Output files generated by code (e.g., images, saved data) cannot be preserved or accessed in VPL

[Yerosius]

Hi,

Currently, any files generated during code execution in VPL (especially in languages like Python) are not preserved or accessible after the run finishes. This causes major limitations for assignments or exercises involving:

• Saving plots/images using libraries like matplotlib, cv2, or PIL
• Writing logs or output data to files for further inspection
• Reading/writing files as part of a program’s normal behavior

Examples:

• A Python script generates a .png image using plt.savefig("plot.png"), but the file is not accessible after execution.
• A program writes data to a .txt or .csv file, but students cannot download or verify the content.
• Any workflow involving persistent intermediate output files is essentially broken.

This is especially problematic for courses in data science, computer vision, or file I/O programming.

Suggested improvement:

• Provide a way to persist or download the output files generated by the user’s code during execution.
• For example, any file created in the working directory (or specific whitelist folders) could be shown in the file tree after execution, or be downloadable.

This feature would greatly enhance VPL's usefulness for real-world programming tasks.

Thanks for your work and consideration!

[Astor-Bizard]

Hello,

Yes, this is currently a limitation of "jailed" execution. I believe such improvements should not be taken too lightly - as we should always remember we are executing arbitrary code written by users.

As an idea of what could be done: we tweaked a bit mod_vpl code on our end to send webservice token for executing user to the jail server. Then, the execution ends up making a webservice call to push these files to the VPL activity.
Though not ideal, this may help you for the time being. :)