def test_negative_coords(self):
print('Negative coords test')
key = CoordsKey(D=1)
key.setTensorStride(1)
cm = CoordsManager(D=1)
coords = torch.IntTensor([[0, -3], [0, -2], [0, -1], [0, 0], [0, 1],
[0, 2], [0, 3]])
# Initialize map
mapping, inverse_mapping = cm.initialize(coords, key)
print(mapping, len(mapping))
cm.print_diagnostics(key)
# Create a strided map
stride_key = cm.stride(key, [2])
strided_coords = cm.get_coords(stride_key).numpy().tolist()
self.assertTrue(len(strided_coords) == 4)
self.assertTrue([0, -4] in strided_coords)
self.assertTrue([0, -2] in strided_coords)
self.assertTrue([0, 2] in strided_coords)
print('Stride: ', cm.get_coords(stride_key))
cm.print_diagnostics(stride_key)
print(cm)
def test_coords_map(self):
coords, _, _ = data_loader(1)
key = CoordsKey(D=2)
key.setTensorStride(1)
# Initialize map
cm = CoordsManager(D=2)
mapping, inverse_mapping = cm.initialize(coords,
key,
force_remap=True,
allow_duplicate_coords=False)
print(mapping, len(mapping))
cm.print_diagnostics(key)
print(cm)
print(cm.get_batch_size())
print(cm.get_batch_indices())
# Create a strided map
stride_key = cm.stride(key, [2, 2])
print('Stride: ', cm.get_coords(stride_key))
cm.print_diagnostics(key)
print(cm)
ins, outs = cm.get_coords_map(1, 2)
inc = cm.get_coords(1)
outc = cm.get_coords(2)
for i, o in zip(ins, outs):
print(f"{i}: ({inc[i]}) -> {o}: ({outc[o]})")
def test_coords_manager(self):
key = CoordsKey(D=1)
key.setTensorStride(1)
cm = CoordsManager(D=1)
coords = (torch.rand(5, 2) * 100).int()
cm.initialize(coords, key)
print(cm.get_row_indices_per_batch(key))
print(key)
print(cm)
def test_hash(self):
N, M = 1000, 1000
I, J = np.meshgrid(np.arange(N), np.arange(M))
I = I.reshape(-1, 1) - 100
J = J.reshape(-1, 1) - 100
K = np.zeros_like(I)
C = np.hstack((I, J, K))
coords_manager = CoordsManager(D=2)
coords_key = CoordsKey(2)
coords_key.setTensorStride(1)
coords_manager.initialize(torch.from_numpy(C).int(), coords_key)
print(coords_manager)
def test_coords_manager(self):
key = CoordsKey(D=1)
key.setTensorStride(1)
cm = CoordsManager(D=1)
coords = torch.IntTensor([[0, 1], [0, 1], [0, 2], [0, 2], [1, 0],
[1, 0], [1, 1]])
unique_coords = torch.unique(coords, dim=0)
# Initialize map
mapping, inverse_mapping = cm.initialize(coords,
key,
force_remap=True,
allow_duplicate_coords=False)
self.assertTrue(len(unique_coords) == len(mapping))
print(mapping, len(mapping))
cm.print_diagnostics(key)
print(cm)
self.assertTrue(cm.get_batch_size() == 2)
self.assertTrue(cm.get_batch_indices() == {0, 1})
# Create a strided map
stride_key = cm.stride(key, [4])
strided_coords = cm.get_coords(stride_key)
self.assertTrue(len(strided_coords) == 2)
cm.print_diagnostics(key)
print(cm)
# Create a transposed stride map
transposed_key = cm.transposed_stride(stride_key, [2], [3], [1])
print('Transposed Stride: ', cm.get_coords(transposed_key))
print(cm)
# Create a transposed stride map
transposed_key = cm.transposed_stride(stride_key, [2], [3], [1],
force_creation=True)
print('Forced Transposed Stride: ', cm.get_coords(transposed_key))
print(cm)
# Create a reduction map
key = cm.reduce()
print('Reduction: ', cm.get_coords(key))
print(cm)
print('Reduction mapping: ', cm.get_row_indices_per_batch(stride_key))
print(cm)
def test_coords_manager(self):
key = CoordsKey(D=1)
key.setTensorStride(1)
cm = CoordsManager(D=1)
coords = (torch.rand(20, 2) * 10).int()
print(coords)
unique_coords = torch.unique(coords, dim=0)
print('Num unique: ', unique_coords.shape)
# Initialize map
mapping = cm.initialize(coords,
key,
force_remap=True,
allow_duplicate_coords=False)
print(mapping, len(mapping))
cm.print_diagnostics(key)
print(cm)
print(cm.get_batch_size())
print(cm.get_batch_indices())
# Create a strided map
stride_key = cm.stride(key, [4])
print('Stride: ', cm.get_coords(stride_key))
cm.print_diagnostics(key)
print(cm)
# Create a transposed stride map
transposed_key = cm.transposed_stride(stride_key, [2], [3], [1])
print('Transposed Stride: ', cm.get_coords(transposed_key))
print(cm)
# Create a transposed stride map
transposed_key = cm.transposed_stride(stride_key, [2], [3], [1],
force_creation=True)
print('Forced Transposed Stride: ', cm.get_coords(transposed_key))
print(cm)
# Create a reduction map
key = cm.reduce()
print('Reduction: ', cm.get_coords(key))
print(cm)
print('Reduction mapping: ', cm.get_row_indices_per_batch(stride_key))
print(cm)
def test_hash(self):
try:
import numpy as np
except:
return
N, M = 1000, 1000
I, J = np.meshgrid(np.arange(N), np.arange(M))
I = I.reshape(-1, 1) - 100
J = J.reshape(-1, 1) - 100
K = np.zeros_like(I)
C = np.hstack((I, J, K))
coords_manager = CoordsManager(D=2)
coords_key = CoordsKey(2)
coords_manager.initialize(torch.from_numpy(C).int(), coords_key)
def test_coords_key(self):
key = CoordsKey(D=1)
key.setKey(1)
self.assertTrue(key.getKey() == 1)
key.setTensorStride([1])
print(key)