Add substitute domains for canonical representation (#6)

jace · Mar 5, 2018 · f85fcb0 · f85fcb0
1 parent a773c02
commit f85fcb0
Show file tree

Hide file tree

Showing 2 changed files with 22 additions and 5 deletions.
diff --git a/mxsniff/__init__.py b/mxsniff/__init__.py
@@ -107,7 +107,7 @@ class MXLookupException(Exception):
     pass
 
 
-def canonical_email(email, lowercase=False, strip_periods=False):
+def canonical_email(email, lowercase=False, strip_periods=False, substitute_domains={}):
     """
     Return a canonical representation of an email address to facilitate string
     comparison::
@@ -132,6 +132,9 @@ def canonical_email(email, lowercase=False, strip_periods=False):
         mailbox = mailbox.lower()
     # Example.com --> example.com
     domain = domain.lower()
+    # googlemail.com --> gmail.com
+    if domain in substitute_domains:
+        domain = substitute_domains[domain]
     # example, example.com --> [email protected]
     return '%s@%s' % (mailbox, domain)
 
@@ -415,9 +418,9 @@ def main_internal(args, name='mxsniff'):
     [
     {"canonical": null, "domain": "example.com", "match": ["nomx"], "mx": [], "providers": [], "public": false, "query": "example.com"}
     ]
-    >>> main_internal(['Example <exam.ple@gmail.com>', '-v'])  # doctest: +ELLIPSIS
+    >>> main_internal(['Example <exam.ple+extra@googlemail.com>', '-v'])  # doctest: +ELLIPSIS
     [
-    {"canonical": "[email protected]", "domain": "gmail.com", "match": ["google-gmail"], "mx": [...], "providers": [...], "public": true, "query": "Example <exam.ple@gmail.com>"}
+    {"canonical": "[email protected]", "domain": "googlemail.com", "match": ["google-gmail"], "mx": [...], "providers": [...], "public": true, "query": "Example <exam.ple+extra@googlemail.com>"}
     ]
     """
     import argparse

diff --git a/mxsniff/providers.py b/mxsniff/providers.py
@@ -158,7 +158,13 @@
         'url': 'https://gmail.com/',
         'domains': ['gmail.com', 'googlemail.com'],
         'public': True,
-        'canonical_flags': {'lowercase': True, 'strip_periods': True},
+        'canonical_flags': {
+            'lowercase': True,
+            'strip_periods': True,
+            'substitute_domains': {
+                'googlemail.com': 'gmail.com'
+                }
+            },
         },
     'google-apps': {'mx': [
         'aspmx.l.google.com',
@@ -433,7 +439,15 @@
         'title': "Yahoo Mail",
         'domains': ['rocketmail.com', 'yahoo.com', 'yahoo.co.uk', 'yahoo.co.in', 'ymail.com'],
         'public': True,
-        'canonical_flags': {'lowercase': True},
+        'canonical_flags': {
+            'lowercase': True,
+            'substitute_domains': {
+                'rocketmail.com': 'yahoo.com',
+                'yahoo.co.uk': 'yahoo.com',
+                'yahoo.co.in': 'yahoo.com',
+                'ymail.com': 'yahoo.com',
+                },
+            },
         },
     'yandex': {'mx': [
         'mx.yandex.net',