Added Extracting open images

open-images/open-images.py

+import argparse
+import pandas as pd
+import numpy as np
+import os
+import shutil
+from tqdm import tqdm
+import sys
+import tarfile
+
+
+##
+parser = argparse.ArgumentParser(description="Arguments for Prep")
+
+parser.add_argument('-d', '--data', type=str, help="Pfad für den Daten Ordner", default="/nue/net_scratch/transfer/open-images-dataset")
+
+parser.print_help()
+
+args = parser.parse_args()
+##
+
+
+
+parent_dir = args.data
+
+assert not os.path.isdir('train'), "Train Ordner schon erstellt. Wurde alles entpackt? Falls nicht, lösche train Ordner und führe Programm erneut aus."
+
+print(os.path.isdir('train')*-1)
+
+os.makedirs("train-temp")
+os.makedirs("train")
+
+print("Extracting folders..")
+
+for file in tqdm(os.listdir(parent_dir)):
+    if 'train_' in file:
+        temp = tarfile.open(parent_dir + file, "r")
+        temp.extractall("train-temp")
+
+
+temp_dir = "train-temp"
+abs_path = os.path.abspath("train")
+print(f'Moving Images into {abs_path}')
+for f in tqdm(os.listdir(temp_dir)):
+    t_dir = temp_dir+"/" +f
+    for img in os.listdir(t_dir):
+        shutil.move(t_dir +"/"+ img, "train")
+
+shutil.rmtree("train-temp")
+
+
+print(f'Nutze diesen Pfad für das Pretraining: {abs_path}')
\ No newline at end of file