Zotero can extract metadata from PDFs, but at present it can't import folder structure. We're always happy to take patches: https://github.com/zotero/zotero