@
MelBr, I add batches of 'books' into a library called Review. In there I first decide whether or not I really want the 'book'. If I do then I fix up title, author, metadata etc and use Find Duplicates to check for duplicates with my Main library. As I fix them I move them to my Main library. If I don't want them then I remove them.
When a batch is complete I manually delete the originals, they are always in a folder that has a name indicating source and date of download, so that's easily done.
I find using a separate library makes processing of batches of new 'books' easier - and it keeps my Main library unpolluted by duplicates, author name variants etc.
I don't do any tagging when 'books' are in Review as I haven't worked out a technique to have a controlled vocabulary for an empty library - but nor have I tried very hard
BR
My 'books' are public domain 'papers' from academia, 'reports' from govt agencies, quangos, think tanks etc, 'reprints' from the media etc.