download_images_from_a_csv.py

"""
python3.6
downloads images from URLs stored in a csv file according to labels
and saves them to a specified directory / subdirectory
e.g. for sorting images into folders by labels for image classification
"""

import pandas as pd 
import urllib 
import os
# for python3: install SSL certificates
# /Applications/Python\ 3.6/Install\ Certificates.command

labels = ['label1', 'label2', 'label3'...
]

df = pd.read_csv("clothing_col_nov_dec2018.csv")

for label in labels:
	directory = "path/and/folder" #+ label
	if not os.path.exists(directory):
		os.makedirs(directory)
	df_temp  = df.query("label_col==@label").reset_index() 

	for idx, row in df_temp.iterrows() :   
		img1 = row['col_1']
		img2 = row['col_2']
		img3 = row['col_3']
		img4 = row['col_4']

		imgs = [img1, img2, img3, img4]
		
		pic_name = row['col']
		subdirectory = row['subdirectory_col']
		directory_label = directory + subdirectory + "/" + label
		if not os.path.exists(subdirectory):
			os.makedirs(subdirectory)
		i = 1
		for img in imgs:
			try:
				# store by category and color
				urllib.request.urlretrieve(img, directory + "/" + subdirectory+"/"+label+"/"+label+"-"+pic_name+"-"+str(idx)+"-"+str(i)+".jpg")
				i = i+1
			except:
				print('failed for '+ pic_name + "-" + str(i) + ' in ' + subdirectory)
				continue