generate_dataset.py 10 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265
  1. #!/usr/bin/env python3
  2. # -*- coding: utf-8 -*-
  3. """
  4. Created on Wed Jun 19 11:47:42 2019
  5. @author: jbuisine
  6. """
  7. # main imports
  8. import sys, os, argparse
  9. import numpy as np
  10. import random
  11. # images processing imports
  12. from PIL import Image
  13. from ipfml.processing.segmentation import divide_in_blocks
  14. # modules imports
  15. sys.path.insert(0, '') # trick to enable import of main folder module
  16. import custom_config as cfg
  17. from modules.utils import data as dt
  18. from modules.classes.Transformation import Transformation
  19. # getting configuration information
  20. zone_folder = cfg.zone_folder
  21. learned_folder = cfg.learned_zones_folder
  22. min_max_filename = cfg.min_max_filename_extension
  23. # define all scenes values
  24. scenes_list = cfg.scenes_names
  25. scenes_indices = cfg.scenes_indices
  26. dataset_path = cfg.dataset_path
  27. zones = cfg.zones_indices
  28. seuil_expe_filename = cfg.seuil_expe_filename
  29. features_choices = cfg.features_choices_labels
  30. output_data_folder = cfg.output_data_folder
  31. generic_output_file_svd = '_random.csv'
  32. def generate_data_model(_filename, _transformations, _scenes_list, _nb_zones = 4, _random=0):
  33. output_train_filename = _filename + ".train"
  34. output_test_filename = _filename + ".val"
  35. if not '/' in output_train_filename:
  36. raise Exception("Please select filename with directory path to save data. Example : data/dataset")
  37. # create path if not exists
  38. if not os.path.exists(output_data_folder):
  39. os.makedirs(output_data_folder)
  40. zones_indices = zones
  41. train_file_data = []
  42. test_file_data = []
  43. scenes = os.listdir(dataset_path)
  44. # remove min max file from scenes folder
  45. scenes = [s for s in scenes if min_max_filename not in s]
  46. # go ahead each scenes
  47. for folder_scene in _scenes_list:
  48. scene_path = os.path.join(dataset_path, folder_scene)
  49. # shuffle list of zones (=> randomly choose zones)
  50. # only in random mode
  51. if _random:
  52. random.shuffle(zones_indices)
  53. # store zones learned
  54. learned_zones_indices = zones_indices[:_nb_zones]
  55. # write into file
  56. folder_learned_path = os.path.join(learned_folder, _filename.split('/')[1])
  57. if not os.path.exists(folder_learned_path):
  58. os.makedirs(folder_learned_path)
  59. file_learned_path = os.path.join(folder_learned_path, folder_scene + '.csv')
  60. with open(file_learned_path, 'w') as f:
  61. for i in learned_zones_indices:
  62. f.write(str(i) + ';')
  63. for id_zone, index_folder in enumerate(zones_indices):
  64. index_str = str(index_folder)
  65. if len(index_str) < 2:
  66. index_str = "0" + index_str
  67. current_zone_folder = "zone" + index_str
  68. zone_path = os.path.join(scene_path, current_zone_folder)
  69. # custom path for interval of reconstruction and metric
  70. features_path = []
  71. for transformation in _transformations:
  72. # check if it's a static content and create augmented images if necessary
  73. if transformation.getName() == 'static':
  74. # {sceneName}/zoneXX/static
  75. static_metric_path = os.path.join(zone_path, transformation.getName())
  76. # img.png
  77. image_name = transformation.getParam().split('/')[-1]
  78. # {sceneName}/zoneXX/static/img
  79. image_prefix_name = image_name.replace('.png', '')
  80. image_folder_path = os.path.join(static_metric_path, image_prefix_name)
  81. if not os.path.exists(image_folder_path):
  82. os.makedirs(image_folder_path)
  83. features_path.append(image_folder_path)
  84. # get image path to manage
  85. # {sceneName}/static/img.png
  86. transform_image_path = os.path.join(scene_path, transformation.getName(), image_name)
  87. static_transform_image = Image.open(transform_image_path)
  88. static_transform_image_block = divide_in_blocks(static_transform_image, cfg.sub_image_size)[id_zone]
  89. dt.augmented_data_image(static_transform_image_block, image_folder_path, image_prefix_name)
  90. else:
  91. metric_interval_path = os.path.join(zone_path, transformation.getTransformationPath())
  92. features_path.append(metric_interval_path)
  93. # as labels are same for each metric
  94. for label in os.listdir(features_path[0]):
  95. label_features_path = []
  96. for path in features_path:
  97. label_path = os.path.join(path, label)
  98. label_features_path.append(label_path)
  99. # getting images list for each metric
  100. features_images_list = []
  101. for index_metric, label_path in enumerate(label_features_path):
  102. if _transformations[index_metric].getName() == 'static':
  103. # by default append nothing..
  104. features_images_list.append([])
  105. else:
  106. images = sorted(os.listdir(label_path))
  107. features_images_list.append(images)
  108. # construct each line using all images path of each
  109. for index_image in range(0, len(features_images_list[0])):
  110. images_path = []
  111. # get information about rotation and flip from first transformation (need to be a not static transformation)
  112. current_post_fix = features_images_list[0][index_image].split(cfg.post_image_name_separator)[-1]
  113. # getting images with same index and hence name for each metric (transformation)
  114. for index_metric in range(0, len(features_path)):
  115. # custom behavior for static transformation (need to check specific image)
  116. if _transformations[index_metric].getName() == 'static':
  117. # add static path with selecting correct data augmented image
  118. image_name = _transformations[index_metric].getParam().split('/')[-1].replace('.png', '')
  119. img_path = os.path.join(features_path[index_metric], image_name + cfg.post_image_name_separator + current_post_fix)
  120. images_path.append(img_path)
  121. else:
  122. img_path = features_images_list[index_metric][index_image]
  123. images_path.append(os.path.join(label_features_path[index_metric], img_path))
  124. if label == cfg.noisy_folder:
  125. line = '1;'
  126. else:
  127. line = '0;'
  128. # compute line information with all images paths
  129. for id_path, img_path in enumerate(images_path):
  130. if id_path < len(images_path) - 1:
  131. line = line + img_path + '::'
  132. else:
  133. line = line + img_path
  134. line = line + '\n'
  135. if id_zone < _nb_zones:
  136. train_file_data.append(line)
  137. else:
  138. test_file_data.append(line)
  139. train_file = open(output_train_filename, 'w')
  140. test_file = open(output_test_filename, 'w')
  141. random.shuffle(train_file_data)
  142. random.shuffle(test_file_data)
  143. for line in train_file_data:
  144. train_file.write(line)
  145. for line in test_file_data:
  146. test_file.write(line)
  147. train_file.close()
  148. test_file.close()
  149. def main():
  150. parser = argparse.ArgumentParser(description="Compute specific dataset for model using of metric")
  151. parser.add_argument('--output', type=str, help='output file name desired (.train and .test)')
  152. parser.add_argument('--features', type=str,
  153. help="list of features choice in order to compute data",
  154. default='svd_reconstruction, ipca_reconstruction',
  155. required=True)
  156. parser.add_argument('--params', type=str,
  157. help="list of specific param for each metric choice (See README.md for further information in 3D mode)",
  158. default='100, 200 :: 50, 25',
  159. required=True)
  160. parser.add_argument('--size', type=str,
  161. help="Size of input images",
  162. default="100, 100")
  163. parser.add_argument('--scenes', type=str, help='List of scenes to use for training data')
  164. parser.add_argument('--nb_zones', type=int, help='Number of zones to use for training data set', choices=list(range(1, 17)))
  165. parser.add_argument('--random', type=int, help='Data will be randomly filled or not', choices=[0, 1])
  166. args = parser.parse_args()
  167. p_filename = args.output
  168. p_features = list(map(str.strip, args.features.split(',')))
  169. p_params = list(map(str.strip, args.params.split('::')))
  170. p_scenes = args.scenes.split(',')
  171. p_size = args.size # not necessary to split here
  172. p_nb_zones = args.nb_zones
  173. p_random = args.random
  174. # create list of Transformation
  175. transformations = []
  176. for id, feature in enumerate(p_features):
  177. if feature not in features_choices:
  178. raise ValueError("Unknown metric, please select a correct metric : ", features_choices)
  179. transformations.append(Transformation(feature, p_params[id], p_size))
  180. if transformations[0].getName() == 'static':
  181. raise ValueError("The first transformation in list cannot be static")
  182. # Update: not use of renderer scenes list
  183. # getting scenes from indexes user selection
  184. scenes_selected = []
  185. for scene_id in p_scenes:
  186. index = scenes_indices.index(scene_id.strip())
  187. scenes_selected.append(scenes_list[index])
  188. # create database using img folder (generate first time only)
  189. generate_data_model(p_filename, transformations, scenes_selected, p_nb_zones, p_random)
  190. if __name__== "__main__":
  191. main()