File size: 11,130 Bytes
2cda712
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
# Importing Libraires
import numpy as np
import pandas as pd
import math
import scipy
from PIL import Image
import random

import torch
import torchvision
from torchvision import transforms

import os,sys,warnings
warnings.filterwarnings("ignore")
from tqdm import tqdm
import pathlib
import functions.utils as utils
import defaults


# Saving Images Paths for UnivFD dataset
def save_all_images_path_UnivFD(
	imgs_dir:str,
	status:str,
	save_path:str,
	replace:bool
):
	if os.path.exists(save_path) == False or replace == True:
		# Dataset
		dataset_images_paths = {}

		# For each UnivFD image-source for the given status
		for _,source in tqdm(enumerate(defaults.All_UnivFD_Sources[status])):
			dataset_images_paths[source] = {}

			# Initializing
			for label in ["fake", "real"]:
				dataset_images_paths[source][label] = []

			# Images Source Directory
			source_images_dir = os.path.join(imgs_dir, status, source)

			# For each label or category
			if ("0_real" in os.listdir(source_images_dir)) and ("1_fake" in os.listdir(source_images_dir)):

				# For each label 
				for _,label in enumerate(os.listdir(source_images_dir)):
					if label == "0_real":
						img_label = "real"
					elif label == "1_fake":
						img_label = "fake"
					else:
						assert False, "Unknown Label encountered."

					for fname in os.listdir(os.path.join(source_images_dir, label)):
						# Image Path and Label
						img_path = os.path.join(source_images_dir, label, fname)

						# Appending
						dataset_images_paths[source][img_label].append(os.path.relpath(img_path, defaults.main_dataset_dir))

			else:
				# For each category
				for _,category in enumerate(os.listdir(source_images_dir)):
					
					# For each label 
					for _,label in enumerate(os.listdir(os.path.join(source_images_dir, category))):
						if label == "0_real":
							img_label = "real"
						elif label == "1_fake":
							img_label = "fake"
						else:
							assert False, "Unknown Label encountered."

						for fname in os.listdir(os.path.join(source_images_dir, category, label)):
							# Image Path and Label
							img_path = os.path.join(source_images_dir, category, label, fname)

							# Appending
							dataset_images_paths[source][img_label].append(os.path.relpath(img_path, defaults.main_dataset_dir))
		
		# Saving
		np.save(save_path, dataset_images_paths)


# Saving Images Paths for GenImage dataset
def save_all_images_path_GenImage(
	imgs_dir:str,
	status:str,
	save_path:str,
	replace:bool
):
	if os.path.exists(save_path) == False or replace == True:
		# Dataset
		dataset_images_paths = {}

		# For each GenImage image-source for the given status
		for _,source in tqdm(enumerate(defaults.All_GenImage_Sources[status])):
			dataset_images_paths[source] = {}

			# Initializing
			for label in ["fake", "real"]:
				dataset_images_paths[source][label] = []

			# Images Source Directory
			source_images_dir = os.path.join(imgs_dir, source, status)

			# For each label
			for _,label in enumerate(os.listdir(source_images_dir)):
				if label == "nature":
					img_label = "real"
				elif label == "ai":
					img_label = "fake"
				elif (label == "ai_reconstructed_inpainting" or label == "nature_reconstructed_inpainting") and status == "train" and source == "sdv4":
					print ("Encountered label:{} for status:{} and source:{}".format(label, status, source))
					img_label = "fake"
				else:
					assert False, "Unknown Label encountered."

				for fname in os.listdir(os.path.join(source_images_dir, label)):
					# Image Path and Label
					img_path = os.path.join(source_images_dir, label, fname)

					# Appending
					dataset_images_paths[source][img_label].append(os.path.relpath(img_path, defaults.main_dataset_dir))
		
		# Saving
		np.save(save_path, dataset_images_paths)


# Saving Images Paths for DRCT dataset
def save_all_images_path_DRCT(
	imgs_dir:str,
	status:str,
	save_path:str,
	replace:bool
):
	if os.path.exists(save_path) == False or replace == True:
		# Dataset
		dataset_images_paths = {}

		# For Training Dataset: Real Images, Fake Images, Real Reconstructed Images and Fake Reconstructed Imagees
		if status == "train":
			# For each image-source
			for _,source in tqdm(enumerate(defaults.All_DRCT_Sources[status])):
				dataset_images_paths[source] = {}

				# Initializing
				for label in ["fake", "real"]:
					dataset_images_paths[source][label] = []

				# Images Source Directory
				real_images_dir = os.path.join(imgs_dir, "real_images", "{}2017".format(status))
				fake_images_dirs = [
					os.path.join(imgs_dir, "fake_images", source, "{}2017".format(status)),
					os.path.join(imgs_dir, "fake_reconstructed_images", source, "{}2017".format(status)),
					os.path.join(imgs_dir, "real_reconstructed_images", source, "{}2017".format(status)),
				]

				# Real Images Paths
				img_label = "real"
				for fname in os.listdir(real_images_dir):
					# Image Path and Label
					img_path = os.path.join(real_images_dir, fname)

					# Appending
					dataset_images_paths[source][img_label].append(os.path.relpath(img_path, defaults.main_dataset_dir))

				# Fake Images Paths
				img_label = "fake"
				for i in range(len(fake_images_dirs)):
					for fname in os.listdir(fake_images_dirs[i]):
						# Image Path and Label
						img_path = os.path.join(fake_images_dirs[i], fname)

						# Appending
						dataset_images_paths[source][img_label].append(os.path.relpath(img_path, defaults.main_dataset_dir))

		# For Validation Dataset: Real Images, Fake Images
		else:
			# For each image-source
			for _,source in tqdm(enumerate(defaults.All_DRCT_Sources[status])):
				dataset_images_paths[source] = {}

				# Initializing
				for label in ["fake", "real"]:
					dataset_images_paths[source][label] = []

				# Images Source Directory
				real_images_dir = os.path.join(imgs_dir, "real_images", "{}2017".format(status))
				fake_images_dir = os.path.join(imgs_dir, "fake_images", source, "{}2017".format(status))

				# Real Images Paths
				img_label = "real"
				for fname in os.listdir(real_images_dir):
					# Image Path and Label
					img_path = os.path.join(real_images_dir, fname)

					# Appending
					dataset_images_paths[source][img_label].append(os.path.relpath(img_path, defaults.main_dataset_dir))

				# Fake Images Paths
				img_label = "fake"
				for fname in os.listdir(fake_images_dir):
					# Image Path and Label
					img_path = os.path.join(fake_images_dir, fname)

					# Appending
					dataset_images_paths[source][img_label].append(os.path.relpath(img_path, defaults.main_dataset_dir))
		
		# Saving
		np.save(save_path, dataset_images_paths)


# Saving all paths of image dataset
def save_all_images_paths(
	imgs_dir:str,
	dataset_type:str,
	status:str,
	save_path:str,
	replace:bool
):
	"""
	Saves path info images of a dataset_type, status, image_sources.
	Args:
		imgs_dir (str): Directory of images.
		dataset_type (str): Type of Dataset. Options: ["UnivFD", "GenImage", "DRCT]
		status (str): ["train", "val"]
		save_path (str): Path to save .npy file.
		replace (bool): Replace File if True.
	"""
	# Assertions
	assert dataset_type in ["UnivFD", "GenImage", "DRCT"], "Invalid dataset"
	assert os.path.exists(imgs_dir), f"Image directory {imgs_dir} is not found."
	assert status in ["train", "val"], "Invalid status"

	if dataset_type == "UnivFD":
		save_all_images_path_UnivFD(
			imgs_dir=imgs_dir,
			status=status,
			save_path=save_path,
			replace=replace
		)
	elif dataset_type == "GenImage":
		save_all_images_path_GenImage(
			imgs_dir=imgs_dir,
			status=status,
			save_path=save_path,
			replace=replace
		)
	else:
		save_all_images_path_DRCT(
			imgs_dir=imgs_dir,
			status=status,
			save_path=save_path,
			replace=replace
		)


# Get Images Paths
def get_image_paths(
	dataset_type:str,
	status:str,
	image_sources:str,
	label:str,
):
	"""
	Get path to all images in the folder based on arguments.
	Args:
		dataset_type (str): Type of Dataset. Options: ["UnivFD", "GenImage", "DRCT]
		status (str): ["train", "val"]
		image_sources (list): Image-Sources to consider for dataset.
		label (str): ["real", "fake"]
	"""
	# Assertions
	assert dataset_type in ["UnivFD", "GenImage", "DRCT"], "Invalid dataset"
	assert status in ["train", "val"], "Invalid status"
	assert label in ["real", "fake"], "Invalid label"


	# Loading Paths
	img_dir = os.path.join(defaults.main_dataset_dir, dataset_type, "dataset")
	info_path = os.path.join(defaults.main_dataset_dir, "Info", "{}_{}_image_Paths.npy".format(dataset_type, status))


	# Saving Info File
	if os.path.exists(info_path) == False:
		print ("Saving Info File")

		save_all_images_paths(
			imgs_dir=img_dir,
			dataset_type=dataset_type,
			status=status,
			save_path=info_path,
			replace=False
		)
	

	# Loading Path Info
	Path_Info = np.load(info_path, allow_pickle=True)[()]
	

	# Dataset
	dataset_images_paths = []
	# For each image-source
	for _, source in enumerate(image_sources):
		for img_path in sorted(Path_Info[source][label]):
			# Image-Path
			dataset_images_paths.append(img_path)

	return dataset_images_paths



# Dataset Paths
def dataset_img_paths(
	dataset_type:str,
	status:str
):
	"""
	Returns real_image_paths and fake_image_paths based on arguments.
	Args:
		dataset_type (str): Type of Dataset. Options: ["UnivFD", "GenImage", "DTCT]
		status (str): ["train", "val"]
	"""
	# Assertions
	assert dataset_type in ["UnivFD", "GenImage", "DRCT"], "Invalid dataset"
	assert status in ["train", "val"], "Invalid status"

	# DRCT Dataset
	if dataset_type == "DRCT":
		train_image_sources, test_image_sources = utils.get_DRCT_options()

		if status == "train":
			image_sources = train_image_sources
		else:
			image_sources = test_image_sources

		real_images_paths = get_image_paths(
			dataset_type=dataset_type,
			status=status,
			image_sources=image_sources,
			label="real"
		)

		fake_images_paths = get_image_paths(
			dataset_type=dataset_type,
			status=status,
			image_sources=image_sources,
			label="fake"
		)

	# GenImage Dataset
	elif dataset_type == "GenImage":
		train_image_sources, test_image_sources = utils.get_GenImage_options()

		if status == "train":
			image_sources = train_image_sources
		else:
			image_sources = test_image_sources

		real_images_paths = get_image_paths(
			dataset_type=dataset_type,
			status=status,
			image_sources=image_sources,
			label="real"
		)

		fake_images_paths = get_image_paths(
			dataset_type=dataset_type,
			status=status,
			image_sources=image_sources,
			label="fake"
		)

	# UnivFD Dataset
	elif dataset_type == "UnivFD":
		train_image_sources, test_image_sources = utils.get_UnivFD_options()

		if status == "train":
			image_sources = train_image_sources
		else:
			image_sources = test_image_sources

		real_images_paths = get_image_paths(
			dataset_type=dataset_type,
			status=status,
			image_sources=image_sources,
			label="real"
		)

		fake_images_paths = get_image_paths(
			dataset_type=dataset_type,
			status=status,
			image_sources=image_sources,
			label="fake"
		)

	else:
		assert False, "Unknown dataset_type: {}".format(dataset_type)

	return real_images_paths, fake_images_paths