ssk 1.1.1

SolidSKeleton tessellation and format tools

SolidSKeleton Python

Python 3.10+.

pip install ssk

Commands

ssk validate model.ssk
ssk convert model.ssk model.sskb
ssk convert model.ssk model.glb
ssk convert model.ssk model.glb --resolution 64
ssk convert model.sskb model.glb
ssk convert model.glb model.ssk
ssk convert model.glb model.sskb --expected-piece-count 42
ssk inspect model.sskb

From this directory without installing:

python -m ssk validate model.ssk

Library

from ssklib import convert, inspect_file, load, validate_file

convert("model.ssk", "model.glb", resolution=64)
result = convert("model.glb", "model.ssk", expected_piece_count=42)
print(result.coverage_percent, result.overfill_percent)

Lower-level GLTF import:

from ssklib import import_gltf_to_ssk

result = import_gltf_to_ssk(
    "model.glb",
    expected_piece_count=42,  # soft guide
    infill_weight=1.18,
    outfill_weight=1.05,
    complexity_weight=1.0,
)
print(result.coverage_percent, result.overfill_percent)

quality = result.score_document(some_doc)

Lower-level functions:

from ssklib import parse_ssk, parse_sskb, resolve, validate, write_sskb

Notes

• Mesh output defaults to resolution 32.
• GLTF/GLB import reconstructs (estimated) SSK-native primitives where practical and reports sampled volume coverage and overfill percentages. expected_piece_count / --expected-piece-count is a soft guide, not an exact target. Import weight options (--infill-weight, --outfill-weight, --complexity-weight) tune the scoring between candidates on a normalised 0–1 scale. Expect slightly different results from the TypeScript package due to numpy vs. JavaScript math differences.
• CSG uses trimesh with Manifold.
• glTF output uses unindexed meshes with flat per-face normals.