Genezio Python Executor

Overview

This repository provides a FastAPI-based service that enables remote execution of Python scripts with optional dependency installation. Users can upload their Python files and execute main.py in a sandboxed environment.

Features

• Remote Code Execution: Upload and execute Python scripts in a controlled environment.
• Dependency Management: Automatically install specified Python dependencies before execution.
• File-Based Input/Output: Supports file uploads and returns execution results, including standard output, errors, and generated files.
• Security: Runs in an isolated /tmp directory with automatic cleanup after execution.
• Multipart Response: Returns execution results and output files in a multipart response format.

Deployment

Deploy your own instance of this executor using Genezio:

API Usage

Endpoint: /execute

• Method: POST
• Description: Accepts Python files and executes main.py in an isolated environment.

Request Parameters

Parameter	Type	Description
dependencies	string (query)	Comma-separated list of Python packages to install. Default is empty.
files	multipart/form-data	One or more Python files (including main.py).

Example Request

curl -X POST "http://localhost:8000/execute?dependencies=requests,numpy" \
    -F "[email protected]" \
    -F "[email protected]"

Example Response

The response is a multipart/form-data containing:

• stdout: Standard output of the script.
• stderr: Standard error output.
• Generated Files (if any) from the script execution.

Example Response Content:

{
  "stdout": "Execution completed successfully.\n",
  "stderr": "",
  "output.txt": "This is the generated file content."
}

Implementation Details

1. Dependency Installation

   • Dependencies are parsed from the query string and installed using pip into /tmp/deps.
   • The PYTHONPATH is updated to include /tmp/deps before executing the script.

2. File Handling

   • Uploaded files are stored in /tmp/code_exec.
   • The script main.py must be present, or execution will fail.

3. Execution Flow

   • Uploaded files are saved.
   • Dependencies (if any) are installed.
   • main.py is executed within the /tmp/code_exec directory.
   • The output is captured and returned in a multipart response.
   • The /tmp/code_exec directory is cleaned up after execution.

Security Considerations

• The execution environment is cleaned after every request.
• Runs in an isolated directory to prevent conflicts.
• Ensure trusted code execution to avoid malicious scripts.

License

This project is licensed under the MIT License.