read: operator could take a URL

[prescod]

Is your feature request related to a problem? Please describe.
Most of the information on the world is not on a local hard drive and an appreciable subset of it is accessible by URL. So URLs should be usable with read and import operators.

Describe the solution you'd like
read: and import: and other file-based operators should accept URLs to allow building distributed systems.

It should be possible to parameterize the URL with data from previous steps. Associating a JWT or other auth is also a non-trivial issue.

Describe alternatives you've considered
I could just use Python functions.

[starpit]

Great suggestion, thank you!