Compare commits
12 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
ab6066eafa | ||
|
|
42258cdd36 | ||
|
|
d3de9e6893 | ||
|
333beb94af | ||
|
|
f3c0942c49 | ||
|
|
02adf53ab9 | ||
|
|
3497b5cab4 | ||
|
|
9c17dca17c | ||
|
|
de342d3177 | ||
|
743b452daf | ||
|
|
c762f3c337 | ||
|
|
31803d41bc |
2
setup.py
2
setup.py
@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
|
||||
|
||||
|
||||
setup(name='talon',
|
||||
version='1.2.2',
|
||||
version='1.2.6',
|
||||
description=("Mailgun library "
|
||||
"to extract message quotations and signatures."),
|
||||
long_description=open("README.rst").read(),
|
||||
|
||||
@@ -12,6 +12,7 @@ CHECKPOINT_PATTERN = re.compile(CHECKPOINT_PREFIX + '\d+' + CHECKPOINT_SUFFIX)
|
||||
|
||||
# HTML quote indicators (tag ids)
|
||||
QUOTE_IDS = ['OLK_SRC_BODY_SECTION']
|
||||
RE_FWD = re.compile("^[-]+[ ]*Forwarded message[ ]*[-]+$", re.I | re.M)
|
||||
|
||||
|
||||
def add_checkpoint(html_note, counter):
|
||||
@@ -77,7 +78,7 @@ def delete_quotation_tags(html_note, counter, quotation_checkpoints):
|
||||
def cut_gmail_quote(html_message):
|
||||
''' Cuts the outermost block element with class gmail_quote. '''
|
||||
gmail_quote = html_message.cssselect('div.gmail_quote')
|
||||
if gmail_quote:
|
||||
if gmail_quote and (gmail_quote[0].text is None or not RE_FWD.match(gmail_quote[0].text)):
|
||||
gmail_quote[0].getparent().remove(gmail_quote[0])
|
||||
return True
|
||||
|
||||
@@ -172,6 +173,7 @@ def cut_from_block(html_message):
|
||||
parent_div_is_all_content = (
|
||||
maybe_body is not None and maybe_body.tag == 'body' and
|
||||
len(maybe_body.getchildren()) == 1)
|
||||
|
||||
if not parent_div_is_all_content:
|
||||
block.getparent().remove(block)
|
||||
return True
|
||||
@@ -185,7 +187,17 @@ def cut_from_block(html_message):
|
||||
"//*[starts-with(mg:tail(), 'Date:')]"))
|
||||
if block:
|
||||
block = block[0]
|
||||
|
||||
if RE_FWD.match(block.getparent().text or ''):
|
||||
return False
|
||||
|
||||
while(block.getnext() is not None):
|
||||
block.getparent().remove(block.getnext())
|
||||
block.getparent().remove(block)
|
||||
return True
|
||||
|
||||
def cut_zimbra_quote(html_message):
|
||||
zDivider = html_message.xpath('//hr[@data-marker="__DIVIDER__"]')
|
||||
if zDivider:
|
||||
zDivider[0].getparent().remove(zDivider[0])
|
||||
return True
|
||||
@@ -148,7 +148,9 @@ SPLITTER_PATTERNS = [
|
||||
re.compile("\d{4}-\d{2}-\d{2}\s+\d{2}:\d{2}\s+GMT.*@", re.S),
|
||||
# Thu, 26 Jun 2014 14:00:51 +0400 Bob <bob@example.com>:
|
||||
re.compile('\S{3,10}, \d\d? \S{3,10} 20\d\d,? \d\d?:\d\d(:\d\d)?'
|
||||
'( \S+){3,6}@\S+:')
|
||||
'( \S+){3,6}@\S+:'),
|
||||
# Sent from Samsung MobileName <address@example.com> wrote:
|
||||
re.compile('Sent from Samsung .*@.*> wrote')
|
||||
]
|
||||
|
||||
|
||||
@@ -350,6 +352,7 @@ def extract_from_html(msg_body):
|
||||
parser=html.HTMLParser(encoding="utf-8")
|
||||
)
|
||||
cut_quotations = (html_quotations.cut_gmail_quote(html_tree) or
|
||||
html_quotations.cut_zimbra_quote(html_tree) or
|
||||
html_quotations.cut_blockquote(html_tree) or
|
||||
html_quotations.cut_microsoft_quote(html_tree) or
|
||||
html_quotations.cut_by_id(html_tree) or
|
||||
|
||||
@@ -131,6 +131,17 @@ def test_gmail_quote():
|
||||
RE_WHITESPACE.sub('', quotations.extract_from_html(msg_body)))
|
||||
|
||||
|
||||
def test_gmail_quote_compact():
|
||||
msg_body = 'Reply' \
|
||||
'<div class="gmail_quote">' \
|
||||
'<div class="gmail_quote">On 11-Apr-2011, at 6:54 PM, Bob <bob@example.com> wrote:' \
|
||||
'<div>Test</div>' \
|
||||
'</div>' \
|
||||
'</div>'
|
||||
eq_("<html><body><p>Reply</p></body></html>",
|
||||
RE_WHITESPACE.sub('', quotations.extract_from_html(msg_body)))
|
||||
|
||||
|
||||
def test_gmail_quote_blockquote():
|
||||
msg_body = """Message
|
||||
<blockquote class="gmail_quote">
|
||||
@@ -340,3 +351,10 @@ def test_CRLF():
|
||||
assert_false(symbol in extracted)
|
||||
eq_("<html><body><p>Reply</p></body></html>",
|
||||
RE_WHITESPACE.sub('', extracted))
|
||||
|
||||
|
||||
def test_gmail_forwarded_msg():
|
||||
msg_body = """<div dir="ltr"><br><div class="gmail_quote">---------- Forwarded message ----------<br>From: <b class="gmail_sendername">Bob</b> <span dir="ltr"><<a href="mailto:bob@example.com">bob@example.com</a>></span><br>Date: Fri, Feb 11, 2010 at 5:59 PM<br>Subject: Bob WFH today<br>To: Mary <<a href="mailto:mary@example.com">mary@example.com</a>><br><br><br><div dir="ltr">eom</div>
|
||||
</div><br></div>"""
|
||||
extracted = quotations.extract_from_html(msg_body)
|
||||
eq_(RE_WHITESPACE.sub('', msg_body), RE_WHITESPACE.sub('', extracted))
|
||||
|
||||
@@ -32,6 +32,19 @@ On 11-Apr-2011, at 6:54 PM, Roman Tkachenko <romant@example.com> wrote:
|
||||
eq_("Test reply", quotations.extract_from_plain(msg_body))
|
||||
|
||||
|
||||
def test_pattern_sent_from_samsung_smb_wrote():
|
||||
msg_body = """Test reply
|
||||
|
||||
Sent from Samsung MobileName <address@example.com> wrote:
|
||||
|
||||
>
|
||||
> Test
|
||||
>
|
||||
> Roman"""
|
||||
|
||||
eq_("Test reply", quotations.extract_from_plain(msg_body))
|
||||
|
||||
|
||||
def test_pattern_on_date_wrote_somebody():
|
||||
eq_('Lorem', quotations.extract_from_plain(
|
||||
"""Lorem
|
||||
|
||||
Reference in New Issue
Block a user