raggle

Overview

raggle is a command-line conversational AI tool built on Retrieval-Augmented Generation (RAG) principles. It searches the web for information based on user queries and processes the results to provide informative responses.

Git branch Index

• main: Main showcase code
• develop: Code under current development [Check out for code]
• idea-pitch : Branch to document Ideas
• A1: Idea 1 of implementation.

Features

• Command-Line Interface: Interact with the application through simple command-line commands.
• Web Search Functionality: Query the web and obtain generated responses.
• Output to File: Save responses in a text file for easy access.
• RAG Implementation: Uses a structured approach to enhance response quality.

Workflow - Approach 1

1. Query Input: Accept the query via command line (future UI support planned).
2. Query Transformation: Rewrite or transform the query as needed (this can be delayed until later stages).
3. Web Search: Perform a web search based on the transformed query.
4. Results Parsing and Chunking:
   • Store search results as a dictionary with URLs as keys for better organization.
5. Storage: Use Chromadb for storing the processed results in a Vector Database.
6. Dense Retrieval: Retrieve relevant chunks from the stored data.
7. Deduplication and Reranking:
   • Implement logic to deduplicate and rerank retrieved elements.
8. Augmentation and Summarization: Use a language model (LLM) to augment and summarize the final output.

Getting Started

Prerequisites

• Python (version 3.10 and above)
• Required Python packages (see below)

Installation

1. Clone the repository: git clone https://github.com/Knight7561/raggle.git
2. Navigate to the project directory: cd raggle
3. Install required packages: pip install -r requirements.txt

Usage

To run the application, use the following command:

python main.py -q "Your query here"

Example: python main.py -q "What is the capital of France?"

The response will be printed to the console and also saved in temp/output.txt.