Merge the IIIF and File imports

A refactoring of the file import could allow importing any amount of images, PDF or IIIF manifests or collections at once without the current restrictions:

• Do not create any 'Import' folder on the backend when importing from files
• Do not have the task make or use a folder for the whole import task, unless the user selected an existing folder to import into (frontend#235 (closed))
• For files that have a single page (images, single-page PDFs), create the element all alone without a folder and use the filename as it's name (importing lol.jpg becomes a lol element)
• For files that have multiple pages (multi-page PDFs), create a folder with the PDF's filename and have numbered pages (lol.pdf becomes a bunch of 1 2 3 pages in a lol folder)
• For files with complex structures (IIIF), let the file-specific function deal with it.
• For files that have existing elements (searching by name), skip them just like the current import. For IIIF files, do not skip anything (will be handled in #13 (closed))

This can have side effects on a few related issues:

• Strange page numbers when importing images (everything is numbered 1), because page numbers were reset with each file
• Lack of error handling causing a file import to crash on any exception for a file