP
• Faiss
• Transformers
• PyPDF2
• Streamlit

Installation to run locally on your pc

1. Clone the repository:

   git clone https://github.com/mohdtalal3/ManaGenie.git
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

Usage:

1. Navigate to the project directory:

   cd Files
   

2. Run the Streamlit app::

   streamlit run app.py
   
   

Installation to run Colab:

1. See or refer to the attached notebook colab_run.ipynb to run on colab.

Project Structure

• app.py: Contains the main streamlit code.
• Rag_note.ipynb: Contains the code of the main strucutre of chat bot .
• colab_run.ipynb: Contains the code to run using colab .
• Files: Contains the files to run locally or to run using colab .
• htmlTemplate.py: Contains the code for the streamlt frontend .