import re from bs4 import BeautifulSoup from bs4.element import SoupStrainer from six import iteritems class BS4Parser(object): def __init__(self, *args, **kwargs): # list type param of "feature" arg is not currently correctly tested by bs4 (r353) # so for now, adjust param to provide possible values until the issue is addressed kwargs_new = {} for k, v in iteritems(kwargs): if 'features' in k and isinstance(v, list): v = [item for item in v if item in ['html5lib', 'html.parser', 'html', 'lxml', 'xml']][0] elif 'parse_only' in k: if isinstance(v, dict): (parse_key, filter_dict), = kwargs[k].items() v = SoupStrainer(parse_key, filter_dict) else: v = SoupStrainer(v) elif 'preclean' in k and v: args = (re.sub(r'(?si)(|)', '', args[0]),) + args[1:] continue kwargs_new[k] = v tag, attr = [x in kwargs_new and kwargs_new.pop(x) or y for (x, y) in [('tag', 'table'), ('attr', '')]] if attr: args = (re.sub(r'(?is).*(<%(tag)s[^>]+%(attr)s[^>]*>.*).*' % {'tag': tag, 'attr': attr}, r'\1', args[0]).strip(),) + args[1:] self.soup = BeautifulSoup(*args, **kwargs_new) def __enter__(self): return self.soup def __exit__(self, exc_ty, exc_val, tb): self.soup.clear(True) self.soup = None