Merge pull request #233 from joe733/workshop

maint: improves `email` module

Author: yozachar

tests/test_email.py

@@ -1,45 +1,54 @@
+"""Test eMail."""
 # -*- coding: utf-8 -*-
+
+# standard
 import pytest
 
+# local
 from validators import email, ValidationFailure
 
 
-@pytest.mark.parametrize(('value', 'whitelist'), [
-    ('email@here.com', None),
-    ('weirder-email@here.and.there.com', None),
-    ('email@[127.0.0.1]', None),
-    ('example@valid-----hyphens.com', None),
-    ('example@valid-with-hyphens.com', None),
-    ('test@domain.with.idn.tld.उदाहरण.परीक्षा', None),
-    ('email@localhost', None),
-    ('email@localdomain', ['localdomain']),
-    ('"test@test"@example.com', None),
-    ('"\\\011"@here.com', None),
-])
-def test_returns_true_on_valid_email(value, whitelist):
-    assert email(value, whitelist=whitelist)
+@pytest.mark.parametrize(
+    ("value",),
+    [
+        ("email@here.com",),
+        ("weirder-email@here.and.there.com",),
+        ("email@127.local.home.arpa",),
+        ("example@valid-----hyphens.com",),
+        ("example@valid-with-hyphens.com",),
+        ("test@domain.with.idn.tld.उदाहरण.परीक्षा",),
+        ("email@localhost.in",),
+        ("email@localdomain.org",),
+        ('"\\\011"@here.com',),
+    ],
+)
+def test_returns_true_on_valid_email(value: str):
+    """Test returns true on valid email."""
+    assert email(value)
 
 
-@pytest.mark.parametrize(('value',), [
-    (None,),
-    ('',),
-    ('abc',),
-    ('abc@',),
-    ('abc@bar',),
-    ('a @x.cz',),
-    ('abc@.com',),
-    ('something@@somewhere.com',),
-    ('email@127.0.0.1',),
-    ('example@invalid-.com',),
-    ('example@-invalid.com',),
-    ('example@inv-.alid-.com',),
-    ('example@inv-.-alid.com',),
-    (
-        'john56789.john56789.john56789.john56789.john56789.john56789.john5'
-        '@example.com',
-    ),
-    # Quoted-string format (CR not allowed)
-    ('"\\\012"@here.com',),
-])
-def test_returns_failed_validation_on_invalid_email(value):
+@pytest.mark.parametrize(
+    ("value",),
+    [
+        (None,),
+        ("",),
+        ("abc",),
+        ("abc@",),
+        ("abc@bar",),
+        ("a @x.cz",),
+        ("abc@.com",),
+        ("something@@somewhere.com",),
+        ("email@127.0.0.1",),
+        ("example@invalid-.com",),
+        ("example@-invalid.com",),
+        ("example@inv-.alid-.com",),
+        ("example@inv-.-alid.com",),
+        ("john56789.john56789.john56789.john56789.john56789.john56789.john5@example.com",),
+        ('"test@test"@example.com',),
+        # Quoted-string format (CR not allowed)
+        ('"\\\012"@here.com',),
+    ],
+)
+def test_returns_failed_validation_on_invalid_email(value: str):
+    """Test returns failed validation on invalid email."""
     assert isinstance(email(value), ValidationFailure)

validators/email.py

@@ -1,75 +1,61 @@
+"""eMail."""
+# -*- coding: utf-8 -*-
+
+# standard
 import re
 
+# local
 from .utils import validator
-
-user_regex = re.compile(
-    # dot-atom
-    r"(^[-!#$%&'*+/=?^_`{}|~0-9A-Z]+"
-    r"(\.[-!#$%&'*+/=?^_`{}|~0-9A-Z]+)*$"
-    # quoted-string
-    r'|^"([\001-\010\013\014\016-\037!#-\[\]-\177]|'
-    r"""\\[\001-\011\013\014\016-\177])*"$)""",
-    re.IGNORECASE
-)
-domain_regex = re.compile(
-    # domain
-    r'(?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+'
-    r'(?:[A-Z]{2,6}\.?|[A-Z0-9-]{2,}\.?$)'
-    # literal form, ipv4 address (SMTP 4.1.3)
-    r'|^\[(25[0-5]|2[0-4]\d|[0-1]?\d?\d)'
-    r'(\.(25[0-5]|2[0-4]\d|[0-1]?\d?\d)){3}\]$',
-    re.IGNORECASE)
-domain_whitelist = ['localhost']
+from .domain import domain
 
 
 @validator
-def email(value, whitelist=None):
-    """
-    Validate an email address.
+def email(value: str, /):
+    """Validate an email address.
 
-    This validator is based on `Django's email validator`_. Returns
-    ``True`` on success and :class:`~validators.utils.ValidationFailure`
-    when validation fails.
+    This was inspired from [Django's email validator][1].
+    Also ref: [RFC 1034][2], [RFC 5321][3] and [RFC 5322][4].
 
-    Examples::
+    [1]: https://github.com/django/django/blob/main/django/core/validators.py#L174
+    [2]: https://www.rfc-editor.org/rfc/rfc1034
+    [3]: https://www.rfc-editor.org/rfc/rfc5321
+    [4]: https://www.rfc-editor.org/rfc/rfc5322
 
+    Examples:
         >>> email('someone@example.com')
-        True
-
+        # Output: True
         >>> email('bogus@@')
-        ValidationFailure(func=email, ...)
-
-    .. _Django's email validator:
-       https://github.com/django/django/blob/master/django/core/validators.py
+        # Output: ValidationFailure(email=email, args={'value': 'bogus@@'})
 
-    .. versionadded:: 0.1
+    Args:
+        value:
+            eMail string to validate.
 
-    :param value: value to validate
-    :param whitelist: domain names to whitelist
+    Returns:
+        (Literal[True]):
+            If `value` is a valid eMail.
+        (ValidationFailure):
+            If `value` is an invalid eMail.
 
-    :copyright: (c) Django Software Foundation and individual contributors.
-    :license: BSD
+    > *New in version 0.1.0*.
     """
-
-    if whitelist is None:
-        whitelist = domain_whitelist
-
-    if not value or '@' not in value:
+    if not value or value.count("@") != 1:
         return False
 
-    user_part, domain_part = value.rsplit('@', 1)
-
-    if not user_regex.match(user_part):
-        return False
+    username_part, domain_part = value.rsplit("@", 1)
 
-    if len(user_part.encode("utf-8")) > 64:
+    if len(username_part) > 64 or len(domain_part) > 253:
+        # ref: RFC 1034 and 5231
         return False
 
-    if domain_part not in whitelist and not domain_regex.match(domain_part):
-        # Try for possible IDN domain-part
-        try:
-            domain_part = domain_part.encode('idna').decode('ascii')
-            return domain_regex.match(domain_part)
-        except UnicodeError:
-            return False
-    return True
+    return (
+        bool(domain(domain_part))
+        if re.compile(
+            # dot-atom
+            r"(^[-!#$%&'*+/=?^_`{}|~0-9A-Z]+(\.[-!#$%&'*+/=?^_`{}|~0-9A-Z]+)*$"
+            # quoted-string
+            + r'|^"([\001-\010\013\014\016-\037!#-\[\]-\177]|\\[\001-\011\013\014\016-\177])*"$)',
+            re.IGNORECASE,
+        ).match(username_part)
+        else False
+    )