Replace now-deprecated imghdr module

[MaggieFero]

Describe the bug
The imghdr module is deprecated as of Python 3.11, and will be removed in Python 3.13. The deprecation notice is here: https://docs.python.org/3/library/imghdr.html but the PEP referenced as featuring details and alternatives has only details, no specific alternative.

We'll need to find an alternative to imghdr by the time we move to 3.13, and when we do we can remove the linter exception introduced into bookwyrm/connectors/abstract_connector.py with PR #3303 as part of the Python 3.11 upgrade.

To Reproduce
See https://docs.python.org/3/library/imghdr.html and https://peps.python.org/pep-0594/#imghdr for details of the deprecation

Expected behavior
We should find a replacement! <3

Screenshots
N/A

Instance
All

Additional context
N/A

[Minnozz]

Pillow (an existing dependency) supports detecting image types by opening the image and looking at the format property: https://pillow.readthedocs.io/en/latest/reference/Image.html#PIL.Image.Image.format

Both places in the code where this functionality is used (that I could find), the purpose is to generate a random filename with an extension that matches the encoding of the image:

bookwyrm/bookwyrm/views/books/books.py

Lines 156 to 162 in 9e7b040

	try:
	image_content, extension = get_image(url)
	except: # pylint: disable=bare-except
	return None
	if not image_content:
	return None
	image_name = str(uuid4()) + "." + extension

bookwyrm/bookwyrm/models/fields.py

Lines 505 to 510 in 9e7b040

	image_content, extension = get_image(url)
	if not image_content:
	return None
	image_name = f"{uuid4()}.{extension}"
	return [image_name, image_content]

Using Pillow this way would parse the entire image instead of just the header, so that makes it less efficient.