Add Discord timestamp converter and transformer

[AbstractUmbra]

Summary

This PR adds both a converter and transformer for use with Discord's timestamp functionality. This is made easier for the end user with the addition of the @time utility in the client, and as such I felt the need to support this.

Checklist

• If code changes were made then they have been tested.
  • I have updated the documentation to reflect the changes.
• This PR fixes an issue.
• This PR adds something new (e.g. new method or parameters).
• This PR is a breaking change (e.g. methods or parameters removed/renamed)
• This PR is not a code change (e.g. documentation, README, ...)

[AbstractUmbra]

Some of my own thoughts on it.

[AbstractUmbra]

Not keen on re-creating the same constant for both transformer and converter, but I wasn't sure where else I could put this to import from outside of utils.

[AbstractUmbra]

As this uses the built-in datetime.datetime class, I opted for naming all resources DatetimeX, even thought this is more for timestamps than datetimes.
Feels unclean to me but naming this eludes me.

[vmphase]

A few suggestions for the implementation:

• The current pattern requires a style flag (e.g., :R), but, iirc, discord timestamps are valid without one (defaulting to :f). Probably, change the pattern to r'<t:(\d+)(?::[tTdDfFsSR])?>' to support both.
• fromtimestamp(n).replace(tzinfo=...) creates a naive local time before converting. It's safer to use datetime.fromtimestamp(int(match[1]), tz=datetime.timezone.utc) directly.

I’d also consider using __slots__ = ('argument',) in BadDatetimeArgument to reduce memory overhead.

[AbstractUmbra]

A few suggestions for the implementation:

• The current pattern requires a style flag (e.g., :R), but, iirc, discord timestamps are valid without one (defaulting to :f). Probably, change the pattern to r'<t:(\d+)(?::[tTdDfFsSR])?>' to support both.
• fromtimestamp(n).replace(tzinfo=...) creates a naive local time before converting. It's safer to use datetime.fromtimestamp(int(match[1]), tz=datetime.timezone.utc) directly.

I’d also consider using __slots__ = ('argument',) in BadDatetimeArgument to reduce memory overhead.

Good calls on those, thank you! 'twas an oversight on my part.

None of the other error classes implement said slots however, so I'll opt out of adding those for now.

[AbstractUmbra]

Additionally I chose .match() over .search() since we'd expect to only parse the <t:xxxxxx:x> input only rather than search an arbitrary string.