Graph Editor - UI support for enum

[inter-sakot4]

Description
In this PR, UI support for enumerator types are added to GraphEditor.

Some assumptions are made to execute this task:

• Fix the issue for only GraphEditor , though similar issue exist also in MaterialXView ( NanoGUI )
• Generalize the solution to int, bool, string, float, double, Vector2, Vector3, Vector4, Color3, Color4, Matrix types

With the above assumptions, scope of this task is limited only to GraphEditor ( IMGUI )

Solution

• renderEnumInput<input_type>(input, uiProperties, temp) is a template function that does the rendering of enumerator input types.

Following edge cases are handled:

• ui.enumerationValues and ui.ui.enumeration are not always populated ( See /MaterialXRender/Util.cpp ). In such cases, the enum and enumvalues are parsed directly from the attribute values to render the enumerators.
• Its common to have string input with enum fields only , but no enumvalues

How to test?

Slightly modified standard_surface_jade.mtlx can be used like below.

<?xml version="1.0"?>
<materialx version="1.39" colorspace="lin_rec709">
  <standard_surface name="SR_jade" type="surfaceshader">
    <input name="base" type="float" value="0.5" />
    <input name="base_color" type="color3" value="0.0603,0.4398,0.1916" />
    <input name="specular_roughness" type="float" value="0.25" enum="0.0, 0.25, 0.5, 0.75" enumvalues="0.0,0.25,0.5,0.75"/>
    <input name="specular_IOR" type="float" value="2.418" />
    <input name="specular_anisotropy" type="float" value="0.5" />
    <input name="subsurface" type="float" value="0.4" enum="0.0,0.4,0.5,0.6,0.7" enumvalues="0.0,0.4,0.5,0.6,0.7" />
    <input name="subsurface_color" type="color3" value="0.0603, 0.4398, 0.1916" enum="DEFAULT_GREEN, RED" enumvalues="0.0603,0.4398,0.1916,0.8,0.04398,0.01916" />
  </standard_surface>
  <surfacematerial name="Jade" type="material">
    <input name="surfaceshader" type="surfaceshader" nodename="SR_jade" />
  </surfacematerial>
</materialx>

Expected Result

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: inter-sakot4 / name: Sathish Kottravel (d87fcb2, d74e1a1, 34acd15, 303be86, 3eba347, 7140665)
• ✅ login: jstone-lucasfilm / name: Jonathan Stone (82453d1, 9c12b05)

[jstone-lucasfilm]

Thanks @inter-sakot4, and I'll link this to GitHub Issue #2366.

[jstone-lucasfilm]

This looks very promising, @inter-sakot4, and I've added a note on the line that appears to be triggering build issues on some platforms.

[jstone-lucasfilm]

Here's the line that likely needs an update for cross-platform compatibility, since the return type of vector::size is size_t.

[inter-sakot4]

Thankyou for taking time to review @jstone-lucasfilm ! Fixed the type issue as suggested and added short comments for clarity.

[jstone-lucasfilm]

Although this is a minor style issue, I notice that this method uses 2-space indentation rather than our standard 4-space indentation, and that would be valuable to fix before we include this change in the codebase.

[jstone-lucasfilm]

I notice some minor inconsistencies in these comments, and it would be valuable to use the same spacing and capitalization for all of them. Usually, we use a single space after the initial // marker, and the first letter of the first sentence is capitalized.

[jstone-lucasfilm]

This looks like a useful note, and I'll CC @lfl-eholthouser for her thoughts, though we should make sure not to include this when the change is ultimately merged to main.

[hybridherbst]

Worth noting that currently MaterialXWebViewer does show enum/enumvalues for non-uniform inputs.
It probably should not, given the spec, but I'm not sure which one is right.

[hybridherbst]

I closed my related PR #2438 because yours looks much more complete, but my test file does not actually show a dropdown UI when I build from this PR @inter-sakot4.

Here's the test file that doesn't bring up a dropdown:
MyKeywordGraph.mtlx.zip

Should look like this:

[jstone-lucasfilm]

@inter-sakot4 Just following up on this PR, to see if you might have a chance to address the issue brought up by @hybridherbst above!