fix connecting to MongoReplicaSetClient

Fix BaseQuerySet#sum and BaseQuerySet#average for fields that specify a db_field

.gitignore

@@ -14,4 +14,6 @@ env/
 .project
 .pydevproject
 tests/test_bugfix.py
-htmlcov/
+htmlcov/
+venv
+venv3

.landscape.yml

@@ -0,0 +1,22 @@
+pylint:
+    disable:
+        # We use this a lot (e.g. via document._meta)
+        - protected-access
+
+    options:
+        additional-builtins:
+            # add xrange and long as valid built-ins. In Python 3, xrange is
+            # translated into range and long is translated into int via 2to3 (see
+            # "use_2to3" in setup.py). This should be removed when we drop Python
+            # 2 support (which probably won't happen any time soon).
+            - xrange
+            - long
+
+pyflakes:
+    disable:
+        # undefined variables are already covered by pylint (and exclude
+        # xrange & long)
+        - F821
+
+ignore-paths:
+    - benchmark.py

.travis.yml

@@ -1,7 +1,6 @@
 language: python
 
 python:
-- '2.6'
 - '2.7'
 - '3.3'
 - '3.4'
@@ -43,7 +42,11 @@ before_script:
 script:
 - tox -e $(echo py$TRAVIS_PYTHON_VERSION-mg$PYMONGO | tr -d . | sed -e 's/pypypy/pypy/') -- --with-coverage
 
-after_script: coveralls --verbose
+# For now only submit coveralls for Python v2.7. Python v3.x currently shows
+# 0% coverage. That's caused by 'use_2to3', which builds the py3-compatible
+# code in a separate dir and runs tests on that.
+after_script:
+- if [[ $TRAVIS_PYTHON_VERSION == '2.7' ]]; then coveralls --verbose; fi
 
 notifications:
   irc: irc.freenode.org#mongoengine

AUTHORS

@@ -242,3 +242,4 @@ that much better:
  * xiaost7 (https://github.com/xiaost7)
  * Victor Varvaryuk
  * Stanislav Kaledin (https://github.com/sallyruthstruik)
+ * Dmitry Yantsen (https://github.com/mrTable)

README.rst

@@ -4,7 +4,7 @@ MongoEngine
 :Info: MongoEngine is an ORM-like layer on top of PyMongo.
 :Repository: https://github.com/MongoEngine/mongoengine
 :Author: Harry Marr (http://github.com/hmarr)
-:Maintainer: Ross Lawley (http://github.com/rozza)
+:Maintainer: Stefan Wójcik (http://github.com/wojcikstefan)
 
 .. image:: https://travis-ci.org/MongoEngine/mongoengine.svg?branch=master
   :target: https://travis-ci.org/MongoEngine/mongoengine
@@ -52,10 +52,14 @@ Some simple examples of what MongoEngine code looks like:
 
 .. code :: python
 
+    from mongoengine import *
+    connect('mydb')
+
     class BlogPost(Document):
         title = StringField(required=True, max_length=200)
         posted = DateTimeField(default=datetime.datetime.now)
         tags = ListField(StringField(max_length=50))
+        meta = {'allow_inheritance': True}
 
     class TextPost(BlogPost):
         content = StringField(required=True)

benchmark.py

@@ -1,118 +1,41 @@
 #!/usr/bin/env python
 
+"""
+Simple benchmark comparing PyMongo and MongoEngine.
+
+Sample run on a mid 2015 MacBook Pro (commit b282511):
+
+Benchmarking...
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries - Pymongo
+2.58979988098
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries - Pymongo write_concern={"w": 0}
+1.26657605171
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries - MongoEngine
+8.4351580143
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries without continual assign - MongoEngine
+7.20191693306
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries - MongoEngine - write_concern={"w": 0}, cascade = True
+6.31104588509
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries - MongoEngine, write_concern={"w": 0}, validate=False, cascade=True
+6.07083487511
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries - MongoEngine, write_concern={"w": 0}, validate=False
+5.97704291344
+----------------------------------------------------------------------------------------------------
+Creating 10000 dictionaries - MongoEngine, force_insert=True, write_concern={"w": 0}, validate=False
+5.9111430645
+"""
+
 import timeit
 
 
-def cprofile_main():
-    from pymongo import Connection
-    connection = Connection()
-    connection.drop_database('timeit_test')
-    connection.disconnect()
-
-    from mongoengine import Document, DictField, connect
-    connect("timeit_test")
-
-    class Noddy(Document):
-        fields = DictField()
-
-    for i in range(1):
-        noddy = Noddy()
-        for j in range(20):
-            noddy.fields["key" + str(j)] = "value " + str(j)
-        noddy.save()
-
-
 def main():
-    """
-    0.4 Performance Figures ...
-
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - Pymongo
-    3.86744189262
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine
-    6.23374891281
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False
-    5.33027005196
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False, cascade=False
-    pass - No Cascade
-
-    0.5.X
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - Pymongo
-    3.89597702026
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine
-    21.7735359669
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False
-    19.8670389652
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False, cascade=False
-    pass - No Cascade
-
-    0.6.X
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - Pymongo
-    3.81559205055
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine
-    10.0446798801
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False
-    9.51354718208
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False, cascade=False
-    9.02567505836
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, force=True
-    8.44933390617
-
-    0.7.X
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - Pymongo
-    3.78801012039
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine
-    9.73050498962
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False
-    8.33456707001
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, safe=False, validate=False, cascade=False
-    8.37778115273
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, force=True
-    8.36906409264
-    0.8.X
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - Pymongo
-    3.69964408875
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - Pymongo write_concern={"w": 0}
-    3.5526599884
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine
-    7.00959801674
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries without continual assign - MongoEngine
-    5.60943293571
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine - write_concern={"w": 0}, cascade=True
-    6.715102911
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, write_concern={"w": 0}, validate=False, cascade=True
-    5.50644683838
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, write_concern={"w": 0}, validate=False
-    4.69851183891
-    ----------------------------------------------------------------------------------------------------
-    Creating 10000 dictionaries - MongoEngine, force_insert=True, write_concern={"w": 0}, validate=False
-    4.68946313858
-    ----------------------------------------------------------------------------------------------------
-    """
     print("Benchmarking...")
 
     setup = """
@@ -131,7 +54,7 @@ noddy = db.noddy
 for i in range(10000):
     example = {'fields': {}}
     for j in range(20):
-        example['fields']["key"+str(j)] = "value "+str(j)
+        example['fields']['key' + str(j)] = 'value ' + str(j)
 
     noddy.save(example)
 
@@ -146,9 +69,10 @@ myNoddys = noddy.find()
 
     stmt = """
 from pymongo import MongoClient
+from pymongo.write_concern import WriteConcern
 connection = MongoClient()
 
-db = connection.timeit_test
+db = connection.get_database('timeit_test', write_concern=WriteConcern(w=0))
 noddy = db.noddy
 
 for i in range(10000):
@@ -156,7 +80,7 @@ for i in range(10000):
     for j in range(20):
         example['fields']["key"+str(j)] = "value "+str(j)
 
-    noddy.save(example, write_concern={"w": 0})
+    noddy.save(example)
 
 myNoddys = noddy.find()
 [n for n in myNoddys] # iterate
@@ -171,10 +95,10 @@ myNoddys = noddy.find()
 from pymongo import MongoClient
 connection = MongoClient()
 connection.drop_database('timeit_test')
-connection.disconnect()
+connection.close()
 
 from mongoengine import Document, DictField, connect
-connect("timeit_test")
+connect('timeit_test')
 
 class Noddy(Document):
     fields = DictField()

docs/changelog.rst

@@ -2,16 +2,38 @@
 Changelog
 =========
 
+Development
+===========
+- (Fill this out as you fix issues and develop you features).
+
+Changes in 0.11.0
+=================
+- BREAKING CHANGE: Renamed `ConnectionError` to `MongoEngineConnectionError` since the former is a built-in exception name in Python v3.x. #1428
+- BREAKING CHANGE: Dropped Python 2.6 support. #1428
+- BREAKING CHANGE: `from mongoengine.base import ErrorClass` won't work anymore for any error from `mongoengine.errors` (e.g. `ValidationError`). Use `from mongoengine.errors import ErrorClass instead`. #1428
+- Fixed absent rounding for DecimalField when `force_string` is set. #1103
+
 Changes in 0.10.8
 =================
-- Fill this in as PRs for v0.10.8 are merged
+- Added support for QuerySet.batch_size (#1426)
+- Fixed query set iteration within iteration #1427
+- Fixed an issue where specifying a MongoDB URI host would override more information than it should #1421
+- Added ability to filter the generic reference field by ObjectId and DBRef #1425
+- Fixed delete cascade for models with a custom primary key field #1247
+- Added ability to specify an authentication mechanism (e.g. X.509) #1333
+- Added support for falsey primary keys (e.g. doc.pk = 0) #1354
+- Fixed QuerySet#sum/average for fields w/ explicit db_field #1417
+- Fixed filtering by embedded_doc=None #1422
+- Added support for cursor.comment #1420
+- Fixed doc.get_<field>_display #1419
+- Fixed __repr__ method of the StrictDict #1424
+- Added a deprecation warning for Python 2.6
 
 Changes in 0.10.7
 =================
 - Dropped Python 3.2 support #1390
 - Fixed the bug where dynamic doc has index inside a dict field #1278
 - Fixed: ListField minus index assignment does not work #1128
-- Fixed not being able to specify `use_db_field=False` on `ListField(EmbeddedDocumentField)` instances
 - Fixed cascade delete mixing among collections #1224
 - Add `signal_kwargs` argument to `Document.save`, `Document.delete` and `BaseQuerySet.insert` to be passed to signals calls #1206
 - Raise `OperationError` when trying to do a `drop_collection` on document with no collection set.
@@ -28,7 +50,7 @@ Changes in 0.10.7
 - Fixed connecting to a list of hosts #1389
 - Fixed a bug where accessing broken references wouldn't raise a DoesNotExist error #1334
 - Fixed not being able to specify use_db_field=False on ListField(EmbeddedDocumentField) instances #1218
-- Improvements to the dictionary fields docs # 1383
+- Improvements to the dictionary fields docs #1383
 
 Changes in 0.10.6
 =================

docs/upgrade.rst

@@ -2,6 +2,39 @@
 Upgrading
 #########
 
+0.11.0
+******
+This release includes a major rehaul of MongoEngine's code quality and
+introduces a few breaking changes. It also touches many different parts of
+the package and although all the changes have been tested and scrutinized,
+you're encouraged to thorougly test the upgrade.
+
+First breaking change involves renaming `ConnectionError` to `MongoEngineConnectionError`.
+If you import or catch this exception, you'll need to rename it in your code.
+
+Second breaking change drops Python v2.6 support. If you run MongoEngine on
+that Python version, you'll need to upgrade it first.
+
+Third breaking change drops an old backward compatibility measure where
+`from mongoengine.base import ErrorClass` would work on top of
+`from mongoengine.errors import ErrorClass` (where `ErrorClass` is e.g.
+`ValidationError`). If you import any exceptions from `mongoengine.base`,
+change it to `mongoengine.errors`.
+
+0.10.8
+******
+This version fixed an issue where specifying a MongoDB URI host would override
+more information than it should. These changes are minor, but they still
+subtly modify the connection logic and thus you're encouraged to test your
+MongoDB connection before shipping v0.10.8 in production.
+
+0.10.7
+******
+
+`QuerySet.aggregate_sum` and `QuerySet.aggregate_average` are dropped. Use
+`QuerySet.sum` and `QuerySet.average` instead which use the aggreation framework
+by default from now on.
+
 0.9.0
 *****
 

mongoengine/__init__.py

@@ -1,25 +1,35 @@
-import connection
-from connection import *
-import document
-from document import *
-import errors
-from errors import *
-import fields
-from fields import *
-import queryset
-from queryset import *
-import signals
-from signals import *
+# Import submodules so that we can expose their __all__
+from mongoengine import connection
+from mongoengine import document
+from mongoengine import errors
+from mongoengine import fields
+from mongoengine import queryset
+from mongoengine import signals
 
-__all__ = (list(document.__all__) + fields.__all__ + connection.__all__ +
-           list(queryset.__all__) + signals.__all__ + list(errors.__all__))
+# Import everything from each submodule so that it can be accessed via
+# mongoengine, e.g. instead of `from mongoengine.connection import connect`,
+# users can simply use `from mongoengine import connect`, or even
+# `from mongoengine import *` and then `connect('testdb')`.
+from mongoengine.connection import *
+from mongoengine.document import *
+from mongoengine.errors import *
+from mongoengine.fields import *
+from mongoengine.queryset import *
+from mongoengine.signals import *
 
-VERSION = (0, 10, 7)
+
+__all__ = (list(document.__all__) + list(fields.__all__) +
+           list(connection.__all__) + list(queryset.__all__) +
+           list(signals.__all__) + list(errors.__all__))
+
+
+VERSION = (0, 11, 0)
 
 
 def get_version():
-    if isinstance(VERSION[-1], basestring):
-        return '.'.join(map(str, VERSION[:-1])) + VERSION[-1]
+    """Return the VERSION as a string, e.g. for VERSION == (0, 10, 7),
+    return '0.10.7'.
+    """
     return '.'.join(map(str, VERSION))
 
 

mongoengine/base/__init__.py

@@ -1,8 +1,28 @@
+# Base module is split into several files for convenience. Files inside of
+# this module should import from a specific submodule (e.g.
+# `from mongoengine.base.document import BaseDocument`), but all of the
+# other modules should import directly from the top-level module (e.g.
+# `from mongoengine.base import BaseDocument`). This approach is cleaner and
+# also helps with cyclical import errors.
 from mongoengine.base.common import *
 from mongoengine.base.datastructures import *
 from mongoengine.base.document import *
 from mongoengine.base.fields import *
 from mongoengine.base.metaclasses import *
 
-# Help with backwards compatibility
-from mongoengine.errors import *
+__all__ = (
+    # common
+    'UPDATE_OPERATORS', '_document_registry', 'get_document',
+
+    # datastructures
+    'BaseDict', 'BaseList', 'EmbeddedDocumentList',
+
+    # document
+    'BaseDocument',
+
+    # fields
+    'BaseField', 'ComplexBaseField', 'ObjectIdField', 'GeoJsonBaseField',
+
+    # metaclasses
+    'DocumentMetaclass', 'TopLevelDocumentMetaclass'
+)

mongoengine/base/common.py

@@ -1,13 +1,18 @@
 from mongoengine.errors import NotRegistered
 
-__all__ = ('ALLOW_INHERITANCE', 'get_document', '_document_registry')
+__all__ = ('UPDATE_OPERATORS', 'get_document', '_document_registry')
+
+
+UPDATE_OPERATORS = set(['set', 'unset', 'inc', 'dec', 'pop', 'push',
+                        'push_all', 'pull', 'pull_all', 'add_to_set',
+                        'set_on_insert', 'min', 'max'])
 
-ALLOW_INHERITANCE = False
 
 _document_registry = {}
 
 
 def get_document(name):
+    """Get a document class by name."""
     doc = _document_registry.get(name, None)
     if not doc:
         # Possible old style name

mongoengine/base/datastructures.py

@@ -1,14 +1,16 @@
 import itertools
 import weakref
 
+import six
+
 from mongoengine.common import _import_class
 from mongoengine.errors import DoesNotExist, MultipleObjectsReturned
 
-__all__ = ("BaseDict", "BaseList", "EmbeddedDocumentList")
+__all__ = ('BaseDict', 'BaseList', 'EmbeddedDocumentList')
 
 
 class BaseDict(dict):
-    """A special dict so we can watch any changes"""
+    """A special dict so we can watch any changes."""
 
     _dereferenced = False
     _instance = None
@@ -93,8 +95,7 @@ class BaseDict(dict):
 
 
 class BaseList(list):
-    """A special list so we can watch any changes
-    """
+    """A special list so we can watch any changes."""
 
     _dereferenced = False
     _instance = None
@@ -209,17 +210,22 @@ class BaseList(list):
 class EmbeddedDocumentList(BaseList):
 
     @classmethod
-    def __match_all(cls, i, kwargs):
-        items = kwargs.items()
-        return all([
-            getattr(i, k) == v or unicode(getattr(i, k)) == v for k, v in items
-        ])
+    def __match_all(cls, embedded_doc, kwargs):
+        """Return True if a given embedded doc matches all the filter
+        kwargs. If it doesn't return False.
+        """
+        for key, expected_value in kwargs.items():
+            doc_val = getattr(embedded_doc, key)
+            if doc_val != expected_value and six.text_type(doc_val) != expected_value:
+                return False
+        return True
 
     @classmethod
-    def __only_matches(cls, obj, kwargs):
+    def __only_matches(cls, embedded_docs, kwargs):
+        """Return embedded docs that match the filter kwargs."""
         if not kwargs:
-            return obj
-        return filter(lambda i: cls.__match_all(i, kwargs), obj)
+            return embedded_docs
+        return [doc for doc in embedded_docs if cls.__match_all(doc, kwargs)]
 
     def __init__(self, list_items, instance, name):
         super(EmbeddedDocumentList, self).__init__(list_items, instance, name)
@@ -285,18 +291,18 @@ class EmbeddedDocumentList(BaseList):
         values = self.__only_matches(self, kwargs)
         if len(values) == 0:
             raise DoesNotExist(
-                "%s matching query does not exist." % self._name
+                '%s matching query does not exist.' % self._name
             )
         elif len(values) > 1:
             raise MultipleObjectsReturned(
-                "%d items returned, instead of 1" % len(values)
+                '%d items returned, instead of 1' % len(values)
             )
 
         return values[0]
 
     def first(self):
-        """
-        Returns the first embedded document in the list, or ``None`` if empty.
+        """Return the first embedded document in the list, or ``None``
+        if empty.
         """
         if len(self) > 0:
             return self[0]
@@ -438,7 +444,7 @@ class StrictDict(object):
                 __slots__ = allowed_keys_tuple
 
                 def __repr__(self):
-                    return "{%s}" % ', '.join('"{0!s}": {0!r}'.format(k) for k in self.iterkeys())
+                    return '{%s}' % ', '.join('"{0!s}": {1!r}'.format(k, v) for k, v in self.items())
 
             cls._classes[allowed_keys] = SpecificStrictDict
         return cls._classes[allowed_keys]

mongoengine/base/document.py

@@ -1,6 +1,5 @@
 import copy
 import numbers
-import operator
 from collections import Hashable
 from functools import partial
 
@@ -8,30 +7,27 @@ from bson import ObjectId, json_util
 from bson.dbref import DBRef
 from bson.son import SON
 import pymongo
+import six
 
 from mongoengine import signals
-from mongoengine.base.common import ALLOW_INHERITANCE, get_document
-from mongoengine.base.datastructures import (
-    BaseDict,
-    BaseList,
-    EmbeddedDocumentList,
-    SemiStrictDict,
-    StrictDict
-)
+from mongoengine.base.common import get_document
+from mongoengine.base.datastructures import (BaseDict, BaseList,
+                                             EmbeddedDocumentList,
+                                             SemiStrictDict, StrictDict)
 from mongoengine.base.fields import ComplexBaseField
 from mongoengine.common import _import_class
 from mongoengine.errors import (FieldDoesNotExist, InvalidDocumentError,
-                                LookUpError, ValidationError)
-from mongoengine.python_support import PY3, txt_type
+                                LookUpError, OperationError, ValidationError)
 
-__all__ = ('BaseDocument', 'NON_FIELD_ERRORS')
+__all__ = ('BaseDocument',)
 
 NON_FIELD_ERRORS = '__all__'
 
 
 class BaseDocument(object):
     __slots__ = ('_changed_fields', '_initialised', '_created', '_data',
-                 '_dynamic_fields', '_auto_id_field', '_db_field_map', '__weakref__')
+                 '_dynamic_fields', '_auto_id_field', '_db_field_map',
+                 '__weakref__')
 
     _dynamic = False
     _dynamic_lock = True
@@ -57,15 +53,15 @@ class BaseDocument(object):
                 name = next(field)
                 if name in values:
                     raise TypeError(
-                        "Multiple values for keyword argument '" + name + "'")
+                        'Multiple values for keyword argument "%s"' % name)
                 values[name] = value
 
-        __auto_convert = values.pop("__auto_convert", True)
+        __auto_convert = values.pop('__auto_convert', True)
 
         # 399: set default values only to fields loaded from DB
-        __only_fields = set(values.pop("__only_fields", values))
+        __only_fields = set(values.pop('__only_fields', values))
 
-        _created = values.pop("_created", True)
+        _created = values.pop('_created', True)
 
         signals.pre_init.send(self.__class__, document=self, values=values)
 
@@ -76,7 +72,7 @@ class BaseDocument(object):
                 self._fields.keys() + ['id', 'pk', '_cls', '_text_score'])
             if _undefined_fields:
                 msg = (
-                    "The fields '{0}' do not exist on the document '{1}'"
+                    'The fields "{0}" do not exist on the document "{1}"'
                 ).format(_undefined_fields, self._class_name)
                 raise FieldDoesNotExist(msg)
 
@@ -95,7 +91,7 @@ class BaseDocument(object):
             value = getattr(self, key, None)
             setattr(self, key, value)
 
-        if "_cls" not in values:
+        if '_cls' not in values:
             self._cls = self._class_name
 
         # Set passed values after initialisation
@@ -121,7 +117,7 @@ class BaseDocument(object):
                 else:
                     self._data[key] = value
 
-        # Set any get_fieldname_display methods
+        # Set any get_<field>_display methods
         self.__set_field_display()
 
         if self._dynamic:
@@ -150,7 +146,7 @@ class BaseDocument(object):
         if self._dynamic and not self._dynamic_lock:
 
             if not hasattr(self, name) and not name.startswith('_'):
-                DynamicField = _import_class("DynamicField")
+                DynamicField = _import_class('DynamicField')
                 field = DynamicField(db_field=name)
                 field.name = name
                 self._dynamic_fields[name] = field
@@ -169,11 +165,13 @@ class BaseDocument(object):
         except AttributeError:
             self__created = True
 
-        if (self._is_document and not self__created and
-                name in self._meta.get('shard_key', tuple()) and
-                self._data.get(name) != value):
-            OperationError = _import_class('OperationError')
-            msg = "Shard Keys are immutable. Tried to update %s" % name
+        if (
+            self._is_document and
+            not self__created and
+            name in self._meta.get('shard_key', tuple()) and
+            self._data.get(name) != value
+        ):
+            msg = 'Shard Keys are immutable. Tried to update %s' % name
             raise OperationError(msg)
 
         try:
@@ -197,8 +195,8 @@ class BaseDocument(object):
         return data
 
     def __setstate__(self, data):
-        if isinstance(data["_data"], SON):
-            data["_data"] = self.__class__._from_son(data["_data"])._data
+        if isinstance(data['_data'], SON):
+            data['_data'] = self.__class__._from_son(data['_data'])._data
         for k in ('_changed_fields', '_initialised', '_created', '_data',
                   '_dynamic_fields'):
             if k in data:
@@ -212,7 +210,7 @@ class BaseDocument(object):
 
         dynamic_fields = data.get('_dynamic_fields') or SON()
         for k in dynamic_fields.keys():
-            setattr(self, k, data["_data"].get(k))
+            setattr(self, k, data['_data'].get(k))
 
     def __iter__(self):
         return iter(self._fields_ordered)
@@ -254,12 +252,13 @@ class BaseDocument(object):
         return repr_type('<%s: %s>' % (self.__class__.__name__, u))
 
     def __str__(self):
+        # TODO this could be simpler?
         if hasattr(self, '__unicode__'):
-            if PY3:
+            if six.PY3:
                 return self.__unicode__()
             else:
-                return unicode(self).encode('utf-8')
-        return txt_type('%s object' % self.__class__.__name__)
+                return six.text_type(self).encode('utf-8')
+        return six.text_type('%s object' % self.__class__.__name__)
 
     def __eq__(self, other):
         if isinstance(other, self.__class__) and hasattr(other, 'id') and other.id is not None:
@@ -308,7 +307,7 @@ class BaseDocument(object):
             fields = []
 
         data = SON()
-        data["_id"] = None
+        data['_id'] = None
         data['_cls'] = self._class_name
 
         # only root fields ['test1.a', 'test2'] => ['test1', 'test2']
@@ -351,18 +350,8 @@ class BaseDocument(object):
                 else:
                     data[field.name] = value
 
-        # If "_id" has not been set, then try and set it
-        Document = _import_class("Document")
-        if isinstance(self, Document):
-            if data["_id"] is None:
-                data["_id"] = self._data.get("id", None)
-
-        if data['_id'] is None:
-            data.pop('_id')
-
         # Only add _cls if allow_inheritance is True
-        if (not hasattr(self, '_meta') or
-                not self._meta.get('allow_inheritance', ALLOW_INHERITANCE)):
+        if not self._meta.get('allow_inheritance'):
             data.pop('_cls')
 
         return data
@@ -376,16 +365,16 @@ class BaseDocument(object):
         if clean:
             try:
                 self.clean()
-            except ValidationError, error:
+            except ValidationError as error:
                 errors[NON_FIELD_ERRORS] = error
 
         # Get a list of tuples of field names and their current values
         fields = [(self._fields.get(name, self._dynamic_fields.get(name)),
                    self._data.get(name)) for name in self._fields_ordered]
 
-        EmbeddedDocumentField = _import_class("EmbeddedDocumentField")
+        EmbeddedDocumentField = _import_class('EmbeddedDocumentField')
         GenericEmbeddedDocumentField = _import_class(
-            "GenericEmbeddedDocumentField")
+            'GenericEmbeddedDocumentField')
 
         for field, value in fields:
             if value is not None:
@@ -395,21 +384,21 @@ class BaseDocument(object):
                         field._validate(value, clean=clean)
                     else:
                         field._validate(value)
-                except ValidationError, error:
+                except ValidationError as error:
                     errors[field.name] = error.errors or error
-                except (ValueError, AttributeError, AssertionError), error:
+                except (ValueError, AttributeError, AssertionError) as error:
                     errors[field.name] = error
             elif field.required and not getattr(field, '_auto_gen', False):
                 errors[field.name] = ValidationError('Field is required',
                                                      field_name=field.name)
 
         if errors:
-            pk = "None"
+            pk = 'None'
             if hasattr(self, 'pk'):
                 pk = self.pk
             elif self._instance and hasattr(self._instance, 'pk'):
                 pk = self._instance.pk
-            message = "ValidationError (%s:%s) " % (self._class_name, pk)
+            message = 'ValidationError (%s:%s) ' % (self._class_name, pk)
             raise ValidationError(message, errors=errors)
 
     def to_json(self, *args, **kwargs):
@@ -426,33 +415,26 @@ class BaseDocument(object):
         return cls._from_son(json_util.loads(json_data), created=created)
 
     def __expand_dynamic_values(self, name, value):
-        """expand any dynamic values to their correct types / values"""
+        """Expand any dynamic values to their correct types / values."""
         if not isinstance(value, (dict, list, tuple)):
             return value
 
-        EmbeddedDocumentListField = _import_class('EmbeddedDocumentListField')
-
-        is_list = False
-        if not hasattr(value, 'items'):
-            is_list = True
-            value = dict([(k, v) for k, v in enumerate(value)])
-
-        if not is_list and '_cls' in value:
+        # If the value is a dict with '_cls' in it, turn it into a document
+        is_dict = isinstance(value, dict)
+        if is_dict and '_cls' in value:
             cls = get_document(value['_cls'])
             return cls(**value)
 
-        data = {}
-        for k, v in value.items():
-            key = name if is_list else k
-            data[k] = self.__expand_dynamic_values(key, v)
-
-        if is_list:  # Convert back to a list
-            data_items = sorted(data.items(), key=operator.itemgetter(0))
-            value = [v for k, v in data_items]
+        if is_dict:
+            value = {
+                k: self.__expand_dynamic_values(k, v)
+                for k, v in value.items()
+            }
         else:
-            value = data
+            value = [self.__expand_dynamic_values(name, v) for v in value]
 
         # Convert lists / values so we can watch for any changes on them
+        EmbeddedDocumentListField = _import_class('EmbeddedDocumentListField')
         if (isinstance(value, (list, tuple)) and
                 not isinstance(value, BaseList)):
             if issubclass(type(self), EmbeddedDocumentListField):
@@ -465,8 +447,7 @@ class BaseDocument(object):
         return value
 
     def _mark_as_changed(self, key):
-        """Marks a key as explicitly changed by the user
-        """
+        """Mark a key as explicitly changed by the user."""
         if not key:
             return
 
@@ -496,10 +477,11 @@ class BaseDocument(object):
                         remove(field)
 
     def _clear_changed_fields(self):
-        """Using get_changed_fields iterate and remove any fields that are
-        marked as changed"""
+        """Using _get_changed_fields iterate and remove any fields that
+        are marked as changed.
+        """
         for changed in self._get_changed_fields():
-            parts = changed.split(".")
+            parts = changed.split('.')
             data = self
             for part in parts:
                 if isinstance(data, list):
@@ -511,10 +493,13 @@ class BaseDocument(object):
                     data = data.get(part, None)
                 else:
                     data = getattr(data, part, None)
-                if hasattr(data, "_changed_fields"):
-                    if hasattr(data, "_is_document") and data._is_document:
+
+                if hasattr(data, '_changed_fields'):
+                    if getattr(data, '_is_document', False):
                         continue
+
                     data._changed_fields = []
+
         self._changed_fields = []
 
     def _nestable_types_changed_fields(self, changed_fields, key, data, inspected):
@@ -526,26 +511,27 @@ class BaseDocument(object):
             iterator = data.iteritems()
 
         for index, value in iterator:
-            list_key = "%s%s." % (key, index)
+            list_key = '%s%s.' % (key, index)
             # don't check anything lower if this key is already marked
             # as changed.
             if list_key[:-1] in changed_fields:
                 continue
             if hasattr(value, '_get_changed_fields'):
                 changed = value._get_changed_fields(inspected)
-                changed_fields += ["%s%s" % (list_key, k)
+                changed_fields += ['%s%s' % (list_key, k)
                                    for k in changed if k]
             elif isinstance(value, (list, tuple, dict)):
                 self._nestable_types_changed_fields(
                     changed_fields, list_key, value, inspected)
 
     def _get_changed_fields(self, inspected=None):
-        """Returns a list of all fields that have explicitly been changed.
+        """Return a list of all fields that have explicitly been changed.
         """
-        EmbeddedDocument = _import_class("EmbeddedDocument")
-        DynamicEmbeddedDocument = _import_class("DynamicEmbeddedDocument")
-        ReferenceField = _import_class("ReferenceField")
-        SortedListField = _import_class("SortedListField")
+        EmbeddedDocument = _import_class('EmbeddedDocument')
+        DynamicEmbeddedDocument = _import_class('DynamicEmbeddedDocument')
+        ReferenceField = _import_class('ReferenceField')
+        SortedListField = _import_class('SortedListField')
+
         changed_fields = []
         changed_fields += getattr(self, '_changed_fields', [])
 
@@ -572,7 +558,7 @@ class BaseDocument(object):
             ):
                 # Find all embedded fields that have been changed
                 changed = data._get_changed_fields(inspected)
-                changed_fields += ["%s%s" % (key, k) for k in changed if k]
+                changed_fields += ['%s%s' % (key, k) for k in changed if k]
             elif (isinstance(data, (list, tuple, dict)) and
                     db_field_name not in changed_fields):
                 if (hasattr(field, 'field') and
@@ -676,21 +662,25 @@ class BaseDocument(object):
 
     @classmethod
     def _get_collection_name(cls):
-        """Returns the collection name for this class. None for abstract class
+        """Return the collection name for this class. None for abstract
+        class.
         """
         return cls._meta.get('collection', None)
 
     @classmethod
     def _from_son(cls, son, _auto_dereference=True, only_fields=None, created=False):
-        """Create an instance of a Document (subclass) from a PyMongo SON.
+        """Create an instance of a Document (subclass) from a PyMongo
+        SON.
         """
         if not only_fields:
             only_fields = []
 
-        # get the class name from the document, falling back to the given
+        # Get the class name from the document, falling back to the given
         # class if unavailable
         class_name = son.get('_cls', cls._class_name)
-        data = dict(("%s" % key, value) for key, value in son.iteritems())
+
+        # Convert SON to a dict, making sure each key is a string
+        data = {str(key): value for key, value in son.iteritems()}
 
         # Return correct subclass for document type
         if class_name != cls._class_name:
@@ -712,19 +702,20 @@ class BaseDocument(object):
                                         else field.to_python(value))
                     if field_name != field.db_field:
                         del data[field.db_field]
-                except (AttributeError, ValueError), e:
+                except (AttributeError, ValueError) as e:
                     errors_dict[field_name] = e
 
         if errors_dict:
-            errors = "\n".join(["%s - %s" % (k, v)
+            errors = '\n'.join(['%s - %s' % (k, v)
                                 for k, v in errors_dict.items()])
-            msg = ("Invalid data to create a `%s` instance.\n%s"
+            msg = ('Invalid data to create a `%s` instance.\n%s'
                    % (cls._class_name, errors))
             raise InvalidDocumentError(msg)
 
+        # In STRICT documents, remove any keys that aren't in cls._fields
         if cls.STRICT:
-            data = dict((k, v)
-                        for k, v in data.iteritems() if k in cls._fields)
+            data = {k: v for k, v in data.iteritems() if k in cls._fields}
+
         obj = cls(__auto_convert=False, _created=created, __only_fields=only_fields, **data)
         obj._changed_fields = changed_fields
         if not _auto_dereference:
@@ -734,37 +725,43 @@ class BaseDocument(object):
 
     @classmethod
     def _build_index_specs(cls, meta_indexes):
-        """Generate and merge the full index specs
-        """
-
+        """Generate and merge the full index specs."""
         geo_indices = cls._geo_indices()
         unique_indices = cls._unique_with_indexes()
-        index_specs = [cls._build_index_spec(spec)
-                       for spec in meta_indexes]
+        index_specs = [cls._build_index_spec(spec) for spec in meta_indexes]
 
         def merge_index_specs(index_specs, indices):
+            """Helper method for merging index specs."""
             if not indices:
                 return index_specs
 
-            spec_fields = [v['fields']
-                           for k, v in enumerate(index_specs)]
-            # Merge unique_indexes with existing specs
-            for k, v in enumerate(indices):
-                if v['fields'] in spec_fields:
-                    index_specs[spec_fields.index(v['fields'])].update(v)
+            # Create a map of index fields to index spec. We're converting
+            # the fields from a list to a tuple so that it's hashable.
+            spec_fields = {
+                tuple(index['fields']): index for index in index_specs
+            }
+
+            # For each new index, if there's an existing index with the same
+            # fields list, update the existing spec with all data from the
+            # new spec.
+            for new_index in indices:
+                candidate = spec_fields.get(tuple(new_index['fields']))
+                if candidate is None:
+                    index_specs.append(new_index)
                 else:
-                    index_specs.append(v)
+                    candidate.update(new_index)
+
             return index_specs
 
+        # Merge geo indexes and unique_with indexes into the meta index specs.
         index_specs = merge_index_specs(index_specs, geo_indices)
         index_specs = merge_index_specs(index_specs, unique_indices)
         return index_specs
 
     @classmethod
     def _build_index_spec(cls, spec):
-        """Build a PyMongo index spec from a MongoEngine index spec.
-        """
-        if isinstance(spec, basestring):
+        """Build a PyMongo index spec from a MongoEngine index spec."""
+        if isinstance(spec, six.string_types):
             spec = {'fields': [spec]}
         elif isinstance(spec, (list, tuple)):
             spec = {'fields': list(spec)}
@@ -775,8 +772,7 @@ class BaseDocument(object):
         direction = None
 
         # Check to see if we need to include _cls
-        allow_inheritance = cls._meta.get('allow_inheritance',
-                                          ALLOW_INHERITANCE)
+        allow_inheritance = cls._meta.get('allow_inheritance')
         include_cls = (
             allow_inheritance and
             not spec.get('sparse', False) and
@@ -786,7 +782,7 @@ class BaseDocument(object):
 
         # 733: don't include cls if index_cls is False unless there is an explicit cls with the index
         include_cls = include_cls and (spec.get('cls', False) or cls._meta.get('index_cls', True))
-        if "cls" in spec:
+        if 'cls' in spec:
             spec.pop('cls')
         for key in spec['fields']:
             # If inherited spec continue
@@ -801,19 +797,19 @@ class BaseDocument(object):
             # GEOHAYSTACK from )
             # GEO2D from *
             direction = pymongo.ASCENDING
-            if key.startswith("-"):
+            if key.startswith('-'):
                 direction = pymongo.DESCENDING
-            elif key.startswith("$"):
+            elif key.startswith('$'):
                 direction = pymongo.TEXT
-            elif key.startswith("#"):
+            elif key.startswith('#'):
                 direction = pymongo.HASHED
-            elif key.startswith("("):
+            elif key.startswith('('):
                 direction = pymongo.GEOSPHERE
-            elif key.startswith(")"):
+            elif key.startswith(')'):
                 direction = pymongo.GEOHAYSTACK
-            elif key.startswith("*"):
+            elif key.startswith('*'):
                 direction = pymongo.GEO2D
-            if key.startswith(("+", "-", "*", "$", "#", "(", ")")):
+            if key.startswith(('+', '-', '*', '$', '#', '(', ')')):
                 key = key[1:]
 
             # Use real field name, do it manually because we need field
@@ -826,7 +822,7 @@ class BaseDocument(object):
                 parts = []
                 for field in fields:
                     try:
-                        if field != "_id":
+                        if field != '_id':
                             field = field.db_field
                     except AttributeError:
                         pass
@@ -845,49 +841,53 @@ class BaseDocument(object):
         return spec
 
     @classmethod
-    def _unique_with_indexes(cls, namespace=""):
-        """
-        Find and set unique indexes
-        """
+    def _unique_with_indexes(cls, namespace=''):
+        """Find unique indexes in the document schema and return them."""
         unique_indexes = []
         for field_name, field in cls._fields.items():
             sparse = field.sparse
+
             # Generate a list of indexes needed by uniqueness constraints
             if field.unique:
                 unique_fields = [field.db_field]
 
                 # Add any unique_with fields to the back of the index spec
                 if field.unique_with:
-                    if isinstance(field.unique_with, basestring):
+                    if isinstance(field.unique_with, six.string_types):
                         field.unique_with = [field.unique_with]
 
                     # Convert unique_with field names to real field names
                     unique_with = []
                     for other_name in field.unique_with:
                         parts = other_name.split('.')
+
                         # Lookup real name
                         parts = cls._lookup_field(parts)
                         name_parts = [part.db_field for part in parts]
                         unique_with.append('.'.join(name_parts))
+
                         # Unique field should be required
                         parts[-1].required = True
                         sparse = (not sparse and
                                   parts[-1].name not in cls.__dict__)
+
                     unique_fields += unique_with
 
                 # Add the new index to the list
-                fields = [("%s%s" % (namespace, f), pymongo.ASCENDING)
-                          for f in unique_fields]
+                fields = [
+                    ('%s%s' % (namespace, f), pymongo.ASCENDING)
+                    for f in unique_fields
+                ]
                 index = {'fields': fields, 'unique': True, 'sparse': sparse}
                 unique_indexes.append(index)
 
-            if field.__class__.__name__ == "ListField":
+            if field.__class__.__name__ == 'ListField':
                 field = field.field
 
             # Grab any embedded document field unique indexes
-            if (field.__class__.__name__ == "EmbeddedDocumentField" and
+            if (field.__class__.__name__ == 'EmbeddedDocumentField' and
                     field.document_type != cls):
-                field_namespace = "%s." % field_name
+                field_namespace = '%s.' % field_name
                 doc_cls = field.document_type
                 unique_indexes += doc_cls._unique_with_indexes(field_namespace)
 
@@ -899,8 +899,9 @@ class BaseDocument(object):
         geo_indices = []
         inspected.append(cls)
 
-        geo_field_type_names = ["EmbeddedDocumentField", "GeoPointField",
-                                "PointField", "LineStringField", "PolygonField"]
+        geo_field_type_names = ('EmbeddedDocumentField', 'GeoPointField',
+                                'PointField', 'LineStringField',
+                                'PolygonField')
 
         geo_field_types = tuple([_import_class(field)
                                  for field in geo_field_type_names])
@@ -908,32 +909,68 @@ class BaseDocument(object):
         for field in cls._fields.values():
             if not isinstance(field, geo_field_types):
                 continue
+
             if hasattr(field, 'document_type'):
                 field_cls = field.document_type
                 if field_cls in inspected:
                     continue
+
                 if hasattr(field_cls, '_geo_indices'):
                     geo_indices += field_cls._geo_indices(
                         inspected, parent_field=field.db_field)
             elif field._geo_index:
                 field_name = field.db_field
                 if parent_field:
-                    field_name = "%s.%s" % (parent_field, field_name)
-                geo_indices.append({'fields':
-                                    [(field_name, field._geo_index)]})
+                    field_name = '%s.%s' % (parent_field, field_name)
+                geo_indices.append({
+                    'fields': [(field_name, field._geo_index)]
+                })
+
         return geo_indices
 
     @classmethod
     def _lookup_field(cls, parts):
-        """Lookup a field based on its attribute and return a list containing
-        the field's parents and the field.
-        """
+        """Given the path to a given field, return a list containing
+        the Field object associated with that field and all of its parent
+        Field objects.
 
-        ListField = _import_class("ListField")
+        Args:
+            parts (str, list, or tuple) - path to the field. Should be a
+            string for simple fields existing on this document or a list
+            of strings for a field that exists deeper in embedded documents.
+
+        Returns:
+            A list of Field instances for fields that were found or
+            strings for sub-fields that weren't.
+
+        Example:
+            >>> user._lookup_field('name')
+            [<mongoengine.fields.StringField at 0x1119bff50>]
+
+            >>> user._lookup_field('roles')
+            [<mongoengine.fields.EmbeddedDocumentListField at 0x1119ec250>]
+
+            >>> user._lookup_field(['roles', 'role'])
+            [<mongoengine.fields.EmbeddedDocumentListField at 0x1119ec250>,
+             <mongoengine.fields.StringField at 0x1119ec050>]
+
+            >>> user._lookup_field('doesnt_exist')
+            raises LookUpError
+
+            >>> user._lookup_field(['roles', 'doesnt_exist'])
+            [<mongoengine.fields.EmbeddedDocumentListField at 0x1119ec250>,
+             'doesnt_exist']
+
+        """
+        # TODO this method is WAY too complicated. Simplify it.
+        # TODO don't think returning a string for embedded non-existent fields is desired
+
+        ListField = _import_class('ListField')
         DynamicField = _import_class('DynamicField')
 
         if not isinstance(parts, (list, tuple)):
             parts = [parts]
+
         fields = []
         field = None
 
@@ -943,16 +980,17 @@ class BaseDocument(object):
                 fields.append(field_name)
                 continue
 
+            # Look up first field from the document
             if field is None:
-                # Look up first field from the document
                 if field_name == 'pk':
                     # Deal with "primary key" alias
                     field_name = cls._meta['id_field']
+
                 if field_name in cls._fields:
                     field = cls._fields[field_name]
                 elif cls._dynamic:
                     field = DynamicField(db_field=field_name)
-                elif cls._meta.get("allow_inheritance", False) or cls._meta.get("abstract", False):
+                elif cls._meta.get('allow_inheritance') or cls._meta.get('abstract', False):
                     # 744: in case the field is defined in a subclass
                     for subcls in cls.__subclasses__():
                         try:
@@ -965,35 +1003,55 @@ class BaseDocument(object):
                     else:
                         raise LookUpError('Cannot resolve field "%s"' % field_name)
                 else:
-                    raise LookUpError('Cannot resolve field "%s"'
-                                      % field_name)
+                    raise LookUpError('Cannot resolve field "%s"' % field_name)
             else:
                 ReferenceField = _import_class('ReferenceField')
                 GenericReferenceField = _import_class('GenericReferenceField')
+
+                # If previous field was a reference, throw an error (we
+                # cannot look up fields that are on references).
                 if isinstance(field, (ReferenceField, GenericReferenceField)):
                     raise LookUpError('Cannot perform join in mongoDB: %s' %
                                       '__'.join(parts))
+
+                # If the parent field has a "field" attribute which has a
+                # lookup_member method, call it to find the field
+                # corresponding to this iteration.
                 if hasattr(getattr(field, 'field', None), 'lookup_member'):
                     new_field = field.field.lookup_member(field_name)
+
+                # If the parent field is a DynamicField or if it's part of
+                # a DynamicDocument, mark current field as a DynamicField
+                # with db_name equal to the field name.
                 elif cls._dynamic and (isinstance(field, DynamicField) or
                                        getattr(getattr(field, 'document_type', None), '_dynamic', None)):
                     new_field = DynamicField(db_field=field_name)
+
+                # Else, try to use the parent field's lookup_member method
+                # to find the subfield.
+                elif hasattr(field, 'lookup_member'):
+                    new_field = field.lookup_member(field_name)
+
+                # Raise a LookUpError if all the other conditions failed.
                 else:
-                    # Look up subfield on the previous field or raise
-                    try:
-                        new_field = field.lookup_member(field_name)
-                    except AttributeError:
-                        raise LookUpError('Cannot resolve subfield or operator {} '
-                                          'on the field {}'.format(
-                                              field_name, field.name))
+                    raise LookUpError(
+                        'Cannot resolve subfield or operator {} '
+                        'on the field {}'.format(field_name, field.name)
+                    )
+
+                # If current field still wasn't found and the parent field
+                # is a ComplexBaseField, add the name current field name and
+                # move on.
                 if not new_field and isinstance(field, ComplexBaseField):
                     fields.append(field_name)
                     continue
                 elif not new_field:
-                    raise LookUpError('Cannot resolve field "%s"'
-                                      % field_name)
+                    raise LookUpError('Cannot resolve field "%s"' % field_name)
+
                 field = new_field  # update field to the new field type
+
             fields.append(field)
+
         return fields
 
     @classmethod
@@ -1005,19 +1063,18 @@ class BaseDocument(object):
         return '.'.join(parts)
 
     def __set_field_display(self):
-        """Dynamically set the display value for a field with choices"""
-        for attr_name, field in self._fields.items():
-            if field.choices:
-                if self._dynamic:
-                    obj = self
-                else:
-                    obj = type(self)
-                setattr(obj,
-                        'get_%s_display' % attr_name,
-                        partial(self.__get_field_display, field=field))
+        """For each field that specifies choices, create a
+        get_<field>_display method.
+        """
+        fields_with_choices = [(n, f) for n, f in self._fields.items()
+                               if f.choices]
+        for attr_name, field in fields_with_choices:
+            setattr(self,
+                    'get_%s_display' % attr_name,
+                    partial(self.__get_field_display, field=field))
 
     def __get_field_display(self, field):
-        """Returns the display value for a choice field"""
+        """Return the display value for a choice field"""
         value = getattr(self, field.name)
         if field.choices and isinstance(field.choices[0], (list, tuple)):
             return dict(field.choices).get(value, value)

mongoengine/base/fields.py

@@ -4,21 +4,17 @@ import weakref
 
 from bson import DBRef, ObjectId, SON
 import pymongo
+import six
 
-from mongoengine.base.common import ALLOW_INHERITANCE
-from mongoengine.base.datastructures import (
-    BaseDict, BaseList, EmbeddedDocumentList
-)
+from mongoengine.base.common import UPDATE_OPERATORS
+from mongoengine.base.datastructures import (BaseDict, BaseList,
+                                             EmbeddedDocumentList)
 from mongoengine.common import _import_class
 from mongoengine.errors import ValidationError
 
-__all__ = ("BaseField", "ComplexBaseField",
-           "ObjectIdField", "GeoJsonBaseField")
 
-
-UPDATE_OPERATORS = set(['set', 'unset', 'inc', 'dec', 'pop', 'push',
-                        'push_all', 'pull', 'pull_all', 'add_to_set',
-                        'set_on_insert', 'min', 'max'])
+__all__ = ('BaseField', 'ComplexBaseField', 'ObjectIdField',
+           'GeoJsonBaseField')
 
 
 class BaseField(object):
@@ -27,7 +23,6 @@ class BaseField(object):
 
     .. versionchanged:: 0.5 - added verbose and help text
     """
-
     name = None
     _geo_index = False
     _auto_gen = False  # Call `generate` to generate a value
@@ -73,7 +68,7 @@ class BaseField(object):
         self.db_field = (db_field or name) if not primary_key else '_id'
 
         if name:
-            msg = "Fields' 'name' attribute deprecated in favour of 'db_field'"
+            msg = 'Field\'s "name" attribute deprecated in favour of "db_field"'
             warnings.warn(msg, DeprecationWarning)
         self.required = required or primary_key
         self.default = default
@@ -89,7 +84,7 @@ class BaseField(object):
         # Detect and report conflicts between metadata and base properties.
         conflicts = set(dir(self)) & set(kwargs)
         if conflicts:
-            raise TypeError("%s already has attribute(s): %s" % (
+            raise TypeError('%s already has attribute(s): %s' % (
                 self.__class__.__name__, ', '.join(conflicts)))
 
         # Assign metadata to the instance
@@ -147,25 +142,21 @@ class BaseField(object):
                     v._instance = weakref.proxy(instance)
         instance._data[self.name] = value
 
-    def error(self, message="", errors=None, field_name=None):
-        """Raises a ValidationError.
-        """
+    def error(self, message='', errors=None, field_name=None):
+        """Raise a ValidationError."""
         field_name = field_name if field_name else self.name
         raise ValidationError(message, errors=errors, field_name=field_name)
 
     def to_python(self, value):
-        """Convert a MongoDB-compatible type to a Python type.
-        """
+        """Convert a MongoDB-compatible type to a Python type."""
         return value
 
     def to_mongo(self, value):
-        """Convert a Python type to a MongoDB-compatible type.
-        """
+        """Convert a Python type to a MongoDB-compatible type."""
         return self.to_python(value)
 
     def _to_mongo_safe_call(self, value, use_db_field=True, fields=None):
-        """A helper method to call to_mongo with proper inputs
-        """
+        """Helper method to call to_mongo with proper inputs."""
         f_inputs = self.to_mongo.__code__.co_varnames
         ex_vars = {}
         if 'fields' in f_inputs:
@@ -177,15 +168,13 @@ class BaseField(object):
         return self.to_mongo(value, **ex_vars)
 
     def prepare_query_value(self, op, value):
-        """Prepare a value that is being used in a query for PyMongo.
-        """
+        """Prepare a value that is being used in a query for PyMongo."""
         if op in UPDATE_OPERATORS:
             self.validate(value)
         return value
 
     def validate(self, value, clean=True):
-        """Perform validation on a value.
-        """
+        """Perform validation on a value."""
         pass
 
     def _validate_choices(self, value):
@@ -200,11 +189,13 @@ class BaseField(object):
         if isinstance(value, (Document, EmbeddedDocument)):
             if not any(isinstance(value, c) for c in choice_list):
                 self.error(
-                    'Value must be instance of %s' % unicode(choice_list)
+                    'Value must be an instance of %s' % (
+                        six.text_type(choice_list)
+                    )
                 )
         # Choices which are types other than Documents
         elif value not in choice_list:
-            self.error('Value must be one of %s' % unicode(choice_list))
+            self.error('Value must be one of %s' % six.text_type(choice_list))
 
     def _validate(self, value, **kwargs):
         # Check the Choices Constraint
@@ -247,8 +238,7 @@ class ComplexBaseField(BaseField):
     field = None
 
     def __get__(self, instance, owner):
-        """Descriptor to automatically dereference references.
-        """
+        """Descriptor to automatically dereference references."""
         if instance is None:
             # Document class being used rather than a document object
             return self
@@ -260,7 +250,7 @@ class ComplexBaseField(BaseField):
                        (self.field is None or isinstance(self.field,
                                                          (GenericReferenceField, ReferenceField))))
 
-        _dereference = _import_class("DeReference")()
+        _dereference = _import_class('DeReference')()
 
         self._auto_dereference = instance._fields[self.name]._auto_dereference
         if instance._initialised and dereference and instance._data.get(self.name):
@@ -295,9 +285,8 @@ class ComplexBaseField(BaseField):
         return value
 
     def to_python(self, value):
-        """Convert a MongoDB-compatible type to a Python type.
-        """
-        if isinstance(value, basestring):
+        """Convert a MongoDB-compatible type to a Python type."""
+        if isinstance(value, six.string_types):
             return value
 
         if hasattr(value, 'to_python'):
@@ -307,14 +296,14 @@ class ComplexBaseField(BaseField):
         if not hasattr(value, 'items'):
             try:
                 is_list = True
-                value = dict([(k, v) for k, v in enumerate(value)])
+                value = {k: v for k, v in enumerate(value)}
             except TypeError:  # Not iterable return the value
                 return value
 
         if self.field:
             self.field._auto_dereference = self._auto_dereference
-            value_dict = dict([(key, self.field.to_python(item))
-                               for key, item in value.items()])
+            value_dict = {key: self.field.to_python(item)
+                          for key, item in value.items()}
         else:
             Document = _import_class('Document')
             value_dict = {}
@@ -337,13 +326,12 @@ class ComplexBaseField(BaseField):
         return value_dict
 
     def to_mongo(self, value, use_db_field=True, fields=None):
-        """Convert a Python type to a MongoDB-compatible type.
-        """
-        Document = _import_class("Document")
-        EmbeddedDocument = _import_class("EmbeddedDocument")
-        GenericReferenceField = _import_class("GenericReferenceField")
+        """Convert a Python type to a MongoDB-compatible type."""
+        Document = _import_class('Document')
+        EmbeddedDocument = _import_class('EmbeddedDocument')
+        GenericReferenceField = _import_class('GenericReferenceField')
 
-        if isinstance(value, basestring):
+        if isinstance(value, six.string_types):
             return value
 
         if hasattr(value, 'to_mongo'):
@@ -360,13 +348,15 @@ class ComplexBaseField(BaseField):
         if not hasattr(value, 'items'):
             try:
                 is_list = True
-                value = dict([(k, v) for k, v in enumerate(value)])
+                value = {k: v for k, v in enumerate(value)}
             except TypeError:  # Not iterable return the value
                 return value
 
         if self.field:
-            value_dict = dict([(key, self.field._to_mongo_safe_call(item, use_db_field, fields))
-                               for key, item in value.iteritems()])
+            value_dict = {
+                key: self.field._to_mongo_safe_call(item, use_db_field, fields)
+                for key, item in value.iteritems()
+            }
         else:
             value_dict = {}
             for k, v in value.iteritems():
@@ -380,9 +370,7 @@ class ComplexBaseField(BaseField):
                     # any _cls data so make it a generic reference allows
                     # us to dereference
                     meta = getattr(v, '_meta', {})
-                    allow_inheritance = (
-                        meta.get('allow_inheritance', ALLOW_INHERITANCE)
-                        is True)
+                    allow_inheritance = meta.get('allow_inheritance')
                     if not allow_inheritance and not self.field:
                         value_dict[k] = GenericReferenceField().to_mongo(v)
                     else:
@@ -404,8 +392,7 @@ class ComplexBaseField(BaseField):
         return value_dict
 
     def validate(self, value):
-        """If field is provided ensure the value is valid.
-        """
+        """If field is provided ensure the value is valid."""
         errors = {}
         if self.field:
             if hasattr(value, 'iteritems') or hasattr(value, 'items'):
@@ -415,9 +402,9 @@ class ComplexBaseField(BaseField):
             for k, v in sequence:
                 try:
                     self.field._validate(v)
-                except ValidationError, error:
+                except ValidationError as error:
                     errors[k] = error.errors or error
-                except (ValueError, AssertionError), error:
+                except (ValueError, AssertionError) as error:
                     errors[k] = error
 
             if errors:
@@ -443,8 +430,7 @@ class ComplexBaseField(BaseField):
 
 
 class ObjectIdField(BaseField):
-    """A field wrapper around MongoDB's ObjectIds.
-    """
+    """A field wrapper around MongoDB's ObjectIds."""
 
     def to_python(self, value):
         try:
@@ -457,10 +443,10 @@ class ObjectIdField(BaseField):
     def to_mongo(self, value):
         if not isinstance(value, ObjectId):
             try:
-                return ObjectId(unicode(value))
-            except Exception, e:
+                return ObjectId(six.text_type(value))
+            except Exception as e:
                 # e.message attribute has been deprecated since Python 2.6
-                self.error(unicode(e))
+                self.error(six.text_type(e))
         return value
 
     def prepare_query_value(self, op, value):
@@ -468,7 +454,7 @@ class ObjectIdField(BaseField):
 
     def validate(self, value):
         try:
-            ObjectId(unicode(value))
+            ObjectId(six.text_type(value))
         except Exception:
             self.error('Invalid Object ID')
 
@@ -480,21 +466,20 @@ class GeoJsonBaseField(BaseField):
     """
 
     _geo_index = pymongo.GEOSPHERE
-    _type = "GeoBase"
+    _type = 'GeoBase'
 
     def __init__(self, auto_index=True, *args, **kwargs):
         """
-        :param bool auto_index: Automatically create a "2dsphere" index.\
+        :param bool auto_index: Automatically create a '2dsphere' index.\
             Defaults to `True`.
         """
-        self._name = "%sField" % self._type
+        self._name = '%sField' % self._type
         if not auto_index:
             self._geo_index = False
         super(GeoJsonBaseField, self).__init__(*args, **kwargs)
 
     def validate(self, value):
-        """Validate the GeoJson object based on its type
-        """
+        """Validate the GeoJson object based on its type."""
         if isinstance(value, dict):
             if set(value.keys()) == set(['type', 'coordinates']):
                 if value['type'] != self._type:
@@ -509,7 +494,7 @@ class GeoJsonBaseField(BaseField):
             self.error('%s can only accept lists of [x, y]' % self._name)
             return
 
-        validate = getattr(self, "_validate_%s" % self._type.lower())
+        validate = getattr(self, '_validate_%s' % self._type.lower())
         error = validate(value)
         if error:
             self.error(error)
@@ -522,7 +507,7 @@ class GeoJsonBaseField(BaseField):
         try:
             value[0][0][0]
         except (TypeError, IndexError):
-            return "Invalid Polygon must contain at least one valid linestring"
+            return 'Invalid Polygon must contain at least one valid linestring'
 
         errors = []
         for val in value:
@@ -533,12 +518,12 @@ class GeoJsonBaseField(BaseField):
                 errors.append(error)
         if errors:
             if top_level:
-                return "Invalid Polygon:\n%s" % ", ".join(errors)
+                return 'Invalid Polygon:\n%s' % ', '.join(errors)
             else:
-                return "%s" % ", ".join(errors)
+                return '%s' % ', '.join(errors)
 
     def _validate_linestring(self, value, top_level=True):
-        """Validates a linestring"""
+        """Validate a linestring."""
         if not isinstance(value, (list, tuple)):
             return 'LineStrings must contain list of coordinate pairs'
 
@@ -546,7 +531,7 @@ class GeoJsonBaseField(BaseField):
         try:
             value[0][0]
         except (TypeError, IndexError):
-            return "Invalid LineString must contain at least one valid point"
+            return 'Invalid LineString must contain at least one valid point'
 
         errors = []
         for val in value:
@@ -555,19 +540,19 @@ class GeoJsonBaseField(BaseField):
                 errors.append(error)
         if errors:
             if top_level:
-                return "Invalid LineString:\n%s" % ", ".join(errors)
+                return 'Invalid LineString:\n%s' % ', '.join(errors)
             else:
-                return "%s" % ", ".join(errors)
+                return '%s' % ', '.join(errors)
 
     def _validate_point(self, value):
         """Validate each set of coords"""
         if not isinstance(value, (list, tuple)):
             return 'Points must be a list of coordinate pairs'
         elif not len(value) == 2:
-            return "Value (%s) must be a two-dimensional point" % repr(value)
+            return 'Value (%s) must be a two-dimensional point' % repr(value)
         elif (not isinstance(value[0], (float, int)) or
               not isinstance(value[1], (float, int))):
-            return "Both values (%s) in point must be float or int" % repr(value)
+            return 'Both values (%s) in point must be float or int' % repr(value)
 
     def _validate_multipoint(self, value):
         if not isinstance(value, (list, tuple)):
@@ -577,7 +562,7 @@ class GeoJsonBaseField(BaseField):
         try:
             value[0][0]
         except (TypeError, IndexError):
-            return "Invalid MultiPoint must contain at least one valid point"
+            return 'Invalid MultiPoint must contain at least one valid point'
 
         errors = []
         for point in value:
@@ -586,7 +571,7 @@ class GeoJsonBaseField(BaseField):
                 errors.append(error)
 
         if errors:
-            return "%s" % ", ".join(errors)
+            return '%s' % ', '.join(errors)
 
     def _validate_multilinestring(self, value, top_level=True):
         if not isinstance(value, (list, tuple)):
@@ -596,7 +581,7 @@ class GeoJsonBaseField(BaseField):
         try:
             value[0][0][0]
         except (TypeError, IndexError):
-            return "Invalid MultiLineString must contain at least one valid linestring"
+            return 'Invalid MultiLineString must contain at least one valid linestring'
 
         errors = []
         for linestring in value:
@@ -606,9 +591,9 @@ class GeoJsonBaseField(BaseField):
 
         if errors:
             if top_level:
-                return "Invalid MultiLineString:\n%s" % ", ".join(errors)
+                return 'Invalid MultiLineString:\n%s' % ', '.join(errors)
             else:
-                return "%s" % ", ".join(errors)
+                return '%s' % ', '.join(errors)
 
     def _validate_multipolygon(self, value):
         if not isinstance(value, (list, tuple)):
@@ -618,7 +603,7 @@ class GeoJsonBaseField(BaseField):
         try:
             value[0][0][0][0]
         except (TypeError, IndexError):
-            return "Invalid MultiPolygon must contain at least one valid Polygon"
+            return 'Invalid MultiPolygon must contain at least one valid Polygon'
 
         errors = []
         for polygon in value:
@@ -627,9 +612,9 @@ class GeoJsonBaseField(BaseField):
                 errors.append(error)
 
         if errors:
-            return "Invalid MultiPolygon:\n%s" % ", ".join(errors)
+            return 'Invalid MultiPolygon:\n%s' % ', '.join(errors)
 
     def to_mongo(self, value):
         if isinstance(value, dict):
             return value
-        return SON([("type", self._type), ("coordinates", value)])
+        return SON([('type', self._type), ('coordinates', value)])

mongoengine/base/metaclasses.py

@@ -1,10 +1,11 @@
 import warnings
 
-from mongoengine.base.common import ALLOW_INHERITANCE, _document_registry
+import six
+
+from mongoengine.base.common import _document_registry
 from mongoengine.base.fields import BaseField, ComplexBaseField, ObjectIdField
 from mongoengine.common import _import_class
 from mongoengine.errors import InvalidDocumentError
-from mongoengine.python_support import PY3
 from mongoengine.queryset import (DO_NOTHING, DoesNotExist,
                                   MultipleObjectsReturned,
                                   QuerySetManager)
@@ -45,7 +46,8 @@ class DocumentMetaclass(type):
             attrs['_meta'] = meta
             attrs['_meta']['abstract'] = False  # 789: EmbeddedDocument shouldn't inherit abstract
 
-        if attrs['_meta'].get('allow_inheritance', ALLOW_INHERITANCE):
+        # If allow_inheritance is True, add a "_cls" string field to the attrs
+        if attrs['_meta'].get('allow_inheritance'):
             StringField = _import_class('StringField')
             attrs['_cls'] = StringField()
 
@@ -87,16 +89,17 @@ class DocumentMetaclass(type):
         # Ensure no duplicate db_fields
         duplicate_db_fields = [k for k, v in field_names.items() if v > 1]
         if duplicate_db_fields:
-            msg = ("Multiple db_fields defined for: %s " %
-                   ", ".join(duplicate_db_fields))
+            msg = ('Multiple db_fields defined for: %s ' %
+                   ', '.join(duplicate_db_fields))
             raise InvalidDocumentError(msg)
 
         # Set _fields and db_field maps
         attrs['_fields'] = doc_fields
-        attrs['_db_field_map'] = dict([(k, getattr(v, 'db_field', k))
-                                       for k, v in doc_fields.iteritems()])
-        attrs['_reverse_db_field_map'] = dict(
-            (v, k) for k, v in attrs['_db_field_map'].iteritems())
+        attrs['_db_field_map'] = {k: getattr(v, 'db_field', k)
+                                  for k, v in doc_fields.items()}
+        attrs['_reverse_db_field_map'] = {
+            v: k for k, v in attrs['_db_field_map'].items()
+        }
 
         attrs['_fields_ordered'] = tuple(i[1] for i in sorted(
                                          (v.creation_counter, v.name)
@@ -116,10 +119,8 @@ class DocumentMetaclass(type):
             if hasattr(base, '_meta'):
                 # Warn if allow_inheritance isn't set and prevent
                 # inheritance of classes where inheritance is set to False
-                allow_inheritance = base._meta.get('allow_inheritance',
-                                                   ALLOW_INHERITANCE)
-                if (allow_inheritance is not True and
-                        not base._meta.get('abstract')):
+                allow_inheritance = base._meta.get('allow_inheritance')
+                if not allow_inheritance and not base._meta.get('abstract'):
                     raise ValueError('Document %s may not be subclassed' %
                                      base.__name__)
 
@@ -161,7 +162,7 @@ class DocumentMetaclass(type):
         # module continues to use im_func and im_self, so the code below
         # copies __func__ into im_func and __self__ into im_self for
         # classmethod objects in Document derived classes.
-        if PY3:
+        if six.PY3:
             for val in new_class.__dict__.values():
                 if isinstance(val, classmethod):
                     f = val.__get__(new_class)
@@ -179,11 +180,11 @@ class DocumentMetaclass(type):
             if isinstance(f, CachedReferenceField):
 
                 if issubclass(new_class, EmbeddedDocument):
-                    raise InvalidDocumentError(
-                        "CachedReferenceFields is not allowed in EmbeddedDocuments")
+                    raise InvalidDocumentError('CachedReferenceFields is not '
+                                               'allowed in EmbeddedDocuments')
                 if not f.document_type:
                     raise InvalidDocumentError(
-                        "Document is not available to sync")
+                        'Document is not available to sync')
 
                 if f.auto_sync:
                     f.start_listener()
@@ -195,8 +196,8 @@ class DocumentMetaclass(type):
                                       'reverse_delete_rule',
                                       DO_NOTHING)
                 if isinstance(f, DictField) and delete_rule != DO_NOTHING:
-                    msg = ("Reverse delete rules are not supported "
-                           "for %s (field: %s)" %
+                    msg = ('Reverse delete rules are not supported '
+                           'for %s (field: %s)' %
                            (field.__class__.__name__, field.name))
                     raise InvalidDocumentError(msg)
 
@@ -204,16 +205,16 @@ class DocumentMetaclass(type):
 
             if delete_rule != DO_NOTHING:
                 if issubclass(new_class, EmbeddedDocument):
-                    msg = ("Reverse delete rules are not supported for "
-                           "EmbeddedDocuments (field: %s)" % field.name)
+                    msg = ('Reverse delete rules are not supported for '
+                           'EmbeddedDocuments (field: %s)' % field.name)
                     raise InvalidDocumentError(msg)
                 f.document_type.register_delete_rule(new_class,
                                                      field.name, delete_rule)
 
             if (field.name and hasattr(Document, field.name) and
                     EmbeddedDocument not in new_class.mro()):
-                msg = ("%s is a document method and not a valid "
-                       "field name" % field.name)
+                msg = ('%s is a document method and not a valid '
+                       'field name' % field.name)
                 raise InvalidDocumentError(msg)
 
         return new_class
@@ -271,6 +272,11 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
                 'index_drop_dups': False,
                 'index_opts': None,
                 'delete_rules': None,
+
+                # allow_inheritance can be True, False, and None. True means
+                # "allow inheritance", False means "don't allow inheritance",
+                # None means "do whatever your parent does, or don't allow
+                # inheritance if you're a top-level class".
                 'allow_inheritance': None,
             }
             attrs['_is_base_cls'] = True
@@ -303,7 +309,7 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
         # If parent wasn't an abstract class
         if (parent_doc_cls and 'collection' in attrs.get('_meta', {}) and
                 not parent_doc_cls._meta.get('abstract', True)):
-            msg = "Trying to set a collection on a subclass (%s)" % name
+            msg = 'Trying to set a collection on a subclass (%s)' % name
             warnings.warn(msg, SyntaxWarning)
             del attrs['_meta']['collection']
 
@@ -311,7 +317,7 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
         if attrs.get('_is_base_cls') or attrs['_meta'].get('abstract'):
             if (parent_doc_cls and
                     not parent_doc_cls._meta.get('abstract', False)):
-                msg = "Abstract document cannot have non-abstract base"
+                msg = 'Abstract document cannot have non-abstract base'
                 raise ValueError(msg)
             return super_new(cls, name, bases, attrs)
 
@@ -334,12 +340,16 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
 
         meta.merge(attrs.get('_meta', {}))  # Top level meta
 
-        # Only simple classes (direct subclasses of Document)
-        # may set allow_inheritance to False
+        # Only simple classes (i.e. direct subclasses of Document) may set
+        # allow_inheritance to False. If the base Document allows inheritance,
+        # none of its subclasses can override allow_inheritance to False.
         simple_class = all([b._meta.get('abstract')
                             for b in flattened_bases if hasattr(b, '_meta')])
-        if (not simple_class and meta['allow_inheritance'] is False and
-                not meta['abstract']):
+        if (
+            not simple_class and
+            meta['allow_inheritance'] is False and
+            not meta['abstract']
+        ):
             raise ValueError('Only direct subclasses of Document may set '
                              '"allow_inheritance" to False')
 

mongoengine/connection.py

@@ -1,11 +1,14 @@
 from pymongo import MongoClient, ReadPreference, uri_parser
-from mongoengine.python_support import (IS_PYMONGO_3, str_types)
+import six
 
-__all__ = ['ConnectionError', 'connect', 'register_connection',
+from mongoengine.python_support import IS_PYMONGO_3
+
+__all__ = ['MongoEngineConnectionError', 'connect', 'register_connection',
            'DEFAULT_CONNECTION_NAME']
 
 
 DEFAULT_CONNECTION_NAME = 'default'
+
 if IS_PYMONGO_3:
     READ_PREFERENCE = ReadPreference.PRIMARY
 else:
@@ -13,7 +16,10 @@ else:
     READ_PREFERENCE = False
 
 
-class ConnectionError(Exception):
+class MongoEngineConnectionError(Exception):
+    """Error raised when the database connection can't be established or
+    when a connection with a requested alias can't be retrieved.
+    """
     pass
 
 
@@ -24,7 +30,9 @@ _dbs = {}
 
 def register_connection(alias, name=None, host=None, port=None,
                         read_preference=READ_PREFERENCE,
-                        username=None, password=None, authentication_source=None,
+                        username=None, password=None,
+                        authentication_source=None,
+                        authentication_mechanism=None,
                         **kwargs):
     """Add a connection.
 
@@ -38,14 +46,15 @@ def register_connection(alias, name=None, host=None, port=None,
     :param username: username to authenticate with
     :param password: password to authenticate with
     :param authentication_source: database to authenticate against
+    :param authentication_mechanism: database authentication mechanisms.
+        By default, use SCRAM-SHA-1 with MongoDB 3.0 and later,
+        MONGODB-CR (MongoDB Challenge Response protocol) for older servers.
     :param is_mock: explicitly use mongomock for this connection
         (can also be done by using `mongomock://` as db host prefix)
     :param kwargs: allow ad-hoc parameters to be passed into the pymongo driver
 
     .. versionchanged:: 0.10.6 - added mongomock support
     """
-    global _connection_settings
-
     conn_settings = {
         'name': name or 'test',
         'host': host or 'localhost',
@@ -53,35 +62,45 @@ def register_connection(alias, name=None, host=None, port=None,
         'read_preference': read_preference,
         'username': username,
         'password': password,
-        'authentication_source': authentication_source
+        'authentication_source': authentication_source,
+        'authentication_mechanism': authentication_mechanism
     }
 
     conn_host = conn_settings['host']
-    # host can be a list or a string, so if string, force to a list
-    if isinstance(conn_host, str_types):
+
+    # Host can be a list or a string, so if string, force to a list.
+    if isinstance(conn_host, six.string_types):
         conn_host = [conn_host]
 
     resolved_hosts = []
     for entity in conn_host:
-        # Handle uri style connections
+
+        # Handle Mongomock
         if entity.startswith('mongomock://'):
             conn_settings['is_mock'] = True
             # `mongomock://` is not a valid url prefix and must be replaced by `mongodb://`
             resolved_hosts.append(entity.replace('mongomock://', 'mongodb://', 1))
+
+        # Handle URI style connections, only updating connection params which
+        # were explicitly specified in the URI.
         elif '://' in entity:
             uri_dict = uri_parser.parse_uri(entity)
             resolved_hosts.append(entity)
-            conn_settings.update({
-                'name': uri_dict.get('database') or name,
-                'username': uri_dict.get('username'),
-                'password': uri_dict.get('password'),
-                'read_preference': read_preference,
-            })
+
+            if uri_dict.get('database'):
+                conn_settings['name'] = uri_dict.get('database')
+
+            for param in ('read_preference', 'username', 'password'):
+                if uri_dict.get(param):
+                    conn_settings[param] = uri_dict[param]
+
             uri_options = uri_dict['options']
             if 'replicaset' in uri_options:
                 conn_settings['replicaSet'] = True
             if 'authsource' in uri_options:
                 conn_settings['authentication_source'] = uri_options['authsource']
+            if 'authmechanism' in uri_options:
+                conn_settings['authentication_mechanism'] = uri_options['authmechanism']
         else:
             resolved_hosts.append(entity)
     conn_settings['host'] = resolved_hosts
@@ -95,9 +114,7 @@ def register_connection(alias, name=None, host=None, port=None,
 
 
 def disconnect(alias=DEFAULT_CONNECTION_NAME):
-    global _connections
-    global _dbs
-
+    """Close the connection with a given alias."""
     if alias in _connections:
         get_connection(alias=alias).close()
         del _connections[alias]
@@ -106,69 +123,99 @@ def disconnect(alias=DEFAULT_CONNECTION_NAME):
 
 
 def get_connection(alias=DEFAULT_CONNECTION_NAME, reconnect=False):
-    global _connections
+    """Return a connection with a given alias."""
+
     # Connect to the database if not already connected
     if reconnect:
         disconnect(alias)
 
-    if alias not in _connections:
-        if alias not in _connection_settings:
-            msg = 'Connection with alias "%s" has not been defined' % alias
-            if alias == DEFAULT_CONNECTION_NAME:
-                msg = 'You have not defined a default connection'
-            raise ConnectionError(msg)
-        conn_settings = _connection_settings[alias].copy()
+    # If the requested alias already exists in the _connections list, return
+    # it immediately.
+    if alias in _connections:
+        return _connections[alias]
 
-        conn_settings.pop('name', None)
-        conn_settings.pop('username', None)
-        conn_settings.pop('password', None)
-        conn_settings.pop('authentication_source', None)
-
-        is_mock = conn_settings.pop('is_mock', None)
-        if is_mock:
-            # Use MongoClient from mongomock
-            try:
-                import mongomock
-            except ImportError:
-                raise RuntimeError('You need mongomock installed '
-                                   'to mock MongoEngine.')
-            connection_class = mongomock.MongoClient
+    # Validate that the requested alias exists in the _connection_settings.
+    # Raise MongoEngineConnectionError if it doesn't.
+    if alias not in _connection_settings:
+        if alias == DEFAULT_CONNECTION_NAME:
+            msg = 'You have not defined a default connection'
         else:
-            # Use MongoClient from pymongo
-            connection_class = MongoClient
+            msg = 'Connection with alias "%s" has not been defined' % alias
+        raise MongoEngineConnectionError(msg)
+
+    def _clean_settings(settings_dict):
+        irrelevant_fields = set([
+            'name', 'username', 'password', 'authentication_source',
+            'authentication_mechanism'
+        ])
+        return {
+            k: v for k, v in settings_dict.items()
+            if k not in irrelevant_fields
+        }
+
+    # Retrieve a copy of the connection settings associated with the requested
+    # alias and remove the database name and authentication info (we don't
+    # care about them at this point).
+    conn_settings = _clean_settings(_connection_settings[alias].copy())
+
+    # Determine if we should use PyMongo's or mongomock's MongoClient.
+    is_mock = conn_settings.pop('is_mock', False)
+    if is_mock:
+        try:
+            import mongomock
+        except ImportError:
+            raise RuntimeError('You need mongomock installed to mock '
+                               'MongoEngine.')
+        connection_class = mongomock.MongoClient
+    else:
+        connection_class = MongoClient
+
+        # For replica set connections with PyMongo 2.x, use
+        # MongoReplicaSetClient.
+        # TODO remove this once we stop supporting PyMongo 2.x.
+        if 'replicaSet' in conn_settings and not IS_PYMONGO_3:
+            connection_class = MongoReplicaSetClient
+            conn_settings['hosts_or_uri'] = conn_settings.pop('host', None)
+
+            # hosts_or_uri has to be a string, so if 'host' was provided
+            # as a list, join its parts and separate them by ','
+            if isinstance(conn_settings['hosts_or_uri'], list):
+                conn_settings['hosts_or_uri'] = ','.join(
+                    conn_settings['hosts_or_uri'])
 
-        if 'replicaSet' in conn_settings:
             # Discard port since it can't be used on MongoReplicaSetClient
             conn_settings.pop('port', None)
-            # Discard replicaSet if not base string
-            if not isinstance(conn_settings['replicaSet'], basestring):
-                conn_settings.pop('replicaSet', None)
-            if not IS_PYMONGO_3:
-                connection_class = MongoReplicaSetClient
-                conn_settings['hosts_or_uri'] = conn_settings.pop('host', None)
 
+    # Iterate over all of the connection settings and if a connection with
+    # the same parameters is already established, use it instead of creating
+    # a new one.
+    existing_connection = None
+    connection_settings_iterator = (
+        (db_alias, settings.copy())
+        for db_alias, settings in _connection_settings.items()
+    )
+    for db_alias, connection_settings in connection_settings_iterator:
+        connection_settings = _clean_settings(connection_settings)
+        if conn_settings == connection_settings and _connections.get(db_alias):
+            existing_connection = _connections[db_alias]
+            break
+
+    # If an existing connection was found, assign it to the new alias
+    if existing_connection:
+        _connections[alias] = existing_connection
+    else:
+        # Otherwise, create the new connection for this alias. Raise
+        # MongoEngineConnectionError if it can't be established.
         try:
-            connection = None
-            # check for shared connections
-            connection_settings_iterator = (
-                (db_alias, settings.copy()) for db_alias, settings in _connection_settings.iteritems())
-            for db_alias, connection_settings in connection_settings_iterator:
-                connection_settings.pop('name', None)
-                connection_settings.pop('username', None)
-                connection_settings.pop('password', None)
-                connection_settings.pop('authentication_source', None)
-                if conn_settings == connection_settings and _connections.get(db_alias, None):
-                    connection = _connections[db_alias]
-                    break
+            _connections[alias] = connection_class(**conn_settings)
+        except Exception as e:
+            raise MongoEngineConnectionError(
+                'Cannot connect to database %s :\n%s' % (alias, e))
 
-            _connections[alias] = connection if connection else connection_class(**conn_settings)
-        except Exception, e:
-            raise ConnectionError("Cannot connect to database %s :\n%s" % (alias, e))
     return _connections[alias]
 
 
 def get_db(alias=DEFAULT_CONNECTION_NAME, reconnect=False):
-    global _dbs
     if reconnect:
         disconnect(alias)
 
@@ -176,11 +223,13 @@ def get_db(alias=DEFAULT_CONNECTION_NAME, reconnect=False):
         conn = get_connection(alias)
         conn_settings = _connection_settings[alias]
         db = conn[conn_settings['name']]
+        auth_kwargs = {'source': conn_settings['authentication_source']}
+        if conn_settings['authentication_mechanism'] is not None:
+            auth_kwargs['mechanism'] = conn_settings['authentication_mechanism']
         # Authenticate if necessary
-        if conn_settings['username'] and conn_settings['password']:
-            db.authenticate(conn_settings['username'],
-                            conn_settings['password'],
-                            source=conn_settings['authentication_source'])
+        if conn_settings['username'] and (conn_settings['password'] or
+                                          conn_settings['authentication_mechanism'] == 'MONGODB-X509'):
+            db.authenticate(conn_settings['username'], conn_settings['password'], **auth_kwargs)
         _dbs[alias] = db
     return _dbs[alias]
 
@@ -197,7 +246,6 @@ def connect(db=None, alias=DEFAULT_CONNECTION_NAME, **kwargs):
 
     .. versionchanged:: 0.6 - added multiple database support.
     """
-    global _connections
     if alias not in _connections:
         register_connection(alias, db, **kwargs)
 

mongoengine/context_managers.py

@@ -2,12 +2,12 @@ from mongoengine.common import _import_class
 from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
 
 
-__all__ = ("switch_db", "switch_collection", "no_dereference",
-           "no_sub_classes", "query_counter")
+__all__ = ('switch_db', 'switch_collection', 'no_dereference',
+           'no_sub_classes', 'query_counter')
 
 
 class switch_db(object):
-    """ switch_db alias context manager.
+    """switch_db alias context manager.
 
     Example ::
 
@@ -18,15 +18,14 @@ class switch_db(object):
         class Group(Document):
             name = StringField()
 
-        Group(name="test").save()  # Saves in the default db
+        Group(name='test').save()  # Saves in the default db
 
         with switch_db(Group, 'testdb-1') as Group:
-            Group(name="hello testdb!").save()  # Saves in testdb-1
-
+            Group(name='hello testdb!').save()  # Saves in testdb-1
     """
 
     def __init__(self, cls, db_alias):
-        """ Construct the switch_db context manager
+        """Construct the switch_db context manager
 
         :param cls: the class to change the registered db
         :param db_alias: the name of the specific database to use
@@ -34,37 +33,36 @@ class switch_db(object):
         self.cls = cls
         self.collection = cls._get_collection()
         self.db_alias = db_alias
-        self.ori_db_alias = cls._meta.get("db_alias", DEFAULT_CONNECTION_NAME)
+        self.ori_db_alias = cls._meta.get('db_alias', DEFAULT_CONNECTION_NAME)
 
     def __enter__(self):
-        """ change the db_alias and clear the cached collection """
-        self.cls._meta["db_alias"] = self.db_alias
+        """Change the db_alias and clear the cached collection."""
+        self.cls._meta['db_alias'] = self.db_alias
         self.cls._collection = None
         return self.cls
 
     def __exit__(self, t, value, traceback):
-        """ Reset the db_alias and collection """
-        self.cls._meta["db_alias"] = self.ori_db_alias
+        """Reset the db_alias and collection."""
+        self.cls._meta['db_alias'] = self.ori_db_alias
         self.cls._collection = self.collection
 
 
 class switch_collection(object):
-    """ switch_collection alias context manager.
+    """switch_collection alias context manager.
 
     Example ::
 
         class Group(Document):
             name = StringField()
 
-        Group(name="test").save()  # Saves in the default db
+        Group(name='test').save()  # Saves in the default db
 
         with switch_collection(Group, 'group1') as Group:
-            Group(name="hello testdb!").save()  # Saves in group1 collection
-
+            Group(name='hello testdb!').save()  # Saves in group1 collection
     """
 
     def __init__(self, cls, collection_name):
-        """ Construct the switch_collection context manager
+        """Construct the switch_collection context manager.
 
         :param cls: the class to change the registered db
         :param collection_name: the name of the collection to use
@@ -75,7 +73,7 @@ class switch_collection(object):
         self.collection_name = collection_name
 
     def __enter__(self):
-        """ change the _get_collection_name and clear the cached collection """
+        """Change the _get_collection_name and clear the cached collection."""
 
         @classmethod
         def _get_collection_name(cls):
@@ -86,24 +84,23 @@ class switch_collection(object):
         return self.cls
 
     def __exit__(self, t, value, traceback):
-        """ Reset the collection """
+        """Reset the collection."""
         self.cls._collection = self.ori_collection
         self.cls._get_collection_name = self.ori_get_collection_name
 
 
 class no_dereference(object):
-    """ no_dereference context manager.
+    """no_dereference context manager.
 
     Turns off all dereferencing in Documents for the duration of the context
     manager::
 
         with no_dereference(Group) as Group:
             Group.objects.find()
-
     """
 
     def __init__(self, cls):
-        """ Construct the no_dereference context manager.
+        """Construct the no_dereference context manager.
 
         :param cls: the class to turn dereferencing off on
         """
@@ -119,103 +116,102 @@ class no_dereference(object):
                                                ComplexBaseField))]
 
     def __enter__(self):
-        """ change the objects default and _auto_dereference values"""
+        """Change the objects default and _auto_dereference values."""
         for field in self.deref_fields:
             self.cls._fields[field]._auto_dereference = False
         return self.cls
 
     def __exit__(self, t, value, traceback):
-        """ Reset the default and _auto_dereference values"""
+        """Reset the default and _auto_dereference values."""
         for field in self.deref_fields:
             self.cls._fields[field]._auto_dereference = True
         return self.cls
 
 
 class no_sub_classes(object):
-    """ no_sub_classes context manager.
+    """no_sub_classes context manager.
 
     Only returns instances of this class and no sub (inherited) classes::
 
         with no_sub_classes(Group) as Group:
             Group.objects.find()
-
     """
 
     def __init__(self, cls):
-        """ Construct the no_sub_classes context manager.
+        """Construct the no_sub_classes context manager.
 
         :param cls: the class to turn querying sub classes on
         """
         self.cls = cls
 
     def __enter__(self):
-        """ change the objects default and _auto_dereference values"""
+        """Change the objects default and _auto_dereference values."""
         self.cls._all_subclasses = self.cls._subclasses
         self.cls._subclasses = (self.cls,)
         return self.cls
 
     def __exit__(self, t, value, traceback):
-        """ Reset the default and _auto_dereference values"""
+        """Reset the default and _auto_dereference values."""
         self.cls._subclasses = self.cls._all_subclasses
         delattr(self.cls, '_all_subclasses')
         return self.cls
 
 
 class query_counter(object):
-    """ Query_counter context manager to get the number of queries. """
+    """Query_counter context manager to get the number of queries."""
 
     def __init__(self):
-        """ Construct the query_counter. """
+        """Construct the query_counter."""
         self.counter = 0
         self.db = get_db()
 
     def __enter__(self):
-        """ On every with block we need to drop the profile collection. """
+        """On every with block we need to drop the profile collection."""
         self.db.set_profiling_level(0)
         self.db.system.profile.drop()
         self.db.set_profiling_level(2)
         return self
 
     def __exit__(self, t, value, traceback):
-        """ Reset the profiling level. """
+        """Reset the profiling level."""
         self.db.set_profiling_level(0)
 
     def __eq__(self, value):
-        """ == Compare querycounter. """
+        """== Compare querycounter."""
         counter = self._get_count()
         return value == counter
 
     def __ne__(self, value):
-        """ != Compare querycounter. """
+        """!= Compare querycounter."""
         return not self.__eq__(value)
 
     def __lt__(self, value):
-        """ < Compare querycounter. """
+        """< Compare querycounter."""
         return self._get_count() < value
 
     def __le__(self, value):
-        """ <= Compare querycounter. """
+        """<= Compare querycounter."""
         return self._get_count() <= value
 
     def __gt__(self, value):
-        """ > Compare querycounter. """
+        """> Compare querycounter."""
         return self._get_count() > value
 
     def __ge__(self, value):
-        """ >= Compare querycounter. """
+        """>= Compare querycounter."""
         return self._get_count() >= value
 
     def __int__(self):
-        """ int representation. """
+        """int representation."""
         return self._get_count()
 
     def __repr__(self):
-        """ repr query_counter as the number of queries. """
+        """repr query_counter as the number of queries."""
         return u"%s" % self._get_count()
 
     def _get_count(self):
-        """ Get the number of queries. """
-        ignore_query = {"ns": {"$ne": "%s.system.indexes" % self.db.name}}
+        """Get the number of queries."""
+        ignore_query = {'ns': {'$ne': '%s.system.indexes' % self.db.name}}
         count = self.db.system.profile.find(ignore_query).count() - self.counter
         self.counter += 1
         return count

mongoengine/dereference.py

@@ -1,14 +1,12 @@
 from bson import DBRef, SON
+import six
 
-from .base import (
-    BaseDict, BaseList, EmbeddedDocumentList,
-    TopLevelDocumentMetaclass, get_document
-)
-from .connection import get_db
-from .document import Document, EmbeddedDocument
-from .fields import DictField, ListField, MapField, ReferenceField
-from .python_support import txt_type
-from .queryset import QuerySet
+from mongoengine.base import (BaseDict, BaseList, EmbeddedDocumentList,
+                              TopLevelDocumentMetaclass, get_document)
+from mongoengine.connection import get_db
+from mongoengine.document import Document, EmbeddedDocument
+from mongoengine.fields import DictField, ListField, MapField, ReferenceField
+from mongoengine.queryset import QuerySet
 
 
 class DeReference(object):
@@ -25,7 +23,7 @@ class DeReference(object):
             :class:`~mongoengine.base.ComplexBaseField`
         :param get: A boolean determining if being called by __get__
         """
-        if items is None or isinstance(items, basestring):
+        if items is None or isinstance(items, six.string_types):
             return items
 
         # cheapest way to convert a queryset to a list
@@ -68,11 +66,11 @@ class DeReference(object):
 
                         items = _get_items(items)
                     else:
-                        items = dict([
-                            (k, field.to_python(v))
-                            if not isinstance(v, (DBRef, Document)) else (k, v)
-                            for k, v in items.iteritems()]
-                        )
+                        items = {
+                            k: (v if isinstance(v, (DBRef, Document))
+                                else field.to_python(v))
+                            for k, v in items.iteritems()
+                        }
 
         self.reference_map = self._find_references(items)
         self.object_map = self._fetch_objects(doc_type=doc_type)
@@ -90,14 +88,14 @@ class DeReference(object):
             return reference_map
 
         # Determine the iterator to use
-        if not hasattr(items, 'items'):
-            iterator = enumerate(items)
+        if isinstance(items, dict):
+            iterator = items.values()
         else:
-            iterator = items.iteritems()
+            iterator = items
 
         # Recursively find dbreferences
         depth += 1
-        for k, item in iterator:
+        for item in iterator:
             if isinstance(item, (Document, EmbeddedDocument)):
                 for field_name, field in item._fields.iteritems():
                     v = item._data.get(field_name, None)
@@ -151,7 +149,7 @@ class DeReference(object):
                     references = get_db()[collection].find({'_id': {'$in': refs}})
                     for ref in references:
                         if '_cls' in ref:
-                            doc = get_document(ref["_cls"])._from_son(ref)
+                            doc = get_document(ref['_cls'])._from_son(ref)
                         elif doc_type is None:
                             doc = get_document(
                                 ''.join(x.capitalize()
@@ -218,7 +216,7 @@ class DeReference(object):
             if k in self.object_map and not is_list:
                 data[k] = self.object_map[k]
             elif isinstance(v, (Document, EmbeddedDocument)):
-                for field_name, field in v._fields.iteritems():
+                for field_name in v._fields:
                     v = data[k]._data.get(field_name, None)
                     if isinstance(v, DBRef):
                         data[k]._data[field_name] = self.object_map.get(
@@ -227,7 +225,7 @@ class DeReference(object):
                         data[k]._data[field_name] = self.object_map.get(
                             (v['_ref'].collection, v['_ref'].id), v)
                     elif isinstance(v, (dict, list, tuple)) and depth <= self.max_depth:
-                        item_name = txt_type("{0}.{1}.{2}").format(name, k, field_name)
+                        item_name = six.text_type('{0}.{1}.{2}').format(name, k, field_name)
                         data[k]._data[field_name] = self._attach_objects(v, depth, instance=instance, name=item_name)
             elif isinstance(v, (dict, list, tuple)) and depth <= self.max_depth:
                 item_name = '%s.%s' % (name, k) if name else name

mongoengine/document.py

@@ -4,18 +4,12 @@ import warnings
 from bson.dbref import DBRef
 import pymongo
 from pymongo.read_preferences import ReadPreference
+import six
 
 from mongoengine import signals
-from mongoengine.base import (
-    ALLOW_INHERITANCE,
-    BaseDict,
-    BaseDocument,
-    BaseList,
-    DocumentMetaclass,
-    EmbeddedDocumentList,
-    TopLevelDocumentMetaclass,
-    get_document
-)
+from mongoengine.base import (BaseDict, BaseDocument, BaseList,
+                              DocumentMetaclass, EmbeddedDocumentList,
+                              TopLevelDocumentMetaclass, get_document)
 from mongoengine.common import _import_class
 from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
 from mongoengine.context_managers import switch_collection, switch_db
@@ -31,12 +25,10 @@ __all__ = ('Document', 'EmbeddedDocument', 'DynamicDocument',
 
 
 def includes_cls(fields):
-    """ Helper function used for ensuring and comparing indexes
-    """
-
+    """Helper function used for ensuring and comparing indexes."""
     first_field = None
     if len(fields):
-        if isinstance(fields[0], basestring):
+        if isinstance(fields[0], six.string_types):
             first_field = fields[0]
         elif isinstance(fields[0], (list, tuple)) and len(fields[0]):
             first_field = fields[0][0]
@@ -57,9 +49,8 @@ class EmbeddedDocument(BaseDocument):
     to create a specialised version of the embedded document that will be
     stored in the same collection. To facilitate this behaviour a `_cls`
     field is added to documents (hidden though the MongoEngine interface).
-    To disable this behaviour and remove the dependence on the presence of
-    `_cls` set :attr:`allow_inheritance` to ``False`` in the :attr:`meta`
-    dictionary.
+    To enable this behaviour set :attr:`allow_inheritance` to ``True`` in the
+    :attr:`meta` dictionary.
     """
 
     __slots__ = ('_instance', )
@@ -82,6 +73,15 @@ class EmbeddedDocument(BaseDocument):
     def __ne__(self, other):
         return not self.__eq__(other)
 
+    def to_mongo(self, *args, **kwargs):
+        data = super(EmbeddedDocument, self).to_mongo(*args, **kwargs)
+
+        # remove _id from the SON if it's in it and it's None
+        if '_id' in data and data['_id'] is None:
+            del data['_id']
+
+        return data
+
     def save(self, *args, **kwargs):
         self._instance.save(*args, **kwargs)
 
@@ -106,9 +106,8 @@ class Document(BaseDocument):
     create a specialised version of the document that will be stored in the
     same collection. To facilitate this behaviour a `_cls`
     field is added to documents (hidden though the MongoEngine interface).
-    To disable this behaviour and remove the dependence on the presence of
-    `_cls` set :attr:`allow_inheritance` to ``False`` in the :attr:`meta`
-    dictionary.
+    To enable this behaviourset :attr:`allow_inheritance` to ``True`` in the
+    :attr:`meta` dictionary.
 
     A :class:`~mongoengine.Document` may use a **Capped Collection** by
     specifying :attr:`max_documents` and :attr:`max_size` in the :attr:`meta`
@@ -149,26 +148,22 @@ class Document(BaseDocument):
 
     __slots__ = ('__objects',)
 
-    def pk():
-        """Primary key alias
-        """
+    @property
+    def pk(self):
+        """Get the primary key."""
+        if 'id_field' not in self._meta:
+            return None
+        return getattr(self, self._meta['id_field'])
 
-        def fget(self):
-            if 'id_field' not in self._meta:
-                return None
-            return getattr(self, self._meta['id_field'])
-
-        def fset(self, value):
-            return setattr(self, self._meta['id_field'], value)
-
-        return property(fget, fset)
-
-    pk = pk()
+    @pk.setter
+    def pk(self, value):
+        """Set the primary key."""
+        return setattr(self, self._meta['id_field'], value)
 
     @classmethod
     def _get_db(cls):
         """Some Model using other db_alias"""
-        return get_db(cls._meta.get("db_alias", DEFAULT_CONNECTION_NAME))
+        return get_db(cls._meta.get('db_alias', DEFAULT_CONNECTION_NAME))
 
     @classmethod
     def _get_collection(cls):
@@ -211,7 +206,20 @@ class Document(BaseDocument):
                 cls.ensure_indexes()
         return cls._collection
 
-    def modify(self, query={}, **update):
+    def to_mongo(self, *args, **kwargs):
+        data = super(Document, self).to_mongo(*args, **kwargs)
+
+        # If '_id' is None, try and set it from self._data. If that
+        # doesn't exist either, remote '_id' from the SON completely.
+        if data['_id'] is None:
+            if self._data.get('id') is None:
+                del data['_id']
+            else:
+                data['_id'] = self._data['id']
+
+        return data
+
+    def modify(self, query=None, **update):
         """Perform an atomic update of the document in the database and reload
         the document object using updated version.
 
@@ -225,17 +233,19 @@ class Document(BaseDocument):
             database matches the query
         :param update: Django-style update keyword arguments
         """
+        if query is None:
+            query = {}
 
         if self.pk is None:
-            raise InvalidDocumentError("The document does not have a primary key.")
+            raise InvalidDocumentError('The document does not have a primary key.')
 
-        id_field = self._meta["id_field"]
+        id_field = self._meta['id_field']
         query = query.copy() if isinstance(query, dict) else query.to_query(self)
 
         if id_field not in query:
             query[id_field] = self.pk
         elif query[id_field] != self.pk:
-            raise InvalidQueryError("Invalid document modify query: it must modify only this document.")
+            raise InvalidQueryError('Invalid document modify query: it must modify only this document.')
 
         updated = self._qs(**query).modify(new=True, **update)
         if updated is None:
@@ -310,7 +320,7 @@ class Document(BaseDocument):
             self.validate(clean=clean)
 
         if write_concern is None:
-            write_concern = {"w": 1}
+            write_concern = {'w': 1}
 
         doc = self.to_mongo()
 
@@ -347,7 +357,7 @@ class Document(BaseDocument):
                 else:
                     select_dict = {}
                 select_dict['_id'] = object_id
-                shard_key = self.__class__._meta.get('shard_key', tuple())
+                shard_key = self._meta.get('shard_key', tuple())
                 for k in shard_key:
                     path = self._lookup_field(k.split('.'))
                     actual_key = [p.db_field for p in path]
@@ -358,7 +368,7 @@ class Document(BaseDocument):
 
                 def is_new_object(last_error):
                     if last_error is not None:
-                        updated = last_error.get("updatedExisting")
+                        updated = last_error.get('updatedExisting')
                         if updated is not None:
                             return not updated
                     return created
@@ -366,14 +376,14 @@ class Document(BaseDocument):
                 update_query = {}
 
                 if updates:
-                    update_query["$set"] = updates
+                    update_query['$set'] = updates
                 if removals:
-                    update_query["$unset"] = removals
+                    update_query['$unset'] = removals
                 if updates or removals:
                     upsert = save_condition is None
                     last_error = collection.update(select_dict, update_query,
                                                    upsert=upsert, **write_concern)
-                    if not upsert and last_error["n"] == 0:
+                    if not upsert and last_error['n'] == 0:
                         raise SaveConditionError('Race condition preventing'
                                                  ' document update detected')
                     created = is_new_object(last_error)
@@ -384,26 +394,27 @@ class Document(BaseDocument):
 
             if cascade:
                 kwargs = {
-                    "force_insert": force_insert,
-                    "validate": validate,
-                    "write_concern": write_concern,
-                    "cascade": cascade
+                    'force_insert': force_insert,
+                    'validate': validate,
+                    'write_concern': write_concern,
+                    'cascade': cascade
                 }
                 if cascade_kwargs:  # Allow granular control over cascades
                     kwargs.update(cascade_kwargs)
                 kwargs['_refs'] = _refs
                 self.cascade_save(**kwargs)
-        except pymongo.errors.DuplicateKeyError, err:
+        except pymongo.errors.DuplicateKeyError as err:
             message = u'Tried to save duplicate unique keys (%s)'
-            raise NotUniqueError(message % unicode(err))
-        except pymongo.errors.OperationFailure, err:
+            raise NotUniqueError(message % six.text_type(err))
+        except pymongo.errors.OperationFailure as err:
             message = 'Could not save document (%s)'
-            if re.match('^E1100[01] duplicate key', unicode(err)):
+            if re.match('^E1100[01] duplicate key', six.text_type(err)):
                 # E11000 - duplicate key error index
                 # E11001 - duplicate key on update
                 message = u'Tried to save duplicate unique keys (%s)'
-                raise NotUniqueError(message % unicode(err))
-            raise OperationError(message % unicode(err))
+                raise NotUniqueError(message % six.text_type(err))
+            raise OperationError(message % six.text_type(err))
+
         id_field = self._meta['id_field']
         if created or id_field not in self._meta.get('shard_key', []):
             self[id_field] = self._fields[id_field].to_python(object_id)
@@ -414,10 +425,11 @@ class Document(BaseDocument):
         self._created = False
         return self
 
-    def cascade_save(self, *args, **kwargs):
-        """Recursively saves any references /
-           generic references on the document"""
-        _refs = kwargs.get('_refs', []) or []
+    def cascade_save(self, **kwargs):
+        """Recursively save any references and generic references on the
+        document.
+        """
+        _refs = kwargs.get('_refs') or []
 
         ReferenceField = _import_class('ReferenceField')
         GenericReferenceField = _import_class('GenericReferenceField')
@@ -443,16 +455,17 @@ class Document(BaseDocument):
 
     @property
     def _qs(self):
-        """
-        Returns the queryset to use for updating / reloading / deletions
-        """
+        """Return the queryset to use for updating / reloading / deletions."""
         if not hasattr(self, '__objects'):
             self.__objects = QuerySet(self, self._get_collection())
         return self.__objects
 
     @property
     def _object_key(self):
-        """Dict to identify object in collection
+        """Get the query dict that can be used to fetch this object from
+        the database. Most of the time it's a simple PK lookup, but in
+        case of a sharded collection with a compound shard key, it can
+        contain a more complex query.
         """
         select_dict = {'pk': self.pk}
         shard_key = self.__class__._meta.get('shard_key', tuple())
@@ -472,11 +485,11 @@ class Document(BaseDocument):
         Raises :class:`OperationError` if called on an object that has not yet
         been saved.
         """
-        if not self.pk:
+        if self.pk is None:
             if kwargs.get('upsert', False):
                 query = self.to_mongo()
-                if "_cls" in query:
-                    del query["_cls"]
+                if '_cls' in query:
+                    del query['_cls']
                 return self._qs.filter(**query).update_one(**kwargs)
             else:
                 raise OperationError(
@@ -513,7 +526,7 @@ class Document(BaseDocument):
         try:
             self._qs.filter(
                 **self._object_key).delete(write_concern=write_concern, _from_doc_delete=True)
-        except pymongo.errors.OperationFailure, err:
+        except pymongo.errors.OperationFailure as err:
             message = u'Could not delete document (%s)' % err.message
             raise OperationError(message)
         signals.post_delete.send(self.__class__, document=self, **signal_kwargs)
@@ -601,11 +614,12 @@ class Document(BaseDocument):
         if fields and isinstance(fields[0], int):
             max_depth = fields[0]
             fields = fields[1:]
-        elif "max_depth" in kwargs:
-            max_depth = kwargs["max_depth"]
+        elif 'max_depth' in kwargs:
+            max_depth = kwargs['max_depth']
+
+        if self.pk is None:
+            raise self.DoesNotExist('Document does not exist')
 
-        if not self.pk:
-            raise self.DoesNotExist("Document does not exist")
         obj = self._qs.read_preference(ReadPreference.PRIMARY).filter(
             **self._object_key).only(*fields).limit(
             1).select_related(max_depth=max_depth)
@@ -613,7 +627,7 @@ class Document(BaseDocument):
         if obj:
             obj = obj[0]
         else:
-            raise self.DoesNotExist("Document does not exist")
+            raise self.DoesNotExist('Document does not exist')
 
         for field in obj._data:
             if not fields or field in fields:
@@ -655,8 +669,8 @@ class Document(BaseDocument):
     def to_dbref(self):
         """Returns an instance of :class:`~bson.dbref.DBRef` useful in
         `__raw__` queries."""
-        if not self.pk:
-            msg = "Only saved documents can have a valid dbref"
+        if self.pk is None:
+            msg = 'Only saved documents can have a valid dbref'
             raise OperationError(msg)
         return DBRef(self.__class__._get_collection_name(), self.pk)
 
@@ -711,7 +725,7 @@ class Document(BaseDocument):
         fields = index_spec.pop('fields')
         drop_dups = kwargs.get('drop_dups', False)
         if IS_PYMONGO_3 and drop_dups:
-            msg = "drop_dups is deprecated and is removed when using PyMongo 3+."
+            msg = 'drop_dups is deprecated and is removed when using PyMongo 3+.'
             warnings.warn(msg, DeprecationWarning)
         elif not IS_PYMONGO_3:
             index_spec['drop_dups'] = drop_dups
@@ -737,7 +751,7 @@ class Document(BaseDocument):
             will be removed if PyMongo3+ is used
         """
         if IS_PYMONGO_3 and drop_dups:
-            msg = "drop_dups is deprecated and is removed when using PyMongo 3+."
+            msg = 'drop_dups is deprecated and is removed when using PyMongo 3+.'
             warnings.warn(msg, DeprecationWarning)
         elif not IS_PYMONGO_3:
             kwargs.update({'drop_dups': drop_dups})
@@ -757,7 +771,7 @@ class Document(BaseDocument):
         index_opts = cls._meta.get('index_opts') or {}
         index_cls = cls._meta.get('index_cls', True)
         if IS_PYMONGO_3 and drop_dups:
-            msg = "drop_dups is deprecated and is removed when using PyMongo 3+."
+            msg = 'drop_dups is deprecated and is removed when using PyMongo 3+.'
             warnings.warn(msg, DeprecationWarning)
 
         collection = cls._get_collection()
@@ -795,8 +809,7 @@ class Document(BaseDocument):
 
         # If _cls is being used (for polymorphism), it needs an index,
         # only if another index doesn't begin with _cls
-        if (index_cls and not cls_indexed and
-                cls._meta.get('allow_inheritance', ALLOW_INHERITANCE) is True):
+        if index_cls and not cls_indexed and cls._meta.get('allow_inheritance'):
 
             # we shouldn't pass 'cls' to the collection.ensureIndex options
             # because of https://jira.mongodb.org/browse/SERVER-769
@@ -866,16 +879,15 @@ class Document(BaseDocument):
         # finish up by appending { '_id': 1 } and { '_cls': 1 }, if needed
         if [(u'_id', 1)] not in indexes:
             indexes.append([(u'_id', 1)])
-        if (cls._meta.get('index_cls', True) and
-                cls._meta.get('allow_inheritance', ALLOW_INHERITANCE) is True):
+        if cls._meta.get('index_cls', True) and cls._meta.get('allow_inheritance'):
             indexes.append([(u'_cls', 1)])
 
         return indexes
 
     @classmethod
     def compare_indexes(cls):
-        """ Compares the indexes defined in MongoEngine with the ones existing
-        in the database. Returns any missing/extra indexes.
+        """ Compares the indexes defined in MongoEngine with the ones
+        existing in the database. Returns any missing/extra indexes.
         """
 
         required = cls.list_indexes()
@@ -919,8 +931,9 @@ class DynamicDocument(Document):
     _dynamic = True
 
     def __delattr__(self, *args, **kwargs):
-        """Deletes the attribute by setting to None and allowing _delta to unset
-        it"""
+        """Delete the attribute by setting to None and allowing _delta
+        to unset it.
+        """
         field_name = args[0]
         if field_name in self._dynamic_fields:
             setattr(self, field_name, None)
@@ -942,8 +955,9 @@ class DynamicEmbeddedDocument(EmbeddedDocument):
     _dynamic = True
 
     def __delattr__(self, *args, **kwargs):
-        """Deletes the attribute by setting to None and allowing _delta to unset
-        it"""
+        """Delete the attribute by setting to None and allowing _delta
+        to unset it.
+        """
         field_name = args[0]
         if field_name in self._fields:
             default = self._fields[field_name].default
@@ -985,10 +999,10 @@ class MapReduceDocument(object):
             try:
                 self.key = id_field_type(self.key)
             except Exception:
-                raise Exception("Could not cast key as %s" %
+                raise Exception('Could not cast key as %s' %
                                 id_field_type.__name__)
 
-        if not hasattr(self, "_key_object"):
+        if not hasattr(self, '_key_object'):
             self._key_object = self._document.objects.with_id(self.key)
             return self._key_object
         return self._key_object

mongoengine/errors.py

@@ -1,7 +1,6 @@
 from collections import defaultdict
 
-from mongoengine.python_support import txt_type
-
+import six
 
 __all__ = ('NotRegistered', 'InvalidDocumentError', 'LookUpError',
            'DoesNotExist', 'MultipleObjectsReturned', 'InvalidQueryError',
@@ -71,13 +70,13 @@ class ValidationError(AssertionError):
     field_name = None
     _message = None
 
-    def __init__(self, message="", **kwargs):
+    def __init__(self, message='', **kwargs):
         self.errors = kwargs.get('errors', {})
         self.field_name = kwargs.get('field_name')
         self.message = message
 
     def __str__(self):
-        return txt_type(self.message)
+        return six.text_type(self.message)
 
     def __repr__(self):
         return '%s(%s,)' % (self.__class__.__name__, self.message)
@@ -111,17 +110,20 @@ class ValidationError(AssertionError):
             errors_dict = {}
             if not source:
                 return errors_dict
+
             if isinstance(source, dict):
                 for field_name, error in source.iteritems():
                     errors_dict[field_name] = build_dict(error)
             elif isinstance(source, ValidationError) and source.errors:
                 return build_dict(source.errors)
             else:
-                return unicode(source)
+                return six.text_type(source)
+
             return errors_dict
 
         if not self.errors:
             return {}
+
         return build_dict(self.errors)
 
     def _format_errors(self):
@@ -134,10 +136,10 @@ class ValidationError(AssertionError):
                 value = ' '.join(
                     [generate_key(v, k) for k, v in value.iteritems()])
 
-            results = "%s.%s" % (prefix, value) if prefix else value
+            results = '%s.%s' % (prefix, value) if prefix else value
             return results
 
         error_dict = defaultdict(list)
         for k, v in self.to_dict().iteritems():
             error_dict[generate_key(v)].append(k)
-        return ' '.join(["%s: %s" % (k, v) for k, v in error_dict.iteritems()])
+        return ' '.join(['%s: %s' % (k, v) for k, v in error_dict.iteritems()])

mongoengine/fields.py

@@ -3,7 +3,6 @@ import decimal
 import itertools
 import re
 import time
-import urllib2
 import uuid
 import warnings
 from operator import itemgetter
@@ -25,13 +24,13 @@ try:
 except ImportError:
     Int64 = long
 
-from .base import (BaseDocument, BaseField, ComplexBaseField, GeoJsonBaseField,
-                   ObjectIdField, get_document)
-from .connection import DEFAULT_CONNECTION_NAME, get_db
-from .document import Document, EmbeddedDocument
-from .errors import DoesNotExist, ValidationError
-from .python_support import PY3, StringIO, bin_type, str_types, txt_type
-from .queryset import DO_NOTHING, QuerySet
+from mongoengine.base import (BaseDocument, BaseField, ComplexBaseField,
+                              GeoJsonBaseField, ObjectIdField, get_document)
+from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
+from mongoengine.document import Document, EmbeddedDocument
+from mongoengine.errors import DoesNotExist, ValidationError
+from mongoengine.python_support import StringIO
+from mongoengine.queryset import DO_NOTHING, QuerySet
 
 try:
     from PIL import Image, ImageOps
@@ -39,7 +38,7 @@ except ImportError:
     Image = None
     ImageOps = None
 
-__all__ = [
+__all__ = (
     'StringField', 'URLField', 'EmailField', 'IntField', 'LongField',
     'FloatField', 'DecimalField', 'BooleanField', 'DateTimeField',
     'ComplexDateTimeField', 'EmbeddedDocumentField', 'ObjectIdField',
@@ -50,14 +49,14 @@ __all__ = [
     'FileField', 'ImageGridFsProxy', 'ImproperlyConfigured', 'ImageField',
     'GeoPointField', 'PointField', 'LineStringField', 'PolygonField',
     'SequenceField', 'UUIDField', 'MultiPointField', 'MultiLineStringField',
-    'MultiPolygonField', 'GeoJsonBaseField']
+    'MultiPolygonField', 'GeoJsonBaseField'
+)
 
 RECURSIVE_REFERENCE_CONSTANT = 'self'
 
 
 class StringField(BaseField):
-    """A unicode string field.
-    """
+    """A unicode string field."""
 
     def __init__(self, regex=None, max_length=None, min_length=None, **kwargs):
         self.regex = re.compile(regex) if regex else None
@@ -66,7 +65,7 @@ class StringField(BaseField):
         super(StringField, self).__init__(**kwargs)
 
     def to_python(self, value):
-        if isinstance(value, unicode):
+        if isinstance(value, six.text_type):
             return value
         try:
             value = value.decode('utf-8')
@@ -75,7 +74,7 @@ class StringField(BaseField):
         return value
 
     def validate(self, value):
-        if not isinstance(value, basestring):
+        if not isinstance(value, six.string_types):
             self.error('StringField only accepts string values')
 
         if self.max_length is not None and len(value) > self.max_length:
@@ -91,7 +90,7 @@ class StringField(BaseField):
         return None
 
     def prepare_query_value(self, op, value):
-        if not isinstance(op, basestring):
+        if not isinstance(op, six.string_types):
             return value
 
         if op.lstrip('i') in ('startswith', 'endswith', 'contains', 'exact'):
@@ -148,17 +147,6 @@ class URLField(StringField):
             self.error('Invalid URL: {}'.format(value))
             return
 
-        if self.verify_exists:
-            warnings.warn(
-                "The URLField verify_exists argument has intractable security "
-                "and performance issues. Accordingly, it has been deprecated.",
-                DeprecationWarning)
-            try:
-                request = urllib2.Request(value)
-                urllib2.urlopen(request)
-            except Exception, e:
-                self.error('This URL appears to be a broken link: %s' % e)
-
 
 class EmailField(StringField):
     """A field that validates input as an email address.
@@ -182,8 +170,7 @@ class EmailField(StringField):
 
 
 class IntField(BaseField):
-    """An 32-bit integer field.
-    """
+    """32-bit integer field."""
 
     def __init__(self, min_value=None, max_value=None, **kwargs):
         self.min_value, self.max_value = min_value, max_value
@@ -216,8 +203,7 @@ class IntField(BaseField):
 
 
 class LongField(BaseField):
-    """An 64-bit integer field.
-    """
+    """64-bit integer field."""
 
     def __init__(self, min_value=None, max_value=None, **kwargs):
         self.min_value, self.max_value = min_value, max_value
@@ -253,8 +239,7 @@ class LongField(BaseField):
 
 
 class FloatField(BaseField):
-    """An floating point number field.
-    """
+    """Floating point number field."""
 
     def __init__(self, min_value=None, max_value=None, **kwargs):
         self.min_value, self.max_value = min_value, max_value
@@ -291,7 +276,7 @@ class FloatField(BaseField):
 
 
 class DecimalField(BaseField):
-    """A fixed-point decimal number field.
+    """Fixed-point decimal number field.
 
     .. versionchanged:: 0.8
     .. versionadded:: 0.3
@@ -332,25 +317,25 @@ class DecimalField(BaseField):
 
         # Convert to string for python 2.6 before casting to Decimal
         try:
-            value = decimal.Decimal("%s" % value)
+            value = decimal.Decimal('%s' % value)
         except decimal.InvalidOperation:
             return value
-        return value.quantize(decimal.Decimal(".%s" % ("0" * self.precision)), rounding=self.rounding)
+        return value.quantize(decimal.Decimal('.%s' % ('0' * self.precision)), rounding=self.rounding)
 
     def to_mongo(self, value):
         if value is None:
             return value
         if self.force_string:
-            return unicode(value)
+            return six.text_type(self.to_python(value))
         return float(self.to_python(value))
 
     def validate(self, value):
         if not isinstance(value, decimal.Decimal):
-            if not isinstance(value, basestring):
-                value = unicode(value)
+            if not isinstance(value, six.string_types):
+                value = six.text_type(value)
             try:
                 value = decimal.Decimal(value)
-            except Exception, exc:
+            except Exception as exc:
                 self.error('Could not convert value to decimal: %s' % exc)
 
         if self.min_value is not None and value < self.min_value:
@@ -364,7 +349,7 @@ class DecimalField(BaseField):
 
 
 class BooleanField(BaseField):
-    """A boolean field type.
+    """Boolean field type.
 
     .. versionadded:: 0.1.2
     """
@@ -382,7 +367,7 @@ class BooleanField(BaseField):
 
 
 class DateTimeField(BaseField):
-    """A datetime field.
+    """Datetime field.
 
     Uses the python-dateutil library if available alternatively use time.strptime
     to parse the dates.  Note: python-dateutil's parser is fully featured and when
@@ -410,7 +395,7 @@ class DateTimeField(BaseField):
         if callable(value):
             return value()
 
-        if not isinstance(value, basestring):
+        if not isinstance(value, six.string_types):
             return None
 
         # Attempt to parse a datetime:
@@ -537,16 +522,19 @@ class EmbeddedDocumentField(BaseField):
     """
 
     def __init__(self, document_type, **kwargs):
-        if not isinstance(document_type, basestring):
-            if not issubclass(document_type, EmbeddedDocument):
-                self.error('Invalid embedded document class provided to an '
-                           'EmbeddedDocumentField')
+        if (
+            not isinstance(document_type, six.string_types) and
+            not issubclass(document_type, EmbeddedDocument)
+        ):
+            self.error('Invalid embedded document class provided to an '
+                       'EmbeddedDocumentField')
+
         self.document_type_obj = document_type
         super(EmbeddedDocumentField, self).__init__(**kwargs)
 
     @property
     def document_type(self):
-        if isinstance(self.document_type_obj, basestring):
+        if isinstance(self.document_type_obj, six.string_types):
             if self.document_type_obj == RECURSIVE_REFERENCE_CONSTANT:
                 self.document_type_obj = self.owner_document
             else:
@@ -577,7 +565,7 @@ class EmbeddedDocumentField(BaseField):
         return self.document_type._fields.get(member_name)
 
     def prepare_query_value(self, op, value):
-        if not isinstance(value, self.document_type):
+        if value is not None and not isinstance(value, self.document_type):
             value = self.document_type._from_son(value)
         super(EmbeddedDocumentField, self).prepare_query_value(op, value)
         return self.to_mongo(value)
@@ -631,7 +619,7 @@ class DynamicField(BaseField):
         """Convert a Python type to a MongoDB compatible type.
         """
 
-        if isinstance(value, basestring):
+        if isinstance(value, six.string_types):
             return value
 
         if hasattr(value, 'to_mongo'):
@@ -639,7 +627,7 @@ class DynamicField(BaseField):
             val = value.to_mongo(use_db_field, fields)
             # If we its a document thats not inherited add _cls
             if isinstance(value, Document):
-                val = {"_ref": value.to_dbref(), "_cls": cls.__name__}
+                val = {'_ref': value.to_dbref(), '_cls': cls.__name__}
             if isinstance(value, EmbeddedDocument):
                 val['_cls'] = cls.__name__
             return val
@@ -650,7 +638,7 @@ class DynamicField(BaseField):
         is_list = False
         if not hasattr(value, 'items'):
             is_list = True
-            value = dict([(k, v) for k, v in enumerate(value)])
+            value = {k: v for k, v in enumerate(value)}
 
         data = {}
         for k, v in value.iteritems():
@@ -674,12 +662,12 @@ class DynamicField(BaseField):
         return member_name
 
     def prepare_query_value(self, op, value):
-        if isinstance(value, basestring):
+        if isinstance(value, six.string_types):
             return StringField().prepare_query_value(op, value)
         return super(DynamicField, self).prepare_query_value(op, self.to_mongo(value))
 
     def validate(self, value, clean=True):
-        if hasattr(value, "validate"):
+        if hasattr(value, 'validate'):
             value.validate(clean=clean)
 
 
@@ -699,21 +687,27 @@ class ListField(ComplexBaseField):
         super(ListField, self).__init__(**kwargs)
 
     def validate(self, value):
-        """Make sure that a list of valid fields is being used.
-        """
+        """Make sure that a list of valid fields is being used."""
         if (not isinstance(value, (list, tuple, QuerySet)) or
-                isinstance(value, basestring)):
+                isinstance(value, six.string_types)):
             self.error('Only lists and tuples may be used in a list field')
         super(ListField, self).validate(value)
 
     def prepare_query_value(self, op, value):
         if self.field:
-            if op in ('set', 'unset', None) and (
-                    not isinstance(value, basestring) and
-                    not isinstance(value, BaseDocument) and
-                    hasattr(value, '__iter__')):
+
+            # If the value is iterable and it's not a string nor a
+            # BaseDocument, call prepare_query_value for each of its items.
+            if (
+                op in ('set', 'unset', None) and
+                hasattr(value, '__iter__') and
+                not isinstance(value, six.string_types) and
+                not isinstance(value, BaseDocument)
+            ):
                 return [self.field.prepare_query_value(op, v) for v in value]
+
             return self.field.prepare_query_value(op, value)
+
         return super(ListField, self).prepare_query_value(op, value)
 
 
@@ -726,7 +720,6 @@ class EmbeddedDocumentListField(ListField):
         :class:`~mongoengine.EmbeddedDocument`.
 
     .. versionadded:: 0.9
-
     """
 
     def __init__(self, document_type, **kwargs):
@@ -775,17 +768,17 @@ class SortedListField(ListField):
 
 
 def key_not_string(d):
-    """ Helper function to recursively determine if any key in a dictionary is
-    not a string.
+    """Helper function to recursively determine if any key in a
+    dictionary is not a string.
     """
     for k, v in d.items():
-        if not isinstance(k, basestring) or (isinstance(v, dict) and key_not_string(v)):
+        if not isinstance(k, six.string_types) or (isinstance(v, dict) and key_not_string(v)):
             return True
 
 
 def key_has_dot_or_dollar(d):
-    """ Helper function to recursively determine if any key in a dictionary
-    contains a dot or a dollar sign.
+    """Helper function to recursively determine if any key in a
+    dictionary contains a dot or a dollar sign.
     """
     for k, v in d.items():
         if ('.' in k or '$' in k) or (isinstance(v, dict) and key_has_dot_or_dollar(v)):
@@ -813,14 +806,13 @@ class DictField(ComplexBaseField):
         super(DictField, self).__init__(*args, **kwargs)
 
     def validate(self, value):
-        """Make sure that a list of valid fields is being used.
-        """
+        """Make sure that a list of valid fields is being used."""
         if not isinstance(value, dict):
             self.error('Only dictionaries may be used in a DictField')
 
         if key_not_string(value):
-            msg = ("Invalid dictionary key - documents must "
-                   "have only string keys")
+            msg = ('Invalid dictionary key - documents must '
+                   'have only string keys')
             self.error(msg)
         if key_has_dot_or_dollar(value):
             self.error('Invalid dictionary key name - keys may not contain "."'
@@ -835,14 +827,15 @@ class DictField(ComplexBaseField):
                            'istartswith', 'endswith', 'iendswith',
                            'exact', 'iexact']
 
-        if op in match_operators and isinstance(value, basestring):
+        if op in match_operators and isinstance(value, six.string_types):
             return StringField().prepare_query_value(op, value)
 
         if hasattr(self.field, 'field'):
             if op in ('set', 'unset') and isinstance(value, dict):
-                return dict(
-                    (k, self.field.prepare_query_value(op, v))
-                    for k, v in value.items())
+                return {
+                    k: self.field.prepare_query_value(op, v)
+                    for k, v in value.items()
+                }
             return self.field.prepare_query_value(op, value)
 
         return super(DictField, self).prepare_query_value(op, value)
@@ -911,10 +904,12 @@ class ReferenceField(BaseField):
             A reference to an abstract document type is always stored as a
             :class:`~pymongo.dbref.DBRef`, regardless of the value of `dbref`.
         """
-        if not isinstance(document_type, basestring):
-            if not issubclass(document_type, (Document, basestring)):
-                self.error('Argument to ReferenceField constructor must be a '
-                           'document class or a string')
+        if (
+            not isinstance(document_type, six.string_types) and
+            not issubclass(document_type, Document)
+        ):
+            self.error('Argument to ReferenceField constructor must be a '
+                       'document class or a string')
 
         self.dbref = dbref
         self.document_type_obj = document_type
@@ -923,7 +918,7 @@ class ReferenceField(BaseField):
 
     @property
     def document_type(self):
-        if isinstance(self.document_type_obj, basestring):
+        if isinstance(self.document_type_obj, six.string_types):
             if self.document_type_obj == RECURSIVE_REFERENCE_CONSTANT:
                 self.document_type_obj = self.owner_document
             else:
@@ -931,8 +926,7 @@ class ReferenceField(BaseField):
         return self.document_type_obj
 
     def __get__(self, instance, owner):
-        """Descriptor to allow lazy dereferencing.
-        """
+        """Descriptor to allow lazy dereferencing."""
         if instance is None:
             # Document class being used rather than a document object
             return self
@@ -989,8 +983,7 @@ class ReferenceField(BaseField):
         return id_
 
     def to_python(self, value):
-        """Convert a MongoDB-compatible type to a Python type.
-        """
+        """Convert a MongoDB-compatible type to a Python type."""
         if (not self.dbref and
                 not isinstance(value, (DBRef, Document, EmbeddedDocument))):
             collection = self.document_type._get_collection_name()
@@ -1006,7 +999,7 @@ class ReferenceField(BaseField):
     def validate(self, value):
 
         if not isinstance(value, (self.document_type, DBRef)):
-            self.error("A ReferenceField only accepts DBRef or documents")
+            self.error('A ReferenceField only accepts DBRef or documents')
 
         if isinstance(value, Document) and value.id is None:
             self.error('You can only reference documents once they have been '
@@ -1030,14 +1023,19 @@ class CachedReferenceField(BaseField):
     .. versionadded:: 0.9
     """
 
-    def __init__(self, document_type, fields=[], auto_sync=True, **kwargs):
+    def __init__(self, document_type, fields=None, auto_sync=True, **kwargs):
         """Initialises the Cached Reference Field.
 
         :param fields:  A list of fields to be cached in document
         :param auto_sync: if True documents are auto updated.
         """
-        if not isinstance(document_type, basestring) and \
-                not issubclass(document_type, (Document, basestring)):
+        if fields is None:
+            fields = []
+
+        if (
+            not isinstance(document_type, six.string_types) and
+            not issubclass(document_type, Document)
+        ):
             self.error('Argument to CachedReferenceField constructor must be a'
                        ' document class or a string')
 
@@ -1053,18 +1051,20 @@ class CachedReferenceField(BaseField):
                                   sender=self.document_type)
 
     def on_document_pre_save(self, sender, document, created, **kwargs):
-        if not created:
-            update_kwargs = dict(
-                ('set__%s__%s' % (self.name, k), v)
-                for k, v in document._delta()[0].items()
-                if k in self.fields)
+        if created:
+            return None
 
-            if update_kwargs:
-                filter_kwargs = {}
-                filter_kwargs[self.name] = document
+        update_kwargs = {
+            'set__%s__%s' % (self.name, key): val
+            for key, val in document._delta()[0].items()
+            if key in self.fields
+        }
+        if update_kwargs:
+            filter_kwargs = {}
+            filter_kwargs[self.name] = document
 
-                self.owner_document.objects(
-                    **filter_kwargs).update(**update_kwargs)
+            self.owner_document.objects(
+                **filter_kwargs).update(**update_kwargs)
 
     def to_python(self, value):
         if isinstance(value, dict):
@@ -1077,7 +1077,7 @@ class CachedReferenceField(BaseField):
 
     @property
     def document_type(self):
-        if isinstance(self.document_type_obj, basestring):
+        if isinstance(self.document_type_obj, six.string_types):
             if self.document_type_obj == RECURSIVE_REFERENCE_CONSTANT:
                 self.document_type_obj = self.owner_document
             else:
@@ -1117,7 +1117,7 @@ class CachedReferenceField(BaseField):
             # TODO: should raise here or will fail next statement
 
         value = SON((
-            ("_id", id_field.to_mongo(id_)),
+            ('_id', id_field.to_mongo(id_)),
         ))
 
         if fields:
@@ -1143,7 +1143,7 @@ class CachedReferenceField(BaseField):
     def validate(self, value):
 
         if not isinstance(value, self.document_type):
-            self.error("A CachedReferenceField only accepts documents")
+            self.error('A CachedReferenceField only accepts documents')
 
         if isinstance(value, Document) and value.id is None:
             self.error('You can only reference documents once they have been '
@@ -1191,13 +1191,13 @@ class GenericReferenceField(BaseField):
         # Keep the choices as a list of allowed Document class names
         if choices:
             for choice in choices:
-                if isinstance(choice, basestring):
+                if isinstance(choice, six.string_types):
                     self.choices.append(choice)
                 elif isinstance(choice, type) and issubclass(choice, Document):
                     self.choices.append(choice._class_name)
                 else:
                     self.error('Invalid choices provided: must be a list of'
-                               'Document subclasses and/or basestrings')
+                               'Document subclasses and/or six.string_typess')
 
     def _validate_choices(self, value):
         if isinstance(value, dict):
@@ -1249,7 +1249,7 @@ class GenericReferenceField(BaseField):
         if document is None:
             return None
 
-        if isinstance(document, (dict, SON)):
+        if isinstance(document, (dict, SON, ObjectId, DBRef)):
             return document
 
         id_field_name = document.__class__._meta['id_field']
@@ -1280,8 +1280,7 @@ class GenericReferenceField(BaseField):
 
 
 class BinaryField(BaseField):
-    """A binary data field.
-    """
+    """A binary data field."""
 
     def __init__(self, max_bytes=None, **kwargs):
         self.max_bytes = max_bytes
@@ -1289,18 +1288,18 @@ class BinaryField(BaseField):
 
     def __set__(self, instance, value):
         """Handle bytearrays in python 3.1"""
-        if PY3 and isinstance(value, bytearray):
-            value = bin_type(value)
+        if six.PY3 and isinstance(value, bytearray):
+            value = six.binary_type(value)
         return super(BinaryField, self).__set__(instance, value)
 
     def to_mongo(self, value):
         return Binary(value)
 
     def validate(self, value):
-        if not isinstance(value, (bin_type, txt_type, Binary)):
-            self.error("BinaryField only accepts instances of "
-                       "(%s, %s, Binary)" % (
-                           bin_type.__name__, txt_type.__name__))
+        if not isinstance(value, (six.binary_type, six.text_type, Binary)):
+            self.error('BinaryField only accepts instances of '
+                       '(%s, %s, Binary)' % (
+                           six.binary_type.__name__, six.text_type.__name__))
 
         if self.max_bytes is not None and len(value) > self.max_bytes:
             self.error('Binary value is too long')
@@ -1384,11 +1383,13 @@ class GridFSProxy(object):
                 get_db(self.db_alias), self.collection_name)
         return self._fs
 
-    def get(self, id=None):
-        if id:
-            self.grid_id = id
+    def get(self, grid_id=None):
+        if grid_id:
+            self.grid_id = grid_id
+
         if self.grid_id is None:
             return None
+
         try:
             if self.gridout is None:
                 self.gridout = self.fs.get(self.grid_id)
@@ -1432,7 +1433,7 @@ class GridFSProxy(object):
             try:
                 return gridout.read(size)
             except Exception:
-                return ""
+                return ''
 
     def delete(self):
         # Delete file from GridFS, FileField still remains
@@ -1464,9 +1465,8 @@ class FileField(BaseField):
     """
     proxy_class = GridFSProxy
 
-    def __init__(self,
-                 db_alias=DEFAULT_CONNECTION_NAME,
-                 collection_name="fs", **kwargs):
+    def __init__(self, db_alias=DEFAULT_CONNECTION_NAME, collection_name='fs',
+                 **kwargs):
         super(FileField, self).__init__(**kwargs)
         self.collection_name = collection_name
         self.db_alias = db_alias
@@ -1488,8 +1488,10 @@ class FileField(BaseField):
 
     def __set__(self, instance, value):
         key = self.name
-        if ((hasattr(value, 'read') and not
-                isinstance(value, GridFSProxy)) or isinstance(value, str_types)):
+        if (
+            (hasattr(value, 'read') and not isinstance(value, GridFSProxy)) or
+            isinstance(value, (six.binary_type, six.string_types))
+        ):
             # using "FileField() = file/string" notation
             grid_file = instance._data.get(self.name)
             # If a file already exists, delete it
@@ -1558,7 +1560,7 @@ class ImageGridFsProxy(GridFSProxy):
         try:
             img = Image.open(file_obj)
             img_format = img.format
-        except Exception, e:
+        except Exception as e:
             raise ValidationError('Invalid image: %s' % e)
 
         # Progressive JPEG
@@ -1667,10 +1669,10 @@ class ImageGridFsProxy(GridFSProxy):
             return self.fs.get(out.thumbnail_id)
 
     def write(self, *args, **kwargs):
-        raise RuntimeError("Please use \"put\" method instead")
+        raise RuntimeError('Please use "put" method instead')
 
     def writelines(self, *args, **kwargs):
-        raise RuntimeError("Please use \"put\" method instead")
+        raise RuntimeError('Please use "put" method instead')
 
 
 class ImproperlyConfigured(Exception):
@@ -1695,14 +1697,17 @@ class ImageField(FileField):
     def __init__(self, size=None, thumbnail_size=None,
                  collection_name='images', **kwargs):
         if not Image:
-            raise ImproperlyConfigured("PIL library was not found")
+            raise ImproperlyConfigured('PIL library was not found')
 
         params_size = ('width', 'height', 'force')
-        extra_args = dict(size=size, thumbnail_size=thumbnail_size)
+        extra_args = {
+            'size': size,
+            'thumbnail_size': thumbnail_size
+        }
         for att_name, att in extra_args.items():
             value = None
             if isinstance(att, (tuple, list)):
-                if PY3:
+                if six.PY3:
                     value = dict(itertools.zip_longest(params_size, att,
                                                        fillvalue=None))
                 else:
@@ -1763,10 +1768,10 @@ class SequenceField(BaseField):
         Generate and Increment the counter
         """
         sequence_name = self.get_sequence_name()
-        sequence_id = "%s.%s" % (sequence_name, self.name)
+        sequence_id = '%s.%s' % (sequence_name, self.name)
         collection = get_db(alias=self.db_alias)[self.collection_name]
-        counter = collection.find_and_modify(query={"_id": sequence_id},
-                                             update={"$inc": {"next": 1}},
+        counter = collection.find_and_modify(query={'_id': sequence_id},
+                                             update={'$inc': {'next': 1}},
                                              new=True,
                                              upsert=True)
         return self.value_decorator(counter['next'])
@@ -1789,9 +1794,9 @@ class SequenceField(BaseField):
         as it is only fixed on set.
         """
         sequence_name = self.get_sequence_name()
-        sequence_id = "%s.%s" % (sequence_name, self.name)
+        sequence_id = '%s.%s' % (sequence_name, self.name)
         collection = get_db(alias=self.db_alias)[self.collection_name]
-        data = collection.find_one({"_id": sequence_id})
+        data = collection.find_one({'_id': sequence_id})
 
         if data:
             return self.value_decorator(data['next'] + 1)
@@ -1861,8 +1866,8 @@ class UUIDField(BaseField):
         if not self._binary:
             original_value = value
             try:
-                if not isinstance(value, basestring):
-                    value = unicode(value)
+                if not isinstance(value, six.string_types):
+                    value = six.text_type(value)
                 return uuid.UUID(value)
             except Exception:
                 return original_value
@@ -1870,8 +1875,8 @@ class UUIDField(BaseField):
 
     def to_mongo(self, value):
         if not self._binary:
-            return unicode(value)
-        elif isinstance(value, basestring):
+            return six.text_type(value)
+        elif isinstance(value, six.string_types):
             return uuid.UUID(value)
         return value
 
@@ -1882,11 +1887,11 @@ class UUIDField(BaseField):
 
     def validate(self, value):
         if not isinstance(value, uuid.UUID):
-            if not isinstance(value, basestring):
+            if not isinstance(value, six.string_types):
                 value = str(value)
             try:
                 uuid.UUID(value)
-            except Exception, exc:
+            except Exception as exc:
                 self.error('Could not convert to UUID: %s' % exc)
 
 
@@ -1904,19 +1909,18 @@ class GeoPointField(BaseField):
     _geo_index = pymongo.GEO2D
 
     def validate(self, value):
-        """Make sure that a geo-value is of type (x, y)
-        """
+        """Make sure that a geo-value is of type (x, y)"""
         if not isinstance(value, (list, tuple)):
             self.error('GeoPointField can only accept tuples or lists '
                        'of (x, y)')
 
         if not len(value) == 2:
-            self.error("Value (%s) must be a two-dimensional point" %
+            self.error('Value (%s) must be a two-dimensional point' %
                        repr(value))
         elif (not isinstance(value[0], (float, int)) or
               not isinstance(value[1], (float, int))):
             self.error(
-                "Both values (%s) in point must be float or int" % repr(value))
+                'Both values (%s) in point must be float or int' % repr(value))
 
 
 class PointField(GeoJsonBaseField):
@@ -1926,8 +1930,8 @@ class PointField(GeoJsonBaseField):
 
     .. code-block:: js
 
-        { "type" : "Point" ,
-          "coordinates" : [x, y]}
+        {'type' : 'Point' ,
+         'coordinates' : [x, y]}
 
     You can either pass a dict with the full information or a list
     to set the value.
@@ -1936,7 +1940,7 @@ class PointField(GeoJsonBaseField):
 
     .. versionadded:: 0.8
     """
-    _type = "Point"
+    _type = 'Point'
 
 
 class LineStringField(GeoJsonBaseField):
@@ -1946,8 +1950,8 @@ class LineStringField(GeoJsonBaseField):
 
     .. code-block:: js
 
-        { "type" : "LineString" ,
-          "coordinates" : [[x1, y1], [x1, y1] ... [xn, yn]]}
+        {'type' : 'LineString' ,
+         'coordinates' : [[x1, y1], [x1, y1] ... [xn, yn]]}
 
     You can either pass a dict with the full information or a list of points.
 
@@ -1955,7 +1959,7 @@ class LineStringField(GeoJsonBaseField):
 
     .. versionadded:: 0.8
     """
-    _type = "LineString"
+    _type = 'LineString'
 
 
 class PolygonField(GeoJsonBaseField):
@@ -1965,9 +1969,9 @@ class PolygonField(GeoJsonBaseField):
 
     .. code-block:: js
 
-        { "type" : "Polygon" ,
-          "coordinates" : [[[x1, y1], [x1, y1] ... [xn, yn]],
-                           [[x1, y1], [x1, y1] ... [xn, yn]]}
+        {'type' : 'Polygon' ,
+         'coordinates' : [[[x1, y1], [x1, y1] ... [xn, yn]],
+                          [[x1, y1], [x1, y1] ... [xn, yn]]}
 
     You can either pass a dict with the full information or a list
     of LineStrings. The first LineString being the outside and the rest being
@@ -1977,7 +1981,7 @@ class PolygonField(GeoJsonBaseField):
 
     .. versionadded:: 0.8
     """
-    _type = "Polygon"
+    _type = 'Polygon'
 
 
 class MultiPointField(GeoJsonBaseField):
@@ -1987,8 +1991,8 @@ class MultiPointField(GeoJsonBaseField):
 
     .. code-block:: js
 
-        { "type" : "MultiPoint" ,
-          "coordinates" : [[x1, y1], [x2, y2]]}
+        {'type' : 'MultiPoint' ,
+         'coordinates' : [[x1, y1], [x2, y2]]}
 
     You can either pass a dict with the full information or a list
     to set the value.
@@ -1997,7 +2001,7 @@ class MultiPointField(GeoJsonBaseField):
 
     .. versionadded:: 0.9
     """
-    _type = "MultiPoint"
+    _type = 'MultiPoint'
 
 
 class MultiLineStringField(GeoJsonBaseField):
@@ -2007,9 +2011,9 @@ class MultiLineStringField(GeoJsonBaseField):
 
     .. code-block:: js
 
-        { "type" : "MultiLineString" ,
-          "coordinates" : [[[x1, y1], [x1, y1] ... [xn, yn]],
-                           [[x1, y1], [x1, y1] ... [xn, yn]]]}
+        {'type' : 'MultiLineString' ,
+         'coordinates' : [[[x1, y1], [x1, y1] ... [xn, yn]],
+                          [[x1, y1], [x1, y1] ... [xn, yn]]]}
 
     You can either pass a dict with the full information or a list of points.
 
@@ -2017,7 +2021,7 @@ class MultiLineStringField(GeoJsonBaseField):
 
     .. versionadded:: 0.9
     """
-    _type = "MultiLineString"
+    _type = 'MultiLineString'
 
 
 class MultiPolygonField(GeoJsonBaseField):
@@ -2027,14 +2031,14 @@ class MultiPolygonField(GeoJsonBaseField):
 
     .. code-block:: js
 
-        { "type" : "MultiPolygon" ,
-          "coordinates" : [[
-                [[x1, y1], [x1, y1] ... [xn, yn]],
-                [[x1, y1], [x1, y1] ... [xn, yn]]
-            ], [
-                [[x1, y1], [x1, y1] ... [xn, yn]],
-                [[x1, y1], [x1, y1] ... [xn, yn]]
-            ]
+        {'type' : 'MultiPolygon' ,
+         'coordinates' : [[
+               [[x1, y1], [x1, y1] ... [xn, yn]],
+               [[x1, y1], [x1, y1] ... [xn, yn]]
+           ], [
+               [[x1, y1], [x1, y1] ... [xn, yn]],
+               [[x1, y1], [x1, y1] ... [xn, yn]]
+           ]
         }
 
     You can either pass a dict with the full information or a list
@@ -2044,4 +2048,4 @@ class MultiPolygonField(GeoJsonBaseField):
 
     .. versionadded:: 0.9
     """
-    _type = "MultiPolygon"
+    _type = 'MultiPolygon'

mongoengine/python_support.py

@@ -1,7 +1,9 @@
-"""Helper functions and types to aid with Python 2.5 - 3 support."""
-
-import sys
+"""
+Helper functions, constants, and types to aid with Python v2.7 - v3.x and
+PyMongo v2.7 - v3.x support.
+"""
 import pymongo
+import six
 
 
 if pymongo.version_tuple[0] < 3:
@@ -9,29 +11,15 @@ if pymongo.version_tuple[0] < 3:
 else:
     IS_PYMONGO_3 = True
 
-PY3 = sys.version_info[0] == 3
 
-if PY3:
-    import codecs
-    from io import BytesIO as StringIO
+# six.BytesIO resolves to StringIO.StringIO in Py2 and io.BytesIO in Py3.
+StringIO = six.BytesIO
 
-    # return s converted to binary.  b('test') should be equivalent to b'test'
-    def b(s):
-        return codecs.latin_1_encode(s)[0]
-
-    bin_type = bytes
-    txt_type = str
-else:
+# Additionally for Py2, try to use the faster cStringIO, if available
+if not six.PY3:
     try:
-        from cStringIO import StringIO
+        import cStringIO
     except ImportError:
-        from StringIO import StringIO
-
-    # Conversion to binary only necessary in Python 3
-    def b(s):
-        return s
-
-    bin_type = str
-    txt_type = unicode
-
-str_types = (bin_type, txt_type)
+        pass
+    else:
+        StringIO = cStringIO.StringIO

mongoengine/queryset/__init__.py

@@ -1,11 +1,17 @@
-from mongoengine.errors import (DoesNotExist, InvalidQueryError,
-                                MultipleObjectsReturned, NotUniqueError,
-                                OperationError)
+from mongoengine.errors import *
 from mongoengine.queryset.field_list import *
 from mongoengine.queryset.manager import *
 from mongoengine.queryset.queryset import *
 from mongoengine.queryset.transform import *
 from mongoengine.queryset.visitor import *
 
-__all__ = (field_list.__all__ + manager.__all__ + queryset.__all__ +
-           transform.__all__ + visitor.__all__)
+# Expose just the public subset of all imported objects and constants.
+__all__ = (
+    'QuerySet', 'QuerySetNoCache', 'Q', 'queryset_manager', 'QuerySetManager',
+    'QueryFieldList', 'DO_NOTHING', 'NULLIFY', 'CASCADE', 'DENY', 'PULL',
+
+    # Errors that might be related to a queryset, mostly here for backward
+    # compatibility
+    'DoesNotExist', 'InvalidQueryError', 'MultipleObjectsReturned',
+    'NotUniqueError', 'OperationError',
+)

mongoengine/queryset/base.py

@@ -12,9 +12,10 @@ from bson.code import Code
 import pymongo
 import pymongo.errors
 from pymongo.common import validate_read_preference
+import six
 
 from mongoengine import signals
-from mongoengine.base.common import get_document
+from mongoengine.base import get_document
 from mongoengine.common import _import_class
 from mongoengine.connection import get_db
 from mongoengine.context_managers import switch_db
@@ -73,15 +74,16 @@ class BaseQuerySet(object):
         # subclasses of the class being used
         if document._meta.get('allow_inheritance') is True:
             if len(self._document._subclasses) == 1:
-                self._initial_query = {"_cls": self._document._subclasses[0]}
+                self._initial_query = {'_cls': self._document._subclasses[0]}
             else:
                 self._initial_query = {
-                    "_cls": {"$in": self._document._subclasses}}
+                    '_cls': {'$in': self._document._subclasses}}
             self._loaded_fields = QueryFieldList(always_include=['_cls'])
         self._cursor_obj = None
         self._limit = None
         self._skip = None
         self._hint = -1  # Using -1 as None is a valid value for hint
+        self._batch_size = None
         self.only_fields = []
         self._max_time_ms = None
 
@@ -104,8 +106,8 @@ class BaseQuerySet(object):
         if q_obj:
             # make sure proper query object is passed
             if not isinstance(q_obj, QNode):
-                msg = ("Not a query object: %s. "
-                       "Did you intend to use key=value?" % q_obj)
+                msg = ('Not a query object: %s. '
+                       'Did you intend to use key=value?' % q_obj)
                 raise InvalidQueryError(msg)
             query &= q_obj
 
@@ -132,10 +134,10 @@ class BaseQuerySet(object):
         obj_dict = self.__dict__.copy()
 
         # don't picke collection, instead pickle collection params
-        obj_dict.pop("_collection_obj")
+        obj_dict.pop('_collection_obj')
 
         # don't pickle cursor
-        obj_dict["_cursor_obj"] = None
+        obj_dict['_cursor_obj'] = None
 
         return obj_dict
 
@@ -146,7 +148,7 @@ class BaseQuerySet(object):
         See https://github.com/MongoEngine/mongoengine/issues/442
         """
 
-        obj_dict["_collection_obj"] = obj_dict["_document"]._get_collection()
+        obj_dict['_collection_obj'] = obj_dict['_document']._get_collection()
 
         # update attributes
         self.__dict__.update(obj_dict)
@@ -165,7 +167,7 @@ class BaseQuerySet(object):
                 queryset._skip, queryset._limit = key.start, key.stop
                 if key.start and key.stop:
                     queryset._limit = key.stop - key.start
-            except IndexError, err:
+            except IndexError as err:
                 # PyMongo raises an error if key.start == key.stop, catch it,
                 # bin it, kill it.
                 start = key.start or 0
@@ -198,19 +200,16 @@ class BaseQuerySet(object):
         raise NotImplementedError
 
     def _has_data(self):
-        """ Retrieves whether cursor has any data. """
-
+        """Return True if cursor has any data."""
         queryset = self.order_by()
         return False if queryset.first() is None else True
 
     def __nonzero__(self):
-        """ Avoid to open all records in an if stmt in Py2. """
-
+        """Avoid to open all records in an if stmt in Py2."""
         return self._has_data()
 
     def __bool__(self):
-        """ Avoid to open all records in an if stmt in Py3. """
-
+        """Avoid to open all records in an if stmt in Py3."""
         return self._has_data()
 
     # Core functions
@@ -238,7 +237,7 @@ class BaseQuerySet(object):
         queryset = self.clone()
         if queryset._search_text:
             raise OperationError(
-                "It is not possible to use search_text two times.")
+                'It is not possible to use search_text two times.')
 
         query_kwargs = SON({'$search': text})
         if language:
@@ -267,7 +266,7 @@ class BaseQuerySet(object):
         try:
             result = queryset.next()
         except StopIteration:
-            msg = ("%s matching query does not exist."
+            msg = ('%s matching query does not exist.'
                    % queryset._document._class_name)
             raise queryset._document.DoesNotExist(msg)
         try:
@@ -275,6 +274,8 @@ class BaseQuerySet(object):
         except StopIteration:
             return result
 
+        # If we were able to retrieve the 2nd doc, rewind the cursor and
+        # raise the MultipleObjectsReturned exception.
         queryset.rewind()
         message = u'%d items returned, instead of 1' % queryset.count()
         raise queryset._document.MultipleObjectsReturned(message)
@@ -287,8 +288,7 @@ class BaseQuerySet(object):
         return self._document(**kwargs).save()
 
     def first(self):
-        """Retrieve the first object matching the query.
-        """
+        """Retrieve the first object matching the query."""
         queryset = self.clone()
         try:
             result = queryset[0]
@@ -337,7 +337,7 @@ class BaseQuerySet(object):
                        % str(self._document))
                 raise OperationError(msg)
             if doc.pk and not doc._created:
-                msg = "Some documents have ObjectIds use doc.update() instead"
+                msg = 'Some documents have ObjectIds use doc.update() instead'
                 raise OperationError(msg)
 
         signal_kwargs = signal_kwargs or {}
@@ -347,17 +347,17 @@ class BaseQuerySet(object):
         raw = [doc.to_mongo() for doc in docs]
         try:
             ids = self._collection.insert(raw, **write_concern)
-        except pymongo.errors.DuplicateKeyError, err:
+        except pymongo.errors.DuplicateKeyError as err:
             message = 'Could not save document (%s)'
-            raise NotUniqueError(message % unicode(err))
-        except pymongo.errors.OperationFailure, err:
+            raise NotUniqueError(message % six.text_type(err))
+        except pymongo.errors.OperationFailure as err:
             message = 'Could not save document (%s)'
-            if re.match('^E1100[01] duplicate key', unicode(err)):
+            if re.match('^E1100[01] duplicate key', six.text_type(err)):
                 # E11000 - duplicate key error index
                 # E11001 - duplicate key on update
                 message = u'Tried to save duplicate unique keys (%s)'
-                raise NotUniqueError(message % unicode(err))
-            raise OperationError(message % unicode(err))
+                raise NotUniqueError(message % six.text_type(err))
+            raise OperationError(message % six.text_type(err))
 
         if not load_bulk:
             signals.post_bulk_insert.send(
@@ -383,7 +383,8 @@ class BaseQuerySet(object):
             return 0
         return self._cursor.count(with_limit_and_skip=with_limit_and_skip)
 
-    def delete(self, write_concern=None, _from_doc_delete=False, cascade_refs=None):
+    def delete(self, write_concern=None, _from_doc_delete=False,
+               cascade_refs=None):
         """Delete the documents matched by the query.
 
         :param write_concern: Extra keyword arguments are passed down which
@@ -406,8 +407,9 @@ class BaseQuerySet(object):
         # Handle deletes where skips or limits have been applied or
         # there is an untriggered delete signal
         has_delete_signal = signals.signals_available and (
-            signals.pre_delete.has_receivers_for(self._document) or
-            signals.post_delete.has_receivers_for(self._document))
+            signals.pre_delete.has_receivers_for(doc) or
+            signals.post_delete.has_receivers_for(doc)
+        )
 
         call_document_delete = (queryset._skip or queryset._limit or
                                 has_delete_signal) and not _from_doc_delete
@@ -420,37 +422,44 @@ class BaseQuerySet(object):
             return cnt
 
         delete_rules = doc._meta.get('delete_rules') or {}
+        delete_rules = list(delete_rules.items())
+
         # Check for DENY rules before actually deleting/nullifying any other
         # references
-        for rule_entry in delete_rules:
+        for rule_entry, rule in delete_rules:
             document_cls, field_name = rule_entry
             if document_cls._meta.get('abstract'):
                 continue
-            rule = doc._meta['delete_rules'][rule_entry]
-            if rule == DENY and document_cls.objects(
-                    **{field_name + '__in': self}).count() > 0:
-                msg = ("Could not delete document (%s.%s refers to it)"
-                       % (document_cls.__name__, field_name))
-                raise OperationError(msg)
 
-        for rule_entry in delete_rules:
+            if rule == DENY:
+                refs = document_cls.objects(**{field_name + '__in': self})
+                if refs.limit(1).count() > 0:
+                    raise OperationError(
+                        'Could not delete document (%s.%s refers to it)'
+                        % (document_cls.__name__, field_name)
+                    )
+
+        # Check all the other rules
+        for rule_entry, rule in delete_rules:
             document_cls, field_name = rule_entry
             if document_cls._meta.get('abstract'):
                 continue
-            rule = doc._meta['delete_rules'][rule_entry]
+
             if rule == CASCADE:
                 cascade_refs = set() if cascade_refs is None else cascade_refs
                 # Handle recursive reference
                 if doc._collection == document_cls._collection:
                     for ref in queryset:
                         cascade_refs.add(ref.id)
-                ref_q = document_cls.objects(**{field_name + '__in': self, 'id__nin': cascade_refs})
-                ref_q_count = ref_q.count()
-                if ref_q_count > 0:
-                    ref_q.delete(write_concern=write_concern, cascade_refs=cascade_refs)
+                refs = document_cls.objects(**{field_name + '__in': self,
+                                               'pk__nin': cascade_refs})
+                if refs.count() > 0:
+                    refs.delete(write_concern=write_concern,
+                                cascade_refs=cascade_refs)
             elif rule == NULLIFY:
                 document_cls.objects(**{field_name + '__in': self}).update(
-                    write_concern=write_concern, **{'unset__%s' % field_name: 1})
+                    write_concern=write_concern,
+                    **{'unset__%s' % field_name: 1})
             elif rule == PULL:
                 document_cls.objects(**{field_name + '__in': self}).update(
                     write_concern=write_concern,
@@ -458,7 +467,7 @@ class BaseQuerySet(object):
 
         result = queryset._collection.remove(queryset._query, **write_concern)
         if result:
-            return result.get("n")
+            return result.get('n')
 
     def update(self, upsert=False, multi=True, write_concern=None,
                full_result=False, **update):
@@ -479,7 +488,7 @@ class BaseQuerySet(object):
         .. versionadded:: 0.2
         """
         if not update and not upsert:
-            raise OperationError("No update parameters, would remove data")
+            raise OperationError('No update parameters, would remove data')
 
         if write_concern is None:
             write_concern = {}
@@ -492,9 +501,9 @@ class BaseQuerySet(object):
         # then ensure we add _cls to the update operation
         if upsert and '_cls' in query:
             if '$set' in update:
-                update["$set"]["_cls"] = queryset._document._class_name
+                update['$set']['_cls'] = queryset._document._class_name
             else:
-                update["$set"] = {"_cls": queryset._document._class_name}
+                update['$set'] = {'_cls': queryset._document._class_name}
         try:
             result = queryset._collection.update(query, update, multi=multi,
                                                  upsert=upsert, **write_concern)
@@ -502,13 +511,13 @@ class BaseQuerySet(object):
                 return result
             elif result:
                 return result['n']
-        except pymongo.errors.DuplicateKeyError, err:
-            raise NotUniqueError(u'Update failed (%s)' % unicode(err))
-        except pymongo.errors.OperationFailure, err:
-            if unicode(err) == u'multi not coded yet':
+        except pymongo.errors.DuplicateKeyError as err:
+            raise NotUniqueError(u'Update failed (%s)' % six.text_type(err))
+        except pymongo.errors.OperationFailure as err:
+            if six.text_type(err) == u'multi not coded yet':
                 message = u'update() method requires MongoDB 1.1.3+'
                 raise OperationError(message)
-            raise OperationError(u'Update failed (%s)' % unicode(err))
+            raise OperationError(u'Update failed (%s)' % six.text_type(err))
 
     def upsert_one(self, write_concern=None, **update):
         """Overwrite or add the first document matched by the query.
@@ -579,11 +588,11 @@ class BaseQuerySet(object):
         """
 
         if remove and new:
-            raise OperationError("Conflicting parameters: remove and new")
+            raise OperationError('Conflicting parameters: remove and new')
 
         if not update and not upsert and not remove:
             raise OperationError(
-                "No update parameters, must either update or remove")
+                'No update parameters, must either update or remove')
 
         queryset = self.clone()
         query = queryset._query
@@ -594,7 +603,7 @@ class BaseQuerySet(object):
         try:
             if IS_PYMONGO_3:
                 if full_response:
-                    msg = "With PyMongo 3+, it is not possible anymore to get the full response."
+                    msg = 'With PyMongo 3+, it is not possible anymore to get the full response.'
                     warnings.warn(msg, DeprecationWarning)
                 if remove:
                     result = queryset._collection.find_one_and_delete(
@@ -612,14 +621,14 @@ class BaseQuerySet(object):
                 result = queryset._collection.find_and_modify(
                     query, update, upsert=upsert, sort=sort, remove=remove, new=new,
                     full_response=full_response, **self._cursor_args)
-        except pymongo.errors.DuplicateKeyError, err:
-            raise NotUniqueError(u"Update failed (%s)" % err)
-        except pymongo.errors.OperationFailure, err:
-            raise OperationError(u"Update failed (%s)" % err)
+        except pymongo.errors.DuplicateKeyError as err:
+            raise NotUniqueError(u'Update failed (%s)' % err)
+        except pymongo.errors.OperationFailure as err:
+            raise OperationError(u'Update failed (%s)' % err)
 
         if full_response:
-            if result["value"] is not None:
-                result["value"] = self._document._from_son(result["value"], only_fields=self.only_fields)
+            if result['value'] is not None:
+                result['value'] = self._document._from_son(result['value'], only_fields=self.only_fields)
         else:
             if result is not None:
                 result = self._document._from_son(result, only_fields=self.only_fields)
@@ -637,7 +646,7 @@ class BaseQuerySet(object):
         """
         queryset = self.clone()
         if not queryset._query_obj.empty:
-            msg = "Cannot use a filter whilst using `with_id`"
+            msg = 'Cannot use a filter whilst using `with_id`'
             raise InvalidQueryError(msg)
         return queryset.filter(pk=object_id).first()
 
@@ -681,7 +690,7 @@ class BaseQuerySet(object):
         Only return instances of this document and not any inherited documents
         """
         if self._document._meta.get('allow_inheritance') is True:
-            self._initial_query = {"_cls": self._document._class_name}
+            self._initial_query = {'_cls': self._document._class_name}
 
         return self
 
@@ -781,6 +790,19 @@ class BaseQuerySet(object):
         queryset._hint = index
         return queryset
 
+    def batch_size(self, size):
+        """Limit the number of documents returned in a single batch (each
+        batch requires a round trip to the server).
+
+        See http://api.mongodb.com/python/current/api/pymongo/cursor.html#pymongo.cursor.Cursor.batch_size
+        for details.
+
+        :param size: desired size of each batch.
+        """
+        queryset = self.clone()
+        queryset._batch_size = size
+        return queryset
+
     def distinct(self, field):
         """Return a list of distinct values for a given field.
 
@@ -794,49 +816,56 @@ class BaseQuerySet(object):
         .. versionchanged:: 0.6 - Improved db_field refrence handling
         """
         queryset = self.clone()
+
         try:
             field = self._fields_to_dbfields([field]).pop()
-        finally:
-            distinct = self._dereference(queryset._cursor.distinct(field), 1,
-                                         name=field, instance=self._document)
+        except LookUpError:
+            pass
 
-            doc_field = self._document._fields.get(field.split('.', 1)[0])
-            instance = False
-            # We may need to cast to the correct type eg. ListField(EmbeddedDocumentField)
-            EmbeddedDocumentField = _import_class('EmbeddedDocumentField')
-            ListField = _import_class('ListField')
-            GenericEmbeddedDocumentField = _import_class('GenericEmbeddedDocumentField')
-            if isinstance(doc_field, ListField):
-                doc_field = getattr(doc_field, "field", doc_field)
-            if isinstance(doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)):
-                instance = getattr(doc_field, "document_type", False)
-            # handle distinct on subdocuments
-            if '.' in field:
-                for field_part in field.split('.')[1:]:
-                    # if looping on embedded document, get the document type instance
-                    if instance and isinstance(doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)):
-                        doc_field = instance
-                    # now get the subdocument
-                    doc_field = getattr(doc_field, field_part, doc_field)
-                    # We may need to cast to the correct type eg. ListField(EmbeddedDocumentField)
-                    if isinstance(doc_field, ListField):
-                        doc_field = getattr(doc_field, "field", doc_field)
-                    if isinstance(doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)):
-                        instance = getattr(doc_field, "document_type", False)
-            if instance and isinstance(doc_field, (EmbeddedDocumentField,
-                                                   GenericEmbeddedDocumentField)):
-                distinct = [instance(**doc) for doc in distinct]
-            return distinct
+        distinct = self._dereference(queryset._cursor.distinct(field), 1,
+                                     name=field, instance=self._document)
+
+        doc_field = self._document._fields.get(field.split('.', 1)[0])
+        instance = None
+
+        # We may need to cast to the correct type eg. ListField(EmbeddedDocumentField)
+        EmbeddedDocumentField = _import_class('EmbeddedDocumentField')
+        ListField = _import_class('ListField')
+        GenericEmbeddedDocumentField = _import_class('GenericEmbeddedDocumentField')
+        if isinstance(doc_field, ListField):
+            doc_field = getattr(doc_field, 'field', doc_field)
+        if isinstance(doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)):
+            instance = getattr(doc_field, 'document_type', None)
+
+        # handle distinct on subdocuments
+        if '.' in field:
+            for field_part in field.split('.')[1:]:
+                # if looping on embedded document, get the document type instance
+                if instance and isinstance(doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)):
+                    doc_field = instance
+                # now get the subdocument
+                doc_field = getattr(doc_field, field_part, doc_field)
+                # We may need to cast to the correct type eg. ListField(EmbeddedDocumentField)
+                if isinstance(doc_field, ListField):
+                    doc_field = getattr(doc_field, 'field', doc_field)
+                if isinstance(doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)):
+                    instance = getattr(doc_field, 'document_type', None)
+
+        if instance and isinstance(doc_field, (EmbeddedDocumentField,
+                                               GenericEmbeddedDocumentField)):
+            distinct = [instance(**doc) for doc in distinct]
+
+        return distinct
 
     def only(self, *fields):
         """Load only a subset of this document's fields. ::
 
-            post = BlogPost.objects(...).only("title", "author.name")
+            post = BlogPost.objects(...).only('title', 'author.name')
 
         .. note :: `only()` is chainable and will perform a union ::
             So with the following it will fetch both: `title` and `author.name`::
 
-                post = BlogPost.objects.only("title").only("author.name")
+                post = BlogPost.objects.only('title').only('author.name')
 
         :func:`~mongoengine.queryset.QuerySet.all_fields` will reset any
         field filters.
@@ -846,19 +875,19 @@ class BaseQuerySet(object):
         .. versionadded:: 0.3
         .. versionchanged:: 0.5 - Added subfield support
         """
-        fields = dict([(f, QueryFieldList.ONLY) for f in fields])
+        fields = {f: QueryFieldList.ONLY for f in fields}
         self.only_fields = fields.keys()
         return self.fields(True, **fields)
 
     def exclude(self, *fields):
         """Opposite to .only(), exclude some document's fields. ::
 
-            post = BlogPost.objects(...).exclude("comments")
+            post = BlogPost.objects(...).exclude('comments')
 
         .. note :: `exclude()` is chainable and will perform a union ::
             So with the following it will exclude both: `title` and `author.name`::
 
-                post = BlogPost.objects.exclude("title").exclude("author.name")
+                post = BlogPost.objects.exclude('title').exclude('author.name')
 
         :func:`~mongoengine.queryset.QuerySet.all_fields` will reset any
         field filters.
@@ -867,7 +896,7 @@ class BaseQuerySet(object):
 
         .. versionadded:: 0.5
         """
-        fields = dict([(f, QueryFieldList.EXCLUDE) for f in fields])
+        fields = {f: QueryFieldList.EXCLUDE for f in fields}
         return self.fields(**fields)
 
     def fields(self, _only_called=False, **kwargs):
@@ -888,7 +917,7 @@ class BaseQuerySet(object):
         """
 
         # Check for an operator and transform to mongo-style if there is
-        operators = ["slice"]
+        operators = ['slice']
         cleaned_fields = []
         for key, value in kwargs.items():
             parts = key.split('__')
@@ -912,7 +941,7 @@ class BaseQuerySet(object):
         """Include all fields. Reset all previously calls of .only() or
         .exclude(). ::
 
-            post = BlogPost.objects.exclude("comments").all_fields()
+            post = BlogPost.objects.exclude('comments').all_fields()
 
         .. versionadded:: 0.5
         """
@@ -933,6 +962,14 @@ class BaseQuerySet(object):
         queryset._ordering = queryset._get_order_by(keys)
         return queryset
 
+    def comment(self, text):
+        """Add a comment to the query.
+
+        See https://docs.mongodb.com/manual/reference/method/cursor.comment/#cursor.comment
+        for details.
+        """
+        return self._chainable_method('comment', text)
+
     def explain(self, format=False):
         """Return an explain plan record for the
         :class:`~mongoengine.queryset.QuerySet`\ 's cursor.
@@ -940,8 +977,15 @@ class BaseQuerySet(object):
         :param format: format the plan before returning it
         """
         plan = self._cursor.explain()
+
+        # TODO remove this option completely - it's useless. If somebody
+        # wants to pretty-print the output, they easily can.
         if format:
+            msg = ('"format" param of BaseQuerySet.explain has been '
+                   'deprecated and will be removed in future versions.')
+            warnings.warn(msg, DeprecationWarning)
             plan = pprint.pformat(plan)
+
         return plan
 
     # DEPRECATED. Has no more impact on PyMongo 3+
@@ -954,7 +998,7 @@ class BaseQuerySet(object):
         .. deprecated:: Ignored with PyMongo 3+
         """
         if IS_PYMONGO_3:
-            msg = "snapshot is deprecated as it has no impact when using PyMongo 3+."
+            msg = 'snapshot is deprecated as it has no impact when using PyMongo 3+.'
             warnings.warn(msg, DeprecationWarning)
         queryset = self.clone()
         queryset._snapshot = enabled
@@ -980,7 +1024,7 @@ class BaseQuerySet(object):
         .. deprecated:: Ignored with PyMongo 3+
         """
         if IS_PYMONGO_3:
-            msg = "slave_okay is deprecated as it has no impact when using PyMongo 3+."
+            msg = 'slave_okay is deprecated as it has no impact when using PyMongo 3+.'
             warnings.warn(msg, DeprecationWarning)
         queryset = self.clone()
         queryset._slave_okay = enabled
@@ -1042,7 +1086,7 @@ class BaseQuerySet(object):
 
         :param ms: the number of milliseconds before killing the query on the server
         """
-        return self._chainable_method("max_time_ms", ms)
+        return self._chainable_method('max_time_ms', ms)
 
     # JSON Helpers
 
@@ -1125,19 +1169,19 @@ class BaseQuerySet(object):
 
         MapReduceDocument = _import_class('MapReduceDocument')
 
-        if not hasattr(self._collection, "map_reduce"):
-            raise NotImplementedError("Requires MongoDB >= 1.7.1")
+        if not hasattr(self._collection, 'map_reduce'):
+            raise NotImplementedError('Requires MongoDB >= 1.7.1')
 
         map_f_scope = {}
         if isinstance(map_f, Code):
             map_f_scope = map_f.scope
-            map_f = unicode(map_f)
+            map_f = six.text_type(map_f)
         map_f = Code(queryset._sub_js_fields(map_f), map_f_scope)
 
         reduce_f_scope = {}
         if isinstance(reduce_f, Code):
             reduce_f_scope = reduce_f.scope
-            reduce_f = unicode(reduce_f)
+            reduce_f = six.text_type(reduce_f)
         reduce_f_code = queryset._sub_js_fields(reduce_f)
         reduce_f = Code(reduce_f_code, reduce_f_scope)
 
@@ -1147,7 +1191,7 @@ class BaseQuerySet(object):
             finalize_f_scope = {}
             if isinstance(finalize_f, Code):
                 finalize_f_scope = finalize_f.scope
-                finalize_f = unicode(finalize_f)
+                finalize_f = six.text_type(finalize_f)
             finalize_f_code = queryset._sub_js_fields(finalize_f)
             finalize_f = Code(finalize_f_code, finalize_f_scope)
             mr_args['finalize'] = finalize_f
@@ -1163,7 +1207,7 @@ class BaseQuerySet(object):
         else:
             map_reduce_function = 'map_reduce'
 
-            if isinstance(output, basestring):
+            if isinstance(output, six.string_types):
                 mr_args['out'] = output
 
             elif isinstance(output, dict):
@@ -1176,7 +1220,7 @@ class BaseQuerySet(object):
                         break
 
                 else:
-                    raise OperationError("actionData not specified for output")
+                    raise OperationError('actionData not specified for output')
 
                 db_alias = output.get('db_alias')
                 remaing_args = ['db', 'sharded', 'nonAtomic']
@@ -1406,7 +1450,7 @@ class BaseQuerySet(object):
             # snapshot is not handled at all by PyMongo 3+
             # TODO: evaluate similar possibilities using modifiers
             if self._snapshot:
-                msg = "The snapshot option is not anymore available with PyMongo 3+"
+                msg = 'The snapshot option is not anymore available with PyMongo 3+'
                 warnings.warn(msg, DeprecationWarning)
             cursor_args = {
                 'no_cursor_timeout': not self._timeout
@@ -1418,7 +1462,7 @@ class BaseQuerySet(object):
             if fields_name not in cursor_args:
                 cursor_args[fields_name] = {}
 
-            cursor_args[fields_name]['_text_score'] = {'$meta': "textScore"}
+            cursor_args[fields_name]['_text_score'] = {'$meta': 'textScore'}
 
         return cursor_args
 
@@ -1459,6 +1503,9 @@ class BaseQuerySet(object):
             if self._hint != -1:
                 self._cursor_obj.hint(self._hint)
 
+            if self._batch_size is not None:
+                self._cursor_obj.batch_size(self._batch_size)
+
         return self._cursor_obj
 
     def __deepcopy__(self, memo):
@@ -1470,8 +1517,8 @@ class BaseQuerySet(object):
         if self._mongo_query is None:
             self._mongo_query = self._query_obj.to_query(self._document)
             if self._class_check and self._initial_query:
-                if "_cls" in self._mongo_query:
-                    self._mongo_query = {"$and": [self._initial_query, self._mongo_query]}
+                if '_cls' in self._mongo_query:
+                    self._mongo_query = {'$and': [self._initial_query, self._mongo_query]}
                 else:
                     self._mongo_query.update(self._initial_query)
         return self._mongo_query
@@ -1483,8 +1530,7 @@ class BaseQuerySet(object):
         return self.__dereference
 
     def no_dereference(self):
-        """Turn off any dereferencing for the results of this queryset.
-        """
+        """Turn off any dereferencing for the results of this queryset."""
         queryset = self.clone()
         queryset._auto_dereference = False
         return queryset
@@ -1513,7 +1559,7 @@ class BaseQuerySet(object):
                     emit(null, 1);
                 }
             }
-        """ % dict(field=field)
+        """ % {'field': field}
         reduce_func = """
             function(key, values) {
                 var total = 0;
@@ -1535,8 +1581,8 @@ class BaseQuerySet(object):
 
         if normalize:
             count = sum(frequencies.values())
-            frequencies = dict([(k, float(v) / count)
-                                for k, v in frequencies.items()])
+            frequencies = {k: float(v) / count
+                           for k, v in frequencies.items()}
 
         return frequencies
 
@@ -1588,10 +1634,10 @@ class BaseQuerySet(object):
             }
         """
         total, data, types = self.exec_js(freq_func, field)
-        values = dict([(types.get(k), int(v)) for k, v in data.iteritems()])
+        values = {types.get(k): int(v) for k, v in data.iteritems()}
 
         if normalize:
-            values = dict([(k, float(v) / total) for k, v in values.items()])
+            values = {k: float(v) / total for k, v in values.items()}
 
         frequencies = {}
         for k, v in values.iteritems():
@@ -1613,14 +1659,14 @@ class BaseQuerySet(object):
                           for x in document._subclasses][1:]
         for field in fields:
             try:
-                field = ".".join(f.db_field for f in
+                field = '.'.join(f.db_field for f in
                                  document._lookup_field(field.split('.')))
                 ret.append(field)
-            except LookUpError, err:
+            except LookUpError as err:
                 found = False
                 for subdoc in subclasses:
                     try:
-                        subfield = ".".join(f.db_field for f in
+                        subfield = '.'.join(f.db_field for f in
                                             subdoc._lookup_field(field.split('.')))
                         ret.append(subfield)
                         found = True
@@ -1633,15 +1679,14 @@ class BaseQuerySet(object):
         return ret
 
     def _get_order_by(self, keys):
-        """Creates a list of order by fields
-        """
+        """Creates a list of order by fields"""
         key_list = []
         for key in keys:
             if not key:
                 continue
 
             if key == '$text_score':
-                key_list.append(('_text_score', {'$meta': "textScore"}))
+                key_list.append(('_text_score', {'$meta': 'textScore'}))
                 continue
 
             direction = pymongo.ASCENDING
@@ -1713,7 +1758,7 @@ class BaseQuerySet(object):
                     # If we need to coerce types, we need to determine the
                     # type of this field and use the corresponding
                     # .to_python(...)
-                    from mongoengine.fields import EmbeddedDocumentField
+                    EmbeddedDocumentField = _import_class('EmbeddedDocumentField')
 
                     obj = self._document
                     for chunk in path.split('.'):
@@ -1747,7 +1792,7 @@ class BaseQuerySet(object):
             field_name = match.group(1).split('.')
             fields = self._document._lookup_field(field_name)
             # Substitute the correct name for the field into the javascript
-            return ".".join([f.db_field for f in fields])
+            return '.'.join([f.db_field for f in fields])
 
         code = re.sub(u'\[\s*~([A-z_][A-z_0-9.]+?)\s*\]', field_sub, code)
         code = re.sub(u'\{\{\s*~([A-z_][A-z_0-9.]+?)\s*\}\}', field_path_sub,
@@ -1758,21 +1803,21 @@ class BaseQuerySet(object):
         queryset = self.clone()
         method = getattr(queryset._cursor, method_name)
         method(val)
-        setattr(queryset, "_" + method_name, val)
+        setattr(queryset, '_' + method_name, val)
         return queryset
 
     # Deprecated
     def ensure_index(self, **kwargs):
         """Deprecated use :func:`Document.ensure_index`"""
-        msg = ("Doc.objects()._ensure_index() is deprecated. "
-               "Use Doc.ensure_index() instead.")
+        msg = ('Doc.objects()._ensure_index() is deprecated. '
+               'Use Doc.ensure_index() instead.')
         warnings.warn(msg, DeprecationWarning)
         self._document.__class__.ensure_index(**kwargs)
         return self
 
     def _ensure_indexes(self):
         """Deprecated use :func:`~Document.ensure_indexes`"""
-        msg = ("Doc.objects()._ensure_indexes() is deprecated. "
-               "Use Doc.ensure_indexes() instead.")
+        msg = ('Doc.objects()._ensure_indexes() is deprecated. '
+               'Use Doc.ensure_indexes() instead.')
         warnings.warn(msg, DeprecationWarning)
         self._document.__class__.ensure_indexes()

mongoengine/queryset/field_list.py

@@ -67,7 +67,7 @@ class QueryFieldList(object):
         return bool(self.fields)
 
     def as_dict(self):
-        field_list = dict((field, self.value) for field in self.fields)
+        field_list = {field: self.value for field in self.fields}
         if self.slice:
             field_list.update(self.slice)
         if self._id is not None:

mongoengine/queryset/queryset.py

@@ -27,9 +27,10 @@ class QuerySet(BaseQuerySet):
         in batches of ``ITER_CHUNK_SIZE``.
 
         If ``self._has_more`` the cursor hasn't been exhausted so cache then
-        batch.  Otherwise iterate the result_cache.
+        batch. Otherwise iterate the result_cache.
         """
         self._iter = True
+
         if self._has_more:
             return self._iter_results()
 
@@ -42,40 +43,56 @@ class QuerySet(BaseQuerySet):
         """
         if self._len is not None:
             return self._len
+
+        # Populate the result cache with *all* of the docs in the cursor
         if self._has_more:
-            # populate the cache
             list(self._iter_results())
 
+        # Cache the length of the complete result cache and return it
         self._len = len(self._result_cache)
         return self._len
 
     def __repr__(self):
-        """Provides the string representation of the QuerySet
-        """
+        """Provide a string representation of the QuerySet"""
         if self._iter:
             return '.. queryset mid-iteration ..'
 
         self._populate_cache()
         data = self._result_cache[:REPR_OUTPUT_SIZE + 1]
         if len(data) > REPR_OUTPUT_SIZE:
-            data[-1] = "...(remaining elements truncated)..."
+            data[-1] = '...(remaining elements truncated)...'
         return repr(data)
 
     def _iter_results(self):
         """A generator for iterating over the result cache.
 
-        Also populates the cache if there are more possible results to yield.
-        Raises StopIteration when there are no more results"""
+        Also populates the cache if there are more possible results to
+        yield. Raises StopIteration when there are no more results.
+        """
         if self._result_cache is None:
             self._result_cache = []
+
         pos = 0
         while True:
-            upper = len(self._result_cache)
-            while pos < upper:
+
+            # For all positions lower than the length of the current result
+            # cache, serve the docs straight from the cache w/o hitting the
+            # database.
+            # XXX it's VERY important to compute the len within the `while`
+            # condition because the result cache might expand mid-iteration
+            # (e.g. if we call len(qs) inside a loop that iterates over the
+            # queryset). Fortunately len(list) is O(1) in Python, so this
+            # doesn't cause performance issues.
+            while pos < len(self._result_cache):
                 yield self._result_cache[pos]
                 pos += 1
+
+            # Raise StopIteration if we already established there were no more
+            # docs in the db cursor.
             if not self._has_more:
                 raise StopIteration
+
+            # Otherwise, populate more of the cache and repeat.
             if len(self._result_cache) <= pos:
                 self._populate_cache()
 
@@ -86,12 +103,22 @@ class QuerySet(BaseQuerySet):
         """
         if self._result_cache is None:
             self._result_cache = []
-        if self._has_more:
-            try:
-                for i in xrange(ITER_CHUNK_SIZE):
-                    self._result_cache.append(self.next())
-            except StopIteration:
-                self._has_more = False
+
+        # Skip populating the cache if we already established there are no
+        # more docs to pull from the database.
+        if not self._has_more:
+            return
+
+        # Pull in ITER_CHUNK_SIZE docs from the database and store them in
+        # the result cache.
+        try:
+            for _ in xrange(ITER_CHUNK_SIZE):
+                self._result_cache.append(self.next())
+        except StopIteration:
+            # Getting this exception means there are no more docs in the
+            # db cursor. Set _has_more to False so that we can use that
+            # information in other places.
+            self._has_more = False
 
     def count(self, with_limit_and_skip=False):
         """Count the selected elements in the query.
@@ -114,7 +141,7 @@ class QuerySet(BaseQuerySet):
         .. versionadded:: 0.8.3 Convert to non caching queryset
         """
         if self._result_cache is not None:
-            raise OperationError("QuerySet already cached")
+            raise OperationError('QuerySet already cached')
         return self.clone_into(QuerySetNoCache(self._document, self._collection))
 
 
@@ -137,13 +164,14 @@ class QuerySetNoCache(BaseQuerySet):
             return '.. queryset mid-iteration ..'
 
         data = []
-        for i in xrange(REPR_OUTPUT_SIZE + 1):
+        for _ in xrange(REPR_OUTPUT_SIZE + 1):
             try:
                 data.append(self.next())
             except StopIteration:
                 break
+
         if len(data) > REPR_OUTPUT_SIZE:
-            data[-1] = "...(remaining elements truncated)..."
+            data[-1] = '...(remaining elements truncated)...'
 
         self.rewind()
         return repr(data)

mongoengine/queryset/transform.py

@@ -1,9 +1,11 @@
 from collections import defaultdict
 
-from bson import SON
+from bson import ObjectId, SON
+from bson.dbref import DBRef
 import pymongo
+import six
 
-from mongoengine.base.fields import UPDATE_OPERATORS
+from mongoengine.base import UPDATE_OPERATORS
 from mongoengine.common import _import_class
 from mongoengine.connection import get_connection
 from mongoengine.errors import InvalidQueryError
@@ -26,13 +28,13 @@ MATCH_OPERATORS = (COMPARISON_OPERATORS + GEO_OPERATORS +
                    STRING_OPERATORS + CUSTOM_OPERATORS)
 
 
+# TODO make this less complex
 def query(_doc_cls=None, **kwargs):
-    """Transform a query from Django-style format to Mongo format.
-    """
+    """Transform a query from Django-style format to Mongo format."""
     mongo_query = {}
     merge_query = defaultdict(list)
     for key, value in sorted(kwargs.items()):
-        if key == "__raw__":
+        if key == '__raw__':
             mongo_query.update(value)
             continue
 
@@ -45,7 +47,7 @@ def query(_doc_cls=None, **kwargs):
             op = parts.pop()
 
         # Allow to escape operator-like field name by __
-        if len(parts) > 1 and parts[-1] == "":
+        if len(parts) > 1 and parts[-1] == '':
             parts.pop()
 
         negate = False
@@ -57,16 +59,17 @@ def query(_doc_cls=None, **kwargs):
             # Switch field names to proper names [set in Field(name='foo')]
             try:
                 fields = _doc_cls._lookup_field(parts)
-            except Exception, e:
+            except Exception as e:
                 raise InvalidQueryError(e)
             parts = []
 
             CachedReferenceField = _import_class('CachedReferenceField')
+            GenericReferenceField = _import_class('GenericReferenceField')
 
             cleaned_fields = []
             for field in fields:
                 append_field = True
-                if isinstance(field, basestring):
+                if isinstance(field, six.string_types):
                     parts.append(field)
                     append_field = False
                 # is last and CachedReferenceField
@@ -84,9 +87,9 @@ def query(_doc_cls=None, **kwargs):
             singular_ops = [None, 'ne', 'gt', 'gte', 'lt', 'lte', 'not']
             singular_ops += STRING_OPERATORS
             if op in singular_ops:
-                if isinstance(field, basestring):
+                if isinstance(field, six.string_types):
                     if (op in STRING_OPERATORS and
-                            isinstance(value, basestring)):
+                            isinstance(value, six.string_types)):
                         StringField = _import_class('StringField')
                         value = StringField.prepare_query_value(op, value)
                     else:
@@ -101,6 +104,16 @@ def query(_doc_cls=None, **kwargs):
                 # 'in', 'nin' and 'all' require a list of values
                 value = [field.prepare_query_value(op, v) for v in value]
 
+            # If we're querying a GenericReferenceField, we need to alter the
+            # key depending on the value:
+            # * If the value is a DBRef, the key should be "field_name._ref".
+            # * If the value is an ObjectId, the key should be "field_name._ref.$id".
+            if isinstance(field, GenericReferenceField):
+                if isinstance(value, DBRef):
+                    parts[-1] += '._ref'
+                elif isinstance(value, ObjectId):
+                    parts[-1] += '._ref.$id'
+
         # if op and op not in COMPARISON_OPERATORS:
         if op:
             if op in GEO_OPERATORS:
@@ -116,10 +129,10 @@ def query(_doc_cls=None, **kwargs):
                     value = query(field.field.document_type, **value)
                 else:
                     value = field.prepare_query_value(op, value)
-                value = {"$elemMatch": value}
+                value = {'$elemMatch': value}
             elif op in CUSTOM_OPERATORS:
-                NotImplementedError("Custom method '%s' has not "
-                                    "been implemented" % op)
+                NotImplementedError('Custom method "%s" has not '
+                                    'been implemented' % op)
             elif op not in STRING_OPERATORS:
                 value = {'$' + op: value}
 
@@ -128,11 +141,13 @@ def query(_doc_cls=None, **kwargs):
 
         for i, part in indices:
             parts.insert(i, part)
+
         key = '.'.join(parts)
+
         if op is None or key not in mongo_query:
             mongo_query[key] = value
         elif key in mongo_query:
-            if key in mongo_query and isinstance(mongo_query[key], dict):
+            if isinstance(mongo_query[key], dict):
                 mongo_query[key].update(value)
                 # $max/minDistance needs to come last - convert to SON
                 value_dict = mongo_query[key]
@@ -182,15 +197,16 @@ def query(_doc_cls=None, **kwargs):
 
 
 def update(_doc_cls=None, **update):
-    """Transform an update spec from Django-style format to Mongo format.
+    """Transform an update spec from Django-style format to Mongo
+    format.
     """
     mongo_update = {}
     for key, value in update.items():
-        if key == "__raw__":
+        if key == '__raw__':
             mongo_update.update(value)
             continue
         parts = key.split('__')
-        # if there is no operator, default to "set"
+        # if there is no operator, default to 'set'
         if len(parts) < 3 and parts[0] not in UPDATE_OPERATORS:
             parts.insert(0, 'set')
         # Check for an operator and transform to mongo-style if there is
@@ -209,21 +225,21 @@ def update(_doc_cls=None, **update):
             elif op == 'add_to_set':
                 op = 'addToSet'
             elif op == 'set_on_insert':
-                op = "setOnInsert"
+                op = 'setOnInsert'
 
         match = None
         if parts[-1] in COMPARISON_OPERATORS:
             match = parts.pop()
 
         # Allow to escape operator-like field name by __
-        if len(parts) > 1 and parts[-1] == "":
+        if len(parts) > 1 and parts[-1] == '':
             parts.pop()
 
         if _doc_cls:
             # Switch field names to proper names [set in Field(name='foo')]
             try:
                 fields = _doc_cls._lookup_field(parts)
-            except Exception, e:
+            except Exception as e:
                 raise InvalidQueryError(e)
             parts = []
 
@@ -231,7 +247,7 @@ def update(_doc_cls=None, **update):
             appended_sub_field = False
             for field in fields:
                 append_field = True
-                if isinstance(field, basestring):
+                if isinstance(field, six.string_types):
                     # Convert the S operator to $
                     if field == 'S':
                         field = '$'
@@ -252,7 +268,7 @@ def update(_doc_cls=None, **update):
             else:
                 field = cleaned_fields[-1]
 
-            GeoJsonBaseField = _import_class("GeoJsonBaseField")
+            GeoJsonBaseField = _import_class('GeoJsonBaseField')
             if isinstance(field, GeoJsonBaseField):
                 value = field.to_mongo(value)
 
@@ -266,7 +282,7 @@ def update(_doc_cls=None, **update):
                     value = [field.prepare_query_value(op, v) for v in value]
                 elif field.required or value is not None:
                     value = field.prepare_query_value(op, value)
-            elif op == "unset":
+            elif op == 'unset':
                 value = 1
 
         if match:
@@ -276,16 +292,16 @@ def update(_doc_cls=None, **update):
         key = '.'.join(parts)
 
         if not op:
-            raise InvalidQueryError("Updates must supply an operation "
-                                    "eg: set__FIELD=value")
+            raise InvalidQueryError('Updates must supply an operation '
+                                    'eg: set__FIELD=value')
 
         if 'pull' in op and '.' in key:
             # Dot operators don't work on pull operations
             # unless they point to a list field
             # Otherwise it uses nested dict syntax
             if op == 'pullAll':
-                raise InvalidQueryError("pullAll operations only support "
-                                        "a single field depth")
+                raise InvalidQueryError('pullAll operations only support '
+                                        'a single field depth')
 
             # Look for the last list field and use dot notation until there
             field_classes = [c.__class__ for c in cleaned_fields]
@@ -296,7 +312,7 @@ def update(_doc_cls=None, **update):
                 # Then process as normal
                 last_listField = len(
                     cleaned_fields) - field_classes.index(ListField)
-                key = ".".join(parts[:last_listField])
+                key = '.'.join(parts[:last_listField])
                 parts = parts[last_listField:]
                 parts.insert(0, key)
 
@@ -304,7 +320,7 @@ def update(_doc_cls=None, **update):
             for key in parts:
                 value = {key: value}
         elif op == 'addToSet' and isinstance(value, list):
-            value = {key: {"$each": value}}
+            value = {key: {'$each': value}}
         else:
             value = {key: value}
         key = '$' + op
@@ -318,78 +334,82 @@ def update(_doc_cls=None, **update):
 
 
 def _geo_operator(field, op, value):
-    """Helper to return the query for a given geo query"""
-    if op == "max_distance":
+    """Helper to return the query for a given geo query."""
+    if op == 'max_distance':
         value = {'$maxDistance': value}
-    elif op == "min_distance":
+    elif op == 'min_distance':
         value = {'$minDistance': value}
     elif field._geo_index == pymongo.GEO2D:
-        if op == "within_distance":
+        if op == 'within_distance':
             value = {'$within': {'$center': value}}
-        elif op == "within_spherical_distance":
+        elif op == 'within_spherical_distance':
             value = {'$within': {'$centerSphere': value}}
-        elif op == "within_polygon":
+        elif op == 'within_polygon':
             value = {'$within': {'$polygon': value}}
-        elif op == "near":
+        elif op == 'near':
             value = {'$near': value}
-        elif op == "near_sphere":
+        elif op == 'near_sphere':
             value = {'$nearSphere': value}
         elif op == 'within_box':
             value = {'$within': {'$box': value}}
         else:
-            raise NotImplementedError("Geo method '%s' has not "
-                                      "been implemented for a GeoPointField" % op)
+            raise NotImplementedError('Geo method "%s" has not been '
+                                      'implemented for a GeoPointField' % op)
     else:
-        if op == "geo_within":
-            value = {"$geoWithin": _infer_geometry(value)}
-        elif op == "geo_within_box":
-            value = {"$geoWithin": {"$box": value}}
-        elif op == "geo_within_polygon":
-            value = {"$geoWithin": {"$polygon": value}}
-        elif op == "geo_within_center":
-            value = {"$geoWithin": {"$center": value}}
-        elif op == "geo_within_sphere":
-            value = {"$geoWithin": {"$centerSphere": value}}
-        elif op == "geo_intersects":
-            value = {"$geoIntersects": _infer_geometry(value)}
-        elif op == "near":
+        if op == 'geo_within':
+            value = {'$geoWithin': _infer_geometry(value)}
+        elif op == 'geo_within_box':
+            value = {'$geoWithin': {'$box': value}}
+        elif op == 'geo_within_polygon':
+            value = {'$geoWithin': {'$polygon': value}}
+        elif op == 'geo_within_center':
+            value = {'$geoWithin': {'$center': value}}
+        elif op == 'geo_within_sphere':
+            value = {'$geoWithin': {'$centerSphere': value}}
+        elif op == 'geo_intersects':
+            value = {'$geoIntersects': _infer_geometry(value)}
+        elif op == 'near':
             value = {'$near': _infer_geometry(value)}
         else:
-            raise NotImplementedError("Geo method '%s' has not "
-                                      "been implemented for a %s " % (op, field._name))
+            raise NotImplementedError(
+                'Geo method "%s" has not been implemented for a %s '
+                % (op, field._name)
+            )
     return value
 
 
 def _infer_geometry(value):
-    """Helper method that tries to infer the $geometry shape for a given value"""
+    """Helper method that tries to infer the $geometry shape for a
+    given value.
+    """
     if isinstance(value, dict):
-        if "$geometry" in value:
+        if '$geometry' in value:
             return value
         elif 'coordinates' in value and 'type' in value:
-            return {"$geometry": value}
-        raise InvalidQueryError("Invalid $geometry dictionary should have "
-                                "type and coordinates keys")
+            return {'$geometry': value}
+        raise InvalidQueryError('Invalid $geometry dictionary should have '
+                                'type and coordinates keys')
     elif isinstance(value, (list, set)):
         # TODO: shouldn't we test value[0][0][0][0] to see if it is MultiPolygon?
         # TODO: should both TypeError and IndexError be alike interpreted?
 
         try:
             value[0][0][0]
-            return {"$geometry": {"type": "Polygon", "coordinates": value}}
+            return {'$geometry': {'type': 'Polygon', 'coordinates': value}}
         except (TypeError, IndexError):
             pass
 
         try:
             value[0][0]
-            return {"$geometry": {"type": "LineString", "coordinates": value}}
+            return {'$geometry': {'type': 'LineString', 'coordinates': value}}
         except (TypeError, IndexError):
             pass
 
         try:
             value[0]
-            return {"$geometry": {"type": "Point", "coordinates": value}}
+            return {'$geometry': {'type': 'Point', 'coordinates': value}}
         except (TypeError, IndexError):
             pass
 
-    raise InvalidQueryError("Invalid $geometry data. Can be either a dictionary "
-                            "or (nested) lists of coordinate(s)")
+    raise InvalidQueryError('Invalid $geometry data. Can be either a '
+                            'dictionary or (nested) lists of coordinate(s)')

mongoengine/queryset/visitor.py

@@ -69,9 +69,9 @@ class QueryCompilerVisitor(QNodeVisitor):
         self.document = document
 
     def visit_combination(self, combination):
-        operator = "$and"
+        operator = '$and'
         if combination.operation == combination.OR:
-            operator = "$or"
+            operator = '$or'
         return {operator: combination.children}
 
     def visit_query(self, query):
@@ -79,8 +79,7 @@ class QueryCompilerVisitor(QNodeVisitor):
 
 
 class QNode(object):
-    """Base class for nodes in query trees.
-    """
+    """Base class for nodes in query trees."""
 
     AND = 0
     OR = 1
@@ -94,7 +93,8 @@ class QNode(object):
         raise NotImplementedError
 
     def _combine(self, other, operation):
-        """Combine this node with another node into a QCombination object.
+        """Combine this node with another node into a QCombination
+        object.
         """
         if getattr(other, 'empty', True):
             return self
@@ -116,8 +116,8 @@ class QNode(object):
 
 
 class QCombination(QNode):
-    """Represents the combination of several conditions by a given logical
-    operator.
+    """Represents the combination of several conditions by a given
+    logical operator.
     """
 
     def __init__(self, operation, children):

mongoengine/signals.py

@@ -1,7 +1,5 @@
-# -*- coding: utf-8 -*-
-
-__all__ = ['pre_init', 'post_init', 'pre_save', 'pre_save_post_validation',
-           'post_save', 'pre_delete', 'post_delete']
+__all__ = ('pre_init', 'post_init', 'pre_save', 'pre_save_post_validation',
+           'post_save', 'pre_delete', 'post_delete')
 
 signals_available = False
 try:
@@ -34,6 +32,7 @@ except ImportError:
             temporarily_connected_to = _fail
         del _fail
 
+
 # the namespace for code signals.  If you are not mongoengine code, do
 # not put signals in here.  Create your own namespace instead.
 _signals = Namespace()

setup.cfg

@@ -1,13 +1,11 @@
 [nosetests]
-verbosity = 2
-detailed-errors = 1
-cover-erase = 1
-cover-branches = 1
-cover-package = mongoengine
-tests = tests
+verbosity=2
+detailed-errors=1
+tests=tests
+cover-package=mongoengine
 
 [flake8]
 ignore=E501,F401,F403,F405,I201
-exclude=build,dist,docs,venv,.tox,.eggs,tests
-max-complexity=42
+exclude=build,dist,docs,venv,venv3,.tox,.eggs,tests
+max-complexity=45
 application-import-names=mongoengine,tests

setup.py

@@ -21,8 +21,9 @@ except Exception:
 
 
 def get_version(version_tuple):
-    if not isinstance(version_tuple[-1], int):
-        return '.'.join(map(str, version_tuple[:-1])) + version_tuple[-1]
+    """Return the version tuple as a string, e.g. for (0, 10, 7),
+    return '0.10.7'.
+    """
     return '.'.join(map(str, version_tuple))
 
 
@@ -41,31 +42,29 @@ CLASSIFIERS = [
     'Operating System :: OS Independent',
     'Programming Language :: Python',
     "Programming Language :: Python :: 2",
-    "Programming Language :: Python :: 2.6",
     "Programming Language :: Python :: 2.7",
     "Programming Language :: Python :: 3",
-    "Programming Language :: Python :: 3.2",
     "Programming Language :: Python :: 3.3",
     "Programming Language :: Python :: 3.4",
+    "Programming Language :: Python :: 3.5",
     "Programming Language :: Python :: Implementation :: CPython",
     "Programming Language :: Python :: Implementation :: PyPy",
     'Topic :: Database',
     'Topic :: Software Development :: Libraries :: Python Modules',
 ]
 
-extra_opts = {"packages": find_packages(exclude=["tests", "tests.*"])}
+extra_opts = {
+    'packages': find_packages(exclude=['tests', 'tests.*']),
+    'tests_require': ['nose', 'coverage==4.2', 'blinker', 'Pillow>=2.0.0']
+}
 if sys.version_info[0] == 3:
     extra_opts['use_2to3'] = True
-    extra_opts['tests_require'] = ['nose', 'coverage==3.7.1', 'blinker', 'Pillow>=2.0.0']
-    if "test" in sys.argv or "nosetests" in sys.argv:
+    if 'test' in sys.argv or 'nosetests' in sys.argv:
         extra_opts['packages'] = find_packages()
-        extra_opts['package_data'] = {"tests": ["fields/mongoengine.png", "fields/mongodb_leaf.png"]}
+        extra_opts['package_data'] = {
+            'tests': ['fields/mongoengine.png', 'fields/mongodb_leaf.png']}
 else:
-    # coverage 4 does not support Python 3.2 anymore
-    extra_opts['tests_require'] = ['nose', 'coverage==3.7.1', 'blinker', 'Pillow>=2.0.0', 'python-dateutil']
-
-    if sys.version_info[0] == 2 and sys.version_info[1] == 6:
-        extra_opts['tests_require'].append('unittest2')
+    extra_opts['tests_require'] += ['python-dateutil']
 
 setup(
     name='mongoengine',

tests/__init__.py

@@ -2,4 +2,3 @@ from all_warnings import AllWarnings
 from document import *
 from queryset import *
 from fields import *
-from migration import *

tests/all_warnings/__init__.py

@@ -3,8 +3,6 @@ This test has been put into a module.  This is because it tests warnings that
 only get triggered on first hit.  This way we can ensure its imported into the
 top level and called first by the test suite.
 """
-import sys
-sys.path[0:0] = [""]
 import unittest
 import warnings
 

tests/document/__init__.py

@@ -1,5 +1,3 @@
-import sys
-sys.path[0:0] = [""]
 import unittest
 
 from class_methods import *

tests/document/class_methods.py

@@ -1,6 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
 import unittest
 
 from mongoengine import *

tests/document/delta.py

@@ -1,6 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
 import unittest
 
 from bson import SON

tests/document/dynamic.py

@@ -1,6 +1,4 @@
 import unittest
-import sys
-sys.path[0:0] = [""]
 
 from mongoengine import *
 from mongoengine.connection import get_db
@@ -143,11 +141,9 @@ class DynamicTest(unittest.TestCase):
 
     def test_three_level_complex_data_lookups(self):
         """Ensure you can query three level document dynamic fields"""
-        p = self.Person()
-        p.misc = {'hello': {'hello2': 'world'}}
-        p.save()
-        # from pprint import pprint as pp; import pdb; pdb.set_trace();
-        print self.Person.objects(misc__hello__hello2='world')
+        p = self.Person.objects.create(
+            misc={'hello': {'hello2': 'world'}}
+        )
         self.assertEqual(1, self.Person.objects(misc__hello__hello2='world').count())
 
     def test_complex_embedded_document_validation(self):

tests/document/indexes.py

@@ -2,10 +2,8 @@
 import unittest
 import sys
 
-sys.path[0:0] = [""]
 
 import pymongo
-from random import randint
 
 from nose.plugins.skip import SkipTest
 from datetime import datetime
@@ -17,11 +15,9 @@ __all__ = ("IndexesTest", )
 
 
 class IndexesTest(unittest.TestCase):
-    _MAX_RAND = 10 ** 10
 
     def setUp(self):
-        self.db_name = 'mongoenginetest_IndexesTest_' + str(randint(0, self._MAX_RAND))
-        self.connection = connect(db=self.db_name)
+        self.connection = connect(db='mongoenginetest')
         self.db = get_db()
 
         class Person(Document):
@@ -560,8 +556,8 @@ class IndexesTest(unittest.TestCase):
 
         BlogPost.drop_collection()
 
-        for i in xrange(0, 10):
-            tags = [("tag %i" % n) for n in xrange(0, i % 2)]
+        for i in range(0, 10):
+            tags = [("tag %i" % n) for n in range(0, i % 2)]
             BlogPost(tags=tags).save()
 
         self.assertEqual(BlogPost.objects.count(), 10)

tests/document/inheritance.py

@@ -1,6 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
 import unittest
 import warnings
 
@@ -253,19 +251,17 @@ class InheritanceTest(unittest.TestCase):
         self.assertEqual(classes, [Human])
 
     def test_allow_inheritance(self):
-        """Ensure that inheritance may be disabled on simple classes and that
-        _cls and _subclasses will not be used.
+        """Ensure that inheritance is disabled by default on simple
+        classes and that _cls will not be used.
         """
-
         class Animal(Document):
             name = StringField()
 
-        def create_dog_class():
+        # can't inherit because Animal didn't explicitly allow inheritance
+        with self.assertRaises(ValueError):
             class Dog(Animal):
                 pass
 
-        self.assertRaises(ValueError, create_dog_class)
-
         # Check that _cls etc aren't present on simple documents
         dog = Animal(name='dog').save()
         self.assertEqual(dog.to_mongo().keys(), ['_id', 'name'])
@@ -275,17 +271,15 @@ class InheritanceTest(unittest.TestCase):
         self.assertFalse('_cls' in obj)
 
     def test_cant_turn_off_inheritance_on_subclass(self):
-        """Ensure if inheritance is on in a subclass you cant turn it off
+        """Ensure if inheritance is on in a subclass you cant turn it off.
         """
-
         class Animal(Document):
             name = StringField()
             meta = {'allow_inheritance': True}
 
-        def create_mammal_class():
+        with self.assertRaises(ValueError):
             class Mammal(Animal):
                 meta = {'allow_inheritance': False}
-        self.assertRaises(ValueError, create_mammal_class)
 
     def test_allow_inheritance_abstract_document(self):
         """Ensure that abstract documents can set inheritance rules and that
@@ -298,10 +292,9 @@ class InheritanceTest(unittest.TestCase):
         class Animal(FinalDocument):
             name = StringField()
 
-        def create_mammal_class():
+        with self.assertRaises(ValueError):
             class Mammal(Animal):
                 pass
-        self.assertRaises(ValueError, create_mammal_class)
 
         # Check that _cls isn't present in simple documents
         doc = Animal(name='dog')
@@ -360,29 +353,26 @@ class InheritanceTest(unittest.TestCase):
         self.assertEqual(berlin.pk, berlin.auto_id_0)
 
     def test_abstract_document_creation_does_not_fail(self):
-
         class City(Document):
             continent = StringField()
             meta = {'abstract': True,
                     'allow_inheritance': False}
+
         bkk = City(continent='asia')
         self.assertEqual(None, bkk.pk)
         # TODO: expected error? Shouldn't we create a new error type?
-        self.assertRaises(KeyError, lambda: setattr(bkk, 'pk', 1))
+        with self.assertRaises(KeyError):
+            setattr(bkk, 'pk', 1)
 
     def test_allow_inheritance_embedded_document(self):
-        """Ensure embedded documents respect inheritance
-        """
-
+        """Ensure embedded documents respect inheritance."""
         class Comment(EmbeddedDocument):
             content = StringField()
 
-        def create_special_comment():
+        with self.assertRaises(ValueError):
             class SpecialComment(Comment):
                 pass
 
-        self.assertRaises(ValueError, create_special_comment)
-
         doc = Comment(content='test')
         self.assertFalse('_cls' in doc.to_mongo())
 
@@ -454,11 +444,11 @@ class InheritanceTest(unittest.TestCase):
         self.assertEqual(Guppy._get_collection_name(), 'fish')
         self.assertEqual(Human._get_collection_name(), 'human')
 
-        def create_bad_abstract():
+        # ensure that a subclass of a non-abstract class can't be abstract
+        with self.assertRaises(ValueError):
             class EvilHuman(Human):
                 evil = BooleanField(default=True)
                 meta = {'abstract': True}
-        self.assertRaises(ValueError, create_bad_abstract)
 
     def test_abstract_embedded_documents(self):
         # 789: EmbeddedDocument shouldn't inherit abstract

tests/document/instance.py

@@ -1,7 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
-
 import bson
 import os
 import pickle
@@ -16,12 +13,12 @@ from tests.fixtures import (PickleEmbedded, PickleTest, PickleSignalsTest,
                             PickleDynamicEmbedded, PickleDynamicTest)
 
 from mongoengine import *
+from mongoengine.base import get_document, _document_registry
+from mongoengine.connection import get_db
 from mongoengine.errors import (NotRegistered, InvalidDocumentError,
                                 InvalidQueryError, NotUniqueError,
                                 FieldDoesNotExist, SaveConditionError)
 from mongoengine.queryset import NULLIFY, Q
-from mongoengine.connection import get_db
-from mongoengine.base import get_document
 from mongoengine.context_managers import switch_db, query_counter
 from mongoengine import signals
 
@@ -102,21 +99,18 @@ class InstanceTest(unittest.TestCase):
         self.assertEqual(options['size'], 4096)
 
         # Check that the document cannot be redefined with different options
-        def recreate_log_document():
-            class Log(Document):
-                date = DateTimeField(default=datetime.now)
-                meta = {
-                    'max_documents': 11,
-                }
-            # Create the collection by accessing Document.objects
-            Log.objects
-        self.assertRaises(InvalidCollectionError, recreate_log_document)
+        class Log(Document):
+            date = DateTimeField(default=datetime.now)
+            meta = {
+                'max_documents': 11,
+            }
 
-        Log.drop_collection()
+        # Accessing Document.objects creates the collection
+        with self.assertRaises(InvalidCollectionError):
+            Log.objects
 
     def test_capped_collection_default(self):
-        """Ensure that capped collections defaults work properly.
-        """
+        """Ensure that capped collections defaults work properly."""
         class Log(Document):
             date = DateTimeField(default=datetime.now)
             meta = {
@@ -134,16 +128,14 @@ class InstanceTest(unittest.TestCase):
         self.assertEqual(options['size'], 10 * 2**20)
 
         # Check that the document with default value can be recreated
-        def recreate_log_document():
-            class Log(Document):
-                date = DateTimeField(default=datetime.now)
-                meta = {
-                    'max_documents': 10,
-                }
-            # Create the collection by accessing Document.objects
-            Log.objects
-        recreate_log_document()
-        Log.drop_collection()
+        class Log(Document):
+            date = DateTimeField(default=datetime.now)
+            meta = {
+                'max_documents': 10,
+            }
+
+        # Create the collection by accessing Document.objects
+        Log.objects
 
     def test_capped_collection_no_max_size_problems(self):
         """Ensure that capped collections with odd max_size work properly.
@@ -166,16 +158,14 @@ class InstanceTest(unittest.TestCase):
         self.assertTrue(options['size'] >= 10000)
 
         # Check that the document with odd max_size value can be recreated
-        def recreate_log_document():
-            class Log(Document):
-                date = DateTimeField(default=datetime.now)
-                meta = {
-                    'max_size': 10000,
-                }
-            # Create the collection by accessing Document.objects
-            Log.objects
-        recreate_log_document()
-        Log.drop_collection()
+        class Log(Document):
+            date = DateTimeField(default=datetime.now)
+            meta = {
+                'max_size': 10000,
+            }
+
+        # Create the collection by accessing Document.objects
+        Log.objects
 
     def test_repr(self):
         """Ensure that unicode representation works
@@ -286,7 +276,7 @@ class InstanceTest(unittest.TestCase):
 
         list_stats = []
 
-        for i in xrange(10):
+        for i in range(10):
             s = Stats()
             s.save()
             list_stats.append(s)
@@ -356,14 +346,14 @@ class InstanceTest(unittest.TestCase):
         self.assertEqual(User._fields['username'].db_field, '_id')
         self.assertEqual(User._meta['id_field'], 'username')
 
-        def create_invalid_user():
-            User(name='test').save()  # no primary key field
-        self.assertRaises(ValidationError, create_invalid_user)
+        # test no primary key field
+        self.assertRaises(ValidationError, User(name='test').save)
 
-        def define_invalid_user():
+        # define a subclass with a different primary key field than the
+        # parent
+        with self.assertRaises(ValueError):
             class EmailUser(User):
                 email = StringField(primary_key=True)
-        self.assertRaises(ValueError, define_invalid_user)
 
         class EmailUser(User):
             email = StringField()
@@ -411,12 +401,10 @@ class InstanceTest(unittest.TestCase):
 
         # Mimic Place and NicePlace definitions being in a different file
         # and the NicePlace model not being imported in at query time.
-        from mongoengine.base import _document_registry
         del(_document_registry['Place.NicePlace'])
 
-        def query_without_importing_nice_place():
-            print Place.objects.all()
-        self.assertRaises(NotRegistered, query_without_importing_nice_place)
+        with self.assertRaises(NotRegistered):
+            list(Place.objects.all())
 
     def test_document_registry_regressions(self):
 
@@ -745,7 +733,7 @@ class InstanceTest(unittest.TestCase):
 
         try:
             t.save()
-        except ValidationError, e:
+        except ValidationError as e:
             expect_msg = "Draft entries may not have a publication date."
             self.assertTrue(expect_msg in e.message)
             self.assertEqual(e.to_dict(), {'__all__': expect_msg})
@@ -784,7 +772,7 @@ class InstanceTest(unittest.TestCase):
         t = TestDocument(doc=TestEmbeddedDocument(x=10, y=25, z=15))
         try:
             t.save()
-        except ValidationError, e:
+        except ValidationError as e:
             expect_msg = "Value of z != x + y"
             self.assertTrue(expect_msg in e.message)
             self.assertEqual(e.to_dict(), {'doc': {'__all__': expect_msg}})
@@ -798,8 +786,10 @@ class InstanceTest(unittest.TestCase):
 
     def test_modify_empty(self):
         doc = self.Person(name="bob", age=10).save()
-        self.assertRaises(
-            InvalidDocumentError, lambda: self.Person().modify(set__age=10))
+
+        with self.assertRaises(InvalidDocumentError):
+            self.Person().modify(set__age=10)
+
         self.assertDbEqual([dict(doc.to_mongo())])
 
     def test_modify_invalid_query(self):
@@ -807,9 +797,8 @@ class InstanceTest(unittest.TestCase):
         doc2 = self.Person(name="jim", age=20).save()
         docs = [dict(doc1.to_mongo()), dict(doc2.to_mongo())]
 
-        self.assertRaises(
-            InvalidQueryError,
-            lambda: doc1.modify(dict(id=doc2.id), set__value=20))
+        with self.assertRaises(InvalidQueryError):
+            doc1.modify({'id': doc2.id}, set__value=20)
 
         self.assertDbEqual(docs)
 
@@ -818,7 +807,7 @@ class InstanceTest(unittest.TestCase):
         doc2 = self.Person(name="jim", age=20).save()
         docs = [dict(doc1.to_mongo()), dict(doc2.to_mongo())]
 
-        assert not doc1.modify(dict(name=doc2.name), set__age=100)
+        assert not doc1.modify({'name': doc2.name}, set__age=100)
 
         self.assertDbEqual(docs)
 
@@ -827,7 +816,7 @@ class InstanceTest(unittest.TestCase):
         doc2 = self.Person(id=ObjectId(), name="jim", age=20)
         docs = [dict(doc1.to_mongo())]
 
-        assert not doc2.modify(dict(name=doc2.name), set__age=100)
+        assert not doc2.modify({'name': doc2.name}, set__age=100)
 
         self.assertDbEqual(docs)
 
@@ -1293,12 +1282,11 @@ class InstanceTest(unittest.TestCase):
 
     def test_document_update(self):
 
-        def update_not_saved_raises():
+        # try updating a non-saved document
+        with self.assertRaises(OperationError):
             person = self.Person(name='dcrosta')
             person.update(set__name='Dan Crosta')
 
-        self.assertRaises(OperationError, update_not_saved_raises)
-
         author = self.Person(name='dcrosta')
         author.save()
 
@@ -1308,19 +1296,17 @@ class InstanceTest(unittest.TestCase):
         p1 = self.Person.objects.first()
         self.assertEqual(p1.name, author.name)
 
-        def update_no_value_raises():
+        # try sending an empty update
+        with self.assertRaises(OperationError):
             person = self.Person.objects.first()
             person.update()
 
-        self.assertRaises(OperationError, update_no_value_raises)
-
-        def update_no_op_should_default_to_set():
-            person = self.Person.objects.first()
-            person.update(name="Dan")
-            person.reload()
-            return person.name
-
-        self.assertEqual("Dan", update_no_op_should_default_to_set())
+        # update that doesn't explicitly specify an operator should default
+        # to 'set__'
+        person = self.Person.objects.first()
+        person.update(name="Dan")
+        person.reload()
+        self.assertEqual("Dan", person.name)
 
     def test_update_unique_field(self):
         class Doc(Document):
@@ -1329,8 +1315,8 @@ class InstanceTest(unittest.TestCase):
         doc1 = Doc(name="first").save()
         doc2 = Doc(name="second").save()
 
-        self.assertRaises(NotUniqueError, lambda:
-                          doc2.update(set__name=doc1.name))
+        with self.assertRaises(NotUniqueError):
+            doc2.update(set__name=doc1.name)
 
     def test_embedded_update(self):
         """
@@ -1848,15 +1834,13 @@ class InstanceTest(unittest.TestCase):
 
     def test_duplicate_db_fields_raise_invalid_document_error(self):
         """Ensure a InvalidDocumentError is thrown if duplicate fields
-        declare the same db_field"""
-
-        def throw_invalid_document_error():
+        declare the same db_field.
+        """
+        with self.assertRaises(InvalidDocumentError):
             class Foo(Document):
                 name = StringField()
                 name2 = StringField(db_field='name')
 
-        self.assertRaises(InvalidDocumentError, throw_invalid_document_error)
-
     def test_invalid_son(self):
         """Raise an error if loading invalid data"""
         class Occurrence(EmbeddedDocument):
@@ -1868,11 +1852,13 @@ class InstanceTest(unittest.TestCase):
             forms = ListField(StringField(), default=list)
             occurs = ListField(EmbeddedDocumentField(Occurrence), default=list)
 
-        def raise_invalid_document():
-            Word._from_son({'stem': [1, 2, 3], 'forms': 1, 'count': 'one',
-                            'occurs': {"hello": None}})
-
-        self.assertRaises(InvalidDocumentError, raise_invalid_document)
+        with self.assertRaises(InvalidDocumentError):
+            Word._from_son({
+                'stem': [1, 2, 3],
+                'forms': 1,
+                'count': 'one',
+                'occurs': {"hello": None}
+            })
 
     def test_reverse_delete_rule_cascade_and_nullify(self):
         """Ensure that a referenced document is also deleted upon deletion.
@@ -2103,8 +2089,7 @@ class InstanceTest(unittest.TestCase):
         self.assertEqual(Bar.objects.get().foo, None)
 
     def test_invalid_reverse_delete_rule_raise_errors(self):
-
-        def throw_invalid_document_error():
+        with self.assertRaises(InvalidDocumentError):
             class Blog(Document):
                 content = StringField()
                 authors = MapField(ReferenceField(
@@ -2114,21 +2099,15 @@ class InstanceTest(unittest.TestCase):
                         self.Person,
                         reverse_delete_rule=NULLIFY))
 
-        self.assertRaises(InvalidDocumentError, throw_invalid_document_error)
-
-        def throw_invalid_document_error_embedded():
+        with self.assertRaises(InvalidDocumentError):
             class Parents(EmbeddedDocument):
                 father = ReferenceField('Person', reverse_delete_rule=DENY)
                 mother = ReferenceField('Person', reverse_delete_rule=DENY)
 
-        self.assertRaises(
-            InvalidDocumentError, throw_invalid_document_error_embedded)
-
     def test_reverse_delete_rule_cascade_recurs(self):
         """Ensure that a chain of documents is also deleted upon cascaded
         deletion.
         """
-
         class BlogPost(Document):
             content = StringField()
             author = ReferenceField(self.Person, reverse_delete_rule=CASCADE)
@@ -2344,15 +2323,14 @@ class InstanceTest(unittest.TestCase):
         pickle_doc.save()
         pickle_doc.delete()
 
-    def test_throw_invalid_document_error(self):
-
-        # test handles people trying to upsert
-        def throw_invalid_document_error():
+    def test_override_method_with_field(self):
+        """Test creating a field with a field name that would override
+        the "validate" method.
+        """
+        with self.assertRaises(InvalidDocumentError):
             class Blog(Document):
                 validate = DictField()
 
-        self.assertRaises(InvalidDocumentError, throw_invalid_document_error)
-
     def test_mutating_documents(self):
 
         class B(EmbeddedDocument):
@@ -2815,11 +2793,10 @@ class InstanceTest(unittest.TestCase):
         log.log = "Saving"
         log.save()
 
-        def change_shard_key():
+        # try to change the shard key
+        with self.assertRaises(OperationError):
             log.machine = "127.0.0.1"
 
-        self.assertRaises(OperationError, change_shard_key)
-
     def test_shard_key_in_embedded_document(self):
         class Foo(EmbeddedDocument):
             foo = StringField()
@@ -2840,12 +2817,11 @@ class InstanceTest(unittest.TestCase):
         bar_doc.bar = 'baz'
         bar_doc.save()
 
-        def change_shard_key():
+        # try to change the shard key
+        with self.assertRaises(OperationError):
             bar_doc.foo.foo = 'something'
             bar_doc.save()
 
-        self.assertRaises(OperationError, change_shard_key)
-
     def test_shard_key_primary(self):
         class LogEntry(Document):
             machine = StringField(primary_key=True)
@@ -2866,11 +2842,10 @@ class InstanceTest(unittest.TestCase):
         log.log = "Saving"
         log.save()
 
-        def change_shard_key():
+        # try to change the shard key
+        with self.assertRaises(OperationError):
             log.machine = "127.0.0.1"
 
-        self.assertRaises(OperationError, change_shard_key)
-
     def test_kwargs_simple(self):
 
         class Embedded(EmbeddedDocument):
@@ -2955,11 +2930,9 @@ class InstanceTest(unittest.TestCase):
     def test_bad_mixed_creation(self):
         """Ensure that document gives correct error when duplicating arguments
         """
-        def construct_bad_instance():
+        with self.assertRaises(TypeError):
             return self.Person("Test User", 42, name="Bad User")
 
-        self.assertRaises(TypeError, construct_bad_instance)
-
     def test_data_contains_id_field(self):
         """Ensure that asking for _data returns 'id'
         """
@@ -3118,17 +3091,17 @@ class InstanceTest(unittest.TestCase):
         p4 = Person.objects()[0]
         p4.save()
         self.assertEquals(p4.height, 189)
-        
+
         # However the default will not be fixed in DB
         self.assertEquals(Person.objects(height=189).count(), 0)
-        
+
         # alter DB for the new default
         coll = Person._get_collection()
         for person in Person.objects.as_pymongo():
             if 'height' not in person:
                 person['height'] = 189
                 coll.save(person)
-                
+
         self.assertEquals(Person.objects(height=189).count(), 1)
 
     def test_from_son(self):
@@ -3202,5 +3175,20 @@ class InstanceTest(unittest.TestCase):
             self.assertEqual(b._instance, a)
         self.assertEqual(idx, 2)
 
+    def test_falsey_pk(self):
+        """Ensure that we can create and update a document with Falsey PK.
+        """
+        class Person(Document):
+            age = IntField(primary_key=True)
+            height = FloatField()
+
+        person = Person()
+        person.age = 0
+        person.height = 1.89
+        person.save()
+
+        person.update(set__height=2.0)
+
+
 if __name__ == '__main__':
     unittest.main()

tests/document/json_serialisation.py

@@ -1,6 +1,3 @@
-import sys
-sys.path[0:0] = [""]
-
 import unittest
 import uuid
 

tests/document/validation.py

@@ -1,7 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
-
 import unittest
 from datetime import datetime
 
@@ -60,7 +57,7 @@ class ValidatorErrorTest(unittest.TestCase):
 
         try:
             User().validate()
-        except ValidationError, e:
+        except ValidationError as e:
             self.assertTrue("User:None" in e.message)
             self.assertEqual(e.to_dict(), {
                 'username': 'Field is required',
@@ -70,7 +67,7 @@ class ValidatorErrorTest(unittest.TestCase):
         user.name = None
         try:
             user.save()
-        except ValidationError, e:
+        except ValidationError as e:
             self.assertTrue("User:RossC0" in e.message)
             self.assertEqual(e.to_dict(), {
                 'name': 'Field is required'})
@@ -118,7 +115,7 @@ class ValidatorErrorTest(unittest.TestCase):
 
         try:
             Doc(id="bad").validate()
-        except ValidationError, e:
+        except ValidationError as e:
             self.assertTrue("SubDoc:None" in e.message)
             self.assertEqual(e.to_dict(), {
                 "e": {'val': 'OK could not be converted to int'}})
@@ -136,7 +133,7 @@ class ValidatorErrorTest(unittest.TestCase):
         doc.e.val = "OK"
         try:
             doc.save()
-        except ValidationError, e:
+        except ValidationError as e:
             self.assertTrue("Doc:test" in e.message)
             self.assertEqual(e.to_dict(), {
                 "e": {'val': 'OK could not be converted to int'}})
@@ -156,14 +153,14 @@ class ValidatorErrorTest(unittest.TestCase):
 
         s = SubDoc()
 
-        self.assertRaises(ValidationError, lambda: s.validate())
+        self.assertRaises(ValidationError, s.validate)
 
         d1.e = s
         d2.e = s
 
         del d1
 
-        self.assertRaises(ValidationError, lambda: d2.validate())
+        self.assertRaises(ValidationError, d2.validate)
 
     def test_parent_reference_in_child_document(self):
         """

tests/fields/fields.py

@@ -1,11 +1,7 @@
 # -*- coding: utf-8 -*-
-import sys
-
 import six
 from nose.plugins.skip import SkipTest
 
-sys.path[0:0] = [""]
-
 import datetime
 import unittest
 import uuid
@@ -29,10 +25,9 @@ except ImportError:
 
 from mongoengine import *
 from mongoengine.connection import get_db
-from mongoengine.base import _document_registry
-from mongoengine.base.datastructures import BaseDict, EmbeddedDocumentList
+from mongoengine.base import (BaseDict, BaseField, EmbeddedDocumentList,
+                              _document_registry)
 from mongoengine.errors import NotRegistered, DoesNotExist
-from mongoengine.python_support import PY3, b, bin_type
 
 __all__ = ("FieldTest", "EmbeddedDocumentListFieldTestCase")
 
@@ -483,27 +478,41 @@ class FieldTest(unittest.TestCase):
 
     def test_decimal_storage(self):
         class Person(Document):
-            btc = DecimalField(precision=4)
+            float_value = DecimalField(precision=4)
+            string_value = DecimalField(precision=4, force_string=True)
 
         Person.drop_collection()
-        Person(btc=10).save()
-        Person(btc=10.1).save()
-        Person(btc=10.11).save()
-        Person(btc="10.111").save()
-        Person(btc=Decimal("10.1111")).save()
-        Person(btc=Decimal("10.11111")).save()
+        values_to_store = [10, 10.1, 10.11, "10.111", Decimal("10.1111"), Decimal("10.11111")]
+        for store_at_creation in [True, False]:
+            for value in values_to_store:
+                # to_python is called explicitly if values were sent in the kwargs of __init__
+                if store_at_creation:
+                    Person(float_value=value, string_value=value).save()
+                else:
+                    person = Person.objects.create()
+                    person.float_value = value
+                    person.string_value = value
+                    person.save()
 
         # How its stored
-        expected = [{'btc': 10.0}, {'btc': 10.1}, {'btc': 10.11},
-                    {'btc': 10.111}, {'btc': 10.1111}, {'btc': 10.1111}]
+        expected = [
+            {'float_value': 10.0, 'string_value': '10.0000'},
+            {'float_value': 10.1, 'string_value': '10.1000'},
+            {'float_value': 10.11, 'string_value': '10.1100'},
+            {'float_value': 10.111, 'string_value': '10.1110'},
+            {'float_value': 10.1111, 'string_value': '10.1111'},
+            {'float_value': 10.1111, 'string_value': '10.1111'}]
+        expected.extend(expected)
         actual = list(Person.objects.exclude('id').as_pymongo())
         self.assertEqual(expected, actual)
 
         # How it comes out locally
         expected = [Decimal('10.0000'), Decimal('10.1000'), Decimal('10.1100'),
                     Decimal('10.1110'), Decimal('10.1111'), Decimal('10.1111')]
-        actual = list(Person.objects().scalar('btc'))
-        self.assertEqual(expected, actual)
+        expected.extend(expected)
+        for field_name in ['float_value', 'string_value']:
+            actual = list(Person.objects().scalar(field_name))
+            self.assertEqual(expected, actual)
 
     def test_boolean_validation(self):
         """Ensure that invalid values cannot be assigned to boolean fields.
@@ -639,8 +648,8 @@ class FieldTest(unittest.TestCase):
 
         # Post UTC - microseconds are rounded (down) nearest millisecond and
         # dropped
-        d1 = datetime.datetime(1970, 01, 01, 00, 00, 01, 999)
-        d2 = datetime.datetime(1970, 01, 01, 00, 00, 01)
+        d1 = datetime.datetime(1970, 1, 1, 0, 0, 1, 999)
+        d2 = datetime.datetime(1970, 1, 1, 0, 0, 1)
         log = LogEntry()
         log.date = d1
         log.save()
@@ -649,15 +658,15 @@ class FieldTest(unittest.TestCase):
         self.assertEqual(log.date, d2)
 
         # Post UTC - microseconds are rounded (down) nearest millisecond
-        d1 = datetime.datetime(1970, 01, 01, 00, 00, 01, 9999)
-        d2 = datetime.datetime(1970, 01, 01, 00, 00, 01, 9000)
+        d1 = datetime.datetime(1970, 1, 1, 0, 0, 1, 9999)
+        d2 = datetime.datetime(1970, 1, 1, 0, 0, 1, 9000)
         log.date = d1
         log.save()
         log.reload()
         self.assertNotEqual(log.date, d1)
         self.assertEqual(log.date, d2)
 
-        if not PY3:
+        if not six.PY3:
             # Pre UTC dates microseconds below 1000 are dropped
             # This does not seem to be true in PY3
             d1 = datetime.datetime(1969, 12, 31, 23, 59, 59, 999)
@@ -677,7 +686,7 @@ class FieldTest(unittest.TestCase):
 
         LogEntry.drop_collection()
 
-        d1 = datetime.datetime(1970, 01, 01, 00, 00, 01)
+        d1 = datetime.datetime(1970, 1, 1, 0, 0, 1)
         log = LogEntry()
         log.date = d1
         log.validate()
@@ -694,8 +703,8 @@ class FieldTest(unittest.TestCase):
         LogEntry.drop_collection()
 
         # create 60 log entries
-        for i in xrange(1950, 2010):
-            d = datetime.datetime(i, 01, 01, 00, 00, 01)
+        for i in range(1950, 2010):
+            d = datetime.datetime(i, 1, 1, 0, 0, 1)
             LogEntry(date=d).save()
 
         self.assertEqual(LogEntry.objects.count(), 60)
@@ -742,7 +751,7 @@ class FieldTest(unittest.TestCase):
 
         # Post UTC - microseconds are rounded (down) nearest millisecond and
         # dropped - with default datetimefields
-        d1 = datetime.datetime(1970, 01, 01, 00, 00, 01, 999)
+        d1 = datetime.datetime(1970, 1, 1, 0, 0, 1, 999)
         log = LogEntry()
         log.date = d1
         log.save()
@@ -751,7 +760,7 @@ class FieldTest(unittest.TestCase):
 
         # Post UTC - microseconds are rounded (down) nearest millisecond - with
         # default datetimefields
-        d1 = datetime.datetime(1970, 01, 01, 00, 00, 01, 9999)
+        d1 = datetime.datetime(1970, 1, 1, 0, 0, 1, 9999)
         log.date = d1
         log.save()
         log.reload()
@@ -768,7 +777,7 @@ class FieldTest(unittest.TestCase):
         # Pre UTC microseconds above 1000 is wonky - with default datetimefields
         # log.date has an invalid microsecond value so I can't construct
         # a date to compare.
-        for i in xrange(1001, 3113, 33):
+        for i in range(1001, 3113, 33):
             d1 = datetime.datetime(1969, 12, 31, 23, 59, 59, i)
             log.date = d1
             log.save()
@@ -778,7 +787,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(log, log1)
 
         # Test string padding
-        microsecond = map(int, [math.pow(10, x) for x in xrange(6)])
+        microsecond = map(int, [math.pow(10, x) for x in range(6)])
         mm = dd = hh = ii = ss = [1, 10]
 
         for values in itertools.product([2014], mm, dd, hh, ii, ss, microsecond):
@@ -800,7 +809,7 @@ class FieldTest(unittest.TestCase):
 
         LogEntry.drop_collection()
 
-        d1 = datetime.datetime(1970, 01, 01, 00, 00, 01, 999)
+        d1 = datetime.datetime(1970, 1, 1, 0, 0, 1, 999)
         log = LogEntry()
         log.date = d1
         log.save()
@@ -811,8 +820,8 @@ class FieldTest(unittest.TestCase):
         LogEntry.drop_collection()
 
         # create 60 log entries
-        for i in xrange(1950, 2010):
-            d = datetime.datetime(i, 01, 01, 00, 00, 01, 999)
+        for i in range(1950, 2010):
+            d = datetime.datetime(i, 1, 1, 0, 0, 1, 999)
             LogEntry(date=d).save()
 
         self.assertEqual(LogEntry.objects.count(), 60)
@@ -1047,7 +1056,7 @@ class FieldTest(unittest.TestCase):
         BlogPost.drop_collection()
 
     def test_list_assignment(self):
-        """Ensure that list field element assignment and slicing work 
+        """Ensure that list field element assignment and slicing work
         """
         class BlogPost(Document):
             info = ListField()
@@ -1057,12 +1066,12 @@ class FieldTest(unittest.TestCase):
         post = BlogPost()
         post.info = ['e1', 'e2', 3, '4', 5]
         post.save()
-        
+
         post.info[0] = 1
         post.save()
         post.reload()
         self.assertEqual(post.info[0], 1)
-        
+
         post.info[1:3] = ['n2', 'n3']
         post.save()
         post.reload()
@@ -1120,12 +1129,11 @@ class FieldTest(unittest.TestCase):
         e.mapping = [1]
         e.save()
 
-        def create_invalid_mapping():
+        # try creating an invalid mapping
+        with self.assertRaises(ValidationError):
             e.mapping = ["abc"]
             e.save()
 
-        self.assertRaises(ValidationError, create_invalid_mapping)
-
         Simple.drop_collection()
 
     def test_list_field_rejects_strings(self):
@@ -1209,7 +1217,7 @@ class FieldTest(unittest.TestCase):
         self.assertEqual(simple.widgets, [4])
 
     def test_list_field_with_negative_indices(self):
-        
+
         class Simple(Document):
             widgets = ListField()
 
@@ -1392,12 +1400,11 @@ class FieldTest(unittest.TestCase):
         e.mapping['someint'] = 1
         e.save()
 
-        def create_invalid_mapping():
+        # try creating an invalid mapping
+        with self.assertRaises(ValidationError):
             e.mapping['somestring'] = "abc"
             e.save()
 
-        self.assertRaises(ValidationError, create_invalid_mapping)
-
         Simple.drop_collection()
 
     def test_dictfield_complex(self):
@@ -1470,11 +1477,10 @@ class FieldTest(unittest.TestCase):
         self.assertEqual(BaseDict, type(e.mapping))
         self.assertEqual({"ints": [3, 4]}, e.mapping)
 
-        def create_invalid_mapping():
+        # try creating an invalid mapping
+        with self.assertRaises(ValueError):
             e.update(set__mapping={"somestrings": ["foo", "bar", ]})
 
-        self.assertRaises(ValueError, create_invalid_mapping)
-
         Simple.drop_collection()
 
     def test_mapfield(self):
@@ -1489,18 +1495,14 @@ class FieldTest(unittest.TestCase):
         e.mapping['someint'] = 1
         e.save()
 
-        def create_invalid_mapping():
+        with self.assertRaises(ValidationError):
             e.mapping['somestring'] = "abc"
             e.save()
 
-        self.assertRaises(ValidationError, create_invalid_mapping)
-
-        def create_invalid_class():
+        with self.assertRaises(ValidationError):
             class NoDeclaredType(Document):
                 mapping = MapField()
 
-        self.assertRaises(ValidationError, create_invalid_class)
-
         Simple.drop_collection()
 
     def test_complex_mapfield(self):
@@ -1529,14 +1531,10 @@ class FieldTest(unittest.TestCase):
         self.assertTrue(isinstance(e2.mapping['somestring'], StringSetting))
         self.assertTrue(isinstance(e2.mapping['someint'], IntegerSetting))
 
-        def create_invalid_mapping():
+        with self.assertRaises(ValidationError):
             e.mapping['someint'] = 123
             e.save()
 
-        self.assertRaises(ValidationError, create_invalid_mapping)
-
-        Extensible.drop_collection()
-
     def test_embedded_mapfield_db_field(self):
 
         class Embedded(EmbeddedDocument):
@@ -1746,8 +1744,8 @@ class FieldTest(unittest.TestCase):
         # Reference is no longer valid
         foo.delete()
         bar = Bar.objects.get()
-        self.assertRaises(DoesNotExist, lambda: getattr(bar, 'ref'))
-        self.assertRaises(DoesNotExist, lambda: getattr(bar, 'generic_ref'))
+        self.assertRaises(DoesNotExist, getattr, bar, 'ref')
+        self.assertRaises(DoesNotExist, getattr, bar, 'generic_ref')
 
         # When auto_dereference is disabled, there is no trouble returning DBRef
         bar = Bar.objects.get()
@@ -1823,7 +1821,7 @@ class FieldTest(unittest.TestCase):
                                'parent': "50a234ea469ac1eda42d347d"})
         mongoed = p1.to_mongo()
         self.assertTrue(isinstance(mongoed['parent'], ObjectId))
-        
+
     def test_cached_reference_field_get_and_save(self):
         """
         Tests #1047: CachedReferenceField creates DBRefs on to_python, but can't save them on to_mongo
@@ -1835,11 +1833,11 @@ class FieldTest(unittest.TestCase):
         class Ocorrence(Document):
             person = StringField()
             animal = CachedReferenceField(Animal)
-        
+
         Animal.drop_collection()
         Ocorrence.drop_collection()
-        
-        Ocorrence(person="testte", 
+
+        Ocorrence(person="testte",
                   animal=Animal(name="Leopard", tag="heavy").save()).save()
         p = Ocorrence.objects.get()
         p.person = 'new_testte'
@@ -2022,7 +2020,7 @@ class FieldTest(unittest.TestCase):
         })
 
     def test_cached_reference_fields_on_embedded_documents(self):
-        def build():
+        with self.assertRaises(InvalidDocumentError):
             class Test(Document):
                 name = StringField()
 
@@ -2031,8 +2029,6 @@ class FieldTest(unittest.TestCase):
                     'test': CachedReferenceField(Test)
             })
 
-        self.assertRaises(InvalidDocumentError, build)
-
     def test_cached_reference_auto_sync(self):
         class Person(Document):
             TYPES = (
@@ -2810,6 +2806,38 @@ class FieldTest(unittest.TestCase):
         Post.drop_collection()
         User.drop_collection()
 
+    def test_generic_reference_filter_by_dbref(self):
+        """Ensure we can search for a specific generic reference by
+        providing its ObjectId.
+        """
+        class Doc(Document):
+            ref = GenericReferenceField()
+
+        Doc.drop_collection()
+
+        doc1 = Doc.objects.create()
+        doc2 = Doc.objects.create(ref=doc1)
+
+        doc = Doc.objects.get(ref=DBRef('doc', doc1.pk))
+        self.assertEqual(doc, doc2)
+
+    def test_generic_reference_filter_by_objectid(self):
+        """Ensure we can search for a specific generic reference by
+        providing its DBRef.
+        """
+        class Doc(Document):
+            ref = GenericReferenceField()
+
+        Doc.drop_collection()
+
+        doc1 = Doc.objects.create()
+        doc2 = Doc.objects.create(ref=doc1)
+
+        self.assertTrue(isinstance(doc1.pk, ObjectId))
+
+        doc = Doc.objects.get(ref=doc1.pk)
+        self.assertEqual(doc, doc2)
+
     def test_binary_fields(self):
         """Ensure that binary fields can be stored and retrieved.
         """
@@ -2817,7 +2845,7 @@ class FieldTest(unittest.TestCase):
             content_type = StringField()
             blob = BinaryField()
 
-        BLOB = b('\xe6\x00\xc4\xff\x07')
+        BLOB = six.b('\xe6\x00\xc4\xff\x07')
         MIME_TYPE = 'application/octet-stream'
 
         Attachment.drop_collection()
@@ -2827,7 +2855,7 @@ class FieldTest(unittest.TestCase):
 
         attachment_1 = Attachment.objects().first()
         self.assertEqual(MIME_TYPE, attachment_1.content_type)
-        self.assertEqual(BLOB, bin_type(attachment_1.blob))
+        self.assertEqual(BLOB, six.binary_type(attachment_1.blob))
 
         Attachment.drop_collection()
 
@@ -2854,13 +2882,13 @@ class FieldTest(unittest.TestCase):
 
         attachment_required = AttachmentRequired()
         self.assertRaises(ValidationError, attachment_required.validate)
-        attachment_required.blob = Binary(b('\xe6\x00\xc4\xff\x07'))
+        attachment_required.blob = Binary(six.b('\xe6\x00\xc4\xff\x07'))
         attachment_required.validate()
 
         attachment_size_limit = AttachmentSizeLimit(
-            blob=b('\xe6\x00\xc4\xff\x07'))
+            blob=six.b('\xe6\x00\xc4\xff\x07'))
         self.assertRaises(ValidationError, attachment_size_limit.validate)
-        attachment_size_limit.blob = b('\xe6\x00\xc4\xff')
+        attachment_size_limit.blob = six.b('\xe6\x00\xc4\xff')
         attachment_size_limit.validate()
 
         Attachment.drop_collection()
@@ -3001,28 +3029,32 @@ class FieldTest(unittest.TestCase):
                 ('S', 'Small'), ('M', 'Medium'), ('L', 'Large'),
                 ('XL', 'Extra Large'), ('XXL', 'Extra Extra Large')))
             style = StringField(max_length=3, choices=(
-                ('S', 'Small'), ('B', 'Baggy'), ('W', 'wide')), default='S')
+                ('S', 'Small'), ('B', 'Baggy'), ('W', 'Wide')), default='W')
 
         Shirt.drop_collection()
 
-        shirt = Shirt()
+        shirt1 = Shirt()
+        shirt2 = Shirt()
 
-        self.assertEqual(shirt.get_size_display(), None)
-        self.assertEqual(shirt.get_style_display(), 'Small')
+        # Make sure get_<field>_display returns the default value (or None)
+        self.assertEqual(shirt1.get_size_display(), None)
+        self.assertEqual(shirt1.get_style_display(), 'Wide')
 
-        shirt.size = "XXL"
-        shirt.style = "B"
-        self.assertEqual(shirt.get_size_display(), 'Extra Extra Large')
-        self.assertEqual(shirt.get_style_display(), 'Baggy')
+        shirt1.size = 'XXL'
+        shirt1.style = 'B'
+        shirt2.size = 'M'
+        shirt2.style = 'S'
+        self.assertEqual(shirt1.get_size_display(), 'Extra Extra Large')
+        self.assertEqual(shirt1.get_style_display(), 'Baggy')
+        self.assertEqual(shirt2.get_size_display(), 'Medium')
+        self.assertEqual(shirt2.get_style_display(), 'Small')
 
         # Set as Z - an invalid choice
-        shirt.size = "Z"
-        shirt.style = "Z"
-        self.assertEqual(shirt.get_size_display(), 'Z')
-        self.assertEqual(shirt.get_style_display(), 'Z')
-        self.assertRaises(ValidationError, shirt.validate)
-
-        Shirt.drop_collection()
+        shirt1.size = 'Z'
+        shirt1.style = 'Z'
+        self.assertEqual(shirt1.get_size_display(), 'Z')
+        self.assertEqual(shirt1.get_style_display(), 'Z')
+        self.assertRaises(ValidationError, shirt1.validate)
 
     def test_simple_choices_validation(self):
         """Ensure that value is in a container of allowed values.
@@ -3102,7 +3134,7 @@ class FieldTest(unittest.TestCase):
 
         try:
             shirt.validate()
-        except ValidationError, error:
+        except ValidationError as error:
             # get the validation rules
             error_dict = error.to_dict()
             self.assertEqual(error_dict['size'], SIZE_MESSAGE)
@@ -3131,7 +3163,7 @@ class FieldTest(unittest.TestCase):
         self.db['mongoengine.counters'].drop()
         Person.drop_collection()
 
-        for x in xrange(10):
+        for x in range(10):
             Person(name="Person %s" % x).save()
 
         c = self.db['mongoengine.counters'].find_one({'_id': 'person.id'})
@@ -3155,7 +3187,7 @@ class FieldTest(unittest.TestCase):
         self.db['mongoengine.counters'].drop()
         Person.drop_collection()
 
-        for x in xrange(10):
+        for x in range(10):
             Person(name="Person %s" % x).save()
 
         self.assertEqual(Person.id.get_next_value(), 11)
@@ -3170,7 +3202,7 @@ class FieldTest(unittest.TestCase):
         self.db['mongoengine.counters'].drop()
         Person.drop_collection()
 
-        for x in xrange(10):
+        for x in range(10):
             Person(name="Person %s" % x).save()
 
         self.assertEqual(Person.id.get_next_value(), '11')
@@ -3186,7 +3218,7 @@ class FieldTest(unittest.TestCase):
         self.db['mongoengine.counters'].drop()
         Person.drop_collection()
 
-        for x in xrange(10):
+        for x in range(10):
             Person(name="Person %s" % x).save()
 
         c = self.db['mongoengine.counters'].find_one({'_id': 'jelly.id'})
@@ -3211,7 +3243,7 @@ class FieldTest(unittest.TestCase):
         self.db['mongoengine.counters'].drop()
         Person.drop_collection()
 
-        for x in xrange(10):
+        for x in range(10):
             Person(name="Person %s" % x).save()
 
         c = self.db['mongoengine.counters'].find_one({'_id': 'person.id'})
@@ -3273,7 +3305,7 @@ class FieldTest(unittest.TestCase):
         Animal.drop_collection()
         Person.drop_collection()
 
-        for x in xrange(10):
+        for x in range(10):
             Animal(name="Animal %s" % x).save()
             Person(name="Person %s" % x).save()
 
@@ -3303,7 +3335,7 @@ class FieldTest(unittest.TestCase):
         self.db['mongoengine.counters'].drop()
         Person.drop_collection()
 
-        for x in xrange(10):
+        for x in range(10):
             p = Person(name="Person %s" % x)
             p.save()
 
@@ -3490,7 +3522,7 @@ class FieldTest(unittest.TestCase):
         self.assertRaises(ValidationError, post.validate)
         try:
             post.validate()
-        except ValidationError, error:
+        except ValidationError as error:
             # ValidationError.errors property
             self.assertTrue(hasattr(error, 'errors'))
             self.assertTrue(isinstance(error.errors, dict))
@@ -3551,8 +3583,6 @@ class FieldTest(unittest.TestCase):
         Ensure that tuples remain tuples when they are
         inside a ComplexBaseField
         """
-        from mongoengine.base import BaseField
-
         class EnumField(BaseField):
 
             def __init__(self, **kwargs):
@@ -3786,9 +3816,7 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         filtered = self.post1.comments.filter()
 
         # Ensure nothing was changed
-        # < 2.6 Incompatible >
-        # self.assertListEqual(filtered, self.post1.comments)
-        self.assertEqual(filtered, self.post1.comments)
+        self.assertListEqual(filtered, self.post1.comments)
 
     def test_single_keyword_filter(self):
         """
@@ -3839,10 +3867,8 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         Tests the filter method of a List of Embedded Documents
         when the keyword is not a known keyword.
         """
-        # < 2.6 Incompatible >
-        # with self.assertRaises(AttributeError):
-        #    self.post2.comments.filter(year=2)
-        self.assertRaises(AttributeError, self.post2.comments.filter, year=2)
+        with self.assertRaises(AttributeError):
+            self.post2.comments.filter(year=2)
 
     def test_no_keyword_exclude(self):
         """
@@ -3852,9 +3878,7 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         filtered = self.post1.comments.exclude()
 
         # Ensure everything was removed
-        # < 2.6 Incompatible >
-        # self.assertListEqual(filtered, [])
-        self.assertEqual(filtered, [])
+        self.assertListEqual(filtered, [])
 
     def test_single_keyword_exclude(self):
         """
@@ -3900,10 +3924,8 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         Tests the exclude method of a List of Embedded Documents
         when the keyword is not a known keyword.
         """
-        # < 2.6 Incompatible >
-        # with self.assertRaises(AttributeError):
-        #    self.post2.comments.exclude(year=2)
-        self.assertRaises(AttributeError, self.post2.comments.exclude, year=2)
+        with self.assertRaises(AttributeError):
+            self.post2.comments.exclude(year=2)
 
     def test_chained_filter_exclude(self):
         """
@@ -3941,10 +3963,7 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         single keyword.
         """
         comment = self.post1.comments.get(author='user1')
-
-        # < 2.6 Incompatible >
-        # self.assertIsInstance(comment, self.Comments)
-        self.assertTrue(isinstance(comment, self.Comments))
+        self.assertIsInstance(comment, self.Comments)
         self.assertEqual(comment.author, 'user1')
 
     def test_multi_keyword_get(self):
@@ -3953,10 +3972,7 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         multiple keywords.
         """
         comment = self.post2.comments.get(author='user2', message='message2')
-
-        # < 2.6 Incompatible >
-        # self.assertIsInstance(comment, self.Comments)
-        self.assertTrue(isinstance(comment, self.Comments))
+        self.assertIsInstance(comment, self.Comments)
         self.assertEqual(comment.author, 'user2')
         self.assertEqual(comment.message, 'message2')
 
@@ -3965,44 +3981,32 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         Tests the get method of a List of Embedded Documents without
         a keyword to return multiple documents.
         """
-        # < 2.6 Incompatible >
-        # with self.assertRaises(MultipleObjectsReturned):
-        #    self.post1.comments.get()
-        self.assertRaises(MultipleObjectsReturned, self.post1.comments.get)
+        with self.assertRaises(MultipleObjectsReturned):
+            self.post1.comments.get()
 
     def test_keyword_multiple_return_get(self):
         """
         Tests the get method of a List of Embedded Documents with a keyword
         to return multiple documents.
         """
-        # < 2.6 Incompatible >
-        # with self.assertRaises(MultipleObjectsReturned):
-        #    self.post2.comments.get(author='user2')
-        self.assertRaises(
-            MultipleObjectsReturned, self.post2.comments.get, author='user2'
-        )
+        with self.assertRaises(MultipleObjectsReturned):
+            self.post2.comments.get(author='user2')
 
     def test_unknown_keyword_get(self):
         """
         Tests the get method of a List of Embedded Documents with an
         unknown keyword.
         """
-        # < 2.6 Incompatible >
-        # with self.assertRaises(AttributeError):
-        #    self.post2.comments.get(year=2020)
-        self.assertRaises(AttributeError, self.post2.comments.get, year=2020)
+        with self.assertRaises(AttributeError):
+            self.post2.comments.get(year=2020)
 
     def test_no_result_get(self):
         """
         Tests the get method of a List of Embedded Documents where get
         returns no results.
         """
-        # < 2.6 Incompatible >
-        # with self.assertRaises(DoesNotExist):
-        #    self.post1.comments.get(author='user3')
-        self.assertRaises(
-            DoesNotExist, self.post1.comments.get, author='user3'
-        )
+        with self.assertRaises(DoesNotExist):
+            self.post1.comments.get(author='user3')
 
     def test_first(self):
         """
@@ -4012,9 +4016,7 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         comment = self.post1.comments.first()
 
         # Ensure a Comment object was returned.
-        # < 2.6 Incompatible >
-        # self.assertIsInstance(comment, self.Comments)
-        self.assertTrue(isinstance(comment, self.Comments))
+        self.assertIsInstance(comment, self.Comments)
         self.assertEqual(comment, self.post1.comments[0])
 
     def test_create(self):
@@ -4027,22 +4029,14 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         self.post1.save()
 
         # Ensure the returned value is the comment object.
-        # < 2.6 Incompatible >
-        # self.assertIsInstance(comment, self.Comments)
-        self.assertTrue(isinstance(comment, self.Comments))
+        self.assertIsInstance(comment, self.Comments)
         self.assertEqual(comment.author, 'user4')
         self.assertEqual(comment.message, 'message1')
 
         # Ensure the new comment was actually saved to the database.
-        # < 2.6 Incompatible >
-        # self.assertIn(
-        #    comment,
-        #    self.BlogPost.objects(comments__author='user4')[0].comments
-        # )
-        self.assertTrue(
-            comment in self.BlogPost.objects(
-                comments__author='user4'
-            )[0].comments
+        self.assertIn(
+            comment,
+            self.BlogPost.objects(comments__author='user4')[0].comments
         )
 
     def test_filtered_create(self):
@@ -4057,22 +4051,14 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         self.post1.save()
 
         # Ensure the returned value is the comment object.
-        # < 2.6 Incompatible >
-        # self.assertIsInstance(comment, self.Comments)
-        self.assertTrue(isinstance(comment, self.Comments))
+        self.assertIsInstance(comment, self.Comments)
         self.assertEqual(comment.author, 'user4')
         self.assertEqual(comment.message, 'message1')
 
         # Ensure the new comment was actually saved to the database.
-        # < 2.6 Incompatible >
-        # self.assertIn(
-        #    comment,
-        #    self.BlogPost.objects(comments__author='user4')[0].comments
-        # )
-        self.assertTrue(
-            comment in self.BlogPost.objects(
-                comments__author='user4'
-            )[0].comments
+        self.assertIn(
+            comment,
+            self.BlogPost.objects(comments__author='user4')[0].comments
         )
 
     def test_no_keyword_update(self):
@@ -4085,22 +4071,14 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         self.post1.save()
 
         # Ensure that nothing was altered.
-        # < 2.6 Incompatible >
-        # self.assertIn(
-        #    original[0],
-        #    self.BlogPost.objects(id=self.post1.id)[0].comments
-        # )
-        self.assertTrue(
-            original[0] in self.BlogPost.objects(id=self.post1.id)[0].comments
+        self.assertIn(
+            original[0],
+            self.BlogPost.objects(id=self.post1.id)[0].comments
         )
 
-        # < 2.6 Incompatible >
-        # self.assertIn(
-        #    original[1],
-        #    self.BlogPost.objects(id=self.post1.id)[0].comments
-        # )
-        self.assertTrue(
-            original[1] in self.BlogPost.objects(id=self.post1.id)[0].comments
+        self.assertIn(
+            original[1],
+            self.BlogPost.objects(id=self.post1.id)[0].comments
         )
 
         # Ensure the method returned 0 as the number of entries
@@ -4146,13 +4124,9 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         comments.save()
 
         # Ensure that the new comment has been added to the database.
-        # < 2.6 Incompatible >
-        # self.assertIn(
-        #    new_comment,
-        #    self.BlogPost.objects(id=self.post1.id)[0].comments
-        # )
-        self.assertTrue(
-            new_comment in self.BlogPost.objects(id=self.post1.id)[0].comments
+        self.assertIn(
+            new_comment,
+            self.BlogPost.objects(id=self.post1.id)[0].comments
         )
 
     def test_delete(self):
@@ -4164,23 +4138,15 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
 
         # Ensure that all the comments under post1 were deleted in the
         # database.
-        # < 2.6 Incompatible >
-        # self.assertListEqual(
-        #    self.BlogPost.objects(id=self.post1.id)[0].comments, []
-        # )
-        self.assertEqual(
+        self.assertListEqual(
             self.BlogPost.objects(id=self.post1.id)[0].comments, []
         )
 
         # Ensure that post1 comments were deleted from the list.
-        # < 2.6 Incompatible >
-        # self.assertListEqual(self.post1.comments, [])
-        self.assertEqual(self.post1.comments, [])
+        self.assertListEqual(self.post1.comments, [])
 
         # Ensure that comments still returned a EmbeddedDocumentList object.
-        # < 2.6 Incompatible >
-        # self.assertIsInstance(self.post1.comments, EmbeddedDocumentList)
-        self.assertTrue(isinstance(self.post1.comments, EmbeddedDocumentList))
+        self.assertIsInstance(self.post1.comments, EmbeddedDocumentList)
 
         # Ensure that the delete method returned 2 as the number of entries
         # deleted from the database
@@ -4220,21 +4186,15 @@ class EmbeddedDocumentListFieldTestCase(unittest.TestCase):
         self.post1.save()
 
         # Ensure that only the user2 comment was deleted.
-        # < 2.6 Incompatible >
-        # self.assertNotIn(
-        #     comment, self.BlogPost.objects(id=self.post1.id)[0].comments
-        # )
-        self.assertTrue(
-            comment not in self.BlogPost.objects(id=self.post1.id)[0].comments
+        self.assertNotIn(
+            comment, self.BlogPost.objects(id=self.post1.id)[0].comments
         )
         self.assertEqual(
             len(self.BlogPost.objects(id=self.post1.id)[0].comments), 1
         )
 
         # Ensure that the user2 comment no longer exists in the list.
-        # < 2.6 Incompatible >
-        # self.assertNotIn(comment, self.post1.comments)
-        self.assertTrue(comment not in self.post1.comments)
+        self.assertNotIn(comment, self.post1.comments)
         self.assertEqual(len(self.post1.comments), 1)
 
         # Ensure that the delete method returned 1 as the number of entries

tests/fields/file_tests.py

@@ -1,18 +1,16 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
-
 import copy
 import os
 import unittest
 import tempfile
 
 import gridfs
+import six
 
 from nose.plugins.skip import SkipTest
 from mongoengine import *
 from mongoengine.connection import get_db
-from mongoengine.python_support import b, StringIO
+from mongoengine.python_support import StringIO
 
 try:
     from PIL import Image
@@ -49,7 +47,7 @@ class FileTest(unittest.TestCase):
 
         PutFile.drop_collection()
 
-        text = b('Hello, World!')
+        text = six.b('Hello, World!')
         content_type = 'text/plain'
 
         putfile = PutFile()
@@ -88,8 +86,8 @@ class FileTest(unittest.TestCase):
 
         StreamFile.drop_collection()
 
-        text = b('Hello, World!')
-        more_text = b('Foo Bar')
+        text = six.b('Hello, World!')
+        more_text = six.b('Foo Bar')
         content_type = 'text/plain'
 
         streamfile = StreamFile()
@@ -123,8 +121,8 @@ class FileTest(unittest.TestCase):
 
         StreamFile.drop_collection()
 
-        text = b('Hello, World!')
-        more_text = b('Foo Bar')
+        text = six.b('Hello, World!')
+        more_text = six.b('Foo Bar')
         content_type = 'text/plain'
 
         streamfile = StreamFile()
@@ -155,8 +153,8 @@ class FileTest(unittest.TestCase):
         class SetFile(Document):
             the_file = FileField()
 
-        text = b('Hello, World!')
-        more_text = b('Foo Bar')
+        text = six.b('Hello, World!')
+        more_text = six.b('Foo Bar')
 
         SetFile.drop_collection()
 
@@ -185,7 +183,7 @@ class FileTest(unittest.TestCase):
         GridDocument.drop_collection()
 
         with tempfile.TemporaryFile() as f:
-            f.write(b("Hello World!"))
+            f.write(six.b("Hello World!"))
             f.flush()
 
             # Test without default
@@ -202,7 +200,7 @@ class FileTest(unittest.TestCase):
             self.assertEqual(doc_b.the_file.grid_id, doc_c.the_file.grid_id)
 
             # Test with default
-            doc_d = GridDocument(the_file=b(''))
+            doc_d = GridDocument(the_file=six.b(''))
             doc_d.save()
 
             doc_e = GridDocument.objects.with_id(doc_d.id)
@@ -228,7 +226,7 @@ class FileTest(unittest.TestCase):
         # First instance
         test_file = TestFile()
         test_file.name = "Hello, World!"
-        test_file.the_file.put(b('Hello, World!'))
+        test_file.the_file.put(six.b('Hello, World!'))
         test_file.save()
 
         # Second instance
@@ -282,7 +280,7 @@ class FileTest(unittest.TestCase):
 
         test_file = TestFile()
         self.assertFalse(bool(test_file.the_file))
-        test_file.the_file.put(b('Hello, World!'), content_type='text/plain')
+        test_file.the_file.put(six.b('Hello, World!'), content_type='text/plain')
         test_file.save()
         self.assertTrue(bool(test_file.the_file))
 
@@ -297,66 +295,66 @@ class FileTest(unittest.TestCase):
         test_file = TestFile()
         self.assertFalse(test_file.the_file in [{"test": 1}])
 
-    def test_file_disk_space(self): 
-        """ Test disk space usage when we delete/replace a file """ 
+    def test_file_disk_space(self):
+        """ Test disk space usage when we delete/replace a file """
         class TestFile(Document):
             the_file = FileField()
-            
-        text = b('Hello, World!')
+
+        text = six.b('Hello, World!')
         content_type = 'text/plain'
 
         testfile = TestFile()
         testfile.the_file.put(text, content_type=content_type, filename="hello")
         testfile.save()
-        
-        # Now check fs.files and fs.chunks 
+
+        # Now check fs.files and fs.chunks
         db = TestFile._get_db()
-        
+
         files = db.fs.files.find()
         chunks = db.fs.chunks.find()
         self.assertEquals(len(list(files)), 1)
         self.assertEquals(len(list(chunks)), 1)
 
-        # Deleting the docoument should delete the files 
+        # Deleting the docoument should delete the files
         testfile.delete()
-        
+
         files = db.fs.files.find()
         chunks = db.fs.chunks.find()
         self.assertEquals(len(list(files)), 0)
         self.assertEquals(len(list(chunks)), 0)
-        
-        # Test case where we don't store a file in the first place 
+
+        # Test case where we don't store a file in the first place
         testfile = TestFile()
         testfile.save()
-        
+
         files = db.fs.files.find()
         chunks = db.fs.chunks.find()
         self.assertEquals(len(list(files)), 0)
         self.assertEquals(len(list(chunks)), 0)
-        
+
         testfile.delete()
-        
+
         files = db.fs.files.find()
         chunks = db.fs.chunks.find()
         self.assertEquals(len(list(files)), 0)
         self.assertEquals(len(list(chunks)), 0)
-        
-        # Test case where we overwrite the file 
+
+        # Test case where we overwrite the file
         testfile = TestFile()
         testfile.the_file.put(text, content_type=content_type, filename="hello")
         testfile.save()
-        
-        text = b('Bonjour, World!')
+
+        text = six.b('Bonjour, World!')
         testfile.the_file.replace(text, content_type=content_type, filename="hello")
         testfile.save()
-        
+
         files = db.fs.files.find()
         chunks = db.fs.chunks.find()
         self.assertEquals(len(list(files)), 1)
         self.assertEquals(len(list(chunks)), 1)
-        
+
         testfile.delete()
-        
+
         files = db.fs.files.find()
         chunks = db.fs.chunks.find()
         self.assertEquals(len(list(files)), 0)
@@ -372,14 +370,14 @@ class FileTest(unittest.TestCase):
         TestImage.drop_collection()
 
         with tempfile.TemporaryFile() as f:
-            f.write(b("Hello World!"))
+            f.write(six.b("Hello World!"))
             f.flush()
 
             t = TestImage()
             try:
                 t.image.put(f)
                 self.fail("Should have raised an invalidation error")
-            except ValidationError, e:
+            except ValidationError as e:
                 self.assertEqual("%s" % e, "Invalid image: cannot identify image file %s" % f)
 
         t = TestImage()
@@ -496,7 +494,7 @@ class FileTest(unittest.TestCase):
         # First instance
         test_file = TestFile()
         test_file.name = "Hello, World!"
-        test_file.the_file.put(b('Hello, World!'),
+        test_file.the_file.put(six.b('Hello, World!'),
                           name="hello.txt")
         test_file.save()
 
@@ -504,16 +502,15 @@ class FileTest(unittest.TestCase):
         self.assertEqual(data.get('name'), 'hello.txt')
 
         test_file = TestFile.objects.first()
-        self.assertEqual(test_file.the_file.read(),
-                          b('Hello, World!'))
+        self.assertEqual(test_file.the_file.read(), six.b('Hello, World!'))
 
         test_file = TestFile.objects.first()
-        test_file.the_file = b('HELLO, WORLD!')
+        test_file.the_file = six.b('HELLO, WORLD!')
         test_file.save()
 
         test_file = TestFile.objects.first()
         self.assertEqual(test_file.the_file.read(),
-                          b('HELLO, WORLD!'))
+                         six.b('HELLO, WORLD!'))
 
     def test_copyable(self):
         class PutFile(Document):
@@ -521,7 +518,7 @@ class FileTest(unittest.TestCase):
 
         PutFile.drop_collection()
 
-        text = b('Hello, World!')
+        text = six.b('Hello, World!')
         content_type = 'text/plain'
 
         putfile = PutFile()

tests/fields/geo.py

@@ -1,7 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
-
 import unittest
 
 from mongoengine import *

tests/migration/__init__.py

@@ -1,11 +0,0 @@
-import unittest
-
-from convert_to_new_inheritance_model import *
-from decimalfield_as_float import *
-from referencefield_dbref_to_object_id import *
-from turn_off_inheritance import *
-from uuidfield_to_binary import *
-
-
-if __name__ == '__main__':
-    unittest.main()

tests/migration/convert_to_new_inheritance_model.py

@@ -1,51 +0,0 @@
-# -*- coding: utf-8 -*-
-import unittest
-
-from mongoengine import Document, connect
-from mongoengine.connection import get_db
-from mongoengine.fields import StringField
-
-__all__ = ('ConvertToNewInheritanceModel', )
-
-
-class ConvertToNewInheritanceModel(unittest.TestCase):
-
-    def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
-
-    def tearDown(self):
-        for collection in self.db.collection_names():
-            if 'system.' in collection:
-                continue
-            self.db.drop_collection(collection)
-
-    def test_how_to_convert_to_the_new_inheritance_model(self):
-        """Demonstrates migrating from 0.7 to 0.8
-        """
-
-        # 1. Declaration of the class
-        class Animal(Document):
-            name = StringField()
-            meta = {
-                'allow_inheritance': True,
-                'indexes': ['name']
-            }
-
-        # 2. Remove _types
-        collection = Animal._get_collection()
-        collection.update({}, {"$unset": {"_types": 1}}, multi=True)
-
-        # 3. Confirm extra data is removed
-        count = collection.find({'_types': {"$exists": True}}).count()
-        self.assertEqual(0, count)
-
-        # 4. Remove indexes
-        info = collection.index_information()
-        indexes_to_drop = [key for key, value in info.iteritems()
-                           if '_types' in dict(value['key'])]
-        for index in indexes_to_drop:
-            collection.drop_index(index)
-
-        # 5. Recreate indexes
-        Animal.ensure_indexes()

tests/migration/decimalfield_as_float.py

@@ -1,50 +0,0 @@
- # -*- coding: utf-8 -*-
-import unittest
-import decimal
-from decimal import Decimal
-
-from mongoengine import Document, connect
-from mongoengine.connection import get_db
-from mongoengine.fields import StringField, DecimalField, ListField
-
-__all__ = ('ConvertDecimalField', )
-
-
-class ConvertDecimalField(unittest.TestCase):
-
-    def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
-
-    def test_how_to_convert_decimal_fields(self):
-        """Demonstrates migrating from 0.7 to 0.8
-        """
-
-        # 1. Old definition - using dbrefs
-        class Person(Document):
-            name = StringField()
-            money = DecimalField(force_string=True)
-            monies = ListField(DecimalField(force_string=True))
-
-        Person.drop_collection()
-        Person(name="Wilson Jr", money=Decimal("2.50"),
-               monies=[Decimal("2.10"), Decimal("5.00")]).save()
-
-        # 2. Start the migration by changing the schema
-        # Change DecimalField - add precision and rounding settings
-        class Person(Document):
-            name = StringField()
-            money = DecimalField(precision=2, rounding=decimal.ROUND_HALF_UP)
-            monies = ListField(DecimalField(precision=2,
-                                            rounding=decimal.ROUND_HALF_UP))
-
-        # 3. Loop all the objects and mark parent as changed
-        for p in Person.objects:
-            p._mark_as_changed('money')
-            p._mark_as_changed('monies')
-            p.save()
-
-        # 4. Confirmation of the fix!
-        wilson = Person.objects(name="Wilson Jr").as_pymongo()[0]
-        self.assertTrue(isinstance(wilson['money'], float))
-        self.assertTrue(all([isinstance(m, float) for m in wilson['monies']]))

tests/migration/referencefield_dbref_to_object_id.py

@@ -1,52 +0,0 @@
-# -*- coding: utf-8 -*-
-import unittest
-
-from mongoengine import Document, connect
-from mongoengine.connection import get_db
-from mongoengine.fields import StringField, ReferenceField, ListField
-
-__all__ = ('ConvertToObjectIdsModel', )
-
-
-class ConvertToObjectIdsModel(unittest.TestCase):
-
-    def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
-
-    def test_how_to_convert_to_object_id_reference_fields(self):
-        """Demonstrates migrating from 0.7 to 0.8
-        """
-
-        # 1. Old definition - using dbrefs
-        class Person(Document):
-            name = StringField()
-            parent = ReferenceField('self', dbref=True)
-            friends = ListField(ReferenceField('self', dbref=True))
-
-        Person.drop_collection()
-
-        p1 = Person(name="Wilson", parent=None).save()
-        f1 = Person(name="John", parent=None).save()
-        f2 = Person(name="Paul", parent=None).save()
-        f3 = Person(name="George", parent=None).save()
-        f4 = Person(name="Ringo", parent=None).save()
-        Person(name="Wilson Jr", parent=p1, friends=[f1, f2, f3, f4]).save()
-
-        # 2. Start the migration by changing the schema
-        # Change ReferenceField as now dbref defaults to False
-        class Person(Document):
-            name = StringField()
-            parent = ReferenceField('self')
-            friends = ListField(ReferenceField('self'))
-
-        # 3. Loop all the objects and mark parent as changed
-        for p in Person.objects:
-            p._mark_as_changed('parent')
-            p._mark_as_changed('friends')
-            p.save()
-
-        # 4. Confirmation of the fix!
-        wilson = Person.objects(name="Wilson Jr").as_pymongo()[0]
-        self.assertEqual(p1.id, wilson['parent'])
-        self.assertEqual([f1.id, f2.id, f3.id, f4.id], wilson['friends'])

tests/migration/turn_off_inheritance.py

@@ -1,62 +0,0 @@
-# -*- coding: utf-8 -*-
-import unittest
-
-from mongoengine import Document, connect
-from mongoengine.connection import get_db
-from mongoengine.fields import StringField
-
-__all__ = ('TurnOffInheritanceTest', )
-
-
-class TurnOffInheritanceTest(unittest.TestCase):
-
-    def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
-
-    def tearDown(self):
-        for collection in self.db.collection_names():
-            if 'system.' in collection:
-                continue
-            self.db.drop_collection(collection)
-
-    def test_how_to_turn_off_inheritance(self):
-        """Demonstrates migrating from allow_inheritance = True to False.
-        """
-
-        # 1. Old declaration of the class
-
-        class Animal(Document):
-            name = StringField()
-            meta = {
-                'allow_inheritance': True,
-                'indexes': ['name']
-            }
-
-        # 2. Turn off inheritance
-        class Animal(Document):
-            name = StringField()
-            meta = {
-                'allow_inheritance': False,
-                'indexes': ['name']
-            }
-
-        # 3. Remove _types and _cls
-        collection = Animal._get_collection()
-        collection.update({}, {"$unset": {"_types": 1, "_cls": 1}}, multi=True)
-
-        # 3. Confirm extra data is removed
-        count = collection.find({"$or": [{'_types': {"$exists": True}},
-                                         {'_cls': {"$exists": True}}]}).count()
-        assert count == 0
-
-        # 4. Remove indexes
-        info = collection.index_information()
-        indexes_to_drop = [key for key, value in info.iteritems()
-                           if '_types' in dict(value['key'])
-                              or '_cls' in dict(value['key'])]
-        for index in indexes_to_drop:
-            collection.drop_index(index)
-
-        # 5. Recreate indexes
-        Animal.ensure_indexes()

tests/migration/uuidfield_to_binary.py

@@ -1,48 +0,0 @@
-# -*- coding: utf-8 -*-
-import unittest
-import uuid
-
-from mongoengine import Document, connect
-from mongoengine.connection import get_db
-from mongoengine.fields import StringField, UUIDField, ListField
-
-__all__ = ('ConvertToBinaryUUID', )
-
-
-class ConvertToBinaryUUID(unittest.TestCase):
-
-    def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
-
-    def test_how_to_convert_to_binary_uuid_fields(self):
-        """Demonstrates migrating from 0.7 to 0.8
-        """
-
-        # 1. Old definition - using dbrefs
-        class Person(Document):
-            name = StringField()
-            uuid = UUIDField(binary=False)
-            uuids = ListField(UUIDField(binary=False))
-
-        Person.drop_collection()
-        Person(name="Wilson Jr", uuid=uuid.uuid4(),
-               uuids=[uuid.uuid4(), uuid.uuid4()]).save()
-
-        # 2. Start the migration by changing the schema
-        # Change UUIDFIeld as now binary defaults to True
-        class Person(Document):
-            name = StringField()
-            uuid = UUIDField()
-            uuids = ListField(UUIDField())
-
-        # 3. Loop all the objects and mark parent as changed
-        for p in Person.objects:
-            p._mark_as_changed('uuid')
-            p._mark_as_changed('uuids')
-            p.save()
-
-        # 4. Confirmation of the fix!
-        wilson = Person.objects(name="Wilson Jr").as_pymongo()[0]
-        self.assertTrue(isinstance(wilson['uuid'], uuid.UUID))
-        self.assertTrue(all([isinstance(u, uuid.UUID) for u in wilson['uuids']]))

tests/queryset/field_list.py

@@ -1,6 +1,3 @@
-import sys
-sys.path[0:0] = [""]
-
 import unittest
 
 from mongoengine import *
@@ -95,7 +92,7 @@ class OnlyExcludeAllTest(unittest.TestCase):
         exclude = ['d', 'e']
         only = ['b', 'c']
 
-        qs = MyDoc.objects.fields(**dict(((i, 1) for i in include)))
+        qs = MyDoc.objects.fields(**{i: 1 for i in include})
         self.assertEqual(qs._loaded_fields.as_dict(),
                          {'a': 1, 'b': 1, 'c': 1, 'd': 1, 'e': 1})
         qs = qs.only(*only)
@@ -103,14 +100,14 @@ class OnlyExcludeAllTest(unittest.TestCase):
         qs = qs.exclude(*exclude)
         self.assertEqual(qs._loaded_fields.as_dict(), {'b': 1, 'c': 1})
 
-        qs = MyDoc.objects.fields(**dict(((i, 1) for i in include)))
+        qs = MyDoc.objects.fields(**{i: 1 for i in include})
         qs = qs.exclude(*exclude)
         self.assertEqual(qs._loaded_fields.as_dict(), {'a': 1, 'b': 1, 'c': 1})
         qs = qs.only(*only)
         self.assertEqual(qs._loaded_fields.as_dict(), {'b': 1, 'c': 1})
 
         qs = MyDoc.objects.exclude(*exclude)
-        qs = qs.fields(**dict(((i, 1) for i in include)))
+        qs = qs.fields(**{i: 1 for i in include})
         self.assertEqual(qs._loaded_fields.as_dict(), {'a': 1, 'b': 1, 'c': 1})
         qs = qs.only(*only)
         self.assertEqual(qs._loaded_fields.as_dict(), {'b': 1, 'c': 1})
@@ -129,7 +126,7 @@ class OnlyExcludeAllTest(unittest.TestCase):
         exclude = ['d', 'e']
         only = ['b', 'c']
 
-        qs = MyDoc.objects.fields(**dict(((i, 1) for i in include)))
+        qs = MyDoc.objects.fields(**{i: 1 for i in include})
         qs = qs.exclude(*exclude)
         qs = qs.only(*only)
         qs = qs.fields(slice__b=5)

tests/queryset/geo.py

@@ -1,9 +1,5 @@
-import sys
-
-sys.path[0:0] = [""]
-
-import unittest
 from datetime import datetime, timedelta
+import unittest
 
 from pymongo.errors import OperationFailure
 from mongoengine import *

tests/queryset/modify.py

@@ -1,6 +1,3 @@
-import sys
-sys.path[0:0] = [""]
-
 import unittest
 
 from mongoengine import connect, Document, IntField
@@ -99,4 +96,4 @@ class FindAndModifyTest(unittest.TestCase):
 
 
 if __name__ == '__main__':
-    unittest.main()
+    unittest.main()

tests/queryset/queryset.py

@@ -9,13 +9,13 @@ from nose.plugins.skip import SkipTest
 import pymongo
 from pymongo.errors import ConfigurationError
 from pymongo.read_preferences import ReadPreference
-
+import six
 
 from mongoengine import *
 from mongoengine.connection import get_connection, get_db
 from mongoengine.context_managers import query_counter, switch_db
 from mongoengine.errors import InvalidQueryError
-from mongoengine.python_support import IS_PYMONGO_3, PY3
+from mongoengine.python_support import IS_PYMONGO_3
 from mongoengine.queryset import (DoesNotExist, MultipleObjectsReturned,
                                   QuerySet, QuerySetManager, queryset_manager)
 
@@ -25,7 +25,10 @@ __all__ = ("QuerySetTest",)
 class db_ops_tracker(query_counter):
 
     def get_ops(self):
-        ignore_query = {"ns": {"$ne": "%s.system.indexes" % self.db.name}}
+        ignore_query = {
+            'ns': {'$ne': '%s.system.indexes' % self.db.name},
+            'command.count': {'$ne': 'system.profile'}
+        }
         return list(self.db.system.profile.find(ignore_query))
 
 
@@ -94,12 +97,12 @@ class QuerySetTest(unittest.TestCase):
             author = ReferenceField(self.Person)
             author2 = GenericReferenceField()
 
-        def test_reference():
+        # test addressing a field from a reference
+        with self.assertRaises(InvalidQueryError):
             list(BlogPost.objects(author__name="test"))
 
-        self.assertRaises(InvalidQueryError, test_reference)
-
-        def test_generic_reference():
+        # should fail for a generic reference as well
+        with self.assertRaises(InvalidQueryError):
             list(BlogPost.objects(author2__name="test"))
 
     def test_find(self):
@@ -174,7 +177,7 @@ class QuerySetTest(unittest.TestCase):
 
         # Test larger slice __repr__
         self.Person.objects.delete()
-        for i in xrange(55):
+        for i in range(55):
             self.Person(name='A%s' % i, age=i).save()
 
         self.assertEqual(self.Person.objects.count(), 55)
@@ -218,14 +221,15 @@ class QuerySetTest(unittest.TestCase):
         person = self.Person.objects[1]
         self.assertEqual(person.name, "User B")
 
-        self.assertRaises(IndexError, self.Person.objects.__getitem__, 2)
+        with self.assertRaises(IndexError):
+            self.Person.objects[2]
 
         # Find a document using just the object id
         person = self.Person.objects.with_id(person1.id)
         self.assertEqual(person.name, "User A")
 
-        self.assertRaises(
-            InvalidQueryError, self.Person.objects(name="User A").with_id, person1.id)
+        with self.assertRaises(InvalidQueryError):
+            self.Person.objects(name="User A").with_id(person1.id)
 
     def test_find_only_one(self):
         """Ensure that a query using ``get`` returns at most one result.
@@ -337,9 +341,37 @@ class QuerySetTest(unittest.TestCase):
         query = query.filter(boolfield=True)
         self.assertEqual(query.count(), 1)
 
+    def test_batch_size(self):
+        """Ensure that batch_size works."""
+        class A(Document):
+            s = StringField()
+
+        A.drop_collection()
+
+        for i in range(100):
+            A.objects.create(s=str(i))
+
+        # test iterating over the result set
+        cnt = 0
+        for a in A.objects.batch_size(10):
+            cnt += 1
+        self.assertEqual(cnt, 100)
+
+        # test chaining
+        qs = A.objects.all()
+        qs = qs.limit(10).batch_size(20).skip(91)
+        cnt = 0
+        for a in qs:
+            cnt += 1
+        self.assertEqual(cnt, 9)
+
+        # test invalid batch size
+        qs = A.objects.batch_size(-1)
+        with self.assertRaises(ValueError):
+            list(qs)
+
     def test_update_write_concern(self):
         """Test that passing write_concern works"""
-
         self.Person.drop_collection()
 
         write_concern = {"fsync": True}
@@ -365,18 +397,14 @@ class QuerySetTest(unittest.TestCase):
         """Test to ensure that update is passed a value to update to"""
         self.Person.drop_collection()
 
-        author = self.Person(name='Test User')
-        author.save()
+        author = self.Person.objects.create(name='Test User')
 
-        def update_raises():
+        with self.assertRaises(OperationError):
             self.Person.objects(pk=author.pk).update({})
 
-        def update_one_raises():
+        with self.assertRaises(OperationError):
             self.Person.objects(pk=author.pk).update_one({})
 
-        self.assertRaises(OperationError, update_raises)
-        self.assertRaises(OperationError, update_one_raises)
-
     def test_update_array_position(self):
         """Ensure that updating by array position works.
 
@@ -404,8 +432,8 @@ class QuerySetTest(unittest.TestCase):
         Blog.objects.create(posts=[post2, post1])
 
         # Update all of the first comments of second posts of all blogs
-        Blog.objects().update(set__posts__1__comments__0__name="testc")
-        testc_blogs = Blog.objects(posts__1__comments__0__name="testc")
+        Blog.objects().update(set__posts__1__comments__0__name='testc')
+        testc_blogs = Blog.objects(posts__1__comments__0__name='testc')
         self.assertEqual(testc_blogs.count(), 2)
 
         Blog.drop_collection()
@@ -414,14 +442,13 @@ class QuerySetTest(unittest.TestCase):
 
         # Update only the first blog returned by the query
         Blog.objects().update_one(
-            set__posts__1__comments__1__name="testc")
-        testc_blogs = Blog.objects(posts__1__comments__1__name="testc")
+            set__posts__1__comments__1__name='testc')
+        testc_blogs = Blog.objects(posts__1__comments__1__name='testc')
         self.assertEqual(testc_blogs.count(), 1)
 
         # Check that using this indexing syntax on a non-list fails
-        def non_list_indexing():
-            Blog.objects().update(set__posts__1__comments__0__name__1="asdf")
-        self.assertRaises(InvalidQueryError, non_list_indexing)
+        with self.assertRaises(InvalidQueryError):
+            Blog.objects().update(set__posts__1__comments__0__name__1='asdf')
 
         Blog.drop_collection()
 
@@ -489,15 +516,12 @@ class QuerySetTest(unittest.TestCase):
         self.assertEqual(simple.x, [1, 2, None, 4, 3, 2, 3, 4])
 
         # Nested updates arent supported yet..
-        def update_nested():
+        with self.assertRaises(OperationError):
             Simple.drop_collection()
             Simple(x=[{'test': [1, 2, 3, 4]}]).save()
             Simple.objects(x__test=2).update(set__x__S__test__S=3)
             self.assertEqual(simple.x, [1, 2, 3, 4])
 
-        self.assertRaises(OperationError, update_nested)
-        Simple.drop_collection()
-
     def test_update_using_positional_operator_embedded_document(self):
         """Ensure that the embedded documents can be updated using the positional
         operator."""
@@ -590,11 +614,11 @@ class QuerySetTest(unittest.TestCase):
             members = DictField()
 
         club = Club()
-        club.members['John'] = dict(gender="M", age=13)
+        club.members['John'] = {'gender': 'M', 'age': 13}
         club.save()
 
         Club.objects().update(
-            set__members={"John": dict(gender="F", age=14)})
+            set__members={"John": {'gender': 'F', 'age': 14}})
 
         club = Club.objects().first()
         self.assertEqual(club.members['John']['gender'], "F")
@@ -775,7 +799,7 @@ class QuerySetTest(unittest.TestCase):
             post2 = Post(comments=[comment2, comment2])
 
             blogs = []
-            for i in xrange(1, 100):
+            for i in range(1, 100):
                 blogs.append(Blog(title="post %s" % i, posts=[post1, post2]))
 
             Blog.objects.insert(blogs, load_bulk=False)
@@ -812,30 +836,31 @@ class QuerySetTest(unittest.TestCase):
 
         self.assertEqual(Blog.objects.count(), 2)
 
-        # test handles people trying to upsert
-        def throw_operation_error():
+        # test inserting an existing document (shouldn't be allowed)
+        with self.assertRaises(OperationError):
+            blog = Blog.objects.first()
+            Blog.objects.insert(blog)
+
+        # test inserting a query set
+        with self.assertRaises(OperationError):
             blogs = Blog.objects
             Blog.objects.insert(blogs)
 
-        self.assertRaises(OperationError, throw_operation_error)
-
-        # Test can insert new doc
+        # insert a new doc
         new_post = Blog(title="code123", id=ObjectId())
         Blog.objects.insert(new_post)
 
-        # test handles other classes being inserted
-        def throw_operation_error_wrong_doc():
-            class Author(Document):
-                pass
+        class Author(Document):
+            pass
+
+        # try inserting a different document class
+        with self.assertRaises(OperationError):
             Blog.objects.insert(Author())
 
-        self.assertRaises(OperationError, throw_operation_error_wrong_doc)
-
-        def throw_operation_error_not_a_document():
+        # try inserting a non-document
+        with self.assertRaises(OperationError):
             Blog.objects.insert("HELLO WORLD")
 
-        self.assertRaises(OperationError, throw_operation_error_not_a_document)
-
         Blog.drop_collection()
 
         blog1 = Blog(title="code", posts=[post1, post2])
@@ -855,14 +880,13 @@ class QuerySetTest(unittest.TestCase):
         blog3 = Blog(title="baz", posts=[post1, post2])
         Blog.objects.insert([blog1, blog2])
 
-        def throw_operation_error_not_unique():
+        with self.assertRaises(NotUniqueError):
             Blog.objects.insert([blog2, blog3])
 
-        self.assertRaises(NotUniqueError, throw_operation_error_not_unique)
         self.assertEqual(Blog.objects.count(), 2)
 
-        Blog.objects.insert([blog2, blog3], write_concern={"w": 0,
-                                                           'continue_on_error': True})
+        Blog.objects.insert([blog2, blog3],
+                            write_concern={"w": 0, 'continue_on_error': True})
         self.assertEqual(Blog.objects.count(), 3)
 
     def test_get_changed_fields_query_count(self):
@@ -995,7 +1019,7 @@ class QuerySetTest(unittest.TestCase):
 
         Doc.drop_collection()
 
-        for i in xrange(1000):
+        for i in range(1000):
             Doc(number=i).save()
 
         docs = Doc.objects.order_by('number')
@@ -1149,7 +1173,7 @@ class QuerySetTest(unittest.TestCase):
         qs = list(qs)
         expected = list(expected)
         self.assertEqual(len(qs), len(expected))
-        for i in xrange(len(qs)):
+        for i in range(len(qs)):
             self.assertEqual(qs[i], expected[i])
 
     def test_ordering(self):
@@ -1189,7 +1213,8 @@ class QuerySetTest(unittest.TestCase):
         self.assertSequence(qs, expected)
 
     def test_clear_ordering(self):
-        """ Ensure that the default ordering can be cleared by calling order_by().
+        """Ensure that the default ordering can be cleared by calling
+        order_by() w/o any arguments.
         """
         class BlogPost(Document):
             title = StringField()
@@ -1205,12 +1230,13 @@ class QuerySetTest(unittest.TestCase):
             BlogPost.objects.filter(title='whatever').first()
             self.assertEqual(len(q.get_ops()), 1)
             self.assertEqual(
-                q.get_ops()[0]['query']['$orderby'], {u'published_date': -1})
+                q.get_ops()[0]['query']['$orderby'],
+                {'published_date': -1}
+            )
 
         with db_ops_tracker() as q:
             BlogPost.objects.filter(title='whatever').order_by().first()
             self.assertEqual(len(q.get_ops()), 1)
-            print q.get_ops()[0]['query']
             self.assertFalse('$orderby' in q.get_ops()[0]['query'])
 
     def test_no_ordering_for_get(self):
@@ -1239,7 +1265,8 @@ class QuerySetTest(unittest.TestCase):
             self.assertFalse('$orderby' in q.get_ops()[0]['query'])
 
     def test_find_embedded(self):
-        """Ensure that an embedded document is properly returned from a query.
+        """Ensure that an embedded document is properly returned from
+        a query.
         """
         class User(EmbeddedDocument):
             name = StringField()
@@ -1250,16 +1277,31 @@ class QuerySetTest(unittest.TestCase):
 
         BlogPost.drop_collection()
 
-        post = BlogPost(content='Had a good coffee today...')
-        post.author = User(name='Test User')
-        post.save()
+        BlogPost.objects.create(
+            author=User(name='Test User'),
+            content='Had a good coffee today...'
+        )
 
         result = BlogPost.objects.first()
         self.assertTrue(isinstance(result.author, User))
         self.assertEqual(result.author.name, 'Test User')
 
+    def test_find_empty_embedded(self):
+        """Ensure that you can save and find an empty embedded document."""
+        class User(EmbeddedDocument):
+            name = StringField()
+
+        class BlogPost(Document):
+            content = StringField()
+            author = EmbeddedDocumentField(User)
+
         BlogPost.drop_collection()
 
+        BlogPost.objects.create(content='Anonymous post...')
+
+        result = BlogPost.objects.get(author=None)
+        self.assertEqual(result.author, None)
+
     def test_find_dict_item(self):
         """Ensure that DictField items may be found.
         """
@@ -1667,7 +1709,7 @@ class QuerySetTest(unittest.TestCase):
 
         Log.drop_collection()
 
-        for i in xrange(10):
+        for i in range(10):
             Log().save()
 
         Log.objects()[3:5].delete()
@@ -1867,12 +1909,10 @@ class QuerySetTest(unittest.TestCase):
         Site.objects(id=s.id).update_one(pull__collaborators__user='Esteban')
         self.assertEqual(Site.objects.first().collaborators, [])
 
-        def pull_all():
+        with self.assertRaises(InvalidQueryError):
             Site.objects(id=s.id).update_one(
                 pull_all__collaborators__user=['Ross'])
 
-        self.assertRaises(InvalidQueryError, pull_all)
-
     def test_pull_from_nested_embedded(self):
 
         class User(EmbeddedDocument):
@@ -1903,12 +1943,10 @@ class QuerySetTest(unittest.TestCase):
             pull__collaborators__unhelpful={'name': 'Frank'})
         self.assertEqual(Site.objects.first().collaborators['unhelpful'], [])
 
-        def pull_all():
+        with self.assertRaises(InvalidQueryError):
             Site.objects(id=s.id).update_one(
                 pull_all__collaborators__helpful__name=['Ross'])
 
-        self.assertRaises(InvalidQueryError, pull_all)
-
     def test_pull_from_nested_mapfield(self):
 
         class Collaborator(EmbeddedDocument):
@@ -1937,12 +1975,10 @@ class QuerySetTest(unittest.TestCase):
             pull__collaborators__unhelpful={'user': 'Frank'})
         self.assertEqual(Site.objects.first().collaborators['unhelpful'], [])
 
-        def pull_all():
+        with self.assertRaises(InvalidQueryError):
             Site.objects(id=s.id).update_one(
                 pull_all__collaborators__helpful__user=['Ross'])
 
-        self.assertRaises(InvalidQueryError, pull_all)
-
     def test_update_one_pop_generic_reference(self):
 
         class BlogTag(Document):
@@ -2199,6 +2235,21 @@ class QuerySetTest(unittest.TestCase):
             a.author.name for a in Author.objects.order_by('-author__age')]
         self.assertEqual(names, ['User A', 'User B', 'User C'])
 
+    def test_comment(self):
+        """Make sure adding a comment to the query works."""
+        class User(Document):
+            age = IntField()
+
+        with db_ops_tracker() as q:
+            adult = (User.objects.filter(age__gte=18)
+                .comment('looking for an adult')
+                .first())
+            ops = q.get_ops()
+            self.assertEqual(len(ops), 1)
+            op = ops[0]
+            self.assertEqual(op['query']['$query'], {'age': {'$gte': 18}})
+            self.assertEqual(op['query']['$comment'], 'looking for an adult')
+
     def test_map_reduce(self):
         """Ensure map/reduce is both mapping and reducing.
         """
@@ -2552,7 +2603,7 @@ class QuerySetTest(unittest.TestCase):
         BlogPost(hits=2, tags=['music', 'actors']).save()
 
         def test_assertions(f):
-            f = dict((key, int(val)) for key, val in f.items())
+            f = {key: int(val) for key, val in f.items()}
             self.assertEqual(
                 set(['music', 'film', 'actors', 'watch']), set(f.keys()))
             self.assertEqual(f['music'], 3)
@@ -2567,7 +2618,7 @@ class QuerySetTest(unittest.TestCase):
 
         # Ensure query is taken into account
         def test_assertions(f):
-            f = dict((key, int(val)) for key, val in f.items())
+            f = {key: int(val) for key, val in f.items()}
             self.assertEqual(set(['music', 'actors', 'watch']), set(f.keys()))
             self.assertEqual(f['music'], 2)
             self.assertEqual(f['actors'], 1)
@@ -2631,7 +2682,7 @@ class QuerySetTest(unittest.TestCase):
         doc.save()
 
         def test_assertions(f):
-            f = dict((key, int(val)) for key, val in f.items())
+            f = {key: int(val) for key, val in f.items()}
             self.assertEqual(
                 set(['62-3331-1656', '62-3332-1656']), set(f.keys()))
             self.assertEqual(f['62-3331-1656'], 2)
@@ -2645,7 +2696,7 @@ class QuerySetTest(unittest.TestCase):
 
         # Ensure query is taken into account
         def test_assertions(f):
-            f = dict((key, int(val)) for key, val in f.items())
+            f = {key: int(val) for key, val in f.items()}
             self.assertEqual(set(['62-3331-1656']), set(f.keys()))
             self.assertEqual(f['62-3331-1656'], 2)
 
@@ -2752,10 +2803,10 @@ class QuerySetTest(unittest.TestCase):
 
         Test.drop_collection()
 
-        for i in xrange(50):
+        for i in range(50):
             Test(val=1).save()
 
-        for i in xrange(20):
+        for i in range(20):
             Test(val=2).save()
 
         freqs = Test.objects.item_frequencies(
@@ -3545,7 +3596,7 @@ class QuerySetTest(unittest.TestCase):
 
         Post.drop_collection()
 
-        for i in xrange(10):
+        for i in range(10):
             Post(title="Post %s" % i).save()
 
         self.assertEqual(5, Post.objects.limit(5).skip(5).count(with_limit_and_skip=True))
@@ -3560,7 +3611,7 @@ class QuerySetTest(unittest.TestCase):
             pass
 
         MyDoc.drop_collection()
-        for i in xrange(0, 10):
+        for i in range(0, 10):
             MyDoc().save()
 
         self.assertEqual(MyDoc.objects.count(), 10)
@@ -3616,7 +3667,7 @@ class QuerySetTest(unittest.TestCase):
 
         Number.drop_collection()
 
-        for i in xrange(1, 101):
+        for i in range(1, 101):
             t = Number(n=i)
             t.save()
 
@@ -3763,11 +3814,9 @@ class QuerySetTest(unittest.TestCase):
         self.assertTrue(a in results)
         self.assertTrue(c in results)
 
-        def invalid_where():
+        with self.assertRaises(TypeError):
             list(IntPair.objects.where(fielda__gte=3))
 
-        self.assertRaises(TypeError, invalid_where)
-
     def test_scalar(self):
 
         class Organization(Document):
@@ -4023,7 +4072,7 @@ class QuerySetTest(unittest.TestCase):
 
         # Test larger slice __repr__
         self.Person.objects.delete()
-        for i in xrange(55):
+        for i in range(55):
             self.Person(name='A%s' % i, age=i).save()
 
         self.assertEqual(self.Person.objects.scalar('name').count(), 55)
@@ -4031,7 +4080,7 @@ class QuerySetTest(unittest.TestCase):
             "A0", "%s" % self.Person.objects.order_by('name').scalar('name').first())
         self.assertEqual(
             "A0", "%s" % self.Person.objects.scalar('name').order_by('name')[0])
-        if PY3:
+        if six.PY3:
             self.assertEqual("['A1', 'A2']", "%s" % self.Person.objects.order_by(
                 'age').scalar('name')[1:3])
             self.assertEqual("['A51', 'A52']", "%s" % self.Person.objects.order_by(
@@ -4049,7 +4098,7 @@ class QuerySetTest(unittest.TestCase):
 
         pks = self.Person.objects.order_by('age').scalar('pk')[1:3]
         names = self.Person.objects.scalar('name').in_bulk(list(pks)).values()
-        if PY3:
+        if six.PY3:
             expected = "['A1', 'A2']"
         else:
             expected = "[u'A1', u'A2']"
@@ -4405,7 +4454,7 @@ class QuerySetTest(unittest.TestCase):
             name = StringField()
 
         Person.drop_collection()
-        for i in xrange(100):
+        for i in range(100):
             Person(name="No: %s" % i).save()
 
         with query_counter() as q:
@@ -4436,7 +4485,7 @@ class QuerySetTest(unittest.TestCase):
             name = StringField()
 
         Person.drop_collection()
-        for i in xrange(100):
+        for i in range(100):
             Person(name="No: %s" % i).save()
 
         with query_counter() as q:
@@ -4480,7 +4529,7 @@ class QuerySetTest(unittest.TestCase):
             fields = DictField()
 
         Noddy.drop_collection()
-        for i in xrange(100):
+        for i in range(100):
             noddy = Noddy()
             for j in range(20):
                 noddy.fields["key" + str(j)] = "value " + str(j)
@@ -4492,7 +4541,9 @@ class QuerySetTest(unittest.TestCase):
         self.assertEqual(counter, 100)
 
         self.assertEqual(len(list(docs)), 100)
-        self.assertRaises(TypeError, lambda: len(docs))
+
+        with self.assertRaises(TypeError):
+            len(docs)
 
         with query_counter() as q:
             self.assertEqual(q, 0)
@@ -4681,7 +4732,7 @@ class QuerySetTest(unittest.TestCase):
             name = StringField()
 
         Person.drop_collection()
-        for i in xrange(100):
+        for i in range(100):
             Person(name="No: %s" % i).save()
 
         with query_counter() as q:
@@ -4805,10 +4856,10 @@ class QuerySetTest(unittest.TestCase):
         ])
 
     def test_delete_count(self):
-        [self.Person(name="User {0}".format(i), age=i * 10).save() for i in xrange(1, 4)]
+        [self.Person(name="User {0}".format(i), age=i * 10).save() for i in range(1, 4)]
         self.assertEqual(self.Person.objects().delete(), 3)  # test ordinary QuerySey delete count
 
-        [self.Person(name="User {0}".format(i), age=i * 10).save() for i in xrange(1, 4)]
+        [self.Person(name="User {0}".format(i), age=i * 10).save() for i in range(1, 4)]
 
         self.assertEqual(self.Person.objects().skip(1).delete(), 2)  # test Document delete with existing documents
 
@@ -4817,12 +4868,14 @@ class QuerySetTest(unittest.TestCase):
 
     def test_max_time_ms(self):
         # 778: max_time_ms can get only int or None as input
-        self.assertRaises(TypeError, self.Person.objects(name="name").max_time_ms, "not a number")
+        self.assertRaises(TypeError,
+                          self.Person.objects(name="name").max_time_ms,
+                          'not a number')
 
     def test_subclass_field_query(self):
         class Animal(Document):
             is_mamal = BooleanField()
-            meta = dict(allow_inheritance=True)
+            meta = {'allow_inheritance': True}
 
         class Cat(Animal):
             whiskers_length = FloatField()
@@ -4860,6 +4913,56 @@ class QuerySetTest(unittest.TestCase):
 
         self.assertEqual(1, Doc.objects(item__type__="axe").count())
 
+    def test_len_during_iteration(self):
+        """Tests that calling len on a queyset during iteration doesn't
+        stop paging.
+        """
+        class Data(Document):
+            pass
+
+        for i in range(300):
+            Data().save()
+
+        records = Data.objects.limit(250)
+
+        # This should pull all 250 docs from mongo and populate the result
+        # cache
+        len(records)
+
+        # Assert that iterating over documents in the qs touches every
+        # document even if we call len(qs) midway through the iteration.
+        for i, r in enumerate(records):
+            if i == 58:
+                len(records)
+        self.assertEqual(i, 249)
+
+        # Assert the same behavior is true even if we didn't pre-populate the
+        # result cache.
+        records = Data.objects.limit(250)
+        for i, r in enumerate(records):
+            if i == 58:
+                len(records)
+        self.assertEqual(i, 249)
+
+    def test_iteration_within_iteration(self):
+        """You should be able to reliably iterate over all the documents
+        in a given queryset even if there are multiple iterations of it
+        happening at the same time.
+        """
+        class Data(Document):
+            pass
+
+        for i in range(300):
+            Data().save()
+
+        qs = Data.objects.limit(250)
+        for i, doc in enumerate(qs):
+            for j, doc2 in enumerate(qs):
+                pass
+
+        self.assertEqual(i, 249)
+        self.assertEqual(j, 249)
+
 
 if __name__ == '__main__':
     unittest.main()

tests/queryset/transform.py

@@ -238,7 +238,8 @@ class TransformTest(unittest.TestCase):
         box = [(35.0, -125.0), (40.0, -100.0)]
         # I *meant* to execute location__within_box=box
         events = Event.objects(location__within=box)
-        self.assertRaises(InvalidQueryError, lambda: events.count())
+        with self.assertRaises(InvalidQueryError):
+            events.count()
 
 
 if __name__ == '__main__':

tests/queryset/visitor.py

@@ -185,7 +185,7 @@ class QTest(unittest.TestCase):
             x = IntField()
 
         TestDoc.drop_collection()
-        for i in xrange(1, 101):
+        for i in range(1, 101):
             t = TestDoc(x=i)
             t.save()
 
@@ -268,14 +268,13 @@ class QTest(unittest.TestCase):
         self.assertEqual(self.Person.objects(Q(age__in=[20, 30])).count(), 3)
 
         # Test invalid query objs
-        def wrong_query_objs():
+        with self.assertRaises(InvalidQueryError):
             self.Person.objects('user1')
 
-        def wrong_query_objs_filter():
-            self.Person.objects('user1')
+        # filter should fail, too
+        with self.assertRaises(InvalidQueryError):
+            self.Person.objects.filter('user1')
 
-        self.assertRaises(InvalidQueryError, wrong_query_objs)
-        self.assertRaises(InvalidQueryError, wrong_query_objs_filter)
 
     def test_q_regex(self):
         """Ensure that Q objects can be queried using regexes.

tests/test_connection.py

@@ -1,9 +1,6 @@
-import sys
 import datetime
 from pymongo.errors import OperationFailure
 
-sys.path[0:0] = [""]
-
 try:
     import unittest2 as unittest
 except ImportError:
@@ -19,7 +16,8 @@ from mongoengine import (
 )
 from mongoengine.python_support import IS_PYMONGO_3
 import mongoengine.connection
-from mongoengine.connection import get_db, get_connection, ConnectionError
+from mongoengine.connection import (MongoEngineConnectionError, get_db,
+                                    get_connection)
 
 
 def get_tz_awareness(connection):
@@ -159,7 +157,10 @@ class ConnectionTest(unittest.TestCase):
         c.mongoenginetest.add_user("username", "password")
 
         if not IS_PYMONGO_3:
-            self.assertRaises(ConnectionError, connect, "testdb_uri_bad", host='mongodb://test:password@localhost')
+            self.assertRaises(
+                MongoEngineConnectionError, connect, 'testdb_uri_bad',
+                host='mongodb://test:password@localhost'
+            )
 
         connect("testdb_uri", host='mongodb://username:password@localhost/mongoenginetest')
 
@@ -174,19 +175,9 @@ class ConnectionTest(unittest.TestCase):
         c.mongoenginetest.system.users.remove({})
 
     def test_connect_uri_without_db(self):
-        """Ensure connect() method works properly with uri's without database_name
+        """Ensure connect() method works properly if the URI doesn't
+        include a database name.
         """
-        c = connect(db='mongoenginetest', alias='admin')
-        c.admin.system.users.remove({})
-        c.mongoenginetest.system.users.remove({})
-
-        c.admin.add_user("admin", "password")
-        c.admin.authenticate("admin", "password")
-        c.mongoenginetest.add_user("username", "password")
-
-        if not IS_PYMONGO_3:
-            self.assertRaises(ConnectionError, connect, "testdb_uri_bad", host='mongodb://test:password@localhost')
-
         connect("mongoenginetest", host='mongodb://localhost/')
 
         conn = get_connection()
@@ -196,8 +187,31 @@ class ConnectionTest(unittest.TestCase):
         self.assertTrue(isinstance(db, pymongo.database.Database))
         self.assertEqual(db.name, 'mongoenginetest')
 
-        c.admin.system.users.remove({})
-        c.mongoenginetest.system.users.remove({})
+    def test_connect_uri_default_db(self):
+        """Ensure connect() defaults to the right database name if
+        the URI and the database_name don't explicitly specify it.
+        """
+        connect(host='mongodb://localhost/')
+
+        conn = get_connection()
+        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+
+        db = get_db()
+        self.assertTrue(isinstance(db, pymongo.database.Database))
+        self.assertEqual(db.name, 'test')
+
+    def test_uri_without_credentials_doesnt_override_conn_settings(self):
+        """Ensure connect() uses the username & password params if the URI
+        doesn't explicitly specify them.
+        """
+        c = connect(host='mongodb://localhost/mongoenginetest',
+                    username='user',
+                    password='pass')
+
+        # OperationFailure means that mongoengine attempted authentication
+        # w/ the provided username/password and failed - that's the desired
+        # behavior. If the MongoDB URI would override the credentials
+        self.assertRaises(OperationFailure, get_db)
 
     def test_connect_uri_with_authsource(self):
         """Ensure that the connect() method works well with
@@ -216,10 +230,11 @@ class ConnectionTest(unittest.TestCase):
             self.assertRaises(OperationFailure, test_conn.server_info)
         else:
             self.assertRaises(
-                ConnectionError, connect, 'mongoenginetest', alias='test1',
+                MongoEngineConnectionError, connect, 'mongoenginetest',
+                alias='test1',
                 host='mongodb://username2:password@localhost/mongoenginetest'
             )
-            self.assertRaises(ConnectionError, get_db, 'test1')
+            self.assertRaises(MongoEngineConnectionError, get_db, 'test1')
 
         # Authentication succeeds with "authSource"
         connect(
@@ -240,7 +255,7 @@ class ConnectionTest(unittest.TestCase):
         """
         register_connection('testdb', 'mongoenginetest2')
 
-        self.assertRaises(ConnectionError, get_connection)
+        self.assertRaises(MongoEngineConnectionError, get_connection)
         conn = get_connection('testdb')
         self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
 

tests/test_context_managers.py

@@ -1,5 +1,3 @@
-import sys
-sys.path[0:0] = [""]
 import unittest
 
 from mongoengine import *
@@ -79,7 +77,7 @@ class ContextManagersTest(unittest.TestCase):
         User.drop_collection()
         Group.drop_collection()
 
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             User(name='user %s' % i).save()
 
         user = User.objects.first()
@@ -117,7 +115,7 @@ class ContextManagersTest(unittest.TestCase):
         User.drop_collection()
         Group.drop_collection()
 
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             User(name='user %s' % i).save()
 
         user = User.objects.first()
@@ -195,7 +193,7 @@ class ContextManagersTest(unittest.TestCase):
         with query_counter() as q:
             self.assertEqual(0, q)
 
-            for i in xrange(1, 51):
+            for i in range(1, 51):
                 db.test.find({}).count()
 
             self.assertEqual(50, q)

tests/test_datastructures.py

@@ -1,5 +1,6 @@
 import unittest
-from mongoengine.base.datastructures import StrictDict, SemiStrictDict 
+
+from mongoengine.base.datastructures import StrictDict, SemiStrictDict
 
 
 class TestStrictDict(unittest.TestCase):
@@ -13,9 +14,18 @@ class TestStrictDict(unittest.TestCase):
         d = self.dtype(a=1, b=1, c=1)
         self.assertEqual((d.a, d.b, d.c), (1, 1, 1))
 
+    def test_repr(self):
+        d = self.dtype(a=1, b=2, c=3)
+        self.assertEqual(repr(d), '{"a": 1, "b": 2, "c": 3}')
+
+        # make sure quotes are escaped properly
+        d = self.dtype(a='"', b="'", c="")
+        self.assertEqual(repr(d), '{"a": \'"\', "b": "\'", "c": \'\'}')
+
     def test_init_fails_on_nonexisting_attrs(self):
-        self.assertRaises(AttributeError, lambda: self.dtype(a=1, b=2, d=3))
-        
+        with self.assertRaises(AttributeError):
+            self.dtype(a=1, b=2, d=3)
+
     def test_eq(self):
         d = self.dtype(a=1, b=1, c=1)
         dd = self.dtype(a=1, b=1, c=1)
@@ -24,7 +34,7 @@ class TestStrictDict(unittest.TestCase):
         g = self.strict_dict_class(("a", "b", "c", "d"))(a=1, b=1, c=1, d=1)
         h = self.strict_dict_class(("a", "c", "b"))(a=1, b=1, c=1)
         i = self.strict_dict_class(("a", "c", "b"))(a=1, b=1, c=2)
-        
+
         self.assertEqual(d, dd)
         self.assertNotEqual(d, e)
         self.assertNotEqual(d, f)
@@ -37,20 +47,18 @@ class TestStrictDict(unittest.TestCase):
         d = self.dtype()
         d.a = 1
         self.assertEqual(d.a, 1)
-        self.assertRaises(AttributeError, lambda: d.b)
-    
+        self.assertRaises(AttributeError, getattr, d, 'b')
+
     def test_setattr_raises_on_nonexisting_attr(self):
         d = self.dtype()
-
-        def _f():
+        with self.assertRaises(AttributeError):
             d.x = 1
-        self.assertRaises(AttributeError, _f)
-    
+
     def test_setattr_getattr_special(self):
         d = self.strict_dict_class(["items"])
         d.items = 1
         self.assertEqual(d.items, 1)
-    
+
     def test_get(self):
         d = self.dtype(a=1)
         self.assertEqual(d.get('a'), 1)
@@ -88,7 +96,7 @@ class TestSemiSrictDict(TestStrictDict):
     def test_init_succeeds_with_nonexisting_attrs(self):
         d = self.dtype(a=1, b=1, c=1, x=2)
         self.assertEqual((d.a, d.b, d.c, d.x), (1, 1, 1, 2))
-   
+
     def test_iter_with_nonexisting_attrs(self):
         d = self.dtype(a=1, b=1, c=1, x=2)
         self.assertEqual(list(d), ['a', 'b', 'c', 'x'])

tests/test_dereference.py

@@ -1,6 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
 import unittest
 
 from bson import DBRef, ObjectId
@@ -32,7 +30,7 @@ class FieldTest(unittest.TestCase):
         User.drop_collection()
         Group.drop_collection()
 
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             user = User(name='user %s' % i)
             user.save()
 
@@ -90,7 +88,7 @@ class FieldTest(unittest.TestCase):
         User.drop_collection()
         Group.drop_collection()
 
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             user = User(name='user %s' % i)
             user.save()
 
@@ -162,7 +160,7 @@ class FieldTest(unittest.TestCase):
         User.drop_collection()
         Group.drop_collection()
 
-        for i in xrange(1, 26):
+        for i in range(1, 26):
             user = User(name='user %s' % i)
             user.save()
 
@@ -440,7 +438,7 @@ class FieldTest(unittest.TestCase):
         Group.drop_collection()
 
         members = []
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             a = UserA(name='User A %s' % i)
             a.save()
 
@@ -531,7 +529,7 @@ class FieldTest(unittest.TestCase):
         Group.drop_collection()
 
         members = []
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             a = UserA(name='User A %s' % i)
             a.save()
 
@@ -614,15 +612,15 @@ class FieldTest(unittest.TestCase):
         Group.drop_collection()
 
         members = []
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             user = User(name='user %s' % i)
             user.save()
             members.append(user)
 
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
 
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
 
         with query_counter() as q:
@@ -687,7 +685,7 @@ class FieldTest(unittest.TestCase):
         Group.drop_collection()
 
         members = []
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             a = UserA(name='User A %s' % i)
             a.save()
 
@@ -699,9 +697,9 @@ class FieldTest(unittest.TestCase):
 
             members += [a, b, c]
 
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
 
         with query_counter() as q:
@@ -783,16 +781,16 @@ class FieldTest(unittest.TestCase):
         Group.drop_collection()
 
         members = []
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             a = UserA(name='User A %s' % i)
             a.save()
 
             members += [a]
 
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
 
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
 
         with query_counter() as q:
@@ -866,7 +864,7 @@ class FieldTest(unittest.TestCase):
         Group.drop_collection()
 
         members = []
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             a = UserA(name='User A %s' % i)
             a.save()
 
@@ -878,9 +876,9 @@ class FieldTest(unittest.TestCase):
 
             members += [a, b, c]
 
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
-        group = Group(members=dict([(str(u.id), u) for u in members]))
+        group = Group(members={str(u.id): u for u in members})
         group.save()
 
         with query_counter() as q:
@@ -1103,7 +1101,7 @@ class FieldTest(unittest.TestCase):
         User.drop_collection()
         Group.drop_collection()
 
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             User(name='user %s' % i).save()
 
         Group(name="Test", members=User.objects).save()
@@ -1132,7 +1130,7 @@ class FieldTest(unittest.TestCase):
         User.drop_collection()
         Group.drop_collection()
 
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             User(name='user %s' % i).save()
 
         Group(name="Test", members=User.objects).save()
@@ -1169,7 +1167,7 @@ class FieldTest(unittest.TestCase):
         Group.drop_collection()
 
         members = []
-        for i in xrange(1, 51):
+        for i in range(1, 51):
             a = UserA(name='User A %s' % i).save()
             b = UserB(name='User B %s' % i).save()
             c = UserC(name='User C %s' % i).save()

tests/test_replicaset_connection.py

@@ -1,6 +1,3 @@
-import sys
-
-sys.path[0:0] = [""]
 import unittest
 
 from pymongo import ReadPreference
@@ -18,7 +15,7 @@ else:
 
 import mongoengine
 from mongoengine import *
-from mongoengine.connection import ConnectionError
+from mongoengine.connection import MongoEngineConnectionError
 
 
 class ConnectionTest(unittest.TestCase):
@@ -41,7 +38,7 @@ class ConnectionTest(unittest.TestCase):
             conn = connect(db='mongoenginetest',
                            host="mongodb://localhost/mongoenginetest?replicaSet=rs",
                            read_preference=READ_PREF)
-        except ConnectionError, e:
+        except MongoEngineConnectionError as e:
             return
 
         if not isinstance(conn, CONN_CLASS):

tests/test_signals.py

@@ -1,6 +1,4 @@
 # -*- coding: utf-8 -*-
-import sys
-sys.path[0:0] = [""]
 import unittest
 
 from mongoengine import *