test_email.py

来自「mallet是自然语言处理、机器学习领域的一个开源项目。」· Python 代码 · 共 1,770 行 · 第 1/5 页

        hstr = """\
from modemcable093.139-201-24.que.mc.videotron.ca ([24.201.139.93] helo=cthulhu.gerg.ca)
\tby kronos.mems-exchange.org with esmtp (Exim 4.05)
\tid 17k4h5-00034i-00
\tfor [e-mail]; Wed, 28 Aug 2002 11:25:20 -0400"""
        h = Header(hstr, maxlinelen=78, header_name='Received',
                   continuation_ws='\t')
        eq(h.encode(), """\
from modemcable093.139-201-24.que.mc.videotron.ca ([24.201.139.93]
\thelo=cthulhu.gerg.ca)
\tby kronos.mems-exchange.org with esmtp (Exim 4.05)
\tid 17k4h5-00034i-00
\tfor [e-mail]; Wed, 28 Aug 2002 11:25:20 -0400""")

    def test_long_8bit_header(self):
        eq = self.ndiffAssertEqual
        msg = Message()
        h = Header('Britische Regierung gibt', 'iso-8859-1',
                    header_name='Subject')
        h.append('gr\xfcnes Licht f\xfcr Offshore-Windkraftprojekte')
        msg['Subject'] = h
        eq(msg.as_string(), """\
Subject: =?iso-8859-1?q?Britische_Regierung_gibt?= =?iso-8859-1?q?gr=FCnes?=
 =?iso-8859-1?q?_Licht_f=FCr_Offshore-Windkraftprojekte?=

""")

    def test_long_8bit_header_no_charset(self):
        eq = self.ndiffAssertEqual
        msg = Message()
        msg['Reply-To'] = 'Britische Regierung gibt gr\xfcnes Licht f\xfcr Offshore-Windkraftprojekte <[e-mail]>'
        eq(msg.as_string(), """\
Reply-To: Britische Regierung gibt gr\xfcnes Licht f\xfcr Offshore-Windkraftprojekte <[e-mail]>

""")

    def test_long_to_header(self):
        eq = self.ndiffAssertEqual
        to = '"Someone Test #A" <[e-mail]>,<[e-mail]>,"Someone Test #B" <[e-mail]>, "Someone Test #C" <[e-mail]>, "Someone Test #D" <[e-mail]>'
        msg = Message()
        msg['To'] = to
        eq(msg.as_string(0), '''\
To: "Someone Test #A" <[e-mail]>, <[e-mail]>,
\t"Someone Test #B" <[e-mail]>,
\t"Someone Test #C" <[e-mail]>,
\t"Someone Test #D" <[e-mail]>

''')

    def test_long_line_after_append(self):
        eq = self.ndiffAssertEqual
        s = 'This is an example of string which has almost the limit of header length.'
        h = Header(s)
        h.append('Add another line.')
        eq(h.encode(), """\
This is an example of string which has almost the limit of header length.
 Add another line.""")

    def test_shorter_line_with_append(self):
        eq = self.ndiffAssertEqual
        s = 'This is a shorter line.'
        h = Header(s)
        h.append('Add another sentence. (Surprise?)')
        eq(h.encode(),
           'This is a shorter line. Add another sentence. (Surprise?)')

    def test_long_field_name(self):
        eq = self.ndiffAssertEqual
        fn = 'X-Very-Very-Very-Long-Header-Name'
        gs = "Die Mieter treten hier ein werden mit einem Foerderband komfortabel den Korridor entlang, an s\xfcdl\xfcndischen Wandgem\xe4lden vorbei, gegen die rotierenden Klingen bef\xf6rdert. "
        h = Header(gs, 'iso-8859-1', header_name=fn)
        # BAW: this seems broken because the first line is too long
        eq(h.encode(), """\
=?iso-8859-1?q?Die_Mieter_treten_hier_?=
 =?iso-8859-1?q?ein_werden_mit_einem_Foerderband_komfortabel_den_Korridor_?=
 =?iso-8859-1?q?entlang=2C_an_s=FCdl=FCndischen_Wandgem=E4lden_vorbei=2C_g?=
 =?iso-8859-1?q?egen_die_rotierenden_Klingen_bef=F6rdert=2E_?=""")

    def test_long_received_header(self):
        h = 'from FOO.TLD (vizworld.acl.foo.tld [123.452.678.9]) by hrothgar.la.mastaler.com (tmda-ofmipd) with ESMTP; Wed, 05 Mar 2003 18:10:18 -0700'
        msg = Message()
        msg['Received-1'] = Header(h, continuation_ws='\t')
        msg['Received-2'] = h
        self.assertEqual(msg.as_string(), """\
Received-1: from FOO.TLD (vizworld.acl.foo.tld [123.452.678.9]) by
\throthgar.la.mastaler.com (tmda-ofmipd) with ESMTP;
\tWed, 05 Mar 2003 18:10:18 -0700
Received-2: from FOO.TLD (vizworld.acl.foo.tld [123.452.678.9]) by
\throthgar.la.mastaler.com (tmda-ofmipd) with ESMTP;
\tWed, 05 Mar 2003 18:10:18 -0700

""")

    def test_string_headerinst_eq(self):
        h = '<[e-mail]> (David Bremner\'s message of "Thu, 6 Mar 2003 13:58:21 +0100")'
        msg = Message()
        msg['Received-1'] = Header(h, header_name='Received-1',
                                   continuation_ws='\t')
        msg['Received-2'] = h
        self.assertEqual(msg.as_string(), """\
Received-1: <[e-mail]>
\t(David Bremner's message of "Thu, 6 Mar 2003 13:58:21 +0100")
Received-2: <[e-mail]>
\t(David Bremner's message of "Thu, 6 Mar 2003 13:58:21 +0100")

""")

    def test_long_unbreakable_lines_with_continuation(self):
        eq = self.ndiffAssertEqual
        msg = Message()
        t = """\
 iVBORw0KGgoAAAANSUhEUgAAADAAAAAwBAMAAAClLOS0AAAAGFBMVEUAAAAkHiJeRUIcGBi9
 locQDQ4zJykFBAXJfWDjAAACYUlEQVR4nF2TQY/jIAyFc6lydlG5x8Nyp1Y69wj1PN2I5gzp"""
        msg['Face-1'] = t
        msg['Face-2'] = Header(t, header_name='Face-2')
        eq(msg.as_string(), """\
Face-1: iVBORw0KGgoAAAANSUhEUgAAADAAAAAwBAMAAAClLOS0AAAAGFBMVEUAAAAkHiJeRUIcGBi9
\tlocQDQ4zJykFBAXJfWDjAAACYUlEQVR4nF2TQY/jIAyFc6lydlG5x8Nyp1Y69wj1PN2I5gzp
Face-2: iVBORw0KGgoAAAANSUhEUgAAADAAAAAwBAMAAAClLOS0AAAAGFBMVEUAAAAkHiJeRUIcGBi9
 locQDQ4zJykFBAXJfWDjAAACYUlEQVR4nF2TQY/jIAyFc6lydlG5x8Nyp1Y69wj1PN2I5gzp

""")

    def test_another_long_multiline_header(self):
        eq = self.ndiffAssertEqual
        m = '''\
Received: from siimage.com ([172.25.1.3]) by zima.siliconimage.com with Microsoft SMTPSVC(5.0.2195.4905);
\tWed, 16 Oct 2002 07:41:11 -0700'''
        msg = email.message_from_string(m)
        eq(msg.as_string(), '''\
Received: from siimage.com ([172.25.1.3]) by zima.siliconimage.com with
\tMicrosoft SMTPSVC(5.0.2195.4905); Wed, 16 Oct 2002 07:41:11 -0700

''')

    def test_long_lines_with_different_header(self):
        eq = self.ndiffAssertEqual
        h = """\
List-Unsubscribe: <https://lists.sourceforge.net/lists/listinfo/spamassassin-talk>,
        <mailto:[e-mail]?subject=unsubscribe>"""
        msg = Message()
        msg['List'] = h
        msg['List'] = Header(h, header_name='List')
        eq(msg.as_string(), """\
List: List-Unsubscribe: <https://lists.sourceforge.net/lists/listinfo/spamassassin-talk>,
\t<mailto:[e-mail]?subject=unsubscribe>
List: List-Unsubscribe: <https://lists.sourceforge.net/lists/listinfo/spamassassin-talk>,
 <mailto:[e-mail]?subject=unsubscribe>

""")



# Test mangling of "From " lines in the body of a message
class TestFromMangling(unittest.TestCase):
    def setUp(self):
        self.msg = Message()
        self.msg['From'] = '[e-mail]'
        self.msg.set_payload("""\
From the desk of A.A.A.:
Blah blah blah
""")

    def test_mangled_from(self):
        s = StringIO()
        g = Generator(s, mangle_from_=True)
        g.flatten(self.msg)
        self.assertEqual(s.getvalue(), """\
From: [e-mail]

>From the desk of A.A.A.:
Blah blah blah
""")

    def test_dont_mangle_from(self):
        s = StringIO()
        g = Generator(s, mangle_from_=False)
        g.flatten(self.msg)
        self.assertEqual(s.getvalue(), """\
From: [e-mail]

From the desk of A.A.A.:
Blah blah blah
""")



# Test the basic MIMEAudio class
class TestMIMEAudio(unittest.TestCase):
    def setUp(self):
        # Make sure we pick up the audiotest.au that lives in email/test/data.
        # In Python, there's an audiotest.au living in Lib/test but that isn't
        # included in some binary distros that don't include the test
        # package.  The trailing empty string on the .join() is significant
        # since findfile() will do a dirname().
        datadir = os.path.join(os.path.dirname(landmark), 'data', '')
        fp = open(findfile('audiotest.au', datadir), 'rb')
        try:
            self._audiodata = fp.read()
        finally:
            fp.close()
        self._au = MIMEAudio(self._audiodata)

    def test_guess_minor_type(self):
        self.assertEqual(self._au.get_type(), 'audio/basic')

    def test_encoding(self):
        payload = self._au.get_payload()
        self.assertEqual(base64.decodestring(payload), self._audiodata)

    def checkSetMinor(self):
        au = MIMEAudio(self._audiodata, 'fish')
        self.assertEqual(im.get_type(), 'audio/fish')

    def test_custom_encoder(self):
        eq = self.assertEqual
        def encoder(msg):
            orig = msg.get_payload()
            msg.set_payload(0)
            msg['Content-Transfer-Encoding'] = 'broken64'
        au = MIMEAudio(self._audiodata, _encoder=encoder)
        eq(au.get_payload(), 0)
        eq(au['content-transfer-encoding'], 'broken64')

    def test_add_header(self):
        eq = self.assertEqual
        unless = self.failUnless
        self._au.add_header('Content-Disposition', 'attachment',
                            filename='audiotest.au')
        eq(self._au['content-disposition'],
           'attachment; filename="audiotest.au"')
        eq(self._au.get_params(header='content-disposition'),
           [('attachment', ''), ('filename', 'audiotest.au')])
        eq(self._au.get_param('filename', header='content-disposition'),
           'audiotest.au')
        missing = []
        eq(self._au.get_param('attachment', header='content-disposition'), '')
        unless(self._au.get_param('foo', failobj=missing,
                                  header='content-disposition') is missing)
        # Try some missing stuff
        unless(self._au.get_param('foobar', missing) is missing)
        unless(self._au.get_param('attachment', missing,
                                  header='foobar') is missing)



# Test the basic MIMEImage class
class TestMIMEImage(unittest.TestCase):
    def setUp(self):
        fp = openfile('PyBanner048.gif')
        try:
            self._imgdata = fp.read()
        finally:
            fp.close()
        self._im = MIMEImage(self._imgdata)

    def test_guess_minor_type(self):
        self.assertEqual(self._im.get_type(), 'image/gif')

    def test_encoding(self):
        payload = self._im.get_payload()
        self.assertEqual(base64.decodestring(payload), self._imgdata)

    def checkSetMinor(self):
        im = MIMEImage(self._imgdata, 'fish')
        self.assertEqual(im.get_type(), 'image/fish')

    def test_custom_encoder(self):
        eq = self.assertEqual
        def encoder(msg):
            orig = msg.get_payload()
            msg.set_payload(0)
            msg['Content-Transfer-Encoding'] = 'broken64'
        im = MIMEImage(self._imgdata, _encoder=encoder)
        eq(im.get_payload(), 0)
        eq(im['content-transfer-encoding'], 'broken64')

    def test_add_header(self):
        eq = self.assertEqual
        unless = self.failUnless
        self._im.add_header('Content-Disposition', 'attachment',
                            filename='dingusfish.gif')
        eq(self._im['content-disposition'],
           'attachment; filename="dingusfish.gif"')
        eq(self._im.get_params(header='content-disposition'),
           [('attachment', ''), ('filename', 'dingusfish.gif')])
        eq(self._im.get_param('filename', header='content-disposition'),
           'dingusfish.gif')
        missing = []
        eq(self._im.get_param('attachment', header='content-disposition'), '')
        unless(self._im.get_param('foo', failobj=missing,
                                  header='content-disposition') is missing)
        # Try some missing stuff
        unless(self._im.get_param('foobar', missing) is missing)
        unless(self._im.get_param('attachment', missing,
                                  header='foobar') is missing)



# Test the basic MIMEText class
class TestMIMEText(unittest.TestCase):
    def setUp(self):
        self._msg = MIMEText('hello there')

    def test_types(self):
        eq = self.assertEqual
        unless = self.failUnless
        eq(self._msg.get_type(), 'text/plain')
        eq(self._msg.get_param('charset'), 'us-ascii')
        missing = []
        unless(self._msg.get_param('foobar', missing) is missing)
        unless(self._msg.get_param('charset', missing, header='foobar')
               is missing)

    def test_payload(self):
        self.assertEqual(self._msg.get_payload(), 'hello there')
        self.failUnless(not self._msg.is_multipart())

    def test_charset(self):
        eq = self.assertEqual
        msg = MIMEText('hello there', _charset='us-ascii')
        eq(msg.get_charset().input_charset, 'us-ascii')
        eq(msg['content-type'], 'text/plain; charset="us-ascii"')



# Test a more complicated multipart/mixed type message
class TestMultipartMixed(TestEmailBase):
    def setUp(self):
        fp = openfile('PyBanner048.gif')
        try:
            data = fp.read()
        finally:
            fp.close()

        container = MIMEBase('multipart', 'mixed', boundary='BOUNDARY')
        image = MIMEImage(data, name='dingusfish.gif')
        image.add_header('content-disposition', 'attachment',
                         filename='dingusfish.gif')
        intro = MIMEText('''\
Hi there,

This is the dingus fish.
''')
        container.attach(intro)
        container.attach(image)
        container['From'] = 'Barry <[e-mail]>'
        container['To'] = 'Dingus Lovers <[e-mail]>'
        container['Subject'] = 'Here is your dingus fish'

        now = 987809702.54848599
        timetuple = time.localtime(now)
        if timetuple[-1] == 0:
            tzsecs = time.timezone