Compare commits
10 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
ab6066eafa | ||
|
|
42258cdd36 | ||
|
|
d3de9e6893 | ||
|
333beb94af | ||
|
|
f3c0942c49 | ||
|
|
02adf53ab9 | ||
|
|
3497b5cab4 | ||
|
|
9c17dca17c | ||
|
|
de342d3177 | ||
|
743b452daf |
2
setup.py
2
setup.py
@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
|
|||||||
|
|
||||||
|
|
||||||
setup(name='talon',
|
setup(name='talon',
|
||||||
version='1.2.3',
|
version='1.2.6',
|
||||||
description=("Mailgun library "
|
description=("Mailgun library "
|
||||||
"to extract message quotations and signatures."),
|
"to extract message quotations and signatures."),
|
||||||
long_description=open("README.rst").read(),
|
long_description=open("README.rst").read(),
|
||||||
|
|||||||
@@ -78,7 +78,7 @@ def delete_quotation_tags(html_note, counter, quotation_checkpoints):
|
|||||||
def cut_gmail_quote(html_message):
|
def cut_gmail_quote(html_message):
|
||||||
''' Cuts the outermost block element with class gmail_quote. '''
|
''' Cuts the outermost block element with class gmail_quote. '''
|
||||||
gmail_quote = html_message.cssselect('div.gmail_quote')
|
gmail_quote = html_message.cssselect('div.gmail_quote')
|
||||||
if gmail_quote and not RE_FWD.match(gmail_quote[0].text):
|
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])
|
gmail_quote[0].getparent().remove(gmail_quote[0])
|
||||||
return True
|
return True
|
||||||
|
|
||||||
@@ -195,3 +195,9 @@ def cut_from_block(html_message):
|
|||||||
block.getparent().remove(block.getnext())
|
block.getparent().remove(block.getnext())
|
||||||
block.getparent().remove(block)
|
block.getparent().remove(block)
|
||||||
return True
|
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),
|
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>:
|
# 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)?'
|
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")
|
parser=html.HTMLParser(encoding="utf-8")
|
||||||
)
|
)
|
||||||
cut_quotations = (html_quotations.cut_gmail_quote(html_tree) or
|
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_blockquote(html_tree) or
|
||||||
html_quotations.cut_microsoft_quote(html_tree) or
|
html_quotations.cut_microsoft_quote(html_tree) or
|
||||||
html_quotations.cut_by_id(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)))
|
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():
|
def test_gmail_quote_blockquote():
|
||||||
msg_body = """Message
|
msg_body = """Message
|
||||||
<blockquote class="gmail_quote">
|
<blockquote class="gmail_quote">
|
||||||
|
|||||||
@@ -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))
|
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():
|
def test_pattern_on_date_wrote_somebody():
|
||||||
eq_('Lorem', quotations.extract_from_plain(
|
eq_('Lorem', quotations.extract_from_plain(
|
||||||
"""Lorem
|
"""Lorem
|
||||||
|
|||||||
Reference in New Issue
Block a user