Add Pelican URL Parsing library to standardize how we handle pelican URLs

[jhiemstrawisc]

This new pelican_url package and its overall integration should feel like our beloved, familiar net/url package, but with some added niceties for Pelican URL parsing. In particular, it handles

• Query param validation
• Metadata discovery
• URL caching (to avoid frequent metadata lookup for URLs with the same discovery host)

A few caveats -- this grew more than I anticipated. The reviewer should pay particular attention to tests that I deleted -- in many cases, I moved the essence of the test to what I thought was a more appropriate location (such as within the pelican_url module), but there isn't a 1-1 relationship here. If I deleted something I shouldn't have, do let me know.

[jhiemstrawisc]

One to-do that came from a discussion with Brian -- go back through this and restore any exported APIs from the client to have the same signature. There were a few places where I changed whether a URL or a string were passed.

[turetske]

One to-do that came from a discussion with Brian -- go back through this and restore any exported APIs from the client to have the same signature. There were a few places where I changed whether a URL or a string were passed.

@jhiemstrawisc Shall I wait until this is done?

[jhiemstrawisc]

One to-do that came from a discussion with Brian -- go back through this and restore any exported APIs from the client to have the same signature. There were a few places where I changed whether a URL or a string were passed.

@jhiemstrawisc Shall I wait until this is done?

Probably a good idea for now, I think I'll have to adjust a few things that are non-trivial