Make pyhf contrib download be able to handle multiple compression types

[matthewfeickert]

Description

@alexander-held has pointed out that at the moment (pyhf v0.6.2) pyhf contrib download can't download the probability models for:

• https://www.hepdata.net/record/ins1827025 (has models as .tar archive)
• https://www.hepdata.net/record/ins1831504 (has models as .zip archive)

as they both don't have the models compressed as tar.gz. At the moment, pyhf contrib download can only handle .tar.gz

pyhf/src/pyhf/contrib/utils.py

Lines 56 to 64 in cc7cf0f

	with requests.get(archive_url) as response:
	if compress:
	with open(output_directory, "wb") as archive:
	archive.write(response.content)
	else:
	with tarfile.open(
	mode="r|gz", fileobj=BytesIO(response.content)
	) as archive:
	archive.extractall(output_directory)

but that should be expanded.

It is reasonable to have pyhf contirb download be more robust, but ti would also be helpful if HEPData archive uploads had a standardized archive compression format.

cc @kratsg

[matthewfeickert]

This is related to Issue #1111.