Add global way of accessing Delta time and Unscaled Delta time

[Cautami]

• This PR aims to close Add improved support for unscaled time godot-proposals#7775
  Which seems to be pretty positively received and additionally seems to have a general consensus that both deltas and unscaled variants should be added to the Engine class

As such, I've added the following methods:

Engine.get_delta_time()
Engine.get_unscaled_delta_time()
Engine.get_physics_delta_time()
Engine.get_unscaled_physics_delta_time()

get_delta provides the exact same information the _process method provides.
get_physics_delta is the same way but for _physicsProcess

The unscaled methods provide the user a unscaled version of both of these deltas, as they are never multiplied by timescale in the first place, this is the most accurate and convenient way of getting unscaled deltas.

Notes

This is a similar to #86766, just with the extra granularity, and I guess also slightly opinionated, it seemed like the consensus on the proposal was to refer to them as Deltas, makes sense to me and matches what the average user would expect I think.

As far as implementation goes, I'm aware im duplicating process_step here, but I'm not entirely sure how to cleanly handle this.

1. I could remove the get_process_step function (its only used a handful of times internally) and replace it with the new get_unscaled_delta, but I'm worried this will break compatibility with something else.
2. Don't add get_unscaled_delta, and have it live through get_process_step (while still being exposed to the scripting api as get_unscaled_delta, but it feels silly to have get_process_step live alongside the more conventionally named deltas.
3. Rename all the _deltas internally to _step to match process_step, this would internally make the most sense but it could be confusing to have it exposed under a different name.
4. Just abandon hope and call them all "steps", which would make the exposed api and the internal api match, at the cost of just being a little more confusing to the average user.

Screencast_20251029_210032.webm

[Cautami]

Added proper documentation (mostly taken from #86766, thank you Calinou), and I believe its functional and ready to review now. I'm still not certain which way I should take the implementation, but as far as my testing is concerned it works and I'm ready to make any changes to the implementation.