🤖 Atlas 🤖

Introduction

Atlas is a multimodal AI-powered chatbot designed to assist users with queries, provide insights, and facilitate learning experiences. It supports multimodal interaction, enabling seamless understanding of different input types.

🚀 Features

• 🖼️ Multimodal Interaction: Atlas allows you to input different kinds of files in a prompt, enabling seamless multimodal understanding.
• 💾 Maintains conversation context
• 📄 Loads knowledge from markdown files in the result directory
• 🗄️ Stores chat history in an SQLite database
• 💡 Provides informative responses about MLSA SRM and its activities

🌟 Microsoft Learn Student Ambassador (MLSA) Program

The Microsoft Learn Student Ambassador (MLSA) program offers:

• 📚 Learning: Access to Microsoft resources, mentorship, and workshops.
• 🌍 Global Impact: Build projects with real-world impact and connect with an international community.
• 🎓 Career Growth: Gain hands-on experience and certifications.
• 💬 Collaboration: Work with peers and experts on innovative projects like Atlas.

Atlas can help users with learning by:

• ✅ Answering complex queries using multiple inputs.
• 🤖 Innovating with AI for community-driven solutions.
• 🔄 Automating tasks and generating insights.

🛠️ How to Use Atlas

1. ⚙️ Setup

# Clone the repository
git clone https://github.com/MLSA-SRM/Atlas-GPT/

# Install dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env.example .env
# Add your TOGETHER_API_KEY to .env file

2. ▶️ Run the Application

# Start the server
python src/run.py

3. 🖥️ Run the Frontend

# Navigate to the frontend directory
cd frontend/atlasfrontend

# Install frontend dependencies
npm install

# Start the frontend application
npm start

🔗 API Endpoints

🗨️ Start a chat

curl -X POST "http://localhost:8000/chat/{chat_id}" \
     -H "Content-Type: application/json" \
     -d '{"query": "Tell me about MLSA program"}'

❌ Delete a chat

curl -X DELETE "http://localhost:8000/chat/{chat_id}"

🛠️ Debugging

• Run the backend in debug mode:

  python src/run.py --debug

• Check frontend logs via your browser console and network tab.
• Examine log files in the “logs” folder for errors.

📲 Placing Calls

• Use curl or any REST client to send requests:

  curl -X POST "http://localhost:8000/chat/{chat_id}" -H "Content-Type: application/json" -d '{"query":"Hello!"}'

• Delete a chat:

  curl -X DELETE "http://localhost:8000/chat/{chat_id}"

🖥️ Navigating the Updated Frontend

• Launch with:

  cd frontend/atlasfrontend
  npm install
  npm start

• Sidebar:
  • Switch between chats, start new ones, and access MLSA links.
• Main Area:
  • View conversation logs; send messages in the input field.

Typing Indicator

• Typing Indicator: When AtlasGPT is processing your query, a typing indicator will be shown.

Markdown Support

• Markdown Rendering: Messages from AtlasGPT can include markdown formatting for better readability.