def download(args, df_val, folder, dataset_dir, class_name, class_code, class_list=None):
'''
Manage the download of the images and the label maker.
:param args: argument parser.
:param df_val: DataFrame Values
:param folder: train, validation or test
:param dataset_dir: self explanatory
:param class_name: self explanatory
:param class_code: self explanatory
:param class_list: list of the class if multiclasses is activated
:return: None
'''
print('-' * 10 + class_name + '-' * 10)
print('[INFO] Downloading {} images.'.format(args.type_csv))
df_val_images = images_options(df_val, args)
images_list = df_val_images['ImageID'][df_val_images.LabelName == class_code].values
images_list = set(images_list)
if class_list is not None:
class_name_list = '_'.join(class_list)
else:
class_name_list = class_name
download_img(folder, dataset_dir, class_name_list, images_list)
get_label(folder, dataset_dir, class_name, class_code, df_val, class_name_list)
def download(args, df_val, folder, dataset_dir, class_name, class_code, class_list=None, threads = 20):
'''
Manage the download of the images and the label maker.
:param args: argument parser.
:param df_val: DataFrame Values
:param folder: train, validation or test
:param dataset_dir: self explanatory
:param class_name: self explanatory
:param class_code: self explanatory
:param class_list: list of the class if multiclasses is activated
:param threads: number of threads
:return: None
'''
print('-' * 10 + class_name + '-' * 10)
print('[INFO] Downloading {} images.'.format(args.type_csv))
df_val_images = images_options(df_val, args)
images_list = df_val_images['ImageID'][df_val_images.LabelName == class_code].values
images_list = set(images_list)
print("[INFO] Found {} online images for {}.".format(len(images_list), folder))
if args.limit is not None:
import itertools
print('[INFO] Limiting to {} images.'.format(args.limit))
images_list = set(itertools.islice(images_list, args.limit))
if class_list is not None:
class_name_list = '_'.join(class_list)
else:
class_name_list = class_name
download_img(folder, dataset_dir, class_name_list, images_list, threads)
get_label(folder, dataset_dir, class_name, class_code, df_val, class_name_list, args)
def download(args,
df_val,
folder,
dataset_dir,
class_name,
class_code,
class_list=None,
threads=20):
'''
Manage the download of the images and the label maker.
:param args: argument parser.
:param df_val: DataFrame Values
:param folder: train, validation or test
:param dataset_dir: self explanatory
:param class_name: self explanatory
:param class_code: self explanatory
:param class_list: list of the class if multiclasses is activated
:param threads: number of threads
:return: None
'''
if os.name == 'posix':
rows, columns = os.popen('stty size', 'r').read().split()
elif os.name == 'nt':
try:
columns, rows = os.get_terminal_size(0)
except OSError:
columns, rows = os.get_terminal_size(1)
else:
columns = 50
l = int((int(columns) - len(class_name)) / 2)
print('\n' + bc.HEADER + '-' * l + class_name + '-' * l + bc.ENDC)
print(bc.INFO + 'Downloading {} images.'.format(args.type_csv) + bc.ENDC)
df_val_images = images_options(df_val, args)
images_list = df_val_images['ImageID'][df_val_images.LabelName ==
class_code].values
images_list = set(images_list)
print(bc.INFO + '[INFO] Found {} online images for {}.'.format(
len(images_list), folder) + bc.ENDC)
if args.limit is not None:
import itertools
print(bc.INFO + 'Limiting to {} images.'.format(args.limit) + bc.ENDC)
images_list = set(itertools.islice(images_list, args.limit))
if class_list is not None:
class_name_list = '_'.join(class_list)
else:
class_name_list = class_name
download_img(folder, dataset_dir, class_name_list, images_list, threads)
if not args.sub:
get_label(folder, dataset_dir, class_name, class_code, df_val,
class_name_list, args)
def download(args,
df_val,
folder,
dataset_dir,
class_name,
class_code,
class_list=None,
threads=20):
'''
Manage the download of the images and the label maker.
:param args: argument parser.
:param df_val: DataFrame Values
:param folder: train, validation or test
:param dataset_dir: self explanatory
:param class_name: self explanatory
:param class_code: self explanatory
:param class_list: list of the class if multiclasses is activated
:param threads: number of threads
:return: None
'''
print('-' * 10 + class_name + '-' * 10)
print('[INFO] Downloading {} images.'.format(args.type_csv))
df_val_images = images_options(df_val, args)
images_list = df_val_images['ImageID'][df_val_images.LabelName ==
class_code].values
images_list = set(images_list)
print("[INFO] Found {} online images for {}.".format(
len(images_list), folder))
if args.limit is not None:
import itertools
print('[INFO] Limiting to {} images.'.format(args.limit))
images_list = set(itertools.islice(images_list, args.limit))
if class_list is not None:
class_name_list = '_'.join(class_list)
else:
class_name_list = class_name
download_img(folder, dataset_dir, class_name_list, images_list, threads)
get_label(folder, dataset_dir, class_name, class_code, df_val,
class_name_list, args)
def download(args,
df_val,
folder,
dataset_dir,
class_name,
class_code,
class_list=None,
threads=20):
'''
Manage the download of the images and the label maker.
:param args: argument parser.
:param df_val: DataFrame Values
:param folder: train, validation or test
:param dataset_dir: self explanatory
:param class_name: self explanatory
:param class_code: self explanatory
:param class_list: list of the class if multiclasses is activated
:param threads: number of threads
:return: None
'''
if os.name == 'posix':
rows, columns = os.popen('stty size', 'r').read().split()
elif os.name == 'nt':
try:
columns, rows = os.get_terminal_size(0)
except OSError:
columns, rows = os.get_terminal_size(1)
else:
columns = 50
l = int((int(columns) - len(class_name)) / 2)
print('\n' + bc.HEADER + '-' * l + class_name + '-' * l + bc.ENDC)
print(bc.INFO + 'Downloading {} images.'.format(args.type_csv) + bc.ENDC)
df_val_images = images_options(df_val, args)
images_list = df_val_images['ImageID'][df_val_images.LabelName ==
class_code].values
images_list = set(images_list)
print(bc.INFO + '[INFO] Found {} online images for {}.'.format(
len(images_list), folder) + bc.ENDC)
if args.limit is not None:
import itertools
print(bc.INFO + 'Limiting to {} images.'.format(args.limit) + bc.ENDC)
images_list = set(itertools.islice(images_list, args.limit))
if class_list is not None:
class_name_list = '_'.join(class_list)
else:
class_name_list = class_name
"""
aws s3 --no-sign-request --only-show-errors cp s3://open-images-dataset/validation/78fc07b95a216074.jpg "OID\Dataset\validation\Coin"
1f990b331e554127_m03qhv5_c2198baf.png
aws s3 --no-sign-request --only-show-errors cp s3://open-images-dataset/validation/1f990b331e554127_m03qhv5_c2198baf.png "OID\Dataset\validation\Heater"
aws s3 --no-sign-request --only-show-errors cp s3://open-images-dataset/1f990b331e554127_m03qhv5_c2198baf.png "OID\Dataset\validation\Heater"
"""
# download the actual image
download_img(folder, dataset_dir, class_name_list, images_list, threads,
".jpg")
# again for the mask image
images_list = df_val_images['MaskPath'][df_val_images.LabelName ==
class_code].values
images_list = set(images_list)
print(bc.INFO + '[INFO] Found {} online images for {}.'.format(
len(images_list), folder) + bc.ENDC)
if args.limit is not None:
import itertools
print(bc.INFO + 'Limiting to {} images.'.format(args.limit) + bc.ENDC)
images_list = set(itertools.islice(images_list, args.limit))
masks_list = images_list
# download_img(folder, dataset_dir, class_name_list, masks_list, threads, "")
# print("\n\n\n\n\n")
# print(mask)
# orig_mask = "D:\oid-seg_masks\0" + mask[0] + "\\" + mask
# orig_mask = "D:\oid-seg_masks"
new_dst = str(os.path.join(dataset_dir, folder, class_name))
# path = image_dir + '/' + str(image) + file_format + ' "' + download_dir + '\\' + mask + '"'
# for mask in masks_list:
# orig_mask = str(os.path.join("D:\oid-seg_masks", '0' + mask[0], mask))
# command = ('xcopy /q /n ' + orig_mask + ' ' + new_dst)
# # command = command + '\\0' + mask[0] + '\\' + mask + ' '
# # print(command)
# subprocess.call(command)
if not args.skip_json_gen:
# get_label(folder, dataset_dir, class_name, class_code, df_val, class_name_list, args)
make_json(MASK_DIR, masks_list, folder, class_name, dataset_dir)
