downloading images from flickr using python-flickr

download_flickr_image.py

# First, you should install flickrapi
# pip install flickrapi

import flickrapi
import urllib
from PIL import Image

# Flickr api access key 
flickr=flickrapi.FlickrAPI('c6a2c45591d4973ff525042472446ca2', '202ffe6f387ce29b', cache=True)


keyword = 'siberian husky'

photos = flickr.walk(text=keyword,
                     tag_mode='all',
                     tags=keyword,
                     extras='url_c',
                     per_page=100,           # may be you can try different numbers..
                     sort='relevance')

urls = []
for i, photo in enumerate(photos):
    print (i)
    
    url = photo.get('url_c')
    urls.append(url)
    
    # get 50 urls
    if i > 50:
        break

print (urls)

# Download image from the url and save it to '00001.jpg'
urllib.urlretrieve(urls[1], '00001.jpg')

# Resize the image and overwrite it
image = Image.open('00001.jpg') 
image = image.resize((256, 256), Image.ANTIALIAS)
image.save('00001.jpg')

Hi,

thanks for the code! I'm quite new to python and if I run the script I only get the first url downloaded. Is there a way to download all images from alle the urls? Like an array or loop?

Also, can you specify if its possible to change the path where images are downloaded?

Thanks a lot!

I think it woud work for you

pip install flickrapi

import flickrapi
import urllib
from PIL import Image
from tqdm import tqdm

Flickr api access key

Flickr = flickrapi.FlickrAPI('c6a2c45591d4973ff525042472446ca2', '202ffe6f387ce29b', cache=True)

N_MAX = 60000
KEYWORD = 'cat'

PHOTOs = flickr.walk(text=KEYWORD,
tag_mode='all',
tags=KEYWORD
extras='url_c',
per_page=100, # may be you can try different numbers..
sort='relevance')

URLs = []
for N, PHOTO in tqdm(enumerate(PHOTOs)):

URL = PHOTO.get('url_c')
URLs.append(URL)

# get N_MAX URLs
if N > N_MAX:
    break

Download image from the url and save it

RESIZE_OPTION = Image.ANTIALIAS #Image.NEAREST, Image.BICUBIC

for i in range(N_MAX):
fname = f'img_{keyword}{i}.jpg'
urllib.request.urlretrieve(urls[i], f'img{keyword}_{i}.jpg')

# Resize the image and overwrite it
image = Image.open(fname) 
image = image.resize((256, 256), RESIZE_OPTION)
image.save(fname)

Can I use this for my repo with a small adaptation?

Hi,

thanks for the code! I'm quite new to python and if I run the script I only get the first url downloaded. Is there a way to download all images from alle the urls? Like an array or loop?

Also, can you specify if its possible to change the path where images are downloaded?

Thanks a lot!

You can check out mine, I made some changes for what I needed at the time.
https://github.com/lenoqt/PyTorch/blob/main/flickrdownloader.py

Hi,
cool stuff . . . how would I possible download photos fom an album ( a set in flickr wording I assume ) ? Actually I want to check a specific set for updates and download them if there are any . . . this I want to do as a cronjob then . .

I post this again . . . I assume I posted it wrongly the first time
cheers
T

if I just what to download some images at random from flicker without explicitly searching a specific keyword, how can I do it??

Hi, here's fix for the error if you run this script:
https://stackoverflow.com/questions/17960942/attributeerror-module-object-has-no-attribute-urlretrieve

HI!,

ive been trying to create a dataset for my final project and your code has been the most helpful out of anything that I've found. I understand it's older code, but might you be willing to share how one might download the results into a specific folder? also, when I try the resize option on more than one, using the fname, it doesn't work :( it says fname has no attribute read. I would greatly appreciate any help.

HI!,

ive been trying to create a dataset for my final project and your code has been the most helpful out of anything that I've found. I understand it's older code, but might you be willing to share how one might download the results into a specific folder? also, when I try the resize option on more than one, using the fname, it doesn't work :( it says fname has no attribute read. I would greatly appreciate any help.

Both things are doable, for the path you can adapt using os.path then join it with the f-string in the script, for changing dimensions or manipulating any of the image you can use PIL.