A python class for verifying login details for any website

gistfile1.py

import mechanize
import cookielib
import re

class Login:
    def __init__(self):
        # set a bunch of properties as default settings
        self.url = False

        self.formIndex = 0
        self.dualPage = False
        self.userField = "email"
        self.passField = "pass"

        self.success = False
        self.invalid = False
        self.rateLimited = False

        # set up the /browser/
        self.br = mechanize.Browser()
        cj = cookielib.LWPCookieJar()
        self.br.set_cookiejar(cj)

        self.br.set_handle_equiv(True)
        self.br.set_handle_redirect(True)
        self.br.set_handle_referer(True)
        self.br.set_handle_robots(False)

        self.br.set_handle_refresh(
            mechanize._http.HTTPRefreshProcessor(),
            max_time=1
        )
        # Google says we're using an outdated browser ;_;
        self.br.addheaders = [(
            'User-agent',
            'Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.1) '
            'Gecko/2008071615 Fedora/3.0.1-1.fc9 Firefox/3.0.1'
        )]

    # this function uses our browsers instance to log into the page
    def login(self, username, password):
        # make sure all of the required arguments are set
        if not self.url or not self.userField or not self.passField:
            raise ArgumentError('Specify properties: url, userField, passField')
        else:
            if not (self.success or self.invalid):
                raise ArgumentError('Specify atleast one of properties'
                                    ' `success` of ` invalid`')
            else:
                # we definitely have all of values we need, let's get to work
                # renamed for simplicity
                br = self.br
                # browse to the page
                r = br.open(self.url)
                # check if the username and password are on different pages
                if self.dualPage:
                    # we're making a bunch of assumtions here mainly to support
                    # google.
                    #
                    # this checks if the form index is a list, and if it is, it
                    # uses both values; one for each page. See also comment !1
                    if type(self.formIndex) == int:
                        formIndex = self.formIndex
                        indexIsList = False
                    else:
                        formIndex = self.formIndex[0]
                        indexIsList = True
                    # select the NRth form on the page, the first if there's one
                    br.select_form(nr=formIndex)
                    # put the username in the username field
                    br.form[self.userField] = username
                    # submit the form, and wait until the next page loads
                    br.submit()
                    # !1
                    if indexIsList:
                        formIndex = self.formIndex[1]
                    # select the correct form on the next page
                    br.select_form(nr=formIndex)
                    # enter the password into the passowrd field
                    br.form[self.passField] = password
                    # submit the form, hopefully logging us in
                    br.submit()
                else:
                    # all the fields are on the same page
                    # select the NRth form on the page, the first if there's one
                    br.select_form(nr=self.formIndex)
                    # put the username and password into their respective field
                    br.form[self.userField] = username
                    br.form[self.passField] = password
                    # submit the form
                    br.submit()
                # get the HTML of the resulting page
                html = br.response().read()
                # check to see if it was the response we were looking for
                return self.verify(
                    html,
                    self.success,
                    self.invalid,
                    self.rateLimited
                )

    # tries to work out if html is of a success, invalid, or rate limited page
    def verify(self, html, success, invalid, rateLimited):
        # make a blank object, just to keep things pretty
        r = ReturnObject()
        # add the html to the object so the developer can use it for debugging
        r.html = html
        # prioritize rate limiting message, as it's the least likely to appear
        if rateLimited and re.search(rateLimited, html):
            r.valid = False
            r.desc = "The rateLimited regex matched in the page."
            r.rateLimited = True
        # check if the invalid regex matches
        elif invalid and re.search(invalid, html):
            r.valid = False
            r.desc = "The invalid regex matched in the page."
            r.rateLimited = False
        # check if the success regex matches
        elif success and re.search(success, html):
            r.valid = True
            r.desc = "The success regex matched in the page."
            r.rateLimited = False
        # if only invalid is specified, and it didn't match, then we're probably
        # authenticated
        elif invalid and not success:
            r.valid = True
            r.desc = "Success. The invalid regex was not found in the page."
            r.rateLimited = False
        # nothing we provided matched, so it prbably failed to authenticate
        else:
            r.valid = False
            r.desc = "None of the regexes matched the page"
            r.rateLimited = False
        return r

# just something to allow us to return an object instead of a dict
class ReturnObject:
    pass

# a custom error to raise when insufficuent parameters are supplied
class ArgumentError(Exception):
    def __init__(self, value):
        self.value = value
    def __str__(self):
        return repr(self.value)

if __name__ == "__main__":
    # a test script that logs into google
    # instantiate the class
    l = Login()

    # supply all of the required parameters
    l.url = "https://accounts.google.com/ServiceLogin"
    l.dualPage = True
    l.userField = "Email"
    l.passField = "Passwd"
    l.success = "To access all of your Google account settings, upgrade to " \
                "the latest version of any of the following browsers"

    # print the description of a login attempt
    print l.login('[email protected]', 'notanactualpassword').desc