diff options
Diffstat (limited to 'python')
| -rw-r--r-- | python/quote_parsing/download_emails.py | 11 |
1 files changed, 3 insertions, 8 deletions
diff --git a/python/quote_parsing/download_emails.py b/python/quote_parsing/download_emails.py index 3a7831c3..c27097e8 100644 --- a/python/quote_parsing/download_emails.py +++ b/python/quote_parsing/download_emails.py @@ -16,15 +16,10 @@ 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 + l.extend([unicodedata.normalize("NFKD", s) for s in pre.strings]) return "\n".join(l) + def print_citi_html(email): soup = BeautifulSoup(email.get_content(), features="lxml") p = soup.find("p") @@ -77,7 +72,7 @@ def save_emails(update=True): except ValueError: logger.error( "Can't parse HTML email with subject: " - "{subject} and id: {msg['id']}" + f"{subject} and id: {msg['id']}" ) continue else: |