diff options
Diffstat (limited to 'python')
| -rw-r--r-- | python/quote_parsing/download_emails.py | 14 |
1 files changed, 14 insertions, 0 deletions
diff --git a/python/quote_parsing/download_emails.py b/python/quote_parsing/download_emails.py index 269d4dd6..3a7831c3 100644 --- a/python/quote_parsing/download_emails.py +++ b/python/quote_parsing/download_emails.py @@ -13,9 +13,23 @@ from gmail_helpers import GmailMessage from email.utils import parsedate_to_datetime +def print_citi_html2(soup): + l = [] + for pre in soup.findAll("pre"): + for br in p.findAll("br"): + if isinstance(s, NavigableString): + l.append(unicodedata.normalize("NFKD", s)) + elif isinstance(s, Tag) and s.name == "br": + l.append("\n") + else: + raise ValueError + return "\n".join(l) + def print_citi_html(email): soup = BeautifulSoup(email.get_content(), features="lxml") p = soup.find("p") + if p is None: + return print_citi_html2(soup) s = p.next if isinstance(s, NavigableString): l = [unicodedata.normalize("NFKD", s)] |
