spritegen/generate.py at master · miltonllera/spritegen · GitHub

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
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import os
import numpy as np
# import matplotlib.pyplot as plt
from functools import partial
from importlib import import_module
from absl import flags, app

from spriteworld.renderers.pil_renderer import PILRenderer
from spriteworld.renderers.color_maps import hsv_to_rgb
from spriteworld.sprite_generators import create_iterator, chain_generators, \
                                          create_sampler, product_iterator


FLAGS = flags.FLAGS
flags.DEFINE_string('config', default='spriteworld.configs.simple',
                    help='Configuration to generate')
flags.DEFINE_integer('image_size', default=64,
                     help='Size of rendered image')
flags.DEFINE_integer('aa', default=5,
                     help='Anti aliasing of rendered image')
flags.DEFINE_boolean('hsv_color', default=True,
                     help='Whether the task config uses HSV as color factors.')
flags.DEFINE_string('save_folder', default='datasets',
                    help='Folder where to save the data')

def main(argv):
    del argv

    config = import_module(FLAGS.config).load()

    name = config['dataset']
    image_size = FLAGS.image_size, FLAGS.image_size
    distribs = config['distribs']
    params = config['params']
    value_maps = config['value_maps']
    gentype = config['generation_type']

    renderer = PILRenderer(image_size=image_size,
                           anti_aliasing=FLAGS.aa,
                           bg_color=(0,0,0),  # black background
                           color_to_rgb=hsv_to_rgb if FLAGS.hsv_color else None)

    if isinstance(distribs, list):
        n_sprites = len(distribs)
    else:
        n_sprites = 1

    if gentype == 'sample':
        create_generator = partial(create_sampler,
                                   num_samples=config['n_samples'])
    else:
        create_generator = create_iterator

    if n_sprites > 1:
        prefix = 'obj{}-'

        obj_generators = [create_generator(d) for d in distribs]

        if gentype == 'sample':
            generator = chain_generators(*obj_generators)
        else:
            generator = product_iterator(*obj_generators)

    else:
        prefix = ''
        generator = create_generator(distribs)

    images, factor_values = [], []

    for sprites in generator():
        img = renderer.render(sprites)

        # plt.imshow(img)
        # plt.show()

        fv = []
        for s in sprites:
            for (k,v) in s.factors(dtype=float).items():
                if k in params:
                    fv.append(v)

        images.append(img)
        factor_values.append(fv)

    images = np.array(images)
    factor_values  = np.array(factor_values)

    factors = []
    unique_values = {}
    factor_sizes = []

    for j in range(n_sprites):
        for fn in params:
            i = len(factors)

            factor_values[:, i] = value_maps[fn](factor_values[:, i])
            uv = np.unique(factor_values[:, i])
            factor_sizes.append(len(uv))

            fn = prefix.format(j) + fn

            unique_values[fn] = uv
            factors.append(fn)

    factor_values, idx = np.unique(factor_values, return_index=True, axis=0)
    images = images[idx]

    print(unique_values)
    print(len(factor_values))

    save_folder = FLAGS.save_folder
    os.makedirs(save_folder, exist_ok=True)

    np.savez_compressed(os.path.join(save_folder,'{}.npz'.format(name)),
                        # metadata
                        name=name,
                        img_size=(3, *image_size),
                        n_factors=len(factors),
                        factors=factors,
                        factor_sizes=tuple(factor_sizes),
                        unique_values=unique_values,
                        # data
                        images=images,
                        factor_values=factor_values)


if __name__ == '__main__':
    app.run(main)