feat: checkbox for cleaning input dirs (#7)

[nautics889]

This update adds a checkbox to the main window. If checked, it cleans up all the files in input directories (decensor_input/ and also decensor_input_original/ in case if it deals with mosaic censorship).

Similar to what we had in #7 (credits to @ghost), but as a checkbox.
And it cleans only input directory, not decensor_ouput (cause i'm not really sure about how to design it, i mean i can add a button like "Remove all pictures in decensor_output/", but it's purpose would be quite obscure; maybe we should rather add setting output directory via UI).

---

CC: @naphteine

---

• (feat) add a checkbox in MainWindow to enable cleaning input directories after processing images.
• (feat) add clean_up_input_dirs flag attribute in Decensor class to manage the cleanup operation.
• (feat) add clean_input_directories() method in Decensor to remove .png, .jpg, and .jpeg files from input directories.
• (feat) add do_post_jobs() method being called after decensoring process is completed (later we can put here any other logic for post-processing as well).
• (feat) integrated logic to set clean_up_input_dirs via a signal emitted from MainWindow.

[nautics889]

Please, let me know if you think we've got to work on anything else within this PR

[naphteine]

Hello,

Thank you for your effort! Everything seems good at first glance. I will keep this open for a week to let people experiment and complain if something is not right, then merge. This will also give me some time to test it on my current Windows setup, as I had some problems installing everything.

[nautics889]

Sure, take your time.

test it on my current Windows setup, as I had some problems installing everything

Absolute truth, I spent quite some time trying to set an appropriate environment and CUDA version on win11. And yet I couldn't get it to work on any Python version newer than 3.7

[naphteine]

I have tested and saw no problems. As nobody said anything else I'm just merging it.

Absolute truth, I spent quite some time trying to set an appropriate environment and CUDA version on win11. And yet I couldn't get it to work on any Python version newer than 3.7

Yeah that's main problem with DCP currently. I had to go with Python 3.6. I'm planning to move from Tensorflow to PyTorch eventually as it was suggested in #9 but worried about if PyTorch would be best for this project. Well, thanks again for your work!