Lift linearity restriction for `Rgb` and `Luma` and update docs

[Ogeon]

This removes the requirement that Rgb and Luma colors have to be linear for arithmetic operators and a few other traits to be available. There are many cases where non-linear colors are fine (or even required) and hard blocking them ended up being a bit too harsh.

The crate root documentation is also updated to hopefully be a bit more usage oriented and give more useful examples up-front.

[github-actions[bot]]

Benchmark for 9a83392

Click to view benchmark

| Test | Base | PR | % | |------|--------------|------------------|---| | Cie family/lab to lch | 2.2±0.01µs | 2.2±0.04µs | 0.00% | | Cie family/lab to xyz | 547.8±4.19ns | 550.3±5.40ns | +0.46% | | Cie family/lch to lab | 1696.8±11.99ns | 1704.3±25.26ns | +0.44% | | Cie family/linsrgb to xyz | 3.1±0.02µs | 3.1±0.03µs | 0.00% | | Cie family/linsrgb to xyz - wide::f32x8 | **543.6±3.86ns** | 557.4±2.96ns | **+2.54%** | | Cie family/xyz to lab | **15.1±0.08µs** | 15.2±0.17µs | **+0.66%** | | Cie family/xyz to yxy | 470.2±3.05ns | 469.3±3.12ns | -0.19% | | Cie family/yxy to xyz | 378.2±2.21ns | 378.3±2.83ns | +0.03% | | Matrix functions/matrix_inverse | **8.2±0.04ns** | 8.3±0.16ns | **+1.22%** | | Matrix functions/multiply_3x3 | 7.5±0.04ns | 7.5±0.06ns | 0.00% | | Matrix functions/multiply_rgb_to_xyz | 4.6±0.03ns | 4.6±0.21ns | 0.00% | | Matrix functions/multiply_xyz | 4.6±0.04ns | 4.6±0.03ns | 0.00% | | Matrix functions/multiply_xyz_to_rgb | 4.6±0.03ns | 4.6±0.03ns | 0.00% | | Matrix functions/rgb_to_xyz_matrix | 21.5±0.17ns | 21.5±0.10ns | 0.00% | | Rgb family/hsl to hsv | 460.7±3.20ns | 461.0±4.73ns | +0.07% | | Rgb family/hsl to linear hsl | 9.4±0.05µs | **9.1±0.06µs** | **-3.19%** | | Rgb family/hsl to rgb | 1457.1±9.42ns | **1422.8±12.01ns** | **-2.35%** | | Rgb family/hsv to hsl | 606.4±9.61ns | 606.8±9.82ns | +0.07% | | Rgb family/hsv to hwb | 162.3±1.46ns | **159.9±1.09ns** | **-1.48%** | | Rgb family/hsv to linear hsv | 8.8±0.07µs | **8.7±0.05µs** | **-1.14%** | | Rgb family/hsv to rgb | 1350.0±10.43ns | **1326.6±8.36ns** | **-1.73%** | | Rgb family/hwb to hsv | 334.2±3.87ns | 333.2±3.78ns | -0.30% | | Rgb family/hwb to linear hwb | **9.1±0.05µs** | 9.2±0.06µs | **+1.10%** | | Rgb family/linear hsl to hsl | **7.6±0.04µs** | 7.7±0.07µs | **+1.32%** | | Rgb family/linear hsv to hsv | 7.2±0.05µs | 7.2±0.06µs | 0.00% | | Rgb family/linear hwb to hwb | 7.7±0.07µs | 7.7±0.04µs | 0.00% | | Rgb family/linsrgb to rgb | 3.9±0.03µs | 3.9±0.03µs | 0.00% | | Rgb family/linsrgb to rgb - wide::f32x8 | 5.3±0.03µs | 5.3±0.04µs | 0.00% | | Rgb family/linsrgb_f32 to rgb_u8 | 654.4±24.03ns | 650.4±4.02ns | -0.61% | | Rgb family/linsrgb_f64 to rgb_u8 | 835.2±4.63ns | 839.8±5.50ns | +0.55% | | Rgb family/rgb to hsl | **568.7±2.63ns** | 572.5±19.95ns | **+0.67%** | | Rgb family/rgb to hsl - wide::f32x8 | 353.9±1.91ns | 355.0±2.57ns | +0.31% | | Rgb family/rgb to hsv | 362.3±3.56ns | **357.5±7.09ns** | **-1.32%** | | Rgb family/rgb to hsv - wide::f32x8 | 315.7±2.13ns | 315.9±1.64ns | +0.06% | | Rgb family/rgb to linsrgb | **5.3±0.05µs** | 5.4±0.06µs | **+1.89%** | | Rgb family/rgb to linsrgb - wide::f32x8 | 4.0±0.04µs | 4.0±0.04µs | 0.00% | | Rgb family/rgb_u8 to linsrgb_f32 | 235.2±3.67ns | 235.5±3.58ns | +0.13% | | Rgb family/rgb_u8 to linsrgb_f64 | 237.0±6.89ns | 236.3±1.43ns | -0.30% | | Rgb family/xyz to linsrgb | **5.7±0.13µs** | 5.9±0.04µs | **+3.51%** | | Rgb family/xyz to linsrgb - wide::f32x8 | 892.0±8.05ns | 892.4±5.90ns | +0.04% |

[Ogeon]

bors r+

[bors[bot]]

Build succeeded:

• ci