ripe data: handle whitespace in localparts with multiple addresses

handle a special case with abuse-mailbox:
if there are multiple addresses and one of them has a white space in the
local part, splitting by white space does not work
getaddresses logic does not handle whitespace-separated email addresses
so use our only simple email address pattern to extract the addresses
and make the proper parsing and validation with getaddresses

Author: wagner-intevation

intelmq_certbund_contact/ripe/ripe_data.py

@@ -28,9 +28,17 @@
 import gzip
 import ipaddress
 
+from re import compile as re_compile
+from itertools import chain
+
 from email.utils import getaddresses
 
 
+# Basic regular expression pattern for email addresses. Local part can either be quoted or it must not contain spaces.
+# More reasoning in function sanitize_role_entry and in 2auto/issue354
+RE_EMAIL_ADDRESS = re_compile('((".*?"|[^ ]+)@[^ ]+)')
+
+
 def add_db_args(parser):
     parser.add_argument("--conninfo",
                         default='dbname=contactdb',
@@ -626,12 +634,20 @@ def sanitize_role_entry(entry):
 
     if 'abuse-mailbox' in entry:
         # extract email addresses using email.utils.parseaddr
-        # role contained also multiple addresses in the past, this is no longer the case as od 2025-03-05
-        # getaddresses may return the same email address twice if the address is in display name format, therefore make a list of a set
-        # getaddresses does not detect email addresses separated by a simple whitespace, so replace it with a comma
+        # role contained also multiple addresses in the past, this is no longer the case and forbidden as of 2025-03-05
+        # needs to be handled anyway (2auto/issue354)
+        # getaddresses may return the same email address twice if the address is in display name format, therefore use a set
+        # getaddresses does not detect email addresses separated by a simple whitespace, so use a simple regular expression to first find all strings looking like an email address
+        # the regular expression is necessary to handle special cases like:
+        # abuse-mailbox: "abuse contact"@example.com abuse@example.com
+        # lines with multiple addresses containing white spaces in the email addresses
         # getaddresses may return empty strings, so filter out zero-length strings
         # sort the list of addresses for reproducability and consistency
-        entry['abuse-mailbox'] = list(filter(len, sorted({address for _, address in getaddresses([element.replace(' ', ', ') for element in entry['abuse-mailbox']], strict=False)})))
+        entry['abuse-mailbox'] = [address for address in
+                                  sorted({address for _, address in
+                                          getaddresses(chain.from_iterable([[finding[0] for finding in RE_EMAIL_ADDRESS.findall(element)] for element in entry['abuse-mailbox']]),
+                                                      strict=False)})
+                                  if len(address)]
     return entry
 
 

tests/ripe/role_multiple_addresses.txt.gz

@@ -79,6 +79,8 @@ def test_role_special_characters(self):
         abuse-mailbox:  N&INetworkOperations@homeoffice.gov.uk
         abuse-mailbox:  sdm.n&c@frieslandcampina.com
         abuse-mailbox:  kevino'connor@merseyfire.gov.uk
+        No examples in RIPE data, but allowed: whitespaces in the local part
+        abuse-mailbox:  "foo bar"@example.com
         """
         role_list = ripe_data.parse_file(
             f"{dirname(__file__)}/role_special_characters.txt.gz",
@@ -95,6 +97,11 @@ def test_role_special_characters(self):
                     "role": ["Abuse-C Role"],
                     'org': ['ORG-EFHB1-RIPE'],
                 },
+                {
+                    "abuse-mailbox": ['"abuse contact"@example.com'],
+                    "nic-hdl": ["DUMY1-RIPE"],
+                    "role": ["Non-existing contact with white space"],
+                },
             ],
         )
 
@@ -131,6 +138,9 @@ def test_multiple_addresses(self):
         """
         There are no examples in RIPE data as of 2025-03-05 but it should still be parseable:
         abuse-mailbox:  abuse@example.com abuse@example.net
+        abuse-mailbox:  abuse@example.com, abuse@example.net
+        One address with whitespace in local part and also a second address separated by a white space:
+        abuse-mailbox:  "abuse contact"@example.com abuse@example.com
         """
         role_list = ripe_data.parse_file(
             f"{dirname(__file__)}/role_multiple_addresses.txt.gz",
@@ -151,6 +161,11 @@ def test_multiple_addresses(self):
                     "nic-hdl": ["DUMY2-RIPE"],
                     "role": ["Non-existing contact 2 with comma"],
                 },
+                {
+                    "abuse-mailbox": ['"abuse contact"@example.com', 'abuse@example.com'],
+                    "nic-hdl": ["DUMY3-RIPE"],
+                    "role": ["Non-existing contacts with white-space"],
+                },
             ],
         )