https://github.com/ravikumarmn/simple-llamaindex-agent
A FastAPI-based application that combines a Retrieval-Augmented Generation (RAG) system with an intelligent agent for enhanced information retrieval and query handling.
https://github.com/ravikumarmn/simple-llamaindex-agent
ai-agents langchain llama-index pinecone rag streamlit
Last synced: 5 months ago
JSON representation
A FastAPI-based application that combines a Retrieval-Augmented Generation (RAG) system with an intelligent agent for enhanced information retrieval and query handling.
- Host: GitHub
- URL: https://github.com/ravikumarmn/simple-llamaindex-agent
- Owner: ravikumarmn
- Created: 2024-12-26T21:48:45.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-12-26T22:10:15.000Z (over 1 year ago)
- Last Synced: 2025-05-19T01:37:46.459Z (about 1 year ago)
- Topics: ai-agents, langchain, llama-index, pinecone, rag, streamlit
- Language: Python
- Homepage: https://ravikumarmn.github.io/
- Size: 4.23 MB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Simple LlamaIndex Agent
A FastAPI-based application that combines a Retrieval-Augmented Generation (RAG) system with an intelligent agent for enhanced information retrieval and query handling.
## Features
- **RAG System**: Retrieves relevant information from a Pinecone vector database using LlamaIndex
- **Intelligent Agent**: Makes decisions on when to use the vector database vs handling queries directly
- **Multiple Interfaces**: FastAPI endpoints and Streamlit UI
- **Content Safety**: Built-in inappropriate content detection
- **Advanced Embedding**: Uses OpenAI's text-embedding-3-large model
- **LLM Integration**: Powered by GPT-4 for high-quality responses
## Prerequisites
- Python 3.11+
- OpenAI API key
- Pinecone API key
- Cohere API key (for reranking)
## Installation
1. Clone the repository and create a virtual environment:
```bash
python -m venv .venv
source .venv/bin/activate # On Windows use: .venv\Scripts\activate
```
2. Install dependencies:
```bash
pip install -r requirements.txt
```
```bash
pip install -e .
```
3. Set up environment variables:
Create a `.env` file based on `example.env`:
```bash
OPENAI_API_KEY=your_openai_api_key
PINECONE_API_KEY=your_pinecone_api_key
COHERE_API_KEY=your_cohere_api_key
```
## Usage
1. Start the FastAPI server:
```bash
uvicorn src.app.main:app --reload
```
2. Launch the Streamlit interface (optional):
```bash
streamlit run src/streamlit_app.py
```
## API Endpoints
### Agent Query
- **POST** `/agent`
- Intelligent query handling with contextual responses
```json
{
"query": "Tell me about waves in physics"
}
```
## Agent Tools
1. **VectorDBTool**: Handles complex information retrieval queries
2. **InappropriateContentDetector**: Filters inappropriate or offensive content
## Configuration
Key settings in `config/ncert_search.json`:
- Embedding model: text-embedding-3-large (3072 dimensions)
- LLM: GPT-4
- Vector similarity: Top 5 results
- Chunk size: 1024 tokens with 50 token overlap
## Project Structure
```
├── src/
│ ├── app/
│ │ ├── main.py # FastAPI application
│ │ └── streamlit_app.py # Streamlit interface
│ ├── agent.py # Agent implementation
│ ├── document_manager.py # Document handling
│ ├── indexer.py # Vector database operations
│ ├── retrieval.py # RAG implementation
│ └── service_config.py # Configuration management
├── config/
│ └── ncert_search.json # System configuration
└── requirements.txt # Project dependencies
```
## Dependencies
Key packages:
- llama-index-core
- llama-index-vector-stores-pinecone
- llama-index-embeddings-openai
- fastapi
- langchain
- streamlit