# danbooru-metadata
Dump of various portions danbooru's metadata, as of Feburary 2023. Everything was taken directly from their JSON API.

The directory structure follows the Danbooru20XX format of each subfolder for a record type being the record's ID modulo 1000. The `.zip` files can sometimes hold hundreds of thousands of small JSON files when put together, so use caution when extracting.