Agents Course

This course comes from the HuggingFace course: course

Features

• In process

Project Setup Guide

This project uses uv as the package manager for faster and more reliable dependency management, and includes a Makefile for streamlined development workflows.

Prerequisites

• Python 3.12 or higher
• Git
• uv package manager (will be installed automatically via Makefile)

Quick Start

1. Install uv (if not already installed)

make install_uv

2. Create Virtual Environment

make venv

3. Activate Virtual Environment

Windows:

.\activate.ps1

Linux/MacOS:

source .venv/bin/activate

4. Install Dependencies

make install

This will install all required packages using uv for faster dependency resolution and installation.

Available Make Commands

The project includes a comprehensive Makefile with the following commands:

Command	Description
make install_uv	Install uv package manager
make venv	Create virtual environment
make install	Install dependencies using uv
make test	Run tests with pytest
make format	Format code with black and isort
make lint	Run linting tools (format, flake8, pyupgrade)
make check-all	Run all checks and tests
make clean	Clean up environment
make build	Build Docker image
make run	Run Docker container
make help	Display all available commands

Running Applications

Streamlit App

To run the Streamlit application, simply execute the provided PowerShell script from the root directory:

.\run_st_fapp.ps1

This will start the Streamlit app located at src/first_agent/st_app.py.

Gradio App (Dockerized)

The Gradio app in src/first_agent/ should be run in a Docker container due to security issues with current library versions that are mitigated in a containerized environment.

To dockerize and run the Gradio app: see README.md in src/first_agent/.

Environment Configuration

1. Create a .env file in the root directory:

touch .env # For Linux/MacOS# or manually create .env file in Windows

2. Add the following environment variables to your .env file:

# Required SettingsDEEPSEEK_API_KEY="your-deepseek-api-key"# Optional Settings (with defaults)MYAPP_DEBUG=falseMYAPP_ENVIRONMENT=dev# Options: dev, prod, stagingMYAPP_APP_NAME="MyApp"

Environment Variables Description

Variable	Required	Default	Description
DEEPSEEK_API_KEY	Yes	-	API key for DEEPSEEK platform
MYAPP_DEBUG	No	false	Enable/disable debug mode
MYAPP_ENVIRONMENT	No	dev	Application environment (dev/prod/staging)
MYAPP_APP_NAME	No	MyApp	Name of the application

Usage

The application uses Pydantic settings for configuration management. You can import and use the settings in your code like this:

fromconfig.settingsimportsettings# Access configuration valuesapi_key=settings.deepseek_api_keydebug_mode=settings.debugcurrent_env=settings.environment

Development

Project Structure

├── config/
│ └── settings.py # Configuration management
├── src/
│ └── first_agent/ # Main agent application
│ ├── app.py # Main application
│ ├── st_app.py # Streamlit app
│ ├── Gradio_UI.py # Gradio interface
│ ├── Streamlit_UI.py # Streamlit interface
│ └── Dockerfile # Docker configuration for Gradio app
├── Makefile # Development commands
├── run_st_fapp.ps1 # Script to run Streamlit app
├── activate.ps1 # Windows activation script
├── .venv/ # Virtual environment (not in repo)
├── .env # Environment variables (not in repo)
├── .gitignore # Git ignore file
└── README.md # This file

Development Workflow

1. Setup: Use make venv and make install for initial setup
2. Code Quality: Use make format and make lint before committing
3. Testing: Run make test to ensure all tests pass
4. All Checks: Use make check-all for comprehensive validation

Adding New Settings

To add new configuration settings:

1. Add the environment variable to your .env file
2. Add the corresponding field in config/settings.py
3. Update this README with the new variable details

Contributing

1. Create a new branch for your feature
2. Make your changes
3. Run make check-all to ensure code quality
4. Submit a pull request

Security Notes

• Never commit the .env file to version control
• Keep your API keys and secrets secure
• Use environment variables for sensitive information
• Important: Use the dockerized Gradio app to avoid security vulnerabilities in current library versions

License

[Your License Here]

Data Structure

The dashboard uses synthetic data generated with the following structure:

Product Nomenclature

• product_id: Unique identifier for each product
• product_type: Category (MEUBLE or DECO)
• product_name: Name of the product

Transactions

• transaction_id: Unique identifier for each transaction
• date: Transaction date
• order_id: Order identifier
• client_id: Customer identifier
• prod_id: Product identifier
• prod_price: Product price
• prod_qty: Quantity sold

Usage

1. The main chart shows product sales distribution
2. Use the animation controls to switch between different views
3. View summary statistics in the metrics section
4. Toggle the "Show Raw Data" checkbox to see the complete dataset

Dependencies

Dependencies are managed using uv for faster installation and resolution. Main dependencies include:

• streamlit
• ipyvizzu
• pandas
• loguru

To see the complete list of dependencies, check the requirements.txt file.