| 1 |
# Website |
| 2 |
# |
| 3 |
# Douglas Thrift |
| 4 |
# |
| 5 |
# $Id$ |
| 6 |
|
| 7 |
from BeautifulSoup import MinimalSoup |
| 8 |
import copy |
| 9 |
import mechanize |
| 10 |
import re |
| 11 |
|
| 12 |
# XXX: hack to make Wells Fargo http-equiv redirects actually work |
| 13 |
mechanize._http.AbstractHeadParser.head_elems = tuple(list(mechanize._http.AbstractHeadParser.head_elems) + ['body']) |
| 14 |
mechanize._http.HTTPRedirectHandler.max_redirections = 20 |
| 15 |
|
| 16 |
class DuckSoup(MinimalSoup): |
| 17 |
MARKUP_MASSAGE = copy.copy(MinimalSoup.MARKUP_MASSAGE) |
| 18 |
|
| 19 |
MARKUP_MASSAGE.append((re.compile(r'''</?[a-z]+['"]\+['"][a-z]+'''), lambda match: '')) |
| 20 |
|
| 21 |
class Factory(mechanize.DefaultFactory): |
| 22 |
def __init__(self, form_parser): |
| 23 |
mechanize.DefaultFactory.__init__(self) |
| 24 |
|
| 25 |
self._forms_factory = mechanize.FormsFactory(form_parser_class = form_parser) |
| 26 |
|
| 27 |
class Website(object): |
| 28 |
Soup = DuckSoup |
| 29 |
|
| 30 |
def __init__(self, debug, *args, **kwargs): |
| 31 |
self.browser = mechanize.Browser(*args, **kwargs) |
| 32 |
self.browser.addheaders = [('User-agent', 'Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US) AppleWebKit/530.5 (KHTML, like Gecko) Chrome/2.0.172.31 Safari/530.5')] |
| 33 |
|
| 34 |
if debug: |
| 35 |
self.browser.set_debug_http(True) |
| 36 |
|
| 37 |
self.browser.set_handle_equiv(True) |
| 38 |
self.browser.set_handle_redirect(True) |
| 39 |
self.browser.set_handle_refresh(True) |
| 40 |
self.browser.set_handle_robots(False) |
| 41 |
|
| 42 |
self._count = 0 |
| 43 |
|
| 44 |
def _back(self): |
| 45 |
if self._count: |
| 46 |
result = self.browser.back(self._count) |
| 47 |
|
| 48 |
self._count = 0 |
| 49 |
|
| 50 |
return result |
| 51 |
|
| 52 |
def _follow_link(self, *args, **kwargs): |
| 53 |
result = self.browser.follow_link(*args, **kwargs) |
| 54 |
|
| 55 |
self._count += 1 |
| 56 |
|
| 57 |
return result |
| 58 |
|
| 59 |
def _submit(self, *args, **kwargs): |
| 60 |
result = self.browser.submit(*args, **kwargs) |
| 61 |
|
| 62 |
self._count += 1 |
| 63 |
|
| 64 |
return result |
