new: Test parsing just email header

pymisp/tools/emailobject.py

@@ -103,11 +103,16 @@ def email(self) -> EmailMessage:
     @property
     def attachments(self) -> List[Tuple[str, BytesIO]]:
         to_return = []
-        for attachment in self.email.iter_attachments():
-            content = attachment.get_content()  # type: ignore
-            if isinstance(content, str):
-                content = content.encode()
-            to_return.append((attachment.get_filename(), BytesIO(content)))
+        try:
+            for attachment in self.email.iter_attachments():
+                content = attachment.get_content()  # type: ignore
+                if isinstance(content, str):
+                    content = content.encode()
+                to_return.append((attachment.get_filename(), BytesIO(content)))
+        except AttributeError:
+            # ignore bug in Python3.6, that cause exception for empty email body,
+            # see https://stackoverflow.com/questions/56391306/attributeerror-str-object-has-no-attribute-copy-when-parsing-multipart-emai
+            pass
         return to_return
 
     def __generate_attributes(self):

tests/email_testfiles/mail_1_headers_only.eml

@@ -0,0 +1,28 @@
+Return-Path: <suvorov.s@nalg.ru>
+Delivered-To: kinney@noth.com
+Received: (qmail 11769 invoked from network); 22 Aug 2016 14:23:01 -0000
+Received: from smtprelay0207.b.hostedemail.com (HELO smtprelay.b.hostedemail.com) (64.98.42.207)
+  by smtp.server.net with SMTP; 22 Aug 2016 14:23:01 -0000
+Received: from filter.hostedemail.com (10.5.19.248.rfc1918.com [10.5.19.248])
+	by smtprelay06.b.hostedemail.com (Postfix) with ESMTP id 2CC378D014
+	for <kinney@noth.com>; Mon, 22 Aug 2016 14:22:58 +0000 (UTC)
+Received: from DM6PR06MB4475.namprd06.prod.outlook.com (2603:10b6:207:3d::31)
+ by BL0PR06MB4465.namprd06.prod.outlook.com with HTTPS id 12345 via
+ BL0PR02CA0054.NAMPRD02.PROD.OUTLOOK.COM; Mon, 1 Oct 2018 09:49:22 +0000
+Received: from DM3NAM03FT035.eop-NAM03.prod.protection.outlook.com
+ (2a01:111:f400:7e49::205) by CY4PR0601CA0051.outlook.office365.com
+ (2603:10b6:910:89::28) with Microsoft SMTP Server (version=TLS1_2,
+ cipher=TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384) id 15.20.1185.23 via Frontend
+ Transport; Mon, 1 Oct 2018 09:49:21 +0000
+X-Session-Marker: 6A64617A657940616C6578616E646572736D6974682E636F6D
+X-Spam-Summary: 69,4.5,0,,d41d8cd98f00b204,suvorov.s@nalg.ru,:,RULES_HIT:46:150:152:379:553:871:967:989:1000:1254:1260:1263:1313:1381:1516:1517:1520:1575:1594:1605:1676:1699:1730:1747:1764:1777:1792:1823:2044:2197:2199:2393:2525:2560:2563:2682:2685:2827:2859:2911:2933:2937:2939:2942:2945:2947:2951:2954:3022:3867:3872:3890:3934:3936:3938:3941:3944:3947:3950:3953:3956:3959:4425:5007:6001:6261:6506:6678:6747:6748:7281:7398:7688:8599:8824:8957:9009:9025:9388:10004:10848:11604:11638:11639:11783:11914:12043:12185:12445:12517:12519:12740:13026:14149:14381:14658:14659:14687:21080:21221:30054:30055:30065:30066,0,RBL:none,CacheIP:none,Bayesian:0.5,0.5,0.5,Netcheck:none,DomainCache:0,MSF:not bulk,SPF:fn,MSBL:0,DNSBL:none,Custom_rules:0:0:0,LFtime:5,LUA_SUMMARY:none
+X-HE-Tag: print38_7083d7fd63e24
+X-Filterd-Recvd-Size: 64695
+Received: from computer_3436 (unknown [43.230.105.145])
+	(Authenticated sender: jdazey@alexandersmith.com)
+	by omf06.b.hostedemail.com (Postfix) with ESMTPA
+	for <kinney@noth.com>; Mon, 22 Aug 2016 14:22:52 +0000 (UTC)
+From: =?UTF-8?B?0YHQu9GD0LbQsdCwINCk0J3QoSDQlNCw0L3QuNC40Lsg0KHRg9Cy0L7RgNC+0LI=?= <suvorov.s@nalg.ru>
+To: kinney@noth.com
+Subject: =?UTF-8?B?0L/QuNGB0YzQvNC+INGD0LLQtdC00L7QvC3QtQ==?=
+Content-Type: multipart/mixed; boundary="2NqJR3m2cLnhEraiqXA4Q9hqnmihx7b7"

tests/test_emailobject.py

@@ -13,6 +13,7 @@ def test_mail_1(self):
         self.assertEqual(self._get_values(email_object, "to")[0], "kinney@noth.com")
         self.assertEqual(self._get_values(email_object, "from")[0], "suvorov.s@nalg.ru")
         self.assertEqual(self._get_values(email_object, "from-display-name")[0], "служба ФНС Даниил Суворов")
+        self.assertEqual(len(self._get_values(email_object, "email-body")), 1)
 
         self.assertEqual(self._get_values(email_object, "received-header-ip")[0], "43.230.105.145")
         self.assertEqual(self._get_values(email_object, "received-header-ip")[1], "2a01:111:f400:7e49::205")
@@ -22,6 +23,17 @@ def test_mail_1(self):
             self.assertIsInstance(file_name, str)
             self.assertIsInstance(file_content, BytesIO)
 
+    def test_mail_1_headers_only(self):
+        email_object = EMailObject(Path("tests/email_testfiles/mail_1_headers_only.eml"))
+        self.assertEqual(self._get_values(email_object, "subject")[0], "письмо уведом-е")
+        self.assertEqual(self._get_values(email_object, "to")[0], "kinney@noth.com")
+        self.assertEqual(self._get_values(email_object, "from")[0], "suvorov.s@nalg.ru")
+
+        self.assertEqual(len(self._get_values(email_object, "email-body")), 0)
+
+        self.assertIsInstance(email_object.email, EmailMessage)
+        self.assertEqual(len(email_object.attachments), 0)
+
     def test_mail_1_msg(self):
         email_object = EMailObject(Path("tests/email_testfiles/mail_1.msg"))
         self.assertEqual(self._get_values(email_object, "subject")[0],