[BUG FIX] Raise on ambiguous morph access for heterogeneous entities

[kunni918]

Description

This PR makes .morph fail closed for heterogeneous entities created from multiple morph variants. Heterogeneous entities already expose .morphs for all variants and .main_morph for the first variant, so .morph should not silently return only the first morph in an order-dependent way.

The PR also updates internal paths that need non-throwing access:

• Rasterizer Plane checks now use .main_morph when preserving the existing first-variant behavior.
• ImGui overlay scene-rebuild capture stores tuple(entity.morphs) for heterogeneous entities so variants are not dropped.
• Brief repr no longer probes .morph through hasattr, so repr(scene.entities) remains usable.

Related Issue

Resolves #2750

Motivation and Context

Accessing .morph on a heterogeneous entity previously returned only the first morph. That made calls such as .morph.size ambiguous and order-dependent when variants had different morph types or dimensions. A maintainer clarified in the issue that heterogeneous entities should be accessed through .morphs, and that .morph should raise in this case.

This keeps the fail-closed public API while moving known internal renderer, overlay, and repr paths to explicit .main_morph / .morphs access.

Per-variant floor/reflection semantics are intentionally out of scope for this PR; this PR preserves the existing first-morph renderer behavior while avoiding accidental .morph property calls.

How Has This Been / Can This Be Tested?

Tested locally with:

.venv/bin/python -m pytest -q \
  tests/test_rigid_physics.py::test_heterogeneous_morph_property_raises \
  tests/test_imgui_overlay.py::test_imgui_overlay_capture_pending_entities_preserves_heterogeneous_morphs \
  tests/test_render.py::test_rasterizer_context_heterogeneous_main_morph_no_throw

.venv/bin/python -m pytest tests/test_rigid_physics.py -q \
  -k 'heterogeneous and (invalid_material or morph_property or fewer_envs or mass_setters or aabb)'

.venv/bin/pre-commit run --files \
  genesis/engine/entities/rigid_entity/rigid_entity.py \
  genesis/repr_base.py \
  genesis/ext/pyrender/overlay/plugin.py \
  genesis/vis/rasterizer_context.py \
  tests/test_rigid_physics.py \
  tests/test_imgui_overlay.py \
  tests/test_render.py

The heterogeneous regression tests were run outside the sandbox because Quadrants writes compilation cache under ~/.cache/quadrants. I did not run the full test suite locally.

Screenshots (if appropriate):

N/A

Checklist:

• I read the CONTRIBUTING document.

• I followed the Submitting Code Changes section of CONTRIBUTING document.

• I tagged the title correctly (including BUG FIX/FEATURE/MISC/BREAKING)

• I updated the documentation accordingly or no change is needed.

• I tested my changes and added instructions on how to test it for reviewers.

• I have added tests to cover my changes.

• All new and existing tests passed.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 598ece81f3

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".