Kelvin to RGB in python

rgb_to_kelvin.py

"""
    Based on: http://www.tannerhelland.com/4435/convert-temperature-rgb-algorithm-code/
    Comments resceived: https://gist.github.com/petrklus/b1f427accdf7438606a6
    Original pseudo code:
    
    Set Temperature = Temperature \ 100
    
    Calculate Red:

    If Temperature <= 66 Then
        Red = 255
    Else
        Red = Temperature - 60
        Red = 329.698727446 * (Red ^ -0.1332047592)
        If Red < 0 Then Red = 0
        If Red > 255 Then Red = 255
    End If
    
    Calculate Green:

    If Temperature <= 66 Then
        Green = Temperature
        Green = 99.4708025861 * Ln(Green) - 161.1195681661
        If Green < 0 Then Green = 0
        If Green > 255 Then Green = 255
    Else
        Green = Temperature - 60
        Green = 288.1221695283 * (Green ^ -0.0755148492)
        If Green < 0 Then Green = 0
        If Green > 255 Then Green = 255
    End If
    
    Calculate Blue:

    If Temperature >= 66 Then
        Blue = 255
    Else

        If Temperature <= 19 Then
            Blue = 0
        Else
            Blue = Temperature - 10
            Blue = 138.5177312231 * Ln(Blue) - 305.0447927307
            If Blue < 0 Then Blue = 0
            If Blue > 255 Then Blue = 255
        End If

    End If
"""

import math
def convert_K_to_RGB(colour_temperature):
    """
    Converts from K to RGB, algorithm courtesy of 
    http://www.tannerhelland.com/4435/convert-temperature-rgb-algorithm-code/
    """
    #range check
    if colour_temperature < 1000: 
        colour_temperature = 1000
    elif colour_temperature > 40000:
        colour_temperature = 40000
    
    tmp_internal = colour_temperature / 100.0
    
    # red 
    if tmp_internal <= 66:
        red = 255
    else:
        tmp_red = 329.698727446 * math.pow(tmp_internal - 60, -0.1332047592)
        if tmp_red < 0:
            red = 0
        elif tmp_red > 255:
            red = 255
        else:
            red = tmp_red
    
    # green
    if tmp_internal <=66:
        tmp_green = 99.4708025861 * math.log(tmp_internal) - 161.1195681661
        if tmp_green < 0:
            green = 0
        elif tmp_green > 255:
            green = 255
        else:
            green = tmp_green
    else:
        tmp_green = 288.1221695283 * math.pow(tmp_internal - 60, -0.0755148492)
        if tmp_green < 0:
            green = 0
        elif tmp_green > 255:
            green = 255
        else:
            green = tmp_green
    
    # blue
    if tmp_internal >=66:
        blue = 255
    elif tmp_internal <= 19:
        blue = 0
    else:
        tmp_blue = 138.5177312231 * math.log(tmp_internal - 10) - 305.0447927307
        if tmp_blue < 0:
            blue = 0
        elif tmp_blue > 255:
            blue = 255
        else:
            blue = tmp_blue
    
    return red, green, blue
    


if __name__ == "__main__":
    print("Preview requires matplotlib")
    from matplotlib import pyplot as plt

    step_size = 100
    for i in range(0, 15000, step_size):
        color = list(map(lambda div: div/255.0, convert_K_to_RGB(i))) + [1]
        print(color)
        plt.plot((i, i), (0, 1), linewidth=step_size/2.0, linestyle="-", color=color)
    
    plt.show()  

With increasing temperature, output seems like red becoming less saturated, passing through white, and becoming more saturated blue. I expected red turning to orange, yellow, becoming less saturated and turning into white, then becoming more saturated blue. I don't see that yellow part here. It does match https://en.wikipedia.org/wiki/Color_temperature so I guess it is correct. But I was expecting something more like the heated metal colour discussed at: https://physics.stackexchange.com/questions/304299/how-heated-metal-colors-relate-to-black-body-color-at-the-same-temperature. The heated metal colour matches colour temperature of light bulbs more closely.

Edit: For using an RGB led to approximate colour temperature, https://github.com/carpdiem/Color-Match gave better results. Note that there you need to know the wavelength of LEDs to get accurate results.

Edit: Another source is https://github.com/jonls/redshift/blob/master/src/colorramp.c. It uses a lookup table and interpolation, and is much faster than Color-Match.

@dreamlayers Thx for all the links, especially the redhift one really helped be get this going for my ESPHome project. For anyone interested, see: https://github.com/markusressel/CustomRGBWLight/blob/master/components/ColorTempRGBWLight/ColorTempRGBWLight.h

I suggest using numpy for a much shorter code:

import numpy as np

def convert_K_to_RGB(colour_temperature: float) -> np.ndarray:
    """
    Converts from K to RGB, algorithm courtesy of
    http://www.tannerhelland.com/4435/convert-temperature-rgb-algorithm-code/
    """
    # Range check.
    colour_temperature = np.clip(colour_temperature, 1000, 40000)

    tmp_internal = colour_temperature / 100.0

    # Red.
    if tmp_internal <= 66:
        red = 255
    else:
        red = 329.698727446 * (tmp_internal - 60)**-0.1332047592

    # Green.
    if tmp_internal <= 66:
        green = 99.4708025861 * np.log(tmp_internal) - 161.1195681661
    else:
        green = 288.1221695283 * (tmp_internal - 60)**-0.0755148492

    # Blue.
    if tmp_internal >= 66:
        blue = 255
    elif tmp_internal <= 19:
        blue = 0
    else:
        blue = 138.5177312231 * np.log(tmp_internal - 10) - 305.0447927307

    return np.clip((red, green, blue), 0, 255)

I have also changed your main so that we can zoom in infinitely without seeing empty space between the color bars.

if __name__ == "__main__":

    from matplotlib import pyplot as plt

    STEP_SIZE = 100
    temperatures = np.arange(0, 15000, STEP_SIZE)
    colors = [convert_K_to_RGB(temperature)/255 for temperature in temperatures]
    fig, ax = plt.subplots()
    ax.set_xlim(temperatures[0], temperatures[-1])
    ax.set_ylim(0, 1)
    ax.bar(temperatures, height=1, width=STEP_SIZE, align="edge", color=colors)
    fig.show()

With 10 more minutes of work, we could make the function support polymorphism so that it works on a whole numpy array at once.

Thank you - useful for cases when np is available, I think both pure-python and numpy based versions are useful for different scenarios