README.md

metadata

title: First Agent Template
emoji: 🤖
colorFrom: pink
colorTo: yellow
sdk: gradio
sdk_version: 5.15.0
app_file: app.py
pinned: false
tags:
  - smolagents
  - agent
  - smolagent
  - tool
  - agent-course

SmoLAgents Conversational Agent

A powerful conversational agent built with SmoLAgents that can connect to various language models, perform web searches, create visualizations, execute code, and much more.

📋 Overview

This project provides a flexible and powerful conversational agent that can:

• Connect to different types of language models (local or cloud-based)
• Perform web searches to retrieve up-to-date information
• Visit and extract content from webpages
• Execute shell commands with appropriate security measures
• Create and modify files
• Generate data visualizations based on natural language requests
• Execute Python code within the chat interface

The agent is available through two interfaces:

• A Gradio interface (original)
• A Streamlit interface (new) with enhanced features and configuration options

🛠️ Prerequisites

• Python 3.8+
• A language model, which can be one of:
  • A local model running through an OpenAI-compatible API server (like LM Studio, Ollama, etc.)
  • A Hugging Face model accessible via API
  • A cloud-based model with API access

🚀 Installation

1. Clone this repository:

   git clone https://github.com/yourusername/smolagents-conversational-agent.git
   cd smolagents-conversational-agent
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

🔧 Setup

Setting Up a Language Model

You have several options for the language model:

Option 1: Local Model with LM Studio (Recommended for beginners)

1. Download and install LM Studio
2. Launch LM Studio and download a model (e.g., Mistral 7B, Llama 2, etc.)
3. Start the local server by clicking "Start Server"
4. Note the server URL (typically http://localhost:1234/v1)

Option 2: Using OpenRouter

1. Create an account on OpenRouter
2. Get your API key from the dashboard
3. Use the OpenRouter URL and your API key in the agent configuration

Option 3: Hugging Face API ( no more tested be careful )

1. If you have access to Hugging Face API endpoints, you can use them directly
2. Configure the URL and parameters in the agent interface

💻 Usage

Streamlit Interface (Recommended)

The Streamlit interface offers a more user-friendly experience with additional features:

1. Launch the Streamlit application:

   streamlit run streamlit_app.py
   

2. Access the interface in your web browser at http://localhost:8501

3. Configure your model in the sidebar:

   • Select the model type (OpenAI Server, Hugging Face API, or Hugging Face Cloud)
   • Enter the required configuration parameters
   • Click "Apply Configuration"

4. Start chatting with the agent in the main interface

Gradio Interface

The original Gradio interface is still available:

1. Launch the Gradio application:

   python app.py
   

2. Access the interface in your web browser at the URL displayed in the terminal (typically http://localhost:7860)

🌟 Features

Streamlit Interface Features

• Interactive Chat Interface: Engage in natural conversations with the agent
• Multiple Model Support:
  • OpenAI Server (LM Studio or other OpenAI-compatible servers)
  • Hugging Face API
  • Hugging Face Cloud
• Real-time Agent Reasoning: See the agent's thought process as it works on your request
• Customizable Configuration: Adjust model parameters without modifying code
• Data Visualization: Request and generate charts directly in the chat
• Code Execution: Run Python code generated by the agent within the chat interface
• Timezone Display: Check current time in different time zones
• Custom Icon: Uses a custom ico.webp icon for the application and sidebar

Agent Tools

The agent comes equipped with several powerful tools:

• Web Search: Search the web via DuckDuckGo to get up-to-date information
• Webpage Visiting: Visit and extract content from specific webpages
• Shell Command Execution: Run commands on your system (with appropriate security)
• File Operations: Create and modify files on your system
• Data Visualization: Generate charts and graphs based on your requests
• Code Execution: Run Python code within the chat interface

🧩 Extending the Agent

Adding Custom Tools

You can extend the agent with your own custom tools by modifying the app.py file:

@tool
def my_custom_tool(arg1: str, arg2: int) -> str:
    """Description of what the tool does
    Args:
        arg1: description of the first argument
        arg2: description of the second argument
    """
    # Your tool implementation
    return "Tool result"

Customizing Prompts

The agent's behavior can be customized by modifying the prompt templates in the prompts.yaml file.

📊 Visualization Examples

The agent can generate visualizations based on natural language requests. Try asking:

• "Show me a line chart of temperature trends over the past year"
• "Create a bar chart of sales by region"
• "Display a scatter plot of age vs. income"

🔍 Troubleshooting

• Agent not responding: Verify that your LLM server is running and accessible
• Connection errors: Check the URL and API key in your configuration
• Slow responses: Consider using a smaller or more efficient model
• Missing dependencies: Ensure all requirements are installed via pip install -r requirements.txt

📚 Examples

Here are some example queries you can try with the agent:

• "What's the current time in Tokyo?"
• "Can you summarize the latest news about AI?"
• "Create a Python function to sort a list of dictionaries by a specific key"
• "Explain how transformer models work in AI"
• "Show me a bar chart of population by continent"
• "Write a simple web scraper to extract headlines from a news website"

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

---

For more information on Hugging Face Spaces configuration, visit https://huggingface.co/docs/hub/spaces-config-reference