Intelligent Digital Assets Assistant (IDAA)
Demo Notebook |
React Demo |
Streamlit Demo |
Docker Image
</p>
> [!NOTE]
> 🚧 **This repository features a React frontend. If you need a Streamlit frontend, visit the [Streamlit version](https://github.com/AdityaNikhil/IDAA-Streamlit-App).**
---
## Table of Contents
1. [Overview](#overview)
2. [Approach](#approach)
3. [Setup](#setup)
4. [Getting Started](#getting-started)
5. [License](#license)
6. [Contributors](#contributors)
---
## Overview
The **Intelligent Digital Asset Advisor (IDAA)** is an AI-powered financial advisory system integrating three core functions:
- **Professor:** Provides educational insights on digital assets.
- **Analyst:** Processes market and on-chain data for analytics.
- **Advisor:** Offers strategic financial guidance.
Built using **GPT-4, LLAMA 3.3, and RAG**, IDAA delivers real-time, informed, and personalized digital asset recommendations. The system orchestrates these components seamlessly, ensuring accurate and timely financial insights.
---
## Approach
### System Workflow
1. **User Query Handling**: Queries are routed to the appropriate agent via the **supervisor node**.
2. **Analyst Agent**: Queries PostgreSQL for historical market data.
3. **Advisor Agent**: Uses Pinecone vector DB for trading insights.
4. **Professor Agent**: Gathers real-time educational content.
5. **Query Processing**:
- Analyst queries PostgreSQL and generates insights.
- Advisor applies **RAG operations** for recommendations.
- Professor fetches relevant educational materials.
This modular approach ensures a well-rounded financial advisory experience.
---
## Setup
> [!IMPORTANT]
> **Ensure the ETL pipeline is configured and PostgreSQL data is loaded before proceeding.**
> Refer to the [ETL setup](etl/) for more details.
> Ensure the ETL pipeline is up and running.
### Prerequisites
- **PostgreSQL Database**: Ensure historical data is loaded.
- **API Keys**: Obtain keys from **Langsmith, Groq, OpenAI, TavilySearch**.
### Installation
Clone the repository:
```sh
git clone https://github.com/adityanikhil/idaa
cd idaa
```
### Configure Environment Variables
Create `setup.yml` and add your API keys:
```yaml
environment:
LANGSMITH_TRACING_V2: "true"
LANGCHAIN_ENDPOINT: "https://api.smith.langchain.com"
LANGSMITH_API_KEY: "YOUR_API_KEY"
TAVILY_API_KEY: "YOUR_API_KEY"
GROQ_API_KEY: "YOUR_API_KEY"
OPENAI_API_KEY: "YOUR_API_KEY"
DATABASE_URI: "postgresql://postgres:postgres@host.docker.internal:5432/postgres"
```
---
## Getting Started
### Setting up the [Docker Image](https://hub.docker.com/repository/docker/adityadigala/idaa-react-app/)
This docker image consists of the FastAPI endpoint responsible for generating chatbot responses.
#### Pull the docker image
```sh
docker pull adityadigala/idaa-react-app:latest
```
#### Run the Fast API service through the docker image
```sh
docker run -p 8000:8000 -v 'path/to/setup.yml':/app/setup.yml adityadigala/idaa-react-app:latest
```
### Running the React Frontend
1. Clone the [frontend repo](https://github.com/yolandawu/IDAA-FN):
```sh
git clone https://github.com/yolandawu/IDAA-FN
cd IDAA-FN
```
2. Install dependencies:
```sh
npm install
```
3. Configure API endpoint in `.env`:
```sh
REACT_APP_BACKEND_URL=http://localhost:8000
```
4. Start the React server:
```sh
npm run dev
```
5. Open your browser at [http://localhost:3000](http://localhost:3000).
For additional frontend setup options, check [this guide](https://github.com/yolandawu/IDAA-FN?tab=readme-ov-file#deployment).
> [!NOTE]
> 🚧 **Need a Streamlit frontend?** Visit the [Streamlit version](https://github.com/AdityaNikhil/IDAA-Streamlit-App).
---
## License
This project is licensed under the **MIT License**. See [LICENSE](/IDAA/LICENSE) for details.
---
## Contributors
- **[Aditya Nikhil Digala](https://linkedin.com/in/aditya-nikhil)**
- **[Yang Wu](https://www.linkedin.com/in/yangywu/)**
- **[Vaeshnavi Vella](https://www.linkedin.com/in/vaeshnavi-vella/)**