How to create a node group to quickly select a color from a list for semantic segmentation

Question

I have a list with 150 colors, all representing a type of object in a scene (for example: walls, ceiling, grass, cabinet, painting, water, chair, table etc).

you can see the list here

I want to create a reusable node group to quickly change an objects color.

What I have done in the past (if i had less than 10 colors):

Node Group

Add every color as RGB node, group them and name every output according to the color

Pros: Colors are available to use in procedural materials, I can hide unneeded colors quickly

Cons: potentially long list and inconvinient to set up for large amount of colors

Seperate material for each color

Pros: quickly apply colors to objects, searchable

Cons: inconvinient to set up for a large amount of colors, not usable for procedural materials

---

Any idea for a more convenient way to achieve this is greatly appreciated. Easy to set up would be great as well as some kind of searchability.

Currently I am thinking about just making a seperate node group for every color/object.

Answer 1

Creating a separate node group for each color is probably one of the most flexible solutions. They are searchable, available for different materials and if you want to change the color and it's inside a group, it will update everywhere it is used.

If we are talking about setting up hundreds of groups, Python will be the way to go. That's how it might look:

import bpy, requests, csv, io

# Your color values are in sRGB color space (I guess and hope so)
# so we need to convert them to Blender's linear colors with 
# some random code I found online. Hopefully it's OK. 
def srgb2lin(s):  
    if s <= 0.0404482362771082:
        lin = s / 12.92
    else:
        lin = pow(((s + 0.055) / 1.055), 2.4)
    return lin

# You can load CSV file from Google sheets
sheet_id = "1se8YEtb2detS7OuPE86fXGyD269pMycAWe2mtKUj2W8" 
#that's the ID I found in the URL you shared
url = "https://docs.google.com/spreadsheets/d/" + sheet_id+ "/gviz/tq?tqx=out:csv"
r = requests.get(url)
reader = csv.reader(io.StringIO(r.text))
next(reader)    
for row in reader:
    group = bpy.data.node_groups.new(row[8], 'ShaderNodeTree')
    group.outputs.new("NodeSocketColor", "Color")
    group.use_fake_user = True 
    color_node = group.nodes.new('ShaderNodeRGB')
    c = []
    for value in row[5][1:-1].split(","):
        c.append(srgb2lin(1/255*float(value)))
    c.append(1) # thats alpha, because the color in the node is in RGBA format
    color_node.outputs[0].default_value = c
    output_node = group.nodes.new("NodeGroupOutput")
    output_node.location = (200, 0)
    group.links.new(color_node.outputs[0], output_node.inputs[0])

You could also generate a material for each node group that would use the node group for it's color at the same time.