Move download datasets logic our of download_utils.py (#9253)

* add modifications to allow for pdf download

* lint

* lint

* separate content vars

* move download datasets into separate sub module

* wip

* update commandline

* move util functions to utils

* use renamed module

* refactor to use new module name

* refactor to use new module name

* cr

* lint

* add missing import

* add missing import

* add missing import; add back download_llama_datasets

* fix command line

* get metadata from hub

* fix command line

* wip