forked from facebookresearch/neuralvolumes
-
Notifications
You must be signed in to change notification settings - Fork 0
/
plot_test.py
164 lines (137 loc) · 6.44 KB
/
plot_test.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
import matplotlib.pyplot as plt
from models.RayMarchingHelper import RayMarchingHelper
from utils.RenderUtils import get_distributed_coords
# worth to have a look: https://github.com/NVIDIAGameWorks/kaolin
import os
from typing import Callable
import pyvista as pv
import torch
from pyvista import Plotter
from models.volsamplers.warpvoxel import VolSampler
import numpy as np
class TestPlot:
def __init__(self):
self.plotter = pv.Plotter()
def __prepare_template_np_plot(self, np_filename: str, density: float):
template = None
with open(np_filename, 'rb') as f:
template = np.load(f)
if template is None:
raise Exception("should load file")
min = -1.0
max = 1.0
distribution = np.arange(min, max, (2.0 / density))
x, y, z = np.meshgrid(distribution, distribution, distribution)
pos = np.stack((x, y, z), axis=3)
dimension = int(density ** 3)
batchsize = template.shape[0]
pos = np.array([pos for i in range(batchsize)])
pos = pos.reshape((batchsize, 1, dimension, 1, 3))
torch.cuda.set_device("cuda:0")
cur_device = torch.cuda.current_device()
pos = torch.from_numpy(pos)
pos = pos.to(cur_device)
template = torch.from_numpy(template)
template = template.to(cur_device)
volsampler = VolSampler()
sample_rgb, sample_alpha = volsampler(pos=pos, template=template)
sample_rgb = sample_rgb.cpu().numpy().reshape((batchsize, dimension, 3))
sample_alpha = sample_alpha.cpu().numpy().reshape((batchsize, dimension, 1))
pos = pos.cpu().numpy().reshape((batchsize, dimension, 3))
print(np.max(sample_rgb))
print(np.min(sample_rgb))
print(np.max(sample_alpha))
print(np.min(sample_alpha))
sample_rgba = np.zeros((batchsize, dimension, 4))
sample_rgba[:, :, 0:3] = sample_rgb
sample_rgba[:, :, 3] = sample_alpha[:, :, 0]
sample_rgba = sample_rgba.reshape((batchsize, int(density), int(density), int(density), 4))
return pos, sample_rgba.astype(float), dimension
def pyvista_3d_from_template_np(
self,
outputfolder: str,
overwrite_color_to_black: bool = False,
add_ground_truth: bool = False,
nof_frames: int = 500):
plotter_test: Plotter = pv.Plotter()
prepare_template: Callable = self.__prepare_template_np_plot
density: float = 16.0
add_ground_truth_to_plotter: Callable = self.plot_stl_pyvista
def slider_callback_create_points(value):
plotter_test.clear_actors()
res: int = int(value)
np_filename: str = os.path.join(outputfolder, "frame{}.npy".format(res))
pos, sample_rgba, dimension = prepare_template(np_filename, density)
min, max = -1.0, 1.0
x = np.arange(min, max, (2.0 / density))
y = np.arange(min, max, (2.0 / density))
z = np.arange(min, max, (2.0 / density))
x, y, z = np.meshgrid(x, y, z)
if overwrite_color_to_black:
sample_rgba[:, :, :, :, 0:3] = np.zeros(sample_rgba[:, :, :, :, 0:3].shape)
# else:
# sample_rgba[:, :, :, :, 0:3] = sample_rgba[:, :, :, :, 0:3] / 255.
# sample_rgba[:, :, :, :, 3] = np.ones(sample_rgba[:, :, :, :, 3].shape) * 0.5
grid = pv.StructuredGrid(x, y, z)
# batchsize = sample_rgba.shape[0]
# for i in range(batchsize):
sample_rgba = sample_rgba[0, :, :, :, 0:4].reshape((dimension, 4))
actor = plotter_test.add_points(grid.points, scalars=sample_rgba, rgb=True)
if add_ground_truth:
grid, mask = add_ground_truth_to_plotter(
"experiments/blenderLegMovement/data/groundtruth_test/frame{:04d}.stl".format(res))
plotter_test.add_points(grid.points, cmap=["#00000000", "#ff00004D"], scalars=mask)
return
numpy_files = [f for f in os.listdir(outputfolder) if
os.path.isfile(os.path.join(outputfolder, f)) and f.endswith(".npy")]
plotter_test.add_slider_widget(callback=slider_callback_create_points, value=0, rng=[0, len(numpy_files)],
title='Time')
plotter_test.show_axes_all()
plotter_test.show()
def matplotlib_2d_from_template_np(self, np_filename: str):
density = 50.0
pos, sample_rgba, dimension = self.__prepare_template_np_plot(np_filename, density)
plt.imshow(sample_rgba[:, :, int(density / 2)].reshape((int(density), int(density), 4)), vmin=0, vmax=255)
plt.show()
def matplotlib_3d_from_template_np(self, np_filename: str):
density = 20.0
pos, sample_rgba, dimension = self.__prepare_template_np_plot(np_filename, density)
shape_plot = (int(density) + 1, int(density) + 1, int(density) + 1)
x, y, z = (np.indices(shape_plot) / density) * 2.0 - 1.0
all_test = np.full((int(density), int(density), int(density)), True)
ax = plt.figure().add_subplot(projection='3d')
ax.voxels(x, y, z, all_test,
facecolors=sample_rgba / 255.0,
edgecolors=[0.0, 0.0, 0.0, 0.0],
linewidth=0.0)
plt.show()
def plot_nv_from_decout(self, decout: dict):
nof_points = 10
batchsize = 4
start_coords = get_distributed_coords(batchsize=batchsize, fixed_value=-1.0, nof_points=nof_points,
fixed_axis=2)
direction_coords = np.full((batchsize, nof_points, nof_points, 3), (0.0, 0.0, 1.0))
dt = 2.0 / float(nof_points)
t = np.ones((batchsize, nof_points, nof_points)) * -1
end = np.ones((batchsize, nof_points, nof_points))
raymarching = RayMarchingHelper(
torch.from_numpy(start_coords).to("cuda"),
torch.from_numpy(direction_coords).to("cuda"),
dt,
torch.from_numpy(t).to("cuda"),
torch.from_numpy(end).to("cuda"),
RayMarchingHelper.OUTPUT_VOLUME
)
rgb, alpha = raymarching.do_raymarching(
VolSampler(),
decout,
False,
0.0
)
print(rgb.size())
print(alpha.size())
print(rgb[0, 0, 0, 0, 0])
print(alpha[0, 0, 0, 0, 0])
def show_plot(self):
self.plotter.show_axes_all()
self.plotter.show()