Some notes about these files: explicit.txt, general.txt, questionable.txt, sensitive.txt
Within those files, the first value is the 'post_id' (which you see in the URL of each danbooru image) and the second value is the 'score', I sorted the lines by score but this doesn't affect anything.
These were scraped slightly randomly with emphasize put on images that have higher scores.
The scraper's only other rule was to avoid images with the "animated" tag, since I doubt their taglist would fit a static image.