Checks a new line delimited url file for https, http redirection, and a valid ssl certificate

check-website-ssl-from-file.py

#!/usr/bin/python
# Description:
#       Checks websites for https from a text file. The text file must be
#     new line delimited.
#
# Usage:
#     python script.py myfileoflinks.txt
#
# Install:
#     pip install tabulate

from requests.exceptions import SSLError, ConnectionError
import requests
import sys
import pdb
try:
    from tabulate import tabulate
except:
    print("Please install tabulate:")
    print("\tpip install tabulate")
    sys.exit(0)

def _checkurl(url):
    """
    Checks url to see if the ssl cert is trusted
    Checks request history to see if the url redirected from http to https
    """
    global data
    global headers
    result = False
    # attempt to connect to the url as a real browser (see above header)
    try:
        res = requests.get(url, allow_redirects=True, headers=header)
        data[-1]['ssl'] = True
    except SSLError:
        print("* The SSL certificate is not trusted.")
        data[-1]['ssl'] = False
        return result
    except ConnectionError:
        print("* Possibly temporarily blocked for making too many requests")
        data[-1]['ssl'] = False
        return result
    # check history to see if the url redirected us to https
    try:
        surl = res.history[0].headers['Location']
        if surl.startswith("https"):
            print("redirected to: " + surl)
            result = True
    except:
        pass
    return result

def main(args):
    global header
    global data
    header = {'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X ' + \
              '10_10_1) AppleWebKit/537.36 (KHTML, like Gecko) ' + \
              'Chrome/39.0.2171.95 Safari/537.36'}
    # default file name
    secureHttpLinks = "secure-links.txt"
    if len(sys.argv) > 1:
        secureHttpLinks = sys.argv[1]

    # attempt to read file
    try:
        f = open(secureHttpLinks, "r")
        print("Reading {0}...".format(secureHttpLinks))
    except:
        print("The file '{0}' was not found".format(secureHttpLinks))
        sys.exit(0)
    n = 0
    data = []
    # loop through file
    for i in f.readlines():
        # strip new lines from url
        url = i.strip()
        print("Checking {0}".format(url))
        # replace https with http
        if url.startswith("https"):
            url = url.replace("https", "http")
        # check if the url begins with www
        elif url.startswith("www"):
            url = "http://" + url
        data.append([])
        data[n] = {'url': url}
        # check if https works
        data[n]['https'] = _checkurl(url.replace("http", "https"))
        if data[n]['https']:
            # check if it redirected from http to https
            data[n]['redirect'] = _checkurl(url)
        else:
            data[n]['redirect'] = False
        n += 1

    f.close()

    # print the list of json data using the json keys as its headers
    print(tabulate(data, headers="keys"))

if __name__ == "__main__":
    try:
        main(sys.argv)
    except KeyboardInterrupt:
        sys.exit(0)