Summary
Add a way to generate capsule colliders for rigged characters driven by vertex/weight groups (e.g. one capsule per bone's weight group), rather than only the existing whole-object/selection-based capsule generation (collider_shapes/add_bounding_capsule.py).
Motivation
Characters are typically already weight-painted per bone for skinning. Reusing those weight groups to auto-generate a capsule per group would give a quick per-limb/per-bone collision rig without manually selecting geometry per bone.
Rough idea
- For each vertex group (or a user-selected subset) on the mesh, isolate the weighted vertices and fit a capsule to that subset, similar to how the current capsule fit works on a full object/selection.
- Likely wants to reuse the armature/bone data (group name <-> bone name convention) to orient the capsule along the bone rather than purely from vertex distribution, though a vertex-distribution-only fit may be a reasonable first pass.
Open questions
- Naming/parenting convention for the generated capsules relative to bones (parent to bone vs. object, existing collider naming scheme in
collider_conversion/regenerate_name.py).
- How to handle vertex groups with very low/zero total weight or that don't correspond to a bone.
- UI: batch-generate for all groups vs. per-group operator.
Summary
Add a way to generate capsule colliders for rigged characters driven by vertex/weight groups (e.g. one capsule per bone's weight group), rather than only the existing whole-object/selection-based capsule generation (
collider_shapes/add_bounding_capsule.py).Motivation
Characters are typically already weight-painted per bone for skinning. Reusing those weight groups to auto-generate a capsule per group would give a quick per-limb/per-bone collision rig without manually selecting geometry per bone.
Rough idea
Open questions
collider_conversion/regenerate_name.py).