Add info about link header relationships to Memento

[Mr0grog]

Wayback Mementos carry additional info about related resources in the Link header. For example, here’s the header from https://web.archive.org/web/20171124151315id_/https://www.fws.gov/birds/:

<https://www.fws.gov/birds/>; rel="original",
<http://web.archive.org/web/timemap/link/https://www.fws.gov/birds/>; rel="timemap"; type="application/link-format",
<http://web.archive.org/web/https://www.fws.gov/birds/>; rel="timegate",
<http://web.archive.org/web/20050323155300/http://www.fws.gov:80/birds>; rel="first memento"; datetime="Wed, 23 Mar 2005 15:53:00 GMT",
<http://web.archive.org/web/20170929002712/https://www.fws.gov/birds/>; rel="prev memento"; datetime="Fri, 29 Sep 2017 00:27:12 GMT",
<http://web.archive.org/web/20171124151315/https://www.fws.gov/birds/>; rel="memento"; datetime="Fri, 24 Nov 2017 15:13:15 GMT",
<http://web.archive.org/web/20171228222143/https://www.fws.gov/birds/>; rel="next memento"; datetime="Thu, 28 Dec 2017 22:21:43 GMT",
<http://web.archive.org/web/20201011123440/http://www.fws.gov/birds>; rel="last memento"; datetime="Sun, 11 Oct 2020 12:34:40 GMT"

(Line breaks added for clarity.)

This follows the standard format for the Link header. The most accessible docs are at MDN.

It would probably be nice to surface this information in the Memento class in some useful way. We originally planned to do this in #2, but it wasn’t critical and there were enough open questions and options that it seemed worth waiting on coming up with a better design for.

Some possibilities:

• Simply parse the data generically, like the Requests package does:

  memento.links = {
      'original': {
          'url': 'https://www.fws.gov/birds/',
          'rel': 'original'
      },
      'timemap': {
          'url': 'http://web.archive.org/web/timemap/link/https://www.fws.gov/birds/',
          'rel': 'timemap',
          'type': 'application/link-format'
      },
      'timegate': {
          'url': 'http://web.archive.org/web/https://www.fws.gov/birds/',
          'rel': 'timegate'
      },
      'first memento': {
          'url': 'http://web.archive.org/web/20050323155300/http://www.fws.gov:80/birds',
          'rel': 'first memento',
          'datetime': datetime(2005, 3, 23, 15, 53, tzinfo=timezone.utc)
      },
      'prev memento': {
          'url': 'http://web.archive.org/web/20170929002712/https://www.fws.gov/birds/',
          'rel': 'prev memento',
          'datetime': datetime(2017, 9, 29, 0, 27, 12, tzinfo=timezone.utc)
      },
      'memento': {
          'url': 'http://web.archive.org/web/20171124151315/https://www.fws.gov/birds/',
          'rel': 'memento',
          'datetime': datetime(2017, 11, 24, 15, 13, 15, tzinfo=timezone.utc)
      },
      'next memento': {
          'url': 'http://web.archive.org/web/20171228222143/https://www.fws.gov/birds/',
          'rel': 'next memento',
          'datetime': datetime(2017, 12, 28, 22, 21, 43, tzinfo=timezone.utc)
      },
      'last memento': {
          'url': 'http://web.archive.org/web/20201011123440/http://www.fws.gov/birds',
          'rel': 'last memento',
          'datetime': datetime(2020, 10, 11, 12, 34, 40, tzinfo=timezone.utc)
      }
  }

• Since the original and memento relationships are redundant (all that info is already on Memento), we could drop them.

• We could make these a more special type than a dict so they can be passed to get_memento(), e.g:

  get_memento(memento.links['next memento'])

• We could add get_next_memento(), etc. as a shortcut to get_memento(), e.g:

  memento.get_next_memento()
  # Same as:
  get_memento(memento.links['next memento']['url'],
              memento.links['next memento']['datetime'],
              memento.mode)

• Since these are known, predictable links, we could add attributes directly to Memento instead of Memento.links:

  memento.first_memento = {...}
  memento.previous_memento = {...}
  memento.next_memento = {...}
  memento.last_memento = {...}

Another thing to consider is that we don’t currently have any special support for timemaps or timegates, so we can’t do anything special for those links. Basic dict parsing would probably be the lowest common denominator here. It’s not very special, but lets us treat everything the same.

Lots of ways we could go here. ¯\_(ツ)_/¯