https://github.com/jokerdii/rag-streamlit-app
Implemented Simple RAG, GraphRAG (Neo4j), AdaptiveRAG (LangGraph), RAG with Reranking (Cohere) in Streamlit app for information summarization
https://github.com/jokerdii/rag-streamlit-app
cohere knowledge-graph langgraph neo4j rag-chatbot streamlit
Last synced: 2 months ago
JSON representation
Implemented Simple RAG, GraphRAG (Neo4j), AdaptiveRAG (LangGraph), RAG with Reranking (Cohere) in Streamlit app for information summarization
- Host: GitHub
- URL: https://github.com/jokerdii/rag-streamlit-app
- Owner: JoKerDii
- Created: 2025-08-09T22:47:55.000Z (11 months ago)
- Default Branch: main
- Last Pushed: 2025-08-09T23:27:44.000Z (11 months ago)
- Last Synced: 2025-08-10T00:21:16.883Z (11 months ago)
- Topics: cohere, knowledge-graph, langgraph, neo4j, rag-chatbot, streamlit
- Language: Python
- Homepage:
- Size: 968 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# RAG Application
This project is a Streamlit application that showcases and compares different Retrieval-Augmented Generation (RAG) techniques for building an interactive chatbot that can answer questions based on your uploaded documents.
-----
### Features
* **Simple RAG**: A foundational RAG implementation using a simple vector store (**FAISS**) for document retrieval. It's a great starting point to understand the RAG workflow.
* **RAG with Neo4j**: An advanced approach that leverages a **Neo4j knowledge graph** to enhance document retrieval. It uses a hybrid search to combine the power of vector embeddings with structured graph data.
* **Adaptive RAG**: A sophisticated strategy powered by **Langgraph** that dynamically analyzes user queries to determine the optimal retrieval path, offering a more intelligent and flexible RAG pipeline.
* **RAG with Reranking**: Improves the quality of retrieved documents by using a **Cohere Rerank** model to reorder search results, ensuring the most relevant information is presented to the LLM for a more accurate response.
---
### Demo
-----
### Technologies Used
* **Streamlit**: For building the interactive web application user interface.
* **LangChain**: The framework for orchestrating the RAG pipelines, handling document loading, splitting, and conversational chains.
* **FAISS**: An efficient library for vector similarity search, used as the primary vector store.
* **Groq**: The LLM provider (**Llama3-70b-8192**) used for generating responses.
* **Neo4j**: An open-source graph database for the knowledge graph implementation.
* **HuggingFace Embeddings**: For converting text into dense vector representations.
* **Cohere Rerank**: A powerful reranking model to improve the relevance of retrieved documents.
-----
### Prerequisites
* **Python 3.8+**: The project is built with Python.
* **Neo4j Desktop or Server**: Required for the "RAG with Neo4j" feature. You must also install the **APOC plugin** and enable `apoc.*` procedures in `neo4j.conf`.
* **API Keys**: You will need API keys for Groq and Cohere.
-----
### Setup and Installation
1. **Clone the repository**:
```bash
git clone
cd
```
2. **Create a virtual environment**:
```bash
python -m venv venv
source venv/bin/activate # On Windows, use `venv\Scripts\activate`
```
3. **Install dependencies**:
```bash
pip install -r requirements.txt
```
*(Note: You will need to create a `requirements.txt` file by listing all the imported libraries in the code.)*
4. **Configure API Keys**:
Create a `.env` file in the root directory and add your API keys:
```
GROQ_API_KEY="your_groq_api_key"
COHERE_API_KEY="your_cohere_api_key"
```
5. **Neo4j Setup (Optional)**:
If you plan to use the Neo4j feature, ensure your Neo4j database is running and accessible with the correct credentials. The application assumes the default bolt URI (`bolt://localhost:7687`) and `neo4j` as the username. You can modify this in the code if needed.
-----
### Usage
To run the application, simply execute the following command from the project's root directory:
```bash
streamlit run main.py
```
This will launch the Streamlit application in your web browser. You can then navigate between the different RAG modes using the sidebar.