Add a PyO3 wrapper?

[Gowee]

Hi!

I am downloading public domain books with their pages cut in tiles from various libraries for archive purpose. The process involves merging these tiles to reconstruct the pages then assembling them into a PDF. I hope the process to be automated considering the amount of books and pages are not trivial.

While exploring options for dezoomifying libraries in Python, I've observed that the existing solutions are not well-maintained. As an alternative, dezoomify-rs appears to be a more reliable choice. However, to use dezoomify-rs from python, I need to resort to subprocess calls, invoking the binary and managing I/O via the filesystems.

I would like to propose adding a simple PyO3 wrapper to integrate with Python. This wrapper would provide functions that accepts configs or URLs and inputs and returns dezoomified image data.

[lovasoa]

Have you run into particular issues with the subprocess solution? I'd expect the cost of spawning a process to be tiny compared to the cost of downloading and assembling the tiles anyway... Isn't it ?