Implemented lazy regex compiling in Field classes to improve 'import mongoengine' performance

mongoengine/base/utils.py

@@ -0,0 +1,22 @@
+import re
+
+
+class LazyRegexCompiler(object):
+    """Descriptor to allow lazy compilation of regex"""
+
+    def __init__(self, pattern, flags=0):
+        self._pattern = pattern
+        self._flags = flags
+        self._compiled_regex = None
+
+    @property
+    def compiled_regex(self):
+        if self._compiled_regex is None:
+            self._compiled_regex = re.compile(self._pattern, self._flags)
+        return self._compiled_regex
+
+    def __get__(self, obj, objtype):
+        return self.compiled_regex
+
+    def __set__(self, instance, value):
+        raise AttributeError("Can not set attribute LazyRegexCompiler")

mongoengine/common.py

@@ -56,3 +56,4 @@ def _import_class(cls_name):
         _class_registry_cache[cls] = getattr(module, cls)
 
     return _class_registry_cache.get(cls_name)
+

mongoengine/fields.py

@@ -5,7 +5,6 @@ import re
 import socket
 import time
 import uuid
-import warnings
 from operator import itemgetter
 
 from bson import Binary, DBRef, ObjectId, SON
@@ -28,6 +27,7 @@ except ImportError:
 from mongoengine.base import (BaseDocument, BaseField, ComplexBaseField,
                               GeoJsonBaseField, LazyReference, ObjectIdField,
                               get_document)
+from mongoengine.base.utils import LazyRegexCompiler
 from mongoengine.common import _import_class
 from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
 from mongoengine.document import Document, EmbeddedDocument
@@ -123,7 +123,7 @@ class URLField(StringField):
     .. versionadded:: 0.3
     """
 
-    _URL_REGEX = re.compile(
+    _URL_REGEX = LazyRegexCompiler(
         r'^(?:[a-z0-9\.\-]*)://'  # scheme is validated separately
         r'(?:(?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+(?:[A-Z]{2,6}\.?|[A-Z0-9-]{2,}(?<!-)\.?)|'  # domain...
         r'localhost|'  # localhost...
@@ -157,7 +157,7 @@ class EmailField(StringField):
 
     .. versionadded:: 0.4
     """
-    USER_REGEX = re.compile(
+    USER_REGEX = LazyRegexCompiler(
         # `dot-atom` defined in RFC 5322 Section 3.2.3.
         r"(^[-!#$%&'*+/=?^_`{}|~0-9A-Z]+(\.[-!#$%&'*+/=?^_`{}|~0-9A-Z]+)*\Z"
         # `quoted-string` defined in RFC 5322 Section 3.2.4.
@@ -165,7 +165,7 @@ class EmailField(StringField):
         re.IGNORECASE
     )
 
-    UTF8_USER_REGEX = re.compile(
+    UTF8_USER_REGEX = LazyRegexCompiler(
         six.u(
             # RFC 6531 Section 3.3 extends `atext` (used by dot-atom) to
             # include `UTF8-non-ascii`.
@@ -175,7 +175,7 @@ class EmailField(StringField):
         ), re.IGNORECASE | re.UNICODE
     )
 
-    DOMAIN_REGEX = re.compile(
+    DOMAIN_REGEX = LazyRegexCompiler(
         r'((?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+)(?:[A-Z0-9-]{2,63}(?<!-))\Z',
         re.IGNORECASE
     )

mongoengine/python_support.py

@@ -6,11 +6,7 @@ import pymongo
 import six
 
 
-if pymongo.version_tuple[0] < 3:
+IS_PYMONGO_3 = pymongo.version_tuple[0] >= 3
-    IS_PYMONGO_3 = False
-else:
-    IS_PYMONGO_3 = True
-
 
 # six.BytesIO resolves to StringIO.StringIO in Py2 and io.BytesIO in Py3.
 StringIO = six.BytesIO

tests/test_utils.py

@@ -0,0 +1,38 @@
+import unittest
+import re
+
+from mongoengine.base.utils import LazyRegexCompiler
+
+signal_output = []
+
+
+class LazyRegexCompilerTest(unittest.TestCase):
+
+    def test_lazy_regex_compiler_verify_laziness_of_descriptor(self):
+        class UserEmail(object):
+            EMAIL_REGEX = LazyRegexCompiler('@', flags=32)
+
+        descriptor = UserEmail.__dict__['EMAIL_REGEX']
+        self.assertIsNone(descriptor._compiled_regex)
+
+        regex = UserEmail.EMAIL_REGEX
+        self.assertEqual(regex, re.compile('@', flags=32))
+        self.assertEqual(regex.search('user@domain.com').group(), '@')
+
+        user_email = UserEmail()
+        self.assertIs(user_email.EMAIL_REGEX, UserEmail.EMAIL_REGEX)
+
+    def test_lazy_regex_compiler_verify_cannot_set_descriptor_on_instance(self):
+        class UserEmail(object):
+            EMAIL_REGEX = LazyRegexCompiler('@')
+
+        user_email = UserEmail()
+        with self.assertRaises(AttributeError):
+            user_email.EMAIL_REGEX = re.compile('@')
+
+    def test_lazy_regex_compiler_verify_can_override_class_attr(self):
+        class UserEmail(object):
+            EMAIL_REGEX = LazyRegexCompiler('@')
+
+        UserEmail.EMAIL_REGEX = re.compile('cookies')
+        self.assertEqual(UserEmail.EMAIL_REGEX.search('Cake & cookies').group(), 'cookies')