Fix Python Dependency Installation in MCP Code Sandbox (Closes #4)

[bardusco]

Summary

This PR resolves issue #4, where Python projects using requirements.txt failed to execute due to missing system dependencies. The error was reported as "undefined", making debugging difficult.

Changes Introduced

• Updated the run_project.go logic to correctly install Python dependencies using uv in a virtual environment.
• Ensured required system packages (apt-utils, build-essential) are installed before running uv pip install.
• Set DEBIAN_FRONTEND=noninteractive to prevent interactive prompts in Docker builds.

Why this Fix?

• 🛠️ Fixes project execution failures by ensuring dependencies are installed correctly before running the Python script.

Steps to Test

1. Create a Python project with main.py and requirements.txt.
2. Add dependencies (e.g., requests==2.31.0).
3. Run the project using:

   mcp__code-sandbox__run_project \
       --projectDir /path/to/project \
       --language python \
       --entrypointCmd "python main.py"

4. Verify that:
   • The sandbox successfully creates a Docker container.
   • Dependencies are installed correctly using uv.
   • The Python script runs without failure.
   • The tool returns the resource URI of the container logs.

Closes Issue:

✅ Closes #4 — Python projects with requirements.txt now execute correctly inside the MCP Code Sandbox.

---

💡 Let me know if you need any modifications! 🚀