CedarandClaude Opus 4.6
0e35c7a575
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
43 lines
1.7 KiB
Python
43 lines
1.7 KiB
Python
# tests/test_rotation_utils.py
|
|
import torch
|
|
import math
|
|
|
|
def test_identity_rotation():
|
|
from utils.rotation_utils import euler_to_rotation_matrix
|
|
R = euler_to_rotation_matrix(0.0, 0.0, 0.0)
|
|
assert R.shape == (3, 3)
|
|
assert torch.allclose(R, torch.eye(3), atol=1e-6)
|
|
|
|
def test_90_degree_yaw():
|
|
from utils.rotation_utils import euler_to_rotation_matrix
|
|
R = euler_to_rotation_matrix(0.0, 90.0, 0.0)
|
|
# Yaw 90° around Y: x→z, z→-x
|
|
point = torch.tensor([1.0, 0.0, 0.0])
|
|
rotated = R @ point
|
|
expected = torch.tensor([0.0, 0.0, -1.0])
|
|
assert torch.allclose(rotated, expected, atol=1e-5)
|
|
|
|
def test_rotation_matrix_is_orthogonal():
|
|
from utils.rotation_utils import euler_to_rotation_matrix
|
|
R = euler_to_rotation_matrix(25.0, -30.0, 15.0)
|
|
# R^T @ R should be identity
|
|
assert torch.allclose(R.T @ R, torch.eye(3), atol=1e-5)
|
|
# det(R) should be 1
|
|
assert torch.allclose(torch.det(R), torch.tensor(1.0), atol=1e-5)
|
|
|
|
def test_compute_vertex_normals_flat_triangle():
|
|
from utils.rotation_utils import compute_vertex_normals
|
|
# Single triangle lying in XY plane
|
|
vertices = torch.tensor([[[0.0, 0.0, 0.0],
|
|
[1.0, 0.0, 0.0],
|
|
[0.0, 1.0, 0.0]]]) # (1, 3, 3)
|
|
triangles = torch.tensor([[0, 1, 2]], dtype=torch.int32) # (1, 3)
|
|
normals = compute_vertex_normals(vertices, triangles)
|
|
assert normals.shape == (1, 3, 3)
|
|
# All normals should point in +Z (or -Z depending on winding)
|
|
for i in range(3):
|
|
n = normals[0, i]
|
|
assert abs(n[2].item()) > 0.99, f"Normal {i} not along Z: {n.tolist()}"
|
|
assert abs(n[0].item()) < 0.01
|
|
assert abs(n[1].item()) < 0.01
|