def test_rotation(coords, num_atoms):
batch_size = num_atoms.size(0)
R = getRandomRotation(batch_size)
rotate = CoordsRotate()
rotated = rotate(coords, R, num_atoms)
print(rotated)
def runTest(self):
num_aa = torch.zeros(self.batch_size,
dtype=torch.int,
device=self.device).random_(
int(self.length / 2), self.length)
sequences = [
''.join(['G' for i in range(num_aa[j].item())])
for j in range(self.batch_size)
]
fa_angles = torch.randn(self.batch_size,
8,
self.length,
dtype=torch.double,
device='cpu')
coords_fa, res_names, atom_names, num_atoms = self.a2c(
fa_angles, sequences)
R = getRandomRotation(self.batch_size)
T = torch.randn(self.batch_size, 3, dtype=torch.double, device='cpu')
rot_coords_fa = self.rotate(coords_fa, R, num_atoms)
tra_rot_coords_fa = self.translate(rot_coords_fa, T, num_atoms)
tra_rot_coords_fa = tra_rot_coords_fa.to(device=self.device,
dtype=self.dtype)
coords_fa = coords_fa.to(device=self.device, dtype=self.dtype)
num_atoms = num_atoms.to(device=self.device, dtype=torch.int)
min_rmsd = self.rmsd(tra_rot_coords_fa, coords_fa, num_atoms)
for i in range(self.batch_size):
self.assertAlmostEqual(min_rmsd[i].item(), 0.0, places=self.places)
def runTest(self): R = getRandomRotation(self.batch_size) #Matrixes should be det = 1 for i in range(R.size(0)): U = R[i].numpy() D = np.linalg.det(U) self.assertAlmostEqual(D, 1.0, places=self.places)
def runTest(self): coords = torch.zeros_like(self.coords).copy_(self.coords).requires_grad_() #Random rotation matrixes R = getRandomRotation(self.batch_size).to(dtype=self.dtype, device=self.device) result = torch.autograd.gradcheck(self.rotate, (coords, R, self.num_atoms)) self.assertTrue(result)
def test_random_rotation(num_rotations): batch_size = num_rotations R = getRandomRotation(batch_size) num_atoms = torch.ones(R.size(0), dtype=torch.int) coords = torch.zeros(R.size(0),3, dtype=torch.double) coords[:,0]=1.0 coords[:,1]=0.0 coords[:,2]=0.0 rotate = CoordsRotate() rotated = rotate(coords, R, num_atoms) return rotated
box_size = 80 resolution = 1.25 sequence = ['GGAGRRRGGWG'] angles = torch.zeros(len(sequence), 7, len(sequence[0]), dtype=torch.double) angles[0,0,:] = -1.047 angles[0,1,:] = -0.698 angles[0,2:,:] = 110.4*np.pi/180.0 a2c = Angles2Coords() c2cc = Coords2CenteredCoords(rotate=False, translate=False, box_size=box_size, resolution=resolution) c2tc = Coords2TypedCoords() tc2v = TypedCoords2Volume(box_size=box_size, resolution=resolution) rotate = CoordsRotate() translate = CoordsTranslate() R = getRandomRotation( len(sequence) ) volume_rotate = VolumeRotation(mode='bilinear') #Generating protein and initial volume protein, res_names, atom_names, num_atoms = a2c(angles, sequence) protein = c2cc(protein, num_atoms) coords, num_atoms_of_type, offsets = c2tc(protein, res_names, atom_names, num_atoms) volume = tc2v(coords.cuda(), num_atoms_of_type.cuda(), offsets.cuda()) #Rotating protein center = torch.zeros(len(sequence), 3, dtype=torch.double, device='cpu').fill_((box_size - 0.5)*resolution/2.0) centered_coords = translate(coords, -center, num_atoms) rotated_centered_coords = rotate(centered_coords, R, num_atoms) rotated_coords = translate(rotated_centered_coords, center, num_atoms) #Reprojecting protein