zev/imbox
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix inline body parsing

Browse Source
This commit is contained in:
Andrey Mozgunov
2017-09-26 15:51:34 +03:00
parent 02cbf574dd
commit ed86228e86
2 changed files with 63 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
imbox/parser.py
View File

@@ -149,7 +149,7 @@ def parse_email(raw_email, policy=None):
for part in email_message.walk():
content_type = part.get_content_type()
part_maintype = part.get_content_maintype()
content_disposition = part.get('Content-Disposition', None)
content_disposition = part.get_content_disposition()
if content_disposition or not part_maintype == "text":
content = part.get_payload(decode=True)
else:

62
tests/parser_tests.py
View File

@@ -82,6 +82,62 @@ Content-Transfer-Encoding: quoted-printable
"""
raw_email_encoded_another_bad_multipart = b"""Delivered-To: receiver@example.com
Return-Path: <sender@example.com>
Mime-Version: 1.0
Date: Wed, 22 Mar 2017 15:21:55 -0500
Message-ID: <58D29693.192A.0075.1@wimort.com>
Subject: Re: Reaching Out About Peoples Home Equity
From: sender@example.com
To: receiver@example.com
Content-Type: multipart/alternative; boundary="____NOIBTUQXSYRVOOAFLCHY____"
--____NOIBTUQXSYRVOOAFLCHY____
Content-Type: text/plain; charset=iso-8859-15
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline;
modification-date="Wed, 22 Mar 2017 15:21:55 -0500"
Chloe,
--____NOIBTUQXSYRVOOAFLCHY____
Content-Type: multipart/related; boundary="____XTSWHCFJMONXSVGPVDLY____"
--____XTSWHCFJMONXSVGPVDLY____
Content-Type: text/html; charset=iso-8859-15
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline;
modification-date="Wed, 22 Mar 2017 15:21:55 -0500"
<HTML xmlns=3D"http://www.w3.org/1999/xhtml">
<BODY style=3D"COLOR: black; FONT: 10pt Segoe UI; MARGIN: 4px 4px 1px" =
leftMargin=3D0 topMargin=3D0 offset=3D"0" marginwidth=3D"0" marginheight=3D=
"0">
<DIV>Chloe,</DIV>
<IMG src=3D"cid:VFXVGHA=
GXNMI.36b3148cbf284ba18d35bdd8386ac266" width=3D1 height=3D1> </BODY></HTML=
>
--____XTSWHCFJMONXSVGPVDLY____
Content-ID: <TLUACRGXVUBY.IMAGE_3.gif>
Content-Type: image/gif
Content-Transfer-Encoding: base64
R0lGODlhHgHCAPf/AIOPr9GvT7SFcZZjVTEuMLS1tZKUlJN0Znp4eEA7PV1aWvz8+8V6Zl1BNYxX
HvOZ1/zmOd95agUEADs=
--____XTSWHCFJMONXSVGPVDLY____
Content-ID: <VFXVGHAGXNMI.36b3148cbf284ba18d35bdd8386ac266>
Content-Type: image/xxx
Content-Transfer-Encoding: base64
R0lGODlhAQABAPAAAAAAAAAAACH5BAEAAAAALAAAAAABAAEAAAICRAEAOw==
--____XTSWHCFJMONXSVGPVDLY____--
--____NOIBTUQXSYRVOOAFLCHY____--
"""
class TestParser(unittest.TestCase):
def test_parse_email(self):
@@ -98,6 +154,12 @@ class TestParser(unittest.TestCase):
self.assertEqual('Выписка по карте', parsed_email.subject)
self.assertEqual('Выписка по карте 1234', parsed_email.body['html'][0])
def test_parse_email_inline_body(self):
parsed_email = parse_email(raw_email_encoded_another_bad_multipart)
self.assertEqual("Re: Reaching Out About Peoples Home Equity", parsed_email.subject)
self.assertTrue(parsed_email.body['plain'])
self.assertTrue(parsed_email.body['html'])
def test_parse_email_ignores_header_casing(self):
self.assertEqual('one', parse_email('Message-ID: one').message_id)
self.assertEqual('one', parse_email('Message-Id: one').message_id)