Geom color scheme

[agheata]

This Pull request:

Introduces an extensible color scheme mechanism for geometry visualization, improving the default coloring of imported geometries (notably GDML, which does not store color information).

Changes or fixes:

TGeoManager::DefaultColors() historically assigns colors based only on material effective Z, resulting in:

• many volumes sharing the same gray color,
• limited visual distinction,
• little resemblance to “natural” material appearance in detector geometries.

Modern detector descriptions contain rich material names (e.g. Copper, FR4, Kapton, Steel, Carbon fiber, Gases), which can be exploited to produce clearer and more meaningful visualizations.

What’s new

• A new runtime-only strategy class TGeoColorScheme that encapsulates color and transparency policies.
• A default “natural” color scheme based on:
  • name-based material classification (metals, polymers, composites, gases, ceramics, etc.),
  • fallback to a Z-binned lookup when no name-based rule applies.

• Two additional built-in schemes:
  • Flashy – high-contrast, presentation-friendly colors.

• High-contrast – darker, saturated colors suited for light backgrounds.

• User extensibility:

  • users can inject custom logic via hooks (std::function) to override colors or transparency,
  • or subclass TGeoColorScheme for full control.

• Backward compatibility:

  • calling TGeoManager::DefaultColors() with no arguments still works and now defaults to the natural scheme.

API overview

void TGeoManager::DefaultColors(const TGeoColorScheme *scheme = nullptr);

if scheme == nullptr a default TGeoColorScheme(EGeoColorSet::kNatural) is used internally.
The color set is selected via a global enum:

enum class EGeoColorSet {
   kNatural,
   kFlashy,
   kHighContrast
};

Examples

• Default behavior (no code changes required):

gGeoManager->DefaultColors();   // uses natural, name-based scheme

• Select a predefined scheme

TGeoColorScheme cs(EGeoColorSet::kFlashy);
gGeoManager->DefaultColors(&cs);

• User override via hook (no subclassing)

TGeoColorScheme cs(EGeoColorSet::kNatural);
cs.SetColorHook([](const TGeoVolume *v) -> Int_t {
   const TGeoMaterial *m = TGeoColorScheme::GetMaterial(v);
   if (!m || !m->GetName()) return -1;
   if (std::string(m->GetName()).find("Copper") != std::string::npos)
      return TColor::GetColor(0.9f, 0.55f, 0.3f);
   return -1; // fallback to defaults
});
gGeoManager->DefaultColors(&cs);

• Override Z-based fallback

cs.SetZFallbackHook([](Int_t Z, EGeoColorSet) -> Int_t {
   float g = std::min(1.f, Z / 100.f);
   return TColor::GetColor(g, g, g);
});

Checklist:

• tested changes locally
• updated the docs (if necessary)

This PR fixes #

[github-actions]

Test Results

    22 files      22 suites   3d 11h 11m 56s ⏱️
 3 774 tests  3 774 ✅ 0 💤 0 ❌
75 059 runs  75 059 ✅ 0 💤 0 ❌

Results for commit 6a462da.

♻️ This comment has been updated with latest results.

[dpiparo]

Thanks for this solid work.

Besides the minor comments on the code, I see an opportunity for improvement, i.e. the possibility to add an example in the geometry tutorials and tests. Presently, the test coverage of the geometry package can be improved, and this seems like a perfect opportunity to kick off this process!

These changes are impactful and would also deserve some words (and a beautiful image, like "before and after") in the Release Notes.

[agheata]

Adding a tutorial for this, and release notes

[agheata]

Thanks for this solid work.

Besides the minor comments on the code, I see an opportunity for improvement, i.e. the possibility to add an example in the geometry tutorials and tests. Presently, the test coverage of the geometry package can be improved, and this seems like a perfect opportunity to kick off this process!

These changes are impactful and would also deserve some words (and a beautiful image, like "before and after") in the Release Notes.

I added a nice tutorial and the entry in the release notes, w/o a picture (don't know where to put it) but a link to the PR. A coverage test for this is difficult to implement because it is pure visualization, but a working tutorial is more meaningful.

[dpiparo]

Thanks! Just merge it at your convenience.