Merge pull request #87 from mailgun/sergey/1.2.6

bump up version
2016-04-07 17:54:12 -07:00 · 2016-04-07 17:51:48 -07:00 · 2016-04-07 17:47:38 -07:00 · 2016-04-04 14:22:50 -07:00 · 2016-03-04 13:33:46 -08:00 · 2016-03-04 13:14:50 -08:00
5 changed files with 36 additions and 3 deletions
--- a/setup.py
+++ b/setup.py
@@ -2,7 +2,7 @@ from setuptools import setup, find_packages


 setup(name='talon',
-      version='1.2.3',
+      version='1.2.6',
      description=("Mailgun library "
                   "to extract message quotations and signatures."),
      long_description=open("README.rst").read(),
--- a/talon/html_quotations.py
+++ b/talon/html_quotations.py
@@ -78,7 +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 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])
        return True

@@ -195,3 +195,9 @@ def cut_from_block(html_message):
            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
--- a/talon/quotations.py
+++ b/talon/quotations.py
@@ -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
--- a/tests/html_quotations_test.py
+++ b/tests/html_quotations_test.py
@@ -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 &lt;bob@example.com&gt; 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">
--- a/tests/text_quotations_test.py
+++ b/tests/text_quotations_test.py
@@ -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
Author	SHA1	Message	Date
Sergey Obukhov	ab6066eafa	Merge pull request #87 from mailgun/sergey/1.2.6 bump up version	2016-04-07 17:54:12 -07:00
Sergey Obukhov	42258cdd36	bump up version	2016-04-07 17:51:48 -07:00
Sergey Obukhov	d3de9e6893	Merge pull request #86 from dougkeen/master Fix #85 (exception when stripping gmail quotes)	2016-04-07 17:47:38 -07:00
Doug Keen	333beb94af	Fix #85 (exception when stripping gmail quotes)	2016-04-04 14:22:50 -07:00
Sergey Obukhov	f3c0942c49	Merge pull request #80 from mailgun/sergey/12 fixes mailgun/talon#12	2016-03-04 13:33:46 -08:00
Sergey Obukhov	02adf53ab9	fixes mailgun/talon#12	2016-03-04 13:14:50 -08:00
Sergey Obukhov	3497b5cab4	Merge pull request #79 from mailgun/sergey/version bump version	2016-02-29 15:13:51 -08:00
Sergey Obukhov	9c17dca17c	bump version	2016-02-29 14:50:52 -08:00
Sergey Obukhov	de342d3177	Merge pull request #78 from defkev/master Added Zimbra HTML quotation extraction	2016-02-29 14:14:09 -08:00
defkev	743b452daf	Added Zimbra HTML quotation extraction	2016-02-21 16:56:52 +01:00