Generalized use case architecture

[jywarren]

The initial demo at https://divy123.github.io/colorimetry/public/ is great. It's specific to the use case at publiclab/image-sequencer#979.

How can we generalize the code? Can we adapt it to run on pH strips?

https://duckduckgo.com/?q=ph+test+strip&atb=v121-6&ia=images&iax=images

Here is how pH strip comparisons work manually:

Maybe pH is too complex as a next step, since it has 4 patches. What about another test that just accepts a min color and a max color, and provides a linear scale between them. So the parameters could be:

color1 (hex)
color2 (hex)
minValue
maxValue
scale (string, "linear" by default for now, eventually maybe exponential, logarithmic, etc)

These could be passed into a constructor like:

// This outputs an object:
var colorimetry = new Colorimetry(color1, color2, minValue, maxValue, scale)

colorimetry.setupUI(); // generates the demo
var output = colorimetry.getValue(color3); // give it a hex value

Something like this for a very basic color comparison?

The initial demo as built would be more like:

colorimetry.compareColors(averageColor1, averageColor2);

Maybe it would need to be compartmentalized somehow, like we load an image into a generic object like:

var image = new ColorimetryImage('url'); // this loads the image into an internal sequencer
var region1 = image.getAverage(x1, y1, h1, w1);
var region2 = image.getAverage(x2, y2, h2, w2);
colorimetry.compareColors(averageColor1, averageColor2);

Or would the all be part of the same object?

I'm just brainstorming a good general-purpose architecture here. What do you think, @Divy123 ? Can this accommodate the original use case, but make it more generalizable to different kinds of colorimetry?

[jywarren]

I could imagine a combination of 3 linear scales as in the first example being used to find a pH value, don't you think?