Add more Line attributes

[itzpr3d4t0r]

Adds the following Line attributes:

• center
• centerx/y
• angle
• slope

Credits

Geometry Project:
For code, docs and tests:
@novialriptide @Emc2356 @itzpr3d4t0r @ScriptLineStudios @avaxar @Matiiss @newpaxonian @maqa41 @blankRiot96
Also thanks to @Starbuck5 for kickstarting the idea and the occasional help!

Functionality added in this PR

• geometry.pyi
  https://github.com/pygame-community/pygame-geometry/blame/main/geometry.pyi
  Credits to @Emc2356 @itzpr3d4t0r @novialriptide @ScriptLineStudios @avaxar

• geometry.rst
  https://github.com/pygame-community/pygame-geometry/blame/main/docs/line.rst
  Credits to @Emc2356 @itzpr3d4t0r @ScriptLineStudios @avaxar

• line.c
  https://github.com/pygame-community/pygame-geometry/blame/main/src_c/line.c
  Credits to @Emc2356 @itzpr3d4t0r @novialriptide @ScriptLineStudios @avaxar

• geometry.h
  https://github.com/pygame-community/pygame-geometry/blame/main/src_c/include/geometry.h
  Credits to @Emc2356 @itzpr3d4t0r @maqa41

• geometry_test.py
  https://github.com/pygame-community/pygame-geometry/blame/main/test/test_line.py
  Credits to @Emc2356 @itzpr3d4t0r @novialriptide @ScriptLineStudios @avaxar

[MyreMylar]

The angle attribute seems like it is a bit off from what I would expect in pygame. Not sure if this is because of the reversed y axis, or the difference in chosen calculations

I was expecting results something like this I guess:

If I apply rotations to a Vector2 of (0,-10) then it lines up with the image above.

However a line of ((0,0),(0,-10)) currently has an angle of 90° and a line of ((0,0),(10,0)) has a reported angle of -0.0°.

How does everyone else feel about this difference?

[MyreMylar]

How does everyone else feel about this difference?

I guess I should also ask - is there a line.rotate(angle, origin) function on the way? And what are the results of it with an origin of (0,0) compared to doing the same rotation on a Vector2?

The way we rotate stuff geometrically doesn't particularly matter - as long as we are internally consistent.

[aatle]

@MyreMylar you are comparing rotating a line by a certain angle, with the inherent "angle" of a line.
The angle of a line corresponds with the proposed Vector2 angle and angle_rad properties in #3222, and should have similar behavior for special cases.

The angle of a line is similar to atan(line.slope) (quadrant incorrect), and is better described as the angle of a vector representing the line, where a is the origin.

The angle should be in degrees, for consistency. There should probably be both a degrees and a radians implementation for every angle method, also for circle too. This would be consistent with Vector2.

[aatle]

Perhaps the vector corresponding to a line may be a useful attribute. (bx - ax, by - ay).

Changes have been addressed

[aatle]

I think this is inaccurate. The result here should either be an exception or +-inf.
If inf, then I would think that for by - ay == 0 case it would be nan instead.

In any case, the vertical line behavior needs to be documented.

[aatle]

I don't think the angle should be negated.
Also it should use some form of atan2 instead of atan, to reach the full 360 deg range, as had been specified in the docs.
There's also special cases like a == b, inf, etc., where it maybe should emulate python math.atan2(line.by - line.ay, line.bx - line.ax) (also see #3222).