Refactor Open Library to use Github for adding New Identifiers

[mekarpeles]

Currently, the way that new Identifiers are added to Open Library books and authors is through editing https://openlibrary.org/config/edition.yml?m=edit via the infogami database, which requires super librarian or admin permission.

This is blocking lots of new identifier types listed in #8271.

Ideally, we'd use /plugin/openlibrary/editions.page which can be edited by community pull request submissions.

UPDATE:

openlibrary/openlibrary/plugins/upstream/utils.py

Lines 1206 to 1220 in ec3e569

	@web.memoize
	def _get_edition_config():
	"""Returns the edition config.
	The results are cached on the first invocation. Any changes to /config/edition page require restarting the app.
	This is is cached because fetching and creating the Thing object was taking about 20ms of time for each book request.
	"""
	thing = web.ctx.site.get('/config/edition')
	classifications = [Storage(t.dict()) for t in thing.classifications if 'name' in t]
	identifiers = [Storage(t.dict()) for t in thing.identifiers if 'name' in t]
	roles = thing.roles
	return Storage(
	classifications=classifications, identifiers=identifiers, roles=roles
	)

• Open Library currently loads editions.yml in our app.
• We want to simply move the current identifiers out of infogami into a yml file in openlibrary repo
• When we load infogami editions.yml, we will add in the identifiers from the book-identifiers.yml
• restrict access of edition.yml to only be admins -- librarians will submit PRs moving forward in github