https://github.com/ankitrajput0096/QueryPal-Personalized-AI-Friend
QueryPal is a RAG system using Meta's Llama 2.0 (via Ollama), ChromaDB, and LangChain for seamless document retrieval and query handling. It offers precise answers to document-based and general queries through an intuitive, user-friendly dashboard.
https://github.com/ankitrajput0096/QueryPal-Personalized-AI-Friend
chromadb docker docker-compose langchain llama ollama python3
Last synced: 10 months ago
JSON representation
QueryPal is a RAG system using Meta's Llama 2.0 (via Ollama), ChromaDB, and LangChain for seamless document retrieval and query handling. It offers precise answers to document-based and general queries through an intuitive, user-friendly dashboard.
- Host: GitHub
- URL: https://github.com/ankitrajput0096/QueryPal-Personalized-AI-Friend
- Owner: ankitrajput0096
- Created: 2024-12-25T22:43:26.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-12-30T09:36:23.000Z (over 1 year ago)
- Last Synced: 2025-06-10T21:50:57.205Z (about 1 year ago)
- Topics: chromadb, docker, docker-compose, langchain, llama, ollama, python3
- Language: JavaScript
- Homepage:
- Size: 6.3 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# QueryPal - Your Personalized AI Companion
**Built with LLAMA 2 (via Ollama), ChromaDB, Docker, Flask, React, and Redux**
---
## Overview
QueryPal is a **Retrieval-Augmented Generation (RAG)** system that leverages Meta's Llama 2.0 model (via **Ollama**), ChromaDB for vector storage, and LangChain for process orchestration. This project seamlessly combines document retrieval and query handling, enabling:
- Contextually relevant responses to questions about uploaded documents.
- General knowledge query answering.
A sleek and intuitive dashboard enhances the user experience, making it easy to interact with the system.
---
## Screenshots
### Homepage
### Chat Interface
---
## RAG Architecture
QueryPal integrates robust document retrieval capabilities with LLM-driven response generation to ensure accurate and context-aware answers. The process involves:
- **Retriever**: Encoding and indexing external documents into vectors for similarity-based searches.
- **Generator**: Using **Llama 2.0 (via Ollama)** to generate responses based on retrieved documents or pre-trained knowledge.
LangChain orchestrates the workflow, ensuring a seamless integration of these components.
---
## Features
- **General Queries**: Answers a wide range of questions using Llama 2.0’s built-in knowledge, even without uploaded documents.
- **Document-Based Q&A**: Delivers precise, context-aware answers by analyzing uploaded documents.
---
## How It Works
1. **Embedding Creation**: Generates document embeddings using HuggingFace.
2. **Data Persistence**: Stores embeddings for future use.
3. **Vector Database**: Builds a ChromaDB-based vector database for efficient retrieval.
4. **Retriever Initialization**: Fetches relevant documents for user queries.
5. **General Query Handling**: Leverages Llama 2.0 to answer questions unrelated to uploaded documents.
6. **LLM Integration**: Ensures deterministic, contextually relevant answers using **Llama 2.0 (via Ollama)**.
7. **Q&A Pipeline**: Combines the retriever and generator components via LangChain.
---
## Project Structure
- **Frontend**: Developed with React, featuring Redux for state management and React Router for navigation. The UI is intuitive and user-friendly.
- **Backend**: Built on Flask, hosting the RAG stack and exposing necessary APIs.
---
## Getting Started
### Clone the Repository
```bash
git clone git@github.com:ankitrajput0096/QueryPal-Personalized-AI-Friend.git
cd QueryPal-Personalized-AI-Friend
```
---
## Building the Application
### Using Docker
1. Build the Docker image:
```bash
docker-compose build
```
2. Start the Docker containers:
```bash
docker-compose up
```
### Running the Pre-Built Docker Image
1. Start the containers with:
```bash
docker-compose -f docker-compose-run.yml up
```
---
## Interacting with the Backend APIs
### Postman API Collection
A [Postman collection](./RAG_backend/RAG_stack.postman_collection.json) is included for easy API interaction. Import the collection and use the following endpoints:
1. **General Query**
- **Description**: Handles questions unrelated to uploaded documents.
- **Endpoint**: `http://127.0.0.1:8090/ask_general_query`
- **Method**: POST
- **Request Body**:
```json
{
"query": "What is the capital of USA?"
}
```
2. **Upload Document**
- **Description**: Uploads documents for embedding and storage.
- **Endpoint**: `http://127.0.0.1:8090/upload_document`
- **Method**: POST
- **Request Body**: Form-data with key `file`.
3. **Similarity Search**
- **Description**: Finds content similar to a query in uploaded documents.
- **Endpoint**: `http://127.0.0.1:8090/similarity_search`
- **Method**: POST
- **Request Body**:
```json
{
"query": "Find content similar to this query."
}
```
4. **Query Document**
- **Description**: Queries uploaded documents for specific information.
- **Endpoint**: `http://127.0.0.1:8090/query_document`
- **Method**: POST
- **Request Body**:
```json
{
"query": "What is the content of the document?"
}
```
5. **Upload and Query Text**
- **Description**: Uploads text and queries it simultaneously.
- **Endpoint**: `http://127.0.0.1:8090/text_and_query`
- **Method**: POST
- **Request Body**:
```json
{
"text": "Summary of A Brief History of Data Visualization...",
"query": "What were the changes during the 1850–1900 Golden Age of statistical graphics?"
}
```
---
This README ensures a clear understanding of QueryPal’s features and usage. Feel free to reach out with questions or suggestions!