Download images from IPFS

Question

I have a nice set of URLs saved in this format

Number  Link
0   https://ipfs.io/ipfs/QmRRPWG96cmgTn2qSzjwr2qvfNEuhunv6FNeMFGa9bx6mQ
1   https://ipfs.io/ipfs/QmPbxeGcXhYQQNgsC6a36dDyYUcHgMLnGKnF8pVFmGsvqi
2   https://ipfs.io/ipfs/QmcJYkCKK7QPmYWjp4FD2e3Lv5WCGFuHNUByvGKBaytif4
3   https://ipfs.io/ipfs/QmYxT4LnK8sqLupjbS6eRvu1si7Ly2wFQAqFebxhWntcf6
4   https://ipfs.io/ipfs/QmSg9bPzW9anFYc3wWU5KnvymwkxQTpmqcRSfYj7UmiBa7
5   https://ipfs.io/ipfs/QmNwbd7ctEhGpVkP8nZvBBQfiNeFKRdxftJAxxEdkUKLcQ
6   https://ipfs.io/ipfs/QmWBgfBhyVmHNhBfEQ7p1P4Mpn7pm5b8KgSab2caELnTuV
7   https://ipfs.io/ipfs/QmRsJLrg27GQ1ZWyrXZFuJFdU5bapfzsyBfm3CAX1V1bw6

I am trying to use a loop to loop through all of the links and save the file

import urllib.request

for x,y in zip(link, num):
    url = str(x)
    name = str(y)
    filename = "%s.png" % name   
    urllib.request.urlretrieve(url, filename)

Everytime I run this code I get this error

URLError: <urlopen error [WinError 10054] An existing connection was forcibly closed by the remote host>

What is weird is that if I just run the code on one URL then it works fine.

import urllib.request

name = 1
filename = "%s.png" % name   
urllib.request.urlretrieve("https://ipfs.io/ipfs/QmcJYkCKK7QPmYWjp4FD2e3Lv5WCGFuHNUByvGKBaytif4", filename)

How can this be fixed so that the code runs in a loop with no errors?

thanks

EDIT

Here is some code that works for 1 image

import pandas as pd 
import urllib.request

links = [['number', 'link'], ['1', 'https://ipfs.io/ipfs/QmPbxeGcXhYQQNgsC6a36dDyYUcHgMLnGKnF8pVFmGsvqi'], ['2', 'https://ipfs.io/ipfs/QmcJYkCKK7QPmYWjp4FD2e3Lv5WCGFuHNUByvGKBaytif4'], ['3', 'https://ipfs.io/ipfs/QmYxT4LnK8sqLupjbS6eRvu1si7Ly2wFQAqFebxhWntcf6']]
data = pd.DataFrame(links)

link = data.get('Link', None) 
num = data.get('Number', None)


name = 1
filename = "%s.png" % name   
urllib.request.urlretrieve("https://ipfs.io/ipfs/QmYxT4LnK8sqLupjbS6eRvu1si7Ly2wFQAqFebxhWntcf6", filename)

Answer 1

You are being throttled by the IPFS service. You need to implement API rate limiting (or see if the service has a premium option that allows you to pay for higher API request rates).

Here's one way to implement client-side rate limiting, using exponential backoff/retry:

1. save this retry code as retry.py
2. fix a couple of Python v2 issues in retry.py (except ExceptionToCheck as e: at line 32 and print(msg) at line 37)
3. modify your client code as follows

import urllib.request
from retry import retry

LINKS = [
    "https://ipfs.io/ipfs/QmRRPWG96cmgTn2qSzjwr2qvfNEuhunv6FNeMFGa9bx6mQ",
    "https://ipfs.io/ipfs/QmPbxeGcXhYQQNgsC6a36dDyYUcHgMLnGKnF8pVFmGsvqi",
    "https://ipfs.io/ipfs/QmcJYkCKK7QPmYWjp4FD2e3Lv5WCGFuHNUByvGKBaytif4",
    "https://ipfs.io/ipfs/QmYxT4LnK8sqLupjbS6eRvu1si7Ly2wFQAqFebxhWntcf6",
    "https://ipfs.io/ipfs/QmSg9bPzW9anFYc3wWU5KnvymwkxQTpmqcRSfYj7UmiBa7",
    "https://ipfs.io/ipfs/QmNwbd7ctEhGpVkP8nZvBBQfiNeFKRdxftJAxxEdkUKLcQ",
    "https://ipfs.io/ipfs/QmWBgfBhyVmHNhBfEQ7p1P4Mpn7pm5b8KgSab2caELnTuV",
    "https://ipfs.io/ipfs/QmRsJLrg27GQ1ZWyrXZFuJFdU5bapfzsyBfm3CAX1V1bw6",
]

@retry(urllib.error.URLError, tries=4)
def download(index, url):
    filename = "%s.png" % index
    urllib.request.urlretrieve(url, filename)

def main():
    for index, link in enumerate(LINKS):
        print(index, link)
        download(index, link)

if __name__ == '__main__':
    main()

I tested this code without retries and it was throttled (as expected). Then I added the retry decorator and it completed successfully (including a couple of expected retries).