mirror of
https://github.com/public-apis/public-apis
synced 2025-04-21 10:59:48 +02:00
50 lines
1.5 KiB
Python
Executable File
50 lines
1.5 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
|
|
import httplib2
|
|
import re
|
|
import socket
|
|
import sys
|
|
|
|
|
|
def parse_links(filename):
|
|
"""Returns a list of URLs from text file"""
|
|
with open(filename) as fp:
|
|
data = fp.read()
|
|
raw_links = re.findall(
|
|
'((?:https?://|www\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}/)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:\'\".,<>?«»“”‘’]))',
|
|
data)
|
|
links = [raw_link[0] for raw_link in raw_links]
|
|
return links
|
|
|
|
|
|
def validate_links(links):
|
|
"""Checks each entry in JSON file for live link"""
|
|
print('Validating {} links...'.format(len(links)))
|
|
errors = []
|
|
for link in links:
|
|
h = httplib2.Http(disable_ssl_certificate_validation=True, timeout=10)
|
|
try:
|
|
resp = h.request(link)
|
|
code = int(resp[0]['status'])
|
|
# check if status code is a client or server error
|
|
if code >= 404:
|
|
errors.append('{}: {}'.format(code, link))
|
|
except TimeoutError:
|
|
errors.append("TMO: " + link)
|
|
except socket.error as socketerror:
|
|
errors.append("SOC: {} : {}".format(socketerror, link))
|
|
except Exception as e:
|
|
errors.append("ERR: {} : {}".format(e, link))
|
|
return errors
|
|
|
|
if __name__ == "__main__":
|
|
num_args = len(sys.argv)
|
|
if num_args < 2:
|
|
print("No .md file passed")
|
|
sys.exit(1)
|
|
errors = validate_links(parse_links(sys.argv[1]))
|
|
if len(errors) > 0:
|
|
for err in errors:
|
|
print(err)
|
|
sys.exit(1)
|