skip uri test for pymongo < v2.9

Python 2.6 is not supported anymore with version 0.11.0

.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

@@ -1,41 +1,61 @@
 language: python
+
 python:
-- '2.6'
 - '2.7'
-- '3.2'
 - '3.3'
 - '3.4'
 - '3.5'
 - pypy
 - pypy3
+
 env:
 - PYMONGO=2.7
 - PYMONGO=2.8
 - PYMONGO=3.0
 - PYMONGO=dev
+
 matrix:
   fast_finish: true
+
 before_install:
 - travis_retry sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv 7F0CEB10
 - echo 'deb http://downloads-distro.mongodb.org/repo/ubuntu-upstart dist 10gen' |
   sudo tee /etc/apt/sources.list.d/mongodb.list
 - travis_retry sudo apt-get update
 - travis_retry sudo apt-get install mongodb-org-server
+
 install:
 - sudo apt-get install python-dev python3-dev libopenjpeg-dev zlib1g-dev libjpeg-turbo8-dev
   libtiff4-dev libjpeg8-dev libfreetype6-dev liblcms2-dev libwebp-dev tcl8.5-dev tk8.5-dev
   python-tk
-- travis_retry pip install tox>=1.9 coveralls
+- travis_retry pip install --upgrade pip
+- travis_retry pip install coveralls
+- travis_retry pip install flake8
+- travis_retry pip install tox>=1.9
+- travis_retry pip install "virtualenv<14.0.0"  # virtualenv>=14.0.0 has dropped Python 3.2 support (and pypy3 is based on py32)
 - travis_retry tox -e $(echo py$TRAVIS_PYTHON_VERSION-mg$PYMONGO | tr -d . | sed -e 's/pypypy/pypy/') -- -e test
+
+# Run flake8 for py27
+before_script:
+- if [[ $TRAVIS_PYTHON_VERSION == '2.7' ]]; then tox -e flake8; fi
+
 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 [[ $TRAVIS_PYTHON_VERSION == '2.7' ]]; then coveralls --verbose; fi
+
 notifications:
   irc: irc.freenode.org#mongoengine
+
 branches:
   only:
   - master
   - /^v.*$/
+
 deploy:
   provider: pypi
   user: the_drow

AUTHORS

@@ -228,5 +228,18 @@ that much better:
  * Vicki Donchenko (https://github.com/kivistein)
  * Emile Caron (https://github.com/emilecaron)
  * Amit Lichtenberg (https://github.com/amitlicht)
+ * Gang Li (https://github.com/iici-gli)
  * Lars Butler (https://github.com/larsbutler)
  * George Macon (https://github.com/gmacon)
+ * Ashley Whetter (https://github.com/AWhetter)
+ * Paul-Armand Verhaegen (https://github.com/paularmand)
+ * Steven Rossiter (https://github.com/BeardedSteve)
+ * Luo Peng (https://github.com/RussellLuo)
+ * Bryan Bennett (https://github.com/bbenne10)
+ * Gilb's Gilb's (https://github.com/gilbsgilbs)
+ * Joshua Nedrud (https://github.com/Neurostack)
+ * Shu Shen (https://github.com/shushen)
+ * xiaost7 (https://github.com/xiaost7)
+ * Victor Varvaryuk
+ * Stanislav Kaledin (https://github.com/sallyruthstruik)
+ * Dmitry Yantsen (https://github.com/mrTable)

CONTRIBUTING.rst

@@ -14,13 +14,13 @@ Before starting to write code, look for existing `tickets
 <https://github.com/MongoEngine/mongoengine/issues?state=open>`_ or `create one
 <https://github.com/MongoEngine/mongoengine/issues>`_ for your specific
 issue or feature request. That way you avoid working on something
-that might not be of interest or that has already been addressed.  If in doubt
+that might not be of interest or that has already been addressed. If in doubt
 post to the `user group <http://groups.google.com/group/mongoengine-users>`
 
 Supported Interpreters
 ----------------------
 
-MongoEngine supports CPython 2.6 and newer. Language
+MongoEngine supports CPython 2.7 and newer. Language
 features not supported by all interpreters can not be used.
 Please also ensure that your code is properly converted by
 `2to3 <http://docs.python.org/library/2to3.html>`_ for Python 3 support.

README.rst

@@ -4,25 +4,25 @@ 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://secure.travis-ci.org/MongoEngine/mongoengine.png?branch=master
-  :target: http://travis-ci.org/MongoEngine/mongoengine
+.. image:: https://travis-ci.org/MongoEngine/mongoengine.svg?branch=master
+  :target: https://travis-ci.org/MongoEngine/mongoengine
 
-.. image:: https://coveralls.io/repos/MongoEngine/mongoengine/badge.png?branch=master
-  :target: https://coveralls.io/r/MongoEngine/mongoengine?branch=master
+.. image:: https://coveralls.io/repos/github/MongoEngine/mongoengine/badge.svg?branch=master
+  :target: https://coveralls.io/github/MongoEngine/mongoengine?branch=master
 
-.. image:: https://landscape.io/github/MongoEngine/mongoengine/master/landscape.png
-   :target: https://landscape.io/github/MongoEngine/mongoengine/master
-   :alt: Code Health
+.. image:: https://landscape.io/github/MongoEngine/mongoengine/master/landscape.svg?style=flat
+  :target: https://landscape.io/github/MongoEngine/mongoengine/master
+  :alt: Code Health
 
 About
 =====
 MongoEngine is a Python Object-Document Mapper for working with MongoDB.
-Documentation available at http://mongoengine-odm.rtfd.org - there is currently
-a `tutorial <http://readthedocs.org/docs/mongoengine-odm/en/latest/tutorial.html>`_, a `user guide
-<https://mongoengine-odm.readthedocs.org/en/latest/guide/index.html>`_ and an `API reference
-<http://readthedocs.org/docs/mongoengine-odm/en/latest/apireference.html>`_.
+Documentation available at https://mongoengine-odm.readthedocs.io - there is currently
+a `tutorial <https://mongoengine-odm.readthedocs.io/tutorial.html>`_, a `user guide
+<https://mongoengine-odm.readthedocs.io/guide/index.html>`_ and an `API reference
+<https://mongoengine-odm.readthedocs.io/apireference.html>`_.
 
 Installation
 ============
@@ -35,25 +35,37 @@ setup.py install``.
 
 Dependencies
 ============
-- pymongo>=2.7.1
-- sphinx (optional - for documentation generation)
+All of the dependencies can easily be installed via `pip <https://pip.pypa.io/>`_. At the very least, you'll need these two packages to use MongoEngine:
+
+- pymongo>=2.7.1
+- six>=1.10.0
+
+If you utilize a ``DateTimeField``, you might also use a more flexible date parser:
 
-Optional Dependencies
----------------------
-- **Image Fields**: Pillow>=2.0.0
 - dateutil>=2.1.0
 
-.. note
-   MongoEngine always runs it's test suite against the latest patch version of each dependecy. e.g.: PyMongo 3.0.1
+If you need to use an ``ImageField`` or ``ImageGridFsProxy``:
+
+- Pillow>=2.0.0
+
+If you want to generate the documentation (e.g. to contribute to it):
+
+- sphinx
 
 Examples
 ========
-Some simple examples of what MongoEngine code looks like::
+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)
+        posted = DateTimeField(default=datetime.datetime.utcnow)
         tags = ListField(StringField(max_length=50))
+        meta = {'allow_inheritance': True}
 
     class TextPost(BlogPost):
         content = StringField(required=True)
@@ -81,23 +93,24 @@ Some simple examples of what MongoEngine code looks like::
     ...     print
     ...
 
-    >>> len(BlogPost.objects)
+    # Count all blog posts and its subtypes
+    >>> BlogPost.objects.count()
     2
-    >>> len(TextPost.objects)
+    >>> TextPost.objects.count()
     1
-    >>> len(LinkPost.objects)
+    >>> LinkPost.objects.count()
     1
 
-    # Find tagged posts
-    >>> len(BlogPost.objects(tags='mongoengine'))
+    # Count tagged posts
+    >>> BlogPost.objects(tags='mongoengine').count()
     2
-    >>> len(BlogPost.objects(tags='mongodb'))
+    >>> BlogPost.objects(tags='mongodb').count()
     1
 
 Tests
 =====
 To run the test suite, ensure you are running a local instance of MongoDB on
-the standard port, and run: ``python setup.py nosetests``.
+the standard port and have ``nose`` installed. Then, run: ``python setup.py nosetests``.
 
 To run the test suite on every supported Python version and every supported PyMongo version,
 you can use ``tox``.
@@ -124,8 +137,7 @@ Community
   <http://groups.google.com/group/mongoengine-users>`_
 - `MongoEngine Developers mailing list
   <http://groups.google.com/group/mongoengine-dev>`_
-- `#mongoengine IRC channel <http://webchat.freenode.net/?channels=mongoengine>`_
 
 Contributing
 ============
-We welcome contributions! see  the `Contribution guidelines <https://github.com/MongoEngine/mongoengine/blob/master/CONTRIBUTING.rst>`_
+We welcome contributions! See the `Contribution guidelines <https://github.com/MongoEngine/mongoengine/blob/master/CONTRIBUTING.rst>`_

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

@@ -2,19 +2,98 @@
 Changelog
 =========
 
+Development
+===========
+- (Fill this out as you fix issues and develop your features).
+- Fixed using sets in field choices #1481
+- POTENTIAL BREAKING CHANGE: Fixed limit/skip/hint/batch_size chaining #1476
+- POTENTIAL BREAKING CHANGE: Changed a public `QuerySet.clone_into` method to a private `QuerySet._clone_into` #1476
+- Fixed connecting to a replica set with PyMongo 2.x #1436
+- Fixed an obscure error message when filtering by `field__in=non_iterable`. #1237
+
+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
+- BREAKING CHANGE: Accessing a broken reference will raise a `DoesNotExist` error. In the past it used to return `None`. #1334
+- Fixed absent rounding for DecimalField when `force_string` is set. #1103
+
+Changes in 0.10.8
+=================
+- 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 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.
+- count on ListField of EmbeddedDocumentField fails. #1187
+- Fixed long fields stored as int32 in Python 3. #1253
+- MapField now handles unicodes keys correctly. #1267
+- ListField now handles negative indicies correctly. #1270
+- Fixed AttributeError when initializing EmbeddedDocument with positional args. #681
+- Fixed no_cursor_timeout error with pymongo 3.0+ #1304
+- Replaced map-reduce based QuerySet.sum/average with aggregation-based implementations #1336
+- Fixed support for `__` to escape field names that match operators names in `update` #1351
+- Fixed BaseDocument#_mark_as_changed #1369
+- Added support for pickling QuerySet instances. #1397
+- 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
+
+Changes in 0.10.6
+=================
+- Add support for mocking MongoEngine based on mongomock. #1151
+- Fixed not being able to run tests on Windows. #1153
+- Allow creation of sparse compound indexes. #1114
+- count on ListField of EmbeddedDocumentField fails. #1187
+
+Changes in 0.10.5
+=================
+- Fix for reloading of strict with special fields. #1156
+
+Changes in 0.10.4
+=================
+- SaveConditionError is now importable from the top level package. #1165
+- upsert_one method added. #1157
+
+Changes in 0.10.3
+=================
+- Fix `read_preference` (it had chaining issues with PyMongo 2.x and it didn't work at all with PyMongo 3.x) #1042
+
 Changes in 0.10.2
 =================
 - Allow shard key to point to a field in an embedded document. #551
 - Allow arbirary metadata in fields. #1129
+- ReferenceFields now support abstract document types. #837
 
 Changes in 0.10.1
-=======================
+=================
 - Fix infinite recursion with CASCADE delete rules under specific conditions. #1046
 - Fix CachedReferenceField bug when loading cached docs as DBRef but failing to save them. #1047
 - Fix ignored chained options #842
 - Document save's save_condition error raises `SaveConditionError` exception #1070
 - Fix Document.reload for DynamicDocument. #1050
 - StrictDict & SemiStrictDict are shadowed at init time. #1105
+- Fix ListField minus index assignment does not work. #1119
+- Remove code that marks field as changed when the field has default but not existed in database #1126
 - Remove test dependencies (nose and rednose) from install dependencies list. #1079
 - Recursively build query when using elemMatch operator. #1130
 - Fix instance back references for lists of embedded documents. #1131

docs/code/tumblelog.py

@@ -17,6 +17,10 @@ class Post(Document):
     tags = ListField(StringField(max_length=30))
     comments = ListField(EmbeddedDocumentField(Comment))
 
+    # bugfix
+    meta = {'allow_inheritance': True}
+
+
 class TextPost(Post):
     content = StringField()
 
@@ -45,7 +49,8 @@ print 'ALL POSTS'
 print
 for post in Post.objects:
     print post.title
-    print '=' * post.title.count()
+    #print '=' * post.title.count()
+    print "=" * 20
 
     if isinstance(post, TextPost):
         print post.content

docs/guide/connecting.rst

@@ -33,7 +33,7 @@ the :attr:`host` to
     corresponding parameters in :func:`~mongoengine.connect`: ::
 
         connect(
-            name='test',
+            db='test',
             username='user',
             password='12345',
             host='mongodb://admin:qwerty@localhost/production'

docs/guide/defining-documents.rst

@@ -29,7 +29,7 @@ documents are serialized based on their field order.
 
 Dynamic document schemas
 ========================
-One of the benefits of MongoDb is dynamic schemas for a collection, whilst data
+One of the benefits of MongoDB is dynamic schemas for a collection, whilst data
 should be planned and organised (after all explicit is better than implicit!)
 there are scenarios where having dynamic / expando style documents is desirable.
 
@@ -75,6 +75,7 @@ are as follows:
 * :class:`~mongoengine.fields.DynamicField`
 * :class:`~mongoengine.fields.EmailField`
 * :class:`~mongoengine.fields.EmbeddedDocumentField`
+* :class:`~mongoengine.fields.EmbeddedDocumentListField`
 * :class:`~mongoengine.fields.FileField`
 * :class:`~mongoengine.fields.FloatField`
 * :class:`~mongoengine.fields.GenericEmbeddedDocumentField`
@@ -149,7 +150,7 @@ arguments can be set on all fields:
     .. note:: If set, this field is also accessible through the `pk` field.
 
 :attr:`choices` (Default: None)
-    An iterable (e.g. a list or tuple) of choices to which the value of this
+    An iterable (e.g. list, tuple or set) of choices to which the value of this
     field should be limited.
 
     Can be either be a nested tuples of value (stored in mongo) and a
@@ -213,9 +214,9 @@ document class as the first argument::
 
 Dictionary Fields
 -----------------
-Often, an embedded document may be used instead of a dictionary -- generally
-this is recommended as dictionaries don't support validation or custom field
-types. However, sometimes you will not know the structure of what you want to
+Often, an embedded document may be used instead of a dictionary – generally
+embedded documents are recommended as dictionaries don’t support validation
+or custom field types. However, sometimes you will not know the structure of what you want to
 store; in this situation a :class:`~mongoengine.fields.DictField` is appropriate::
 
     class SurveyResponse(Document):
@@ -360,11 +361,6 @@ Its value can take any of the following constants:
    In Django, be sure to put all apps that have such delete rule declarations in
    their :file:`models.py` in the :const:`INSTALLED_APPS` tuple.
 
-
-.. warning::
-   Signals are not triggered when doing cascading updates / deletes - if this
-   is required you must manually handle the update / delete.
-
 Generic reference fields
 ''''''''''''''''''''''''
 A second kind of reference field also exists,

docs/guide/index.rst

@@ -13,3 +13,4 @@ User Guide
    gridfs
    signals
    text-indexes
+   mongomock

docs/guide/mongomock.rst

@@ -0,0 +1,21 @@
+==============================
+Use mongomock for testing
+==============================
+
+`mongomock <https://github.com/vmalloc/mongomock/>`_ is a package to do just 
+what the name implies, mocking a mongo database.
+
+To use with mongoengine, simply specify mongomock when connecting with 
+mongoengine:
+
+.. code-block:: python
+
+    connect('mongoenginetest', host='mongomock://localhost')
+    conn = get_connection()
+
+or with an alias:
+
+.. code-block:: python
+
+    connect('mongoenginetest', host='mongomock://localhost', alias='testdb')
+    conn = get_connection('testdb')

docs/guide/querying.rst

@@ -237,7 +237,7 @@ is preferred for achieving this::
     # All except for the first 5 people
     users = User.objects[5:]
 
-    # 5 users, starting from the 10th user found
+    # 5 users, starting from the 11th user found
     users = User.objects[10:15]
 
 You may also index the query to retrieve a single result. If an item at that
@@ -479,6 +479,8 @@ operators. To use a :class:`~mongoengine.queryset.Q` object, pass it in as the
 first positional argument to :attr:`Document.objects` when you filter it by
 calling it with keyword arguments::
 
+    from mongoengine.queryset.visitor import Q
+
     # Get published posts
     Post.objects(Q(published=True) | Q(publish_date__lte=datetime.now()))
 

docs/guide/signals.rst

@@ -142,11 +142,4 @@ cleaner looking while still allowing manual execution of the callback::
         modified = DateTimeField()
 
 
-ReferenceFields and Signals
----------------------------
-
-Currently `reverse_delete_rule` does not trigger signals on the other part of
-the relationship.  If this is required you must manually handle the
-reverse deletion.
-
 .. _blinker: http://pypi.python.org/pypi/blinker

docs/upgrade.rst

@@ -2,6 +2,53 @@
 Upgrading
 #########
 
+Development
+***********
+(Fill this out whenever you introduce breaking changes to MongoEngine)
+
+This release includes various fixes for the `BaseQuerySet` methods and how they
+are chained together. Since version 0.10.1 applying limit/skip/hint/batch_size
+to an already-existing queryset wouldn't modify the underlying PyMongo cursor.
+This has been fixed now, so you'll need to make sure that your code didn't rely
+on the broken implementation.
+
+Additionally, a public `BaseQuerySet.clone_into` has been renamed to a private
+`_clone_into`. If you directly used that method in your code, you'll need to
+rename its occurrences.
+
+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 +1,36 @@
-import document
-from document import *
-import fields
-from fields import *
-import connection
-from connection import *
-import queryset
-from queryset import *
-import signals
-from signals import *
-from errors import *
-import errors
+# 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, 1)
+
+__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))
 
+
 __version__ = get_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', 'rename'])
 
-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 weakref
 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
@@ -137,10 +138,7 @@ class BaseList(list):
         return super(BaseList, self).__setitem__(key, value)
 
     def __delitem__(self, key, *args, **kwargs):
-        if isinstance(key, slice):
-            self._mark_as_changed()
-        else:
-            self._mark_as_changed(key)
+        self._mark_as_changed()
         return super(BaseList, self).__delitem__(key)
 
     def __setslice__(self, *args, **kwargs):
@@ -199,7 +197,9 @@ class BaseList(list):
     def _mark_as_changed(self, key=None):
         if hasattr(self._instance, '_mark_as_changed'):
             if key:
-                self._instance._mark_as_changed('%s.%s' % (self._name, key))
+                self._instance._mark_as_changed(
+                    '%s.%s' % (self._name, key % len(self))
+                )
             else:
                 self._instance._mark_as_changed(self._name)
 
@@ -207,17 +207,22 @@ class BaseList(list):
 class EmbeddedDocumentList(BaseList):
 
     @classmethod
-    def __match_all(cls, i, kwargs):
-        items = kwargs.items()
-        return all([
-            getattr(i, k) == v or str(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)
@@ -283,18 +288,18 @@ 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]
@@ -424,7 +429,7 @@ class StrictDict(object):
     def __eq__(self, other):
         return self.items() == other.items()
 
-    def __neq__(self, other):
+    def __ne__(self, other):
         return self.items() != other.items()
 
     @classmethod
@@ -436,7 +441,7 @@ 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

@@ -1,37 +1,33 @@
 import copy
-import operator
 import numbers
 from collections import Hashable
 from functools import partial
 
-import pymongo
-from bson import json_util, ObjectId
+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.common import _import_class
-from mongoengine.errors import (ValidationError, InvalidDocumentError,
-                                LookUpError, FieldDoesNotExist)
-from mongoengine.python_support import PY3, txt_type
-from mongoengine.base.common import get_document, ALLOW_INHERITANCE
-from mongoengine.base.datastructures import (
-    BaseDict,
-    BaseList,
-    EmbeddedDocumentList,
-    StrictDict,
-    SemiStrictDict
-)
+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, 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
@@ -51,33 +47,34 @@ class BaseDocument(object):
             # We only want named arguments.
             field = iter(self._fields_ordered)
             # If its an automatic id field then skip to the first defined field
-            if self._auto_id_field:
+            if getattr(self, '_auto_id_field', False):
                 next(field)
             for value in args:
                 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)
 
         # Check if there are undefined fields supplied to the constructor,
         # if so raise an Exception.
         if not self._dynamic and (self._meta.get('strict', True) or _created):
-            for var in values.keys():
-                if var not in self._fields.keys() + ['id', 'pk', '_cls', '_text_score']:
-                    msg = (
-                        "The field '{0}' does not exist on the document '{1}'"
-                    ).format(var, self._class_name)
-                    raise FieldDoesNotExist(msg)
+            _undefined_fields = set(values.keys()) - set(
+                self._fields.keys() + ['id', 'pk', '_cls', '_text_score'])
+            if _undefined_fields:
+                msg = (
+                    'The fields "{0}" do not exist on the document "{1}"'
+                ).format(_undefined_fields, self._class_name)
+                raise FieldDoesNotExist(msg)
 
         if self.STRICT and not self._dynamic:
             self._data = StrictDict.create(allowed_keys=self._fields_ordered)()
@@ -94,7 +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
@@ -120,7 +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:
@@ -149,7 +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
@@ -168,11 +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:
@@ -196,8 +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:
@@ -211,7 +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)
@@ -253,12 +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:
@@ -307,9 +307,9 @@ class BaseDocument(object):
             fields = []
 
         data = SON()
-        data["_id"] = None
+        data['_id'] = None
         data['_cls'] = self._class_name
-        EmbeddedDocumentField = _import_class("EmbeddedDocumentField")
+
         # only root fields ['test1.a', 'test2'] => ['test1', 'test2']
         root_fields = set([f.split('.')[0] for f in fields])
 
@@ -324,21 +324,20 @@ class BaseDocument(object):
                 field = self._dynamic_fields.get(field_name)
 
             if value is not None:
+                f_inputs = field.to_mongo.__code__.co_varnames
+                ex_vars = {}
+                if fields and 'fields' in f_inputs:
+                    key = '%s.' % field_name
+                    embedded_fields = [
+                        i.replace(key, '') for i in fields
+                        if i.startswith(key)]
 
-                if isinstance(field, EmbeddedDocumentField):
-                    if fields:
-                        key = '%s.' % field_name
-                        embedded_fields = [
-                            i.replace(key, '') for i in fields
-                            if i.startswith(key)]
+                    ex_vars['fields'] = embedded_fields
 
-                    else:
-                        embedded_fields = []
+                if 'use_db_field' in f_inputs:
+                    ex_vars['use_db_field'] = use_db_field
 
-                    value = field.to_mongo(value, use_db_field=use_db_field,
-                                           fields=embedded_fields)
-                else:
-                    value = field.to_mongo(value)
+                value = field.to_mongo(value, **ex_vars)
 
             # Handle self generating fields
             if value is None and field._auto_gen:
@@ -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:
@@ -395,27 +384,29 @@ 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):
-        """Converts a document to JSON.
-        :param use_db_field: Set to True by default but enables the output of the json structure with the field names
-            and not the mongodb store db_names in case of set to False
+        """Convert this document to JSON.
+
+        :param use_db_field: Serialize field names as they appear in
+            MongoDB (as opposed to attribute names on this document).
+            Defaults to True.
         """
         use_db_field = kwargs.pop('use_db_field', True)
         return json_util.dumps(self.to_mongo(use_db_field), *args, **kwargs)
@@ -426,33 +417,26 @@ 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 +449,7 @@ 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
 
@@ -491,15 +474,16 @@ class BaseDocument(object):
                 # remove lower level changed fields
                 level = '.'.join(levels[:idx]) + '.'
                 remove = self._changed_fields.remove
-                for field in self._changed_fields:
+                for field in self._changed_fields[:]:
                     if field.startswith(level):
                         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 +495,13 @@ 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 +513,27 @@ 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', [])
 
@@ -566,11 +554,13 @@ class BaseDocument(object):
                     continue
             if isinstance(field, ReferenceField):
                 continue
-            elif (isinstance(data, (EmbeddedDocument, DynamicEmbeddedDocument))
-                  and db_field_name not in changed_fields):
+            elif (
+                isinstance(data, (EmbeddedDocument, DynamicEmbeddedDocument)) and
+                db_field_name not in changed_fields
+            ):
                 # 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
@@ -606,7 +596,9 @@ class BaseDocument(object):
                 for p in parts:
                     if isinstance(d, (ObjectId, DBRef)):
                         break
-                    elif isinstance(d, list) and p.isdigit():
+                    elif isinstance(d, list) and p.lstrip('-').isdigit():
+                        if p[0] == '-':
+                            p = str(len(d) + int(p))
                         try:
                             d = d[int(p)]
                         except IndexError:
@@ -640,7 +632,9 @@ class BaseDocument(object):
                 parts = path.split('.')
                 db_field_name = parts.pop()
                 for p in parts:
-                    if isinstance(d, list) and p.isdigit():
+                    if isinstance(d, list) and p.lstrip('-').isdigit():
+                        if p[0] == '-':
+                            p = str(len(d) + int(p))
                         d = d[int(p)]
                     elif (hasattr(d, '__getattribute__') and
                           not isinstance(d, dict)):
@@ -670,21 +664,28 @@ 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
+        if son and not isinstance(son, dict):
+            raise ValueError("The source SON object needs to be of type 'dict'")
+
+        # 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:
@@ -706,27 +707,20 @@ 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
-            elif field.default:
-                default = field.default
-                if callable(default):
-                    default = default()
-                if isinstance(default, BaseDocument):
-                    changed_fields.append(field_name)
-                elif not only_fields or field_name in only_fields:
-                    changed_fields.append(field_name)
 
         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:
@@ -736,37 +730,43 @@ 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)}
@@ -777,14 +777,17 @@ class BaseDocument(object):
         direction = None
 
         # Check to see if we need to include _cls
-        allow_inheritance = cls._meta.get('allow_inheritance',
-                                          ALLOW_INHERITANCE)
-        include_cls = (allow_inheritance and not spec.get('sparse', False) and
-                       spec.get('cls',  True) and '_cls' not in spec['fields'])
+        allow_inheritance = cls._meta.get('allow_inheritance')
+        include_cls = (
+            allow_inheritance and
+            not spec.get('sparse', False) and
+            spec.get('cls', True) and
+            '_cls' not in spec['fields']
+        )
 
         # 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
@@ -799,19 +802,19 @@ 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
@@ -824,7 +827,7 @@ class BaseDocument(object):
                 parts = []
                 for field in fields:
                     try:
-                        if field != "_id":
+                        if field != '_id':
                             field = field.db_field
                     except AttributeError:
                         pass
@@ -839,57 +842,57 @@ class BaseDocument(object):
 
         if index_list:
             spec['fields'] = index_list
-        if spec.get('sparse', False) and len(spec['fields']) > 1:
-            raise ValueError(
-                'Sparse indexes can only have one field in them. '
-                'See https://jira.mongodb.org/browse/SERVER-2193')
 
         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)
 
@@ -901,8 +904,9 @@ 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])
@@ -910,32 +914,68 @@ 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
 
@@ -945,16 +985,17 @@ 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:
@@ -967,35 +1008,55 @@ 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'), '_dynamic')):
+                                       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
@@ -1007,19 +1068,18 @@ 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 UPDATE_OPERATORS
+from mongoengine.base.datastructures import (BaseDict, BaseList,
+                                             EmbeddedDocumentList)
 from mongoengine.common import _import_class
 from mongoengine.errors import ValidationError
-from mongoengine.base.common import ALLOW_INHERITANCE
-from mongoengine.base.datastructures import (
-    BaseDict, BaseList, EmbeddedDocumentList
-)
-
-__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
@@ -46,7 +41,7 @@ class BaseField(object):
         """
         :param db_field: The database field to store this field in
             (defaults to the name of the field)
-        :param name: Depreciated - use db_field
+        :param name: Deprecated - use db_field
         :param required: If the field is required. Whether it has to have a
             value or not. Defaults to False.
         :param default: (optional) The default value for this field if no 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
@@ -85,13 +80,24 @@ class BaseField(object):
         self.null = null
         self.sparse = sparse
         self._owner_document = None
-        
+
+        # Validate the db_field
+        if isinstance(self.db_field, six.string_types) and (
+            '.' in self.db_field or
+            '\0' in self.db_field or
+            self.db_field.startswith('$')
+        ):
+            raise ValueError(
+                'field names cannot contain dots (".") or null characters '
+                '("\\0"), and they must not start with a dollar sign ("$").'
+            )
+
         # 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" % (
-                self.__class__.__name__, ', '.join(conflicts) ))
-        
+            raise TypeError('%s already has attribute(s): %s' % (
+                self.__class__.__name__, ', '.join(conflicts)))
+
         # Assign metadata to the instance
         # This efficient method is available because no __slots__ are defined.
         self.__dict__.update(kwargs)
@@ -133,7 +139,7 @@ class BaseField(object):
                 if (self.name not in instance._data or
                         instance._data[self.name] != value):
                     instance._mark_as_changed(self.name)
-            except:
+            except Exception:
                 # Values cant be compared eg: naive and tz datetimes
                 # So mark it as changed
                 instance._mark_as_changed(self.name)
@@ -147,32 +153,39 @@ 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):
+        """Helper method to call to_mongo with proper inputs."""
+        f_inputs = self.to_mongo.__code__.co_varnames
+        ex_vars = {}
+        if 'fields' in f_inputs:
+            ex_vars['fields'] = fields
+
+        if 'use_db_field' in f_inputs:
+            ex_vars['use_db_field'] = use_db_field
+
+        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):
@@ -180,19 +193,21 @@ class BaseField(object):
         EmbeddedDocument = _import_class('EmbeddedDocument')
 
         choice_list = self.choices
-        if isinstance(choice_list[0], (list, tuple)):
+        if isinstance(next(iter(choice_list)), (list, tuple)):
+            # next(iter) is useful for sets
             choice_list = [k for k, _ in choice_list]
 
         # Choices which are other types of Documents
         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
@@ -235,8 +250,7 @@ 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
@@ -248,7 +262,7 @@ 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):
@@ -283,11 +297,8 @@ class ComplexBaseField(BaseField):
         return value
 
     def to_python(self, value):
-        """Convert a MongoDB-compatible type to a Python type.
-        """
-        Document = _import_class('Document')
-
-        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'):
@@ -297,15 +308,16 @@ 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 = {}
             for k, v in value.items():
                 if isinstance(v, Document):
@@ -325,21 +337,20 @@ class ComplexBaseField(BaseField):
                                          key=operator.itemgetter(0))]
         return value_dict
 
-    def to_mongo(self, value):
-        """Convert a Python type to a MongoDB-compatible type.
-        """
-        Document = _import_class("Document")
-        EmbeddedDocument = _import_class("EmbeddedDocument")
-        GenericReferenceField = _import_class("GenericReferenceField")
+    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')
 
-        if isinstance(value, basestring):
+        if isinstance(value, six.string_types):
             return value
 
         if hasattr(value, 'to_mongo'):
             if isinstance(value, Document):
                 return GenericReferenceField().to_mongo(value)
             cls = value.__class__
-            val = value.to_mongo()
+            val = value.to_mongo(use_db_field, fields)
             # If it's a document that is not inherited add _cls
             if isinstance(value, EmbeddedDocument):
                 val['_cls'] = cls.__name__
@@ -349,13 +360,15 @@ 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(item))
-                               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():
@@ -369,9 +382,7 @@ 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:
@@ -379,13 +390,13 @@ class ComplexBaseField(BaseField):
                         value_dict[k] = DBRef(collection, v.pk)
                 elif hasattr(v, 'to_mongo'):
                     cls = v.__class__
-                    val = v.to_mongo()
+                    val = v.to_mongo(use_db_field, fields)
                     # If it's a document that is not inherited add _cls
                     if isinstance(v, (Document, EmbeddedDocument)):
                         val['_cls'] = cls.__name__
                     value_dict[k] = val
                 else:
-                    value_dict[k] = self.to_mongo(v)
+                    value_dict[k] = self.to_mongo(v, use_db_field, fields)
 
         if is_list:  # Convert back to a list
             return [v for _, v in sorted(value_dict.items(),
@@ -393,8 +404,7 @@ 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'):
@@ -404,9 +414,9 @@ 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:
@@ -432,24 +442,23 @@ 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:
             if not isinstance(value, ObjectId):
                 value = ObjectId(value)
-        except:
+        except Exception:
             pass
         return value
 
     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):
@@ -457,8 +466,8 @@ class ObjectIdField(BaseField):
 
     def validate(self, value):
         try:
-            ObjectId(unicode(value))
-        except:
+            ObjectId(six.text_type(value))
+        except Exception:
             self.error('Invalid Object ID')
 
 
@@ -469,21 +478,20 @@ 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:
@@ -498,7 +506,7 @@ 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)
@@ -510,8 +518,8 @@ class GeoJsonBaseField(BaseField):
         # Quick and dirty validator
         try:
             value[0][0][0]
-        except:
-            return "Invalid Polygon must contain at least one valid linestring"
+        except (TypeError, IndexError):
+            return 'Invalid Polygon must contain at least one valid linestring'
 
         errors = []
         for val in value:
@@ -522,20 +530,20 @@ 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'
 
         # Quick and dirty validator
         try:
             value[0][0]
-        except:
-            return "Invalid LineString must contain at least one valid point"
+        except (TypeError, IndexError):
+            return 'Invalid LineString must contain at least one valid point'
 
         errors = []
         for val in value:
@@ -544,19 +552,19 @@ 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)):
@@ -565,8 +573,8 @@ class GeoJsonBaseField(BaseField):
         # Quick and dirty validator
         try:
             value[0][0]
-        except:
-            return "Invalid MultiPoint must contain at least one valid point"
+        except (TypeError, IndexError):
+            return 'Invalid MultiPoint must contain at least one valid point'
 
         errors = []
         for point in value:
@@ -575,7 +583,7 @@ 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)):
@@ -584,8 +592,8 @@ class GeoJsonBaseField(BaseField):
         # Quick and dirty validator
         try:
             value[0][0][0]
-        except:
-            return "Invalid MultiLineString must contain at least one valid linestring"
+        except (TypeError, IndexError):
+            return 'Invalid MultiLineString must contain at least one valid linestring'
 
         errors = []
         for linestring in value:
@@ -595,9 +603,9 @@ 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)):
@@ -606,8 +614,8 @@ class GeoJsonBaseField(BaseField):
         # Quick and dirty validator
         try:
             value[0][0][0][0]
-        except:
-            return "Invalid MultiPolygon must contain at least one valid Polygon"
+        except (TypeError, IndexError):
+            return 'Invalid MultiPolygon must contain at least one valid Polygon'
 
         errors = []
         for polygon in value:
@@ -616,9 +624,9 @@ 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

@@ -1,22 +1,23 @@
 import warnings
 
+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)
 
-from mongoengine.base.common import _document_registry, ALLOW_INHERITANCE
-from mongoengine.base.fields import BaseField, ComplexBaseField, ObjectIdField
 
 __all__ = ('DocumentMetaclass', 'TopLevelDocumentMetaclass')
 
 
 class DocumentMetaclass(type):
-    """Metaclass for all documents.
-    """
+    """Metaclass for all documents."""
 
+    # TODO lower complexity of this method
     def __new__(cls, name, bases, attrs):
         flattened_bases = cls._get_bases(bases)
         super_new = super(DocumentMetaclass, cls).__new__
@@ -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__)
 
@@ -161,8 +162,8 @@ 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:
-            for key, val in new_class.__dict__.items():
+        if six.PY3:
+            for val in new_class.__dict__.values():
                 if isinstance(val, classmethod):
                     f = val.__get__(new_class)
                     if hasattr(f, '__func__') and not hasattr(f, 'im_func'):
@@ -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/common.py

@@ -34,7 +34,10 @@ def _import_class(cls_name):
     queryset_classes = ('OperationError',)
     deref_classes = ('DeReference',)
 
-    if cls_name in doc_classes:
+    if cls_name == 'BaseDocument':
+        from mongoengine.base import document as module
+        import_classes = ['BaseDocument']
+    elif cls_name in doc_classes:
         from mongoengine import document as module
         import_classes = doc_classes
     elif cls_name in field_classes:

mongoengine/connection.py

@@ -1,11 +1,14 @@
 from pymongo import MongoClient, ReadPreference, uri_parser
+import six
+
 from mongoengine.python_support import IS_PYMONGO_3
 
-__all__ = ['ConnectionError', 'connect', 'register_connection',
+__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.
 
@@ -38,11 +46,17 @@ 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 kwargs: allow ad-hoc parameters to be passed into the pymongo driver
+    :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: ad-hoc parameters to be passed into the pymongo driver,
+        for example maxpoolsize, tz_aware, etc. See the documentation
+        for pymongo's `MongoClient` for a full list.
 
+    .. versionchanged:: 0.10.6 - added mongomock support
     """
-    global _connection_settings
-
     conn_settings = {
         'name': name or 'test',
         'host': host or 'localhost',
@@ -50,23 +64,48 @@ 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
     }
 
-    # Handle uri style connections
-    if "://" in conn_settings['host']:
-        uri_dict = uri_parser.parse_uri(conn_settings['host'])
-        conn_settings.update({
-            'name': uri_dict.get('database') or name,
-            'username': uri_dict.get('username'),
-            'password': uri_dict.get('password'),
-            'read_preference': read_preference,
-        })
-        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']
+    conn_host = conn_settings['host']
+
+    # 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 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)
+
+            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'] = uri_options['replicaset']
+            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
 
     # Deprecated parameters that should not be passed on
     kwargs.pop('slaves', None)
@@ -77,9 +116,7 @@ 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]
@@ -88,56 +125,99 @@ 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:
+    # If the requested alias already exists in the _connections list, return
+    # it immediately.
+    if alias in _connections:
+        return _connections[alias]
+
+    # 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:
             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()
+        raise MongoEngineConnectionError(msg)
 
-        conn_settings.pop('name', None)
-        conn_settings.pop('username', None)
-        conn_settings.pop('password', None)
-        conn_settings.pop('authentication_source', None)
+    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
-        if 'replicaSet' in conn_settings:
+
+        # 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'])
+
             # 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)
-                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)
 
@@ -145,11 +225,13 @@ 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]
 
@@ -161,12 +243,14 @@ def connect(db=None, alias=DEFAULT_CONNECTION_NAME, **kwargs):
     running on the default port on localhost. If authentication is needed,
     provide username and password arguments as well.
 
-    Multiple databases are supported by using aliases.  Provide a separate
+    Multiple databases are supported by using aliases. Provide a separate
     `alias` to connect to a different instance of :program:`mongod`.
 
+    See the docstring for `register_connection` for more details about all
+    supported 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

@@ -1,13 +1,12 @@
 from bson import DBRef, SON
+import six
 
-from base import (
-    BaseDict, BaseList, EmbeddedDocumentList,
-    TopLevelDocumentMetaclass, get_document
-)
-from fields import (ReferenceField, ListField, DictField, MapField)
-from connection import get_db
-from queryset import QuerySet
-from document import Document, EmbeddedDocument
+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):
@@ -24,7 +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
@@ -67,11 +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)
@@ -89,14 +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)
@@ -150,7 +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()
@@ -217,7 +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(
@@ -226,7 +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 = "{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

@@ -1,28 +1,23 @@
-import warnings
-import pymongo
 import re
+import warnings
 
-from pymongo.read_preferences import ReadPreference
 from bson.dbref import DBRef
+import pymongo
+from pymongo.read_preferences import ReadPreference
+import six
+
 from mongoengine import signals
+from mongoengine.base import (BaseDict, BaseDocument, BaseList,
+                              DocumentMetaclass, EmbeddedDocumentList,
+                              TopLevelDocumentMetaclass, get_document)
 from mongoengine.common import _import_class
-from mongoengine.base import (
-    DocumentMetaclass,
-    TopLevelDocumentMetaclass,
-    BaseDocument,
-    BaseDict,
-    BaseList,
-    EmbeddedDocumentList,
-    ALLOW_INHERITANCE,
-    get_document
-)
-from mongoengine.errors import (InvalidQueryError, InvalidDocumentError,
+from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
+from mongoengine.context_managers import switch_collection, switch_db
+from mongoengine.errors import (InvalidDocumentError, InvalidQueryError,
                                 SaveConditionError)
 from mongoengine.python_support import IS_PYMONGO_3
-from mongoengine.queryset import (OperationError, NotUniqueError,
+from mongoengine.queryset import (NotUniqueError, OperationError,
                                   QuerySet, transform)
-from mongoengine.connection import get_db, DEFAULT_CONNECTION_NAME
-from mongoengine.context_managers import switch_db, switch_collection
 
 __all__ = ('Document', 'EmbeddedDocument', 'DynamicDocument',
            'DynamicEmbeddedDocument', 'OperationError',
@@ -30,12 +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]
@@ -56,9 +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', )
@@ -81,6 +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)
 
@@ -105,9 +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`
@@ -148,26 +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):
@@ -210,31 +206,46 @@ 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.
 
         Returns True if the document has been updated or False if the document
         in the database doesn't match the query.
 
-        .. note:: All unsaved changes that has been made to the document are
+        .. note:: All unsaved changes that have been made to the document are
             rejected if the method returns True.
 
         :param query: the update will be performed only if the document in the
             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:
@@ -250,7 +261,7 @@ class Document(BaseDocument):
 
     def save(self, force_insert=False, validate=True, clean=True,
              write_concern=None, cascade=None, cascade_kwargs=None,
-             _refs=None, save_condition=None, **kwargs):
+             _refs=None, save_condition=None, signal_kwargs=None, **kwargs):
         """Save the :class:`~mongoengine.Document` to the database. If the
         document already exists, it will be updated, otherwise it will be
         created.
@@ -276,6 +287,8 @@ class Document(BaseDocument):
         :param save_condition: only perform save if matching record in db
             satisfies condition(s) (e.g. version number).
             Raises :class:`OperationError` if the conditions are not satisfied
+        :parm signal_kwargs: (optional) kwargs dictionary to be passed to
+            the signal calls.
 
         .. versionchanged:: 0.5
             In existing documents it only saves changed fields using
@@ -297,118 +310,157 @@ class Document(BaseDocument):
             :class:`OperationError` exception raised if save_condition fails.
         .. versionchanged:: 0.10.1
             :class: save_condition failure now raises a `SaveConditionError`
+        .. versionchanged:: 0.10.7
+            Add signal_kwargs argument
         """
-        signals.pre_save.send(self.__class__, document=self)
+        if self._meta.get('abstract'):
+            raise InvalidDocumentError('Cannot save an abstract document.')
+
+        signal_kwargs = signal_kwargs or {}
+        signals.pre_save.send(self.__class__, document=self, **signal_kwargs)
 
         if validate:
             self.validate(clean=clean)
 
         if write_concern is None:
-            write_concern = {"w": 1}
+            write_concern = {'w': 1}
 
         doc = self.to_mongo()
 
         created = ('_id' not in doc or self._created or force_insert)
 
         signals.pre_save_post_validation.send(self.__class__, document=self,
-                                              created=created)
+                                              created=created, **signal_kwargs)
+
+        if self._meta.get('auto_create_index', True):
+            self.ensure_indexes()
 
         try:
-            collection = self._get_collection()
-            if self._meta.get('auto_create_index', True):
-                self.ensure_indexes()
+            # Save a new document or update an existing one
             if created:
-                if force_insert:
-                    object_id = collection.insert(doc, **write_concern)
-                else:
-                    object_id = collection.save(doc, **write_concern)
-                    # In PyMongo 3.0, the save() call calls internally the _update() call
-                    # but they forget to return the _id value passed back, therefore getting it back here
-                    # Correct behaviour in 2.X and in 3.0.1+ versions
-                    if not object_id and pymongo.version_tuple == (3, 0):
-                        pk_as_mongo_obj = self._fields.get(self._meta['id_field']).to_mongo(self.pk)
-                        object_id = self._qs.filter(pk=pk_as_mongo_obj).first() and \
-                                    self._qs.filter(pk=pk_as_mongo_obj).first().pk
+                object_id = self._save_create(doc, force_insert, write_concern)
             else:
-                object_id = doc['_id']
-                updates, removals = self._delta()
-                # Need to add shard key to query, or you get an error
-                if save_condition is not None:
-                    select_dict = transform.query(self.__class__,
-                                                  **save_condition)
-                else:
-                    select_dict = {}
-                select_dict['_id'] = object_id
-                shard_key = self.__class__._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]
-                    val = doc
-                    for ak in actual_key:
-                        val = val[ak]
-                    select_dict['.'.join(actual_key)] = val
-
-                def is_new_object(last_error):
-                    if last_error is not None:
-                        updated = last_error.get("updatedExisting")
-                        if updated is not None:
-                            return not updated
-                    return created
-
-                update_query = {}
-
-                if updates:
-                    update_query["$set"] = updates
-                if removals:
-                    update_query["$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:
-                        raise SaveConditionError('Race condition preventing'
-                                                 ' document update detected')
-                    created = is_new_object(last_error)
+                object_id, created = self._save_update(doc, save_condition,
+                                                       write_concern)
 
             if cascade is None:
-                cascade = self._meta.get(
-                    'cascade', False) or cascade_kwargs is not None
+                cascade = (self._meta.get('cascade', False) or
+                           cascade_kwargs is not None)
 
             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))
+
+        # Make sure we store the PK on this document now that it's saved
         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)
 
-        signals.post_save.send(self.__class__, document=self, created=created)
+        signals.post_save.send(self.__class__, document=self,
+                               created=created, **signal_kwargs)
+
         self._clear_changed_fields()
         self._created = False
+
         return self
 
-    def cascade_save(self, *args, **kwargs):
-        """Recursively saves any references /
-           generic references on an objects"""
-        _refs = kwargs.get('_refs', []) or []
+    def _save_create(self, doc, force_insert, write_concern):
+        """Save a new document.
+
+        Helper method, should only be used inside save().
+        """
+        collection = self._get_collection()
+
+        if force_insert:
+            return collection.insert(doc, **write_concern)
+
+        object_id = collection.save(doc, **write_concern)
+
+        # In PyMongo 3.0, the save() call calls internally the _update() call
+        # but they forget to return the _id value passed back, therefore getting it back here
+        # Correct behaviour in 2.X and in 3.0.1+ versions
+        if not object_id and pymongo.version_tuple == (3, 0):
+            pk_as_mongo_obj = self._fields.get(self._meta['id_field']).to_mongo(self.pk)
+            object_id = (
+                self._qs.filter(pk=pk_as_mongo_obj).first() and
+                self._qs.filter(pk=pk_as_mongo_obj).first().pk
+            )  # TODO doesn't this make 2 queries?
+
+        return object_id
+
+    def _save_update(self, doc, save_condition, write_concern):
+        """Update an existing document.
+
+        Helper method, should only be used inside save().
+        """
+        collection = self._get_collection()
+        object_id = doc['_id']
+        created = False
+
+        select_dict = {}
+        if save_condition is not None:
+            select_dict = transform.query(self.__class__, **save_condition)
+
+        select_dict['_id'] = object_id
+
+        # Need to add shard key to query, or you get an error
+        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]
+            val = doc
+            for ak in actual_key:
+                val = val[ak]
+            select_dict['.'.join(actual_key)] = val
+
+        updates, removals = self._delta()
+        update_query = {}
+        if updates:
+            update_query['$set'] = updates
+        if removals:
+            update_query['$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:
+                raise SaveConditionError('Race condition preventing'
+                                         ' document update detected')
+            if last_error is not None:
+                updated_existing = last_error.get('updatedExisting')
+                if updated_existing is False:
+                    created = True
+                    # !!! This is bad, means we accidentally created a new,
+                    # potentially corrupted document. See
+                    # https://github.com/MongoEngine/mongoengine/issues/564
+
+        return object_id, created
+
+    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')
@@ -434,16 +486,17 @@ 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())
@@ -463,11 +516,11 @@ 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(
@@ -476,32 +529,38 @@ class Document(BaseDocument):
         # Need to add shard key to query, or you get an error
         return self._qs.filter(**self._object_key).update_one(**kwargs)
 
-    def delete(self, **write_concern):
+    def delete(self, signal_kwargs=None, **write_concern):
         """Delete the :class:`~mongoengine.Document` from the database. This
         will only take effect if the document has been previously saved.
 
+        :parm signal_kwargs: (optional) kwargs dictionary to be passed to
+            the signal calls.
         :param write_concern: Extra keyword arguments are passed down which
             will be used as options for the resultant
             ``getLastError`` command.  For example,
             ``save(..., write_concern={w: 2, fsync: True}, ...)`` will
             wait until at least two servers have recorded the write and
             will force an fsync on the primary server.
-        """
-        signals.pre_delete.send(self.__class__, document=self)
 
-        # Delete FileFields separately 
+        .. versionchanged:: 0.10.7
+            Add signal_kwargs argument
+        """
+        signal_kwargs = signal_kwargs or {}
+        signals.pre_delete.send(self.__class__, document=self, **signal_kwargs)
+
+        # Delete FileFields separately
         FileField = _import_class('FileField')
         for name, field in self._fields.iteritems():
-            if isinstance(field, FileField): 
+            if isinstance(field, FileField):
                 getattr(self, name).delete()
 
         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)
+        signals.post_delete.send(self.__class__, document=self, **signal_kwargs)
 
     def switch_db(self, db_alias, keep_created=True):
         """
@@ -586,11 +645,12 @@ 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)
@@ -598,17 +658,22 @@ 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:
                 try:
                     setattr(self, field, self._reload(field, obj[field]))
-                except KeyError:
-                    # If field is removed from the database while the object
-                    # is in memory, a reload would cause a KeyError
-                    # i.e. obj.update(unset__field=1) followed by obj.reload()
-                    delattr(self, field)
+                except (KeyError, AttributeError):
+                    try:
+                        # If field is a special field, e.g. items is stored as _reserved_items,
+                        # an KeyError is thrown. So try to retrieve the field from _data
+                        setattr(self, field, self._reload(field, obj._data.get(field)))
+                    except KeyError:
+                        # If field is removed from the database while the object
+                        # is in memory, a reload would cause a KeyError
+                        # i.e. obj.update(unset__field=1) followed by obj.reload()
+                        delattr(self, field)
 
         self._changed_fields = obj._changed_fields
         self._created = False
@@ -635,8 +700,8 @@ 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)
 
@@ -662,10 +727,20 @@ class Document(BaseDocument):
     def drop_collection(cls):
         """Drops the entire collection associated with this
         :class:`~mongoengine.Document` type from the database.
+
+        Raises :class:`OperationError` if the document has no collection set
+        (i.g. if it is `abstract`)
+
+        .. versionchanged:: 0.10.7
+            :class:`OperationError` exception raised if no collection available
         """
+        col_name = cls._get_collection_name()
+        if not col_name:
+            raise OperationError('Document %s has no collection defined '
+                                 '(is it abstract ?)' % cls)
         cls._collection = None
         db = cls._get_db()
-        db.drop_collection(cls._get_collection_name())
+        db.drop_collection(col_name)
 
     @classmethod
     def create_index(cls, keys, background=False, **kwargs):
@@ -681,7 +756,7 @@ 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
@@ -707,7 +782,7 @@ 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})
@@ -727,7 +802,7 @@ 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()
@@ -765,8 +840,7 @@ 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
@@ -785,7 +859,6 @@ class Document(BaseDocument):
         """ Lists all of the indexes that should be created for given
         collection. It includes all the indexes from super- and sub-classes.
         """
-
         if cls._meta.get('abstract'):
             return []
 
@@ -836,16 +909,15 @@ 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()
@@ -889,8 +961,9 @@ 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)
@@ -912,8 +985,9 @@ 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
@@ -954,11 +1028,11 @@ class MapReduceDocument(object):
         if not isinstance(self.key, id_field_type):
             try:
                 self.key = id_field_type(self.key)
-            except:
-                raise Exception("Could not cast key as %s" %
+            except Exception:
+                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

@@ -1,12 +1,11 @@
 from collections import defaultdict
 
-from mongoengine.python_support import txt_type
-
+import six
 
 __all__ = ('NotRegistered', 'InvalidDocumentError', 'LookUpError',
            'DoesNotExist', 'MultipleObjectsReturned', 'InvalidQueryError',
            'OperationError', 'NotUniqueError', 'FieldDoesNotExist',
-           'ValidationError')
+           'ValidationError', 'SaveConditionError')
 
 
 class NotRegistered(Exception):
@@ -51,8 +50,8 @@ class FieldDoesNotExist(Exception):
     or an :class:`~mongoengine.EmbeddedDocument`.
 
     To avoid this behavior on data loading,
-    you should the :attr:`strict` to ``False``
-    in the :attr:`meta` dictionnary.
+    you should set the :attr:`strict` to ``False``
+    in the :attr:`meta` dictionary.
     """
 
 
@@ -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/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, MultipleObjectsReturned,
-                                InvalidQueryError, OperationError,
-                                NotUniqueError)
+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/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/manager.py

@@ -29,7 +29,7 @@ class QuerySetManager(object):
         Document.objects is accessed.
         """
         if instance is not None:
-            # Document class being used rather than a document object
+            # Document object being used rather than a document class
             return self
 
         # owner is the document that contains the QuerySetManager

mongoengine/queryset/queryset.py

@@ -1,6 +1,6 @@
 from mongoengine.errors import OperationError
-from mongoengine.queryset.base import (BaseQuerySet, DO_NOTHING, NULLIFY,
-                                       CASCADE, DENY, PULL)
+from mongoengine.queryset.base import (BaseQuerySet, CASCADE, DENY, DO_NOTHING,
+                                       NULLIFY, PULL)
 
 __all__ = ('QuerySet', 'QuerySetNoCache', 'DO_NOTHING', 'NULLIFY', 'CASCADE',
            'DENY', 'PULL')
@@ -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()
 
@@ -38,44 +39,60 @@ class QuerySet(BaseQuerySet):
 
     def __len__(self):
         """Since __len__ is called quite frequently (for example, as part of
-        list(qs) we populate the result cache and cache the length.
+        list(qs)), we populate the result cache and cache the length.
         """
         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.
@@ -109,13 +136,15 @@ class QuerySet(BaseQuerySet):
         return self._len
 
     def no_cache(self):
-        """Convert to a non_caching queryset
+        """Convert to a non-caching queryset
 
         .. versionadded:: 0.8.3 Convert to non caching queryset
         """
         if self._result_cache is not None:
-            raise OperationError("QuerySet already cached")
-        return self.clone_into(QuerySetNoCache(self._document, self._collection))
+            raise OperationError('QuerySet already cached')
+
+        return self._clone_into(QuerySetNoCache(self._document,
+                                                self._collection))
 
 
 class QuerySetNoCache(BaseQuerySet):
@@ -126,7 +155,7 @@ class QuerySetNoCache(BaseQuerySet):
 
         .. versionadded:: 0.8.3 Convert to caching queryset
         """
-        return self.clone_into(QuerySet(self._document, self._collection))
+        return self._clone_into(QuerySet(self._document, self._collection))
 
     def __repr__(self):
         """Provides the string representation of the QuerySet
@@ -137,13 +166,14 @@ 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

@@ -1,11 +1,13 @@
 from collections import defaultdict
 
+from bson import ObjectId, SON
+from bson.dbref import DBRef
 import pymongo
-from bson import SON
+import six
 
-from mongoengine.base.fields import UPDATE_OPERATORS
-from mongoengine.connection import get_connection
+from mongoengine.base import UPDATE_OPERATORS
 from mongoengine.common import _import_class
+from mongoengine.connection import get_connection
 from mongoengine.errors import InvalidQueryError
 from mongoengine.python_support import IS_PYMONGO_3
 
@@ -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
 
@@ -44,8 +46,8 @@ def query(_doc_cls=None, **kwargs):
         if len(parts) > 1 and parts[-1] in MATCH_OPERATORS:
             op = parts.pop()
 
-        # Allw to escape operator-like field name by __
-        if len(parts) > 1 and parts[-1] == "":
+        # Allow to escape operator-like field name by __
+        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:
@@ -98,8 +101,31 @@ def query(_doc_cls=None, **kwargs):
                         value = value['_id']
 
             elif op in ('in', 'nin', 'all', 'near') and not isinstance(value, dict):
-                # 'in', 'nin' and 'all' require a list of values
-                value = [field.prepare_query_value(op, v) for v in value]
+                # Raise an error if the in/nin/all/near param is not iterable. We need a
+                # special check for BaseDocument, because - although it's iterable - using
+                # it as such in the context of this method is most definitely a mistake.
+                BaseDocument = _import_class('BaseDocument')
+                if isinstance(value, BaseDocument):
+                    raise TypeError("When using the `in`, `nin`, `all`, or "
+                                    "`near`-operators you can\'t use a "
+                                    "`Document`, you must wrap your object "
+                                    "in a list (object -> [object]).")
+                elif not hasattr(value, '__iter__'):
+                    raise TypeError("The `in`, `nin`, `all`, or "
+                                    "`near`-operators must be applied to an "
+                                    "iterable (e.g. a list).")
+                else:
+                    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:
@@ -108,15 +134,18 @@ def query(_doc_cls=None, **kwargs):
             elif op in ('match', 'elemMatch'):
                 ListField = _import_class('ListField')
                 EmbeddedDocumentField = _import_class('EmbeddedDocumentField')
-                if (isinstance(value, dict) and isinstance(field, ListField) and
-                    isinstance(field.field, EmbeddedDocumentField)):
+                if (
+                    isinstance(value, dict) and
+                    isinstance(field, ListField) and
+                    isinstance(field.field, EmbeddedDocumentField)
+                ):
                     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}
 
@@ -125,11 +154,13 @@ 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]
@@ -179,15 +210,16 @@ 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
@@ -201,22 +233,25 @@ def update(_doc_cls=None, **update):
                 # Support decrement by flipping a positive value's sign
                 # and using 'inc'
                 op = 'inc'
-                if value > 0:
-                    value = -value
+                value = -value
             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] == '':
+            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 = []
 
@@ -224,7 +259,7 @@ 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 = '$'
@@ -245,7 +280,7 @@ 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)
 
@@ -259,7 +294,7 @@ 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:
@@ -269,16 +304,16 @@ 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]
@@ -289,7 +324,7 @@ 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)
 
@@ -297,7 +332,7 @@ 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
@@ -311,74 +346,82 @@ 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}}
-        except:
-            pass
-        try:
-            value[0][0]
-            return {"$geometry": {"type": "LineString", "coordinates": value}}
-        except:
-            pass
-        try:
-            value[0]
-            return {"$geometry": {"type": "Point", "coordinates": value}}
-        except:
+            return {'$geometry': {'type': 'Polygon', 'coordinates': value}}
+        except (TypeError, IndexError):
             pass
 
-    raise InvalidQueryError("Invalid $geometry data. Can be either a dictionary "
-                            "or (nested) lists of coordinate(s)")
+        try:
+            value[0][0]
+            return {'$geometry': {'type': 'LineString', 'coordinates': value}}
+        except (TypeError, IndexError):
+            pass
+
+        try:
+            value[0]
+            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)')

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:
@@ -29,11 +27,12 @@ except ImportError:
                                'because the blinker library is '
                                'not installed.')
 
-        send = lambda *a, **kw: None
+        send = lambda *a, **kw: None  # noqa
         connect = disconnect = has_receivers_for = receivers_for = \
             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()

requirements.txt

@@ -1,2 +1,5 @@
-pymongo>=2.7.1
 nose
+pymongo>=2.7.1
+six==1.10.0
+flake8
+flake8-import-order

setup.cfg

@@ -1,8 +1,11 @@
 [nosetests]
-rednose = 1
-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,venv3,.tox,.eggs,tests
+max-complexity=47
+application-import-names=mongoengine,tests

setup.py

@@ -1,6 +1,6 @@
 import os
 import sys
-from setuptools import setup, find_packages
+from setuptools import find_packages, setup
 
 # Hack to silence atexit traceback in newer python versions
 try:
@@ -8,20 +8,25 @@ try:
 except ImportError:
     pass
 
-DESCRIPTION = 'MongoEngine is a Python Object-Document ' + \
-'Mapper for working with MongoDB.'
-LONG_DESCRIPTION = None
+DESCRIPTION = (
+    'MongoEngine is a Python Object-Document '
+    'Mapper for working with MongoDB.'
+)
+
 try:
-    LONG_DESCRIPTION = open('README.rst').read()
-except:
-    pass
+    with open('README.rst') as fin:
+        LONG_DESCRIPTION = fin.read()
+except Exception:
+    LONG_DESCRIPTION = None
 
 
 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))
 
+
 # Dirty hack to get version number from monogengine/__init__.py - we can't
 # import it as it depends on PyMongo and PyMongo isn't installed until this
 # file is read
@@ -37,47 +42,46 @@ 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', 'rednose', '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', 'rednose', 'coverage==3.7.1', 'blinker', 'Pillow>=2.0.0', 'python-dateutil']
+    extra_opts['tests_require'] += ['python-dateutil']
 
-    if sys.version_info[0] == 2 and sys.version_info[1] == 6:
-        extra_opts['tests_require'].append('unittest2')
-
-setup(name='mongoengine',
-      version=VERSION,
-      author='Harry Marr',
-      author_email='harry.marr@{nospam}gmail.com',
-      maintainer="Ross Lawley",
-      maintainer_email="ross.lawley@{nospam}gmail.com",
-      url='http://mongoengine.org/',
-      download_url='https://github.com/MongoEngine/mongoengine/tarball/master',
-      license='MIT',
-      include_package_data=True,
-      description=DESCRIPTION,
-      long_description=LONG_DESCRIPTION,
-      platforms=['any'],
-      classifiers=CLASSIFIERS,
-      install_requires=['pymongo>=2.7.1'],
-      test_suite='nose.collector',
-      **extra_opts
+setup(
+    name='mongoengine',
+    version=VERSION,
+    author='Harry Marr',
+    author_email='harry.marr@{nospam}gmail.com',
+    maintainer="Ross Lawley",
+    maintainer_email="ross.lawley@{nospam}gmail.com",
+    url='http://mongoengine.org/',
+    download_url='https://github.com/MongoEngine/mongoengine/tarball/master',
+    license='MIT',
+    include_package_data=True,
+    description=DESCRIPTION,
+    long_description=LONG_DESCRIPTION,
+    platforms=['any'],
+    classifiers=CLASSIFIERS,
+    install_requires=['pymongo>=2.7.1', 'six'],
+    test_suite='nose.collector',
+    **extra_opts
 )

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

@@ -2,7 +2,6 @@
 import unittest
 import sys
 
-sys.path[0:0] = [""]
 
 import pymongo
 
@@ -32,10 +31,7 @@ class IndexesTest(unittest.TestCase):
         self.Person = Person
 
     def tearDown(self):
-        for collection in self.db.collection_names():
-            if 'system.' in collection:
-                continue
-            self.db.drop_collection(collection)
+        self.connection.drop_database(self.db)
 
     def test_indexes_document(self):
         """Ensure that indexes are used when meta[indexes] is specified for
@@ -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)
@@ -822,33 +818,34 @@ class IndexesTest(unittest.TestCase):
             name = StringField(required=True)
             term = StringField(required=True)
 
-        class Report(Document):
+        class ReportEmbedded(Document):
             key = EmbeddedDocumentField(CompoundKey, primary_key=True)
             text = StringField()
 
-        Report.drop_collection()
-
         my_key = CompoundKey(name="n", term="ok")
-        report = Report(text="OK", key=my_key).save()
+        report = ReportEmbedded(text="OK", key=my_key).save()
 
         self.assertEqual({'text': 'OK', '_id': {'term': 'ok', 'name': 'n'}},
                          report.to_mongo())
-        self.assertEqual(report, Report.objects.get(pk=my_key))
+        self.assertEqual(report, ReportEmbedded.objects.get(pk=my_key))
 
     def test_compound_key_dictfield(self):
 
-        class Report(Document):
+        class ReportDictField(Document):
             key = DictField(primary_key=True)
             text = StringField()
 
-        Report.drop_collection()
-
         my_key = {"name": "n", "term": "ok"}
-        report = Report(text="OK", key=my_key).save()
+        report = ReportDictField(text="OK", key=my_key).save()
 
         self.assertEqual({'text': 'OK', '_id': {'term': 'ok', 'name': 'n'}},
                          report.to_mongo())
-        self.assertEqual(report, Report.objects.get(pk=my_key))
+
+        # We can't directly call ReportDictField.objects.get(pk=my_key),
+        # because dicts are unordered, and if the order in MongoDB is
+        # different than the one in `my_key`, this test will fail.
+        self.assertEqual(report, ReportDictField.objects.get(pk__name=my_key['name']))
+        self.assertEqual(report, ReportDictField.objects.get(pk__term=my_key['term']))
 
     def test_string_indexes(self):
 
@@ -863,6 +860,20 @@ class IndexesTest(unittest.TestCase):
         self.assertTrue([('provider_ids.foo', 1)] in info)
         self.assertTrue([('provider_ids.bar', 1)] in info)
 
+    def test_sparse_compound_indexes(self):
+
+        class MyDoc(Document):
+            provider_ids = DictField()
+            meta = {
+                "indexes": [{'fields': ("provider_ids.foo", "provider_ids.bar"),
+                             'sparse': True}],
+            }
+
+        info = MyDoc.objects._collection.index_information()
+        self.assertEqual([('provider_ids.foo', 1), ('provider_ids.bar', 1)],
+                         info['provider_ids.foo_1_provider_ids.bar_1']['key'])
+        self.assertTrue(info['provider_ids.foo_1_provider_ids.bar_1']['sparse'])
+
     def test_text_indexes(self):
 
         class Book(Document):
@@ -895,26 +906,38 @@ class IndexesTest(unittest.TestCase):
 
         Issue #812
         """
+        # Use a new connection and database since dropping the database could
+        # cause concurrent tests to fail.
+        connection = connect(db='tempdatabase',
+                             alias='test_indexes_after_database_drop')
+
         class BlogPost(Document):
             title = StringField()
             slug = StringField(unique=True)
 
-        BlogPost.drop_collection()
+            meta = {'db_alias': 'test_indexes_after_database_drop'}
 
-        # Create Post #1
-        post1 = BlogPost(title='test1', slug='test')
-        post1.save()
+        try:
+            BlogPost.drop_collection()
 
-        # Drop the Database
-        self.connection.drop_database(BlogPost._get_db().name)
+            # Create Post #1
+            post1 = BlogPost(title='test1', slug='test')
+            post1.save()
 
-        # Re-create Post #1
-        post1 = BlogPost(title='test1', slug='test')
-        post1.save()
+            # Drop the Database
+            connection.drop_database('tempdatabase')
+
+            # Re-create Post #1
+            post1 = BlogPost(title='test1', slug='test')
+            post1.save()
+
+            # Create Post #2
+            post2 = BlogPost(title='test2', slug='test')
+            self.assertRaises(NotUniqueError, post2.save)
+        finally:
+            # Drop the temporary database at the end
+            connection.drop_database('tempdatabase')
 
-        # Create Post #2
-        post2 = BlogPost(title='test2', slug='test')
-        self.assertRaises(NotUniqueError, post2.save)
 
     def test_index_dont_send_cls_option(self):
         """

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())
 
@@ -411,7 +401,7 @@ class InheritanceTest(unittest.TestCase):
         try:
             class MyDocument(DateCreatedDocument, DateUpdatedDocument):
                 pass
-        except:
+        except Exception:
             self.assertTrue(False, "Couldn't create MyDocument class")
 
     def test_abstract_documents(self):
@@ -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
@@ -13,15 +10,15 @@ from datetime import datetime
 from bson import DBRef, ObjectId
 from tests import fixtures
 from tests.fixtures import (PickleEmbedded, PickleTest, PickleSignalsTest,
-                            PickleDyanmicEmbedded, PickleDynamicTest)
+                            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):
 
@@ -447,6 +435,15 @@ class InstanceTest(unittest.TestCase):
 
         person.to_dbref()
 
+    def test_save_abstract_document(self):
+        """Saving an abstract document should fail."""
+        class Doc(Document):
+            name = StringField()
+            meta = {'abstract': True}
+
+        with self.assertRaises(InvalidDocumentError):
+            Doc(name='aaa').save()
+
     def test_reload(self):
         """Ensure that attributes may be reloaded.
         """
@@ -571,6 +568,28 @@ class InstanceTest(unittest.TestCase):
         except Exception:
             self.assertFalse("Threw wrong exception")
 
+    def test_reload_of_non_strict_with_special_field_name(self):
+        """Ensures reloading works for documents with meta strict == False
+        """
+        class Post(Document):
+            meta = {
+                'strict': False
+            }
+            title = StringField()
+            items = ListField()
+
+        Post.drop_collection()
+
+        Post._get_collection().insert({
+            "title": "Items eclipse",
+            "items": ["more lorem", "even more ipsum"]
+        })
+
+        post = Post.objects.first()
+        post.reload()
+        self.assertEqual(post.title, "Items eclipse")
+        self.assertEqual(post.items, ["more lorem", "even more ipsum"])
+
     def test_dictionary_access(self):
         """Ensure that dictionary-style field access works properly.
         """
@@ -657,6 +676,19 @@ class InstanceTest(unittest.TestCase):
         doc = Doc.objects.get()
         self.assertHasInstance(doc.embedded_field[0], doc)
 
+    def test_embedded_document_complex_instance_no_use_db_field(self):
+        """Ensure that use_db_field is propagated to list of Emb Docs
+        """
+        class Embedded(EmbeddedDocument):
+            string = StringField(db_field='s')
+
+        class Doc(Document):
+            embedded_field = ListField(EmbeddedDocumentField(Embedded))
+
+        d = Doc(embedded_field=[Embedded(string="Hi")]).to_mongo(
+            use_db_field=False).to_dict()
+        self.assertEqual(d['embedded_field'], [{'string': 'Hi'}])
+
     def test_instance_is_set_on_setattr(self):
 
         class Email(EmbeddedDocument):
@@ -710,7 +742,7 @@ 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})
@@ -749,7 +781,7 @@ 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}})
@@ -763,8 +795,10 @@ 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):
@@ -772,9 +806,8 @@ 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)
 
@@ -783,7 +816,7 @@ 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)
 
@@ -792,7 +825,7 @@ 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)
 
@@ -1199,6 +1232,19 @@ class InstanceTest(unittest.TestCase):
         self.assertEqual(person.name, None)
         self.assertEqual(person.age, None)
 
+    def test_update_rename_operator(self):
+        """Test the $rename operator."""
+        coll = self.Person._get_collection()
+        doc = self.Person(name='John').save()
+        raw_doc = coll.find_one({'_id': doc.pk})
+        self.assertEqual(set(raw_doc.keys()), set(['_id', '_cls', 'name']))
+
+        doc.update(rename__name='first_name')
+        raw_doc = coll.find_one({'_id': doc.pk})
+        self.assertEqual(set(raw_doc.keys()),
+                         set(['_id', '_cls', 'first_name']))
+        self.assertEqual(raw_doc['first_name'], 'John')
+
     def test_inserts_if_you_set_the_pk(self):
         p1 = self.Person(name='p1', id=bson.ObjectId()).save()
         p2 = self.Person(name='p2')
@@ -1258,12 +1304,11 @@ 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()
 
@@ -1273,19 +1318,17 @@ 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):
@@ -1294,8 +1337,8 @@ 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):
         """
@@ -1813,15 +1856,13 @@ 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):
@@ -1833,11 +1874,17 @@ 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}})
+        with self.assertRaises(InvalidDocumentError):
+            Word._from_son({
+                'stem': [1, 2, 3],
+                'forms': 1,
+                'count': 'one',
+                'occurs': {"hello": None}
+            })
 
-        self.assertRaises(InvalidDocumentError, raise_invalid_document)
+        # Tests for issue #1438: https://github.com/MongoEngine/mongoengine/issues/1438
+        with self.assertRaises(ValueError):
+            Word._from_son('this is not a valid SON dict')
 
     def test_reverse_delete_rule_cascade_and_nullify(self):
         """Ensure that a referenced document is also deleted upon deletion.
@@ -1871,6 +1918,62 @@ class InstanceTest(unittest.TestCase):
         author.delete()
         self.assertEqual(BlogPost.objects.count(), 0)
 
+    def test_reverse_delete_rule_with_custom_id_field(self):
+        """Ensure that a referenced document with custom primary key
+        is also deleted upon deletion.
+        """
+        class User(Document):
+            name = StringField(primary_key=True)
+
+        class Book(Document):
+            author = ReferenceField(User, reverse_delete_rule=CASCADE)
+            reviewer = ReferenceField(User, reverse_delete_rule=NULLIFY)
+
+        User.drop_collection()
+        Book.drop_collection()
+
+        user = User(name='Mike').save()
+        reviewer = User(name='John').save()
+        book = Book(author=user, reviewer=reviewer).save()
+
+        reviewer.delete()
+        self.assertEqual(Book.objects.count(), 1)
+        self.assertEqual(Book.objects.get().reviewer, None)
+
+        user.delete()
+        self.assertEqual(Book.objects.count(), 0)
+
+    def test_reverse_delete_rule_with_shared_id_among_collections(self):
+        """Ensure that cascade delete rule doesn't mix id among collections.
+        """
+        class User(Document):
+            id = IntField(primary_key=True)
+
+        class Book(Document):
+            id = IntField(primary_key=True)
+            author = ReferenceField(User, reverse_delete_rule=CASCADE)
+
+        User.drop_collection()
+        Book.drop_collection()
+
+        user_1 = User(id=1).save()
+        user_2 = User(id=2).save()
+        book_1 = Book(id=1, author=user_2).save()
+        book_2 = Book(id=2, author=user_1).save()
+
+        user_2.delete()
+        # Deleting user_2 should also delete book_1 but not book_2
+        self.assertEqual(Book.objects.count(), 1)
+        self.assertEqual(Book.objects.get(), book_2)
+
+        user_3 = User(id=3).save()
+        book_3 = Book(id=3, author=user_3).save()
+
+        user_3.delete()
+        # Deleting user_3 should also delete book_3
+        self.assertEqual(Book.objects.count(), 1)
+        self.assertEqual(Book.objects.get(), book_2)
+
     def test_reverse_delete_rule_with_document_inheritance(self):
         """Ensure that a referenced document is also deleted upon deletion
         of a child document.
@@ -2012,8 +2115,7 @@ 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(
@@ -2023,21 +2125,15 @@ 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)
@@ -2226,7 +2322,7 @@ class InstanceTest(unittest.TestCase):
 
         pickle_doc = PickleDynamicTest(
             name="test", number=1, string="One", lists=['1', '2'])
-        pickle_doc.embedded = PickleDyanmicEmbedded(foo="Bar")
+        pickle_doc.embedded = PickleDynamicEmbedded(foo="Bar")
         pickled_doc = pickle.dumps(pickle_doc)  # make sure pickling works even before the doc is saved
 
         pickle_doc.save()
@@ -2253,15 +2349,14 @@ 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):
@@ -2724,11 +2819,10 @@ 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()
@@ -2749,12 +2843,11 @@ 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)
@@ -2775,11 +2868,10 @@ 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):
@@ -2837,6 +2929,20 @@ class InstanceTest(unittest.TestCase):
         self.assertEqual(person.name, "Test User")
         self.assertEqual(person.age, 42)
 
+    def test_positional_creation_embedded(self):
+        """Ensure that embedded document may be created using positional arguments.
+        """
+        job = self.Job("Test Job", 4)
+        self.assertEqual(job.name, "Test Job")
+        self.assertEqual(job.years, 4)
+
+    def test_mixed_creation_embedded(self):
+        """Ensure that embedded document may be created using mixed arguments.
+        """
+        job = self.Job("Test Job", years=4)
+        self.assertEqual(job.name, "Test Job")
+        self.assertEqual(job.years, 4)
+
     def test_mixed_creation_dynamic(self):
         """Ensure that document may be created using mixed arguments.
         """
@@ -2850,11 +2956,9 @@ 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'
         """
@@ -3013,6 +3117,17 @@ 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):
@@ -3086,5 +3201,20 @@ 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/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
@@ -20,15 +18,13 @@ try:
 except ImportError:
     HAS_PIL = False
 
+from tests.utils import MongoDBTestCase
+
 TEST_IMAGE_PATH = os.path.join(os.path.dirname(__file__), 'mongoengine.png')
 TEST_IMAGE2_PATH = os.path.join(os.path.dirname(__file__), 'mongodb_leaf.png')
 
 
-class FileTest(unittest.TestCase):
-
-    def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
+class FileTest(MongoDBTestCase):
 
     def tearDown(self):
         self.db.drop_collection('fs.files')
@@ -49,7 +45,7 @@ class FileTest(unittest.TestCase):
 
         PutFile.drop_collection()
 
-        text = b('Hello, World!')
+        text = six.b('Hello, World!')
         content_type = 'text/plain'
 
         putfile = PutFile()
@@ -88,8 +84,8 @@ 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 +119,8 @@ 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 +151,8 @@ 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 +181,7 @@ 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 +198,7 @@ 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 +224,7 @@ 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 +278,7 @@ 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 +293,66 @@ 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 +368,14 @@ 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 +492,7 @@ 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 +500,15 @@ 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 +516,7 @@ 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/fixtures.py

@@ -26,7 +26,7 @@ class NewDocumentPickleTest(Document):
     new_field = StringField()
 
 
-class PickleDyanmicEmbedded(DynamicEmbeddedDocument):
+class PickleDynamicEmbedded(DynamicEmbeddedDocument):
     date = DateTimeField(default=datetime.now)
 
 

tests/migration/__init__.py

@@ -1,8 +0,0 @@
-from convert_to_new_inheritance_model import *
-from decimalfield_as_float import *
-from refrencefield_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/refrencefield_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)
@@ -144,6 +141,16 @@ class OnlyExcludeAllTest(unittest.TestCase):
         self.assertEqual(qs._loaded_fields.as_dict(),
                          {'b': {'$slice': 5}})
 
+    def test_mix_slice_with_other_fields(self):
+        class MyDoc(Document):
+            a = ListField()
+            b = ListField()
+            c = ListField()
+
+        qs = MyDoc.objects.fields(a=1, b=0, slice__c=2)
+        self.assertEqual(qs._loaded_fields.as_dict(),
+                         {'c': {'$slice': 2}, 'a': 1})
+
     def test_only(self):
         """Ensure that QuerySet.only only returns the requested fields.
         """

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/pickable.py

@@ -0,0 +1,78 @@
+import pickle
+import unittest
+from pymongo.mongo_client import MongoClient
+from mongoengine import Document, StringField, IntField
+from mongoengine.connection import connect
+
+__author__ = 'stas'
+
+class Person(Document):
+    name = StringField()
+    age = IntField()
+
+class TestQuerysetPickable(unittest.TestCase):
+    """
+    Test for adding pickling support for QuerySet instances
+    See issue https://github.com/MongoEngine/mongoengine/issues/442
+    """
+    def setUp(self):
+        super(TestQuerysetPickable, self).setUp()
+
+        connection = connect(db="test") #type: pymongo.mongo_client.MongoClient
+
+        connection.drop_database("test")
+
+        self.john = Person.objects.create(
+            name="John",
+            age=21
+        )
+
+
+    def test_picke_simple_qs(self):
+
+        qs = Person.objects.all()
+
+        pickle.dumps(qs)
+
+    def _get_loaded(self, qs):
+        s = pickle.dumps(qs)
+
+        return pickle.loads(s)
+
+    def test_unpickle(self):
+        qs = Person.objects.all()
+
+        loadedQs = self._get_loaded(qs)
+
+        self.assertEqual(qs.count(), loadedQs.count())
+
+        #can update loadedQs
+        loadedQs.update(age=23)
+
+        #check
+        self.assertEqual(Person.objects.first().age, 23)
+
+    def test_pickle_support_filtration(self):
+        Person.objects.create(
+            name="Alice",
+            age=22
+        )
+
+        Person.objects.create(
+            name="Bob",
+            age=23
+        )
+
+        qs = Person.objects.filter(age__gte=22)
+        self.assertEqual(qs.count(), 2)
+
+        loaded = self._get_loaded(qs)
+
+        self.assertEqual(loaded.count(), 2)
+        self.assertEqual(loaded.filter(name="Bob").first().age, 23)
+    
+
+
+
+
+

tests/queryset/transform.py

@@ -1,11 +1,7 @@
-import sys
-sys.path[0:0] = [""]
-
 import unittest
 
 from mongoengine import *
-from mongoengine.queryset import Q
-from mongoengine.queryset import transform
+from mongoengine.queryset import Q, transform
 
 __all__ = ("TransformTest",)
 
@@ -41,8 +37,8 @@ class TransformTest(unittest.TestCase):
         DicDoc.drop_collection()
         Doc.drop_collection()
 
+        DicDoc().save()
         doc = Doc().save()
-        dic_doc = DicDoc().save()
 
         for k, v in (("set", "$set"), ("set_on_insert", "$setOnInsert"), ("push", "$push")):
             update = transform.update(DicDoc, **{"%s__dictField__test" % k: doc})
@@ -55,7 +51,6 @@ class TransformTest(unittest.TestCase):
         update = transform.update(DicDoc, pull__dictField__test=doc)
         self.assertTrue(isinstance(update["$pull"]["dictField"]["test"], dict))
 
-
     def test_query_field_name(self):
         """Ensure that the correct field name is used when querying.
         """
@@ -156,26 +151,33 @@ class TransformTest(unittest.TestCase):
         class Doc(Document):
             meta = {'allow_inheritance': False}
 
-        raw_query = Doc.objects(__raw__={'deleted': False,
-                                'scraped': 'yes',
-                                '$nor': [{'views.extracted': 'no'},
-                                         {'attachments.views.extracted':'no'}]
-                                })._query
+        raw_query = Doc.objects(__raw__={
+            'deleted': False,
+            'scraped': 'yes',
+            '$nor': [
+                {'views.extracted': 'no'},
+                {'attachments.views.extracted': 'no'}
+            ]
+        })._query
 
-        expected = {'deleted': False, 'scraped': 'yes',
-                    '$nor': [{'views.extracted': 'no'},
-                             {'attachments.views.extracted': 'no'}]}
-        self.assertEqual(expected, raw_query)
+        self.assertEqual(raw_query, {
+            'deleted': False,
+            'scraped': 'yes',
+            '$nor': [
+                {'views.extracted': 'no'},
+                {'attachments.views.extracted': 'no'}
+            ]
+        })
 
     def test_geojson_PointField(self):
         class Location(Document):
             loc = PointField()
 
         update = transform.update(Location, set__loc=[1, 2])
-        self.assertEqual(update, {'$set': {'loc': {"type": "Point", "coordinates": [1,2]}}})
+        self.assertEqual(update, {'$set': {'loc': {"type": "Point", "coordinates": [1, 2]}}})
 
-        update = transform.update(Location, set__loc={"type": "Point", "coordinates": [1,2]})
-        self.assertEqual(update, {'$set': {'loc': {"type": "Point", "coordinates": [1,2]}}})
+        update = transform.update(Location, set__loc={"type": "Point", "coordinates": [1, 2]})
+        self.assertEqual(update, {'$set': {'loc': {"type": "Point", "coordinates": [1, 2]}}})
 
     def test_geojson_LineStringField(self):
         class Location(Document):
@@ -224,6 +226,10 @@ class TransformTest(unittest.TestCase):
         self.assertEqual(1, Doc.objects(item__type__="axe").count())
         self.assertEqual(1, Doc.objects(item__name__="Heroic axe").count())
 
+        Doc.objects(id=doc.id).update(set__item__type__='sword')
+        self.assertEqual(1, Doc.objects(item__type__="sword").count())
+        self.assertEqual(0, Doc.objects(item__type__="axe").count())
+
     def test_understandable_error_raised(self):
         class Event(Document):
             title = StringField()
@@ -232,7 +238,9 @@ 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__':
     unittest.main()

tests/queryset/visitor.py

@@ -1,14 +1,12 @@
-import sys
-sys.path[0:0] = [""]
-
+import datetime
+import re
 import unittest
 
 from bson import ObjectId
-from datetime import datetime
 
 from mongoengine import *
-from mongoengine.queryset import Q
 from mongoengine.errors import InvalidQueryError
+from mongoengine.queryset import Q
 
 __all__ = ("QTest",)
 
@@ -132,12 +130,12 @@ class QTest(unittest.TestCase):
         TestDoc(x=10).save()
         TestDoc(y=True).save()
 
-        self.assertEqual(query,
-        {'$and': [
-            {'$or': [{'x': {'$gt': 0}}, {'x': {'$exists': False}}]},
-            {'$or': [{'x': {'$lt': 100}}, {'y': True}]}
-        ]})
-
+        self.assertEqual(query, {
+            '$and': [
+                {'$or': [{'x': {'$gt': 0}}, {'x': {'$exists': False}}]},
+                {'$or': [{'x': {'$lt': 100}}, {'y': True}]}
+            ]
+        })
         self.assertEqual(2, TestDoc.objects(q1 & q2).count())
 
     def test_or_and_or_combination(self):
@@ -157,15 +155,14 @@ class QTest(unittest.TestCase):
         q2 = (Q(x__lt=100) & (Q(y=False) | Q(y__exists=False)))
         query = (q1 | q2).to_query(TestDoc)
 
-        self.assertEqual(query,
-            {'$or': [
+        self.assertEqual(query, {
+            '$or': [
                 {'$and': [{'x': {'$gt': 0}},
                           {'$or': [{'y': True}, {'y': {'$exists': False}}]}]},
                 {'$and': [{'x': {'$lt': 100}},
                           {'$or': [{'y': False}, {'y': {'$exists': False}}]}]}
-            ]}
-        )
-
+            ]
+        })
         self.assertEqual(2, TestDoc.objects(q1 | q2).count())
 
     def test_multiple_occurence_in_field(self):
@@ -188,7 +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()
 
@@ -215,19 +212,19 @@ class QTest(unittest.TestCase):
 
         BlogPost.drop_collection()
 
-        post1 = BlogPost(title='Test 1', publish_date=datetime(2010, 1, 8), published=False)
+        post1 = BlogPost(title='Test 1', publish_date=datetime.datetime(2010, 1, 8), published=False)
         post1.save()
 
-        post2 = BlogPost(title='Test 2', publish_date=datetime(2010, 1, 15), published=True)
+        post2 = BlogPost(title='Test 2', publish_date=datetime.datetime(2010, 1, 15), published=True)
         post2.save()
 
         post3 = BlogPost(title='Test 3', published=True)
         post3.save()
 
-        post4 = BlogPost(title='Test 4', publish_date=datetime(2010, 1, 8))
+        post4 = BlogPost(title='Test 4', publish_date=datetime.datetime(2010, 1, 8))
         post4.save()
 
-        post5 = BlogPost(title='Test 1', publish_date=datetime(2010, 1, 15))
+        post5 = BlogPost(title='Test 1', publish_date=datetime.datetime(2010, 1, 15))
         post5.save()
 
         post6 = BlogPost(title='Test 1', published=False)
@@ -250,7 +247,7 @@ class QTest(unittest.TestCase):
         self.assertTrue(all(obj.id in posts for obj in published_posts))
 
         # Check Q object combination
-        date = datetime(2010, 1, 10)
+        date = datetime.datetime(2010, 1, 10)
         q = BlogPost.objects(Q(publish_date__lte=date) | Q(published=True))
         posts = [post.id for post in q]
 
@@ -271,12 +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')
-        self.assertRaises(InvalidQueryError, wrong_query_objs)
-        self.assertRaises(InvalidQueryError, wrong_query_objs_filter)
+
+        # filter should fail, too
+        with self.assertRaises(InvalidQueryError):
+            self.Person.objects.filter('user1')
+
 
     def test_q_regex(self):
         """Ensure that Q objects can be queried using regexes.
@@ -284,7 +282,6 @@ class QTest(unittest.TestCase):
         person = self.Person(name='Guido van Rossum')
         person.save()
 
-        import re
         obj = self.Person.objects(Q(name=re.compile('^Gui'))).first()
         self.assertEqual(obj, person)
         obj = self.Person.objects(Q(name=re.compile('^gui'))).first()

tests/test_connection.py

@@ -1,13 +1,11 @@
-import sys
 import datetime
 from pymongo.errors import OperationFailure
 
-sys.path[0:0] = [""]
-
 try:
     import unittest2 as unittest
 except ImportError:
     import unittest
+from nose.plugins.skip import SkipTest
 
 import pymongo
 from bson.tz_util import utc
@@ -18,7 +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):
@@ -51,6 +50,76 @@ class ConnectionTest(unittest.TestCase):
         conn = get_connection('testdb')
         self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
 
+    def test_connect_in_mocking(self):
+        """Ensure that the connect() method works properly in mocking.
+        """
+        try:
+            import mongomock
+        except ImportError:
+            raise SkipTest('you need mongomock installed to run this testcase')
+
+        connect('mongoenginetest', host='mongomock://localhost')
+        conn = get_connection()
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect('mongoenginetest2', host='mongomock://localhost', alias='testdb2')
+        conn = get_connection('testdb2')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect('mongoenginetest3', host='mongodb://localhost', is_mock=True, alias='testdb3')
+        conn = get_connection('testdb3')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect('mongoenginetest4', is_mock=True, alias='testdb4')
+        conn = get_connection('testdb4')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host='mongodb://localhost:27017/mongoenginetest5', is_mock=True, alias='testdb5')
+        conn = get_connection('testdb5')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host='mongomock://localhost:27017/mongoenginetest6', alias='testdb6')
+        conn = get_connection('testdb6')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host='mongomock://localhost:27017/mongoenginetest7', is_mock=True, alias='testdb7')
+        conn = get_connection('testdb7')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+    def test_connect_with_host_list(self):
+        """Ensure that the connect() method works when host is a list
+
+        Uses mongomock to test w/o needing multiple mongod/mongos processes
+        """
+        try:
+            import mongomock
+        except ImportError:
+            raise SkipTest('you need mongomock installed to run this testcase')
+
+        connect(host=['mongomock://localhost'])
+        conn = get_connection()
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host=['mongodb://localhost'], is_mock=True,  alias='testdb2')
+        conn = get_connection('testdb2')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host=['localhost'], is_mock=True,  alias='testdb3')
+        conn = get_connection('testdb3')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host=['mongomock://localhost:27017', 'mongomock://localhost:27018'], alias='testdb4')
+        conn = get_connection('testdb4')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host=['mongodb://localhost:27017', 'mongodb://localhost:27018'], is_mock=True,  alias='testdb5')
+        conn = get_connection('testdb5')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
+        connect(host=['localhost:27017', 'localhost:27018'], is_mock=True,  alias='testdb6')
+        conn = get_connection('testdb6')
+        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+
     def test_disconnect(self):
         """Ensure that the disconnect() method works properly
         """
@@ -88,7 +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')
 
@@ -103,19 +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()
@@ -125,8 +187,44 @@ 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_connect_uri_with_replicaset(self):
+        """Ensure connect() works when specifying a replicaSet."""
+        if IS_PYMONGO_3:
+            c = connect(host='mongodb://localhost/test?replicaSet=local-rs')
+            db = get_db()
+            self.assertTrue(isinstance(db, pymongo.database.Database))
+            self.assertEqual(db.name, 'test')
+        else:
+            # PyMongo < v3.x raises an exception:
+            # "localhost:27017 is not a member of replica set local-rs"
+            with self.assertRaises(MongoEngineConnectionError):
+                c = connect(host='mongodb://localhost/test?replicaSet=local-rs')
+
+    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
@@ -145,13 +243,14 @@ 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"
-        test_conn2 = connect(
+        connect(
             'mongoenginetest', alias='test2',
             host=('mongodb://username2:password@localhost/'
                   'mongoenginetest?authSource=admin')
@@ -169,7 +268,7 @@ 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))
 
@@ -186,8 +285,7 @@ class ConnectionTest(unittest.TestCase):
         self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
 
     def test_connection_kwargs(self):
-        """Ensure that connection kwargs get passed to pymongo.
-        """
+        """Ensure that connection kwargs get passed to pymongo."""
         connect('mongoenginetest', alias='t1', tz_aware=True)
         conn = get_connection('t1')
 
@@ -197,6 +295,45 @@ class ConnectionTest(unittest.TestCase):
         conn = get_connection('t2')
         self.assertFalse(get_tz_awareness(conn))
 
+    def test_connection_pool_via_kwarg(self):
+        """Ensure we can specify a max connection pool size using
+        a connection kwarg.
+        """
+        # Use "max_pool_size" or "maxpoolsize" depending on PyMongo version
+        # (former was changed to the latter as described in
+        # https://jira.mongodb.org/browse/PYTHON-854).
+        # TODO remove once PyMongo < 3.0 support is dropped
+        if pymongo.version_tuple[0] >= 3:
+            pool_size_kwargs = {'maxpoolsize': 100}
+        else:
+            pool_size_kwargs = {'max_pool_size': 100}
+
+        conn = connect('mongoenginetest', alias='max_pool_size_via_kwarg', **pool_size_kwargs)
+        self.assertEqual(conn.max_pool_size, 100)
+
+    def test_connection_pool_via_uri(self):
+        """Ensure we can specify a max connection pool size using
+        an option in a connection URI.
+        """
+        if pymongo.version_tuple[0] == 2 and pymongo.version_tuple[1] < 9:
+            raise SkipTest('maxpoolsize as a URI option is only supported in PyMongo v2.9+')
+
+        conn = connect(host='mongodb://localhost/test?maxpoolsize=100', alias='max_pool_size_via_uri')
+        self.assertEqual(conn.max_pool_size, 100)
+
+    def test_write_concern(self):
+        """Ensure write concern can be specified in connect() via
+        a kwarg or as part of the connection URI.
+        """
+        conn1 = connect(alias='conn1', host='mongodb://localhost/testing?w=1&j=true')
+        conn2 = connect('testing', alias='conn2', w=1, j=True)
+        if IS_PYMONGO_3:
+            self.assertEqual(conn1.write_concern.document, {'w': 1, 'j': True})
+            self.assertEqual(conn2.write_concern.document, {'w': 1, 'j': True})
+        else:
+            self.assertEqual(dict(conn1.write_concern), {'w': 1, 'j': True})
+            self.assertEqual(dict(conn2.write_concern), {'w': 1, 'j': True})
+
     def test_datetime(self):
         connect('mongoenginetest', tz_aware=True)
         d = datetime.datetime(2010, 5, 5, tzinfo=utc)

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
@@ -12,9 +10,13 @@ from mongoengine.context_managers import query_counter
 
 class FieldTest(unittest.TestCase):
 
-    def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
+    @classmethod
+    def setUpClass(cls):
+        cls.db = connect(db='mongoenginetest')
+
+    @classmethod
+    def tearDownClass(cls):
+        cls.db.drop_database('mongoenginetest')
 
     def test_list_item_dereference(self):
         """Ensure that DBRef items in ListFields are dereferenced.
@@ -28,7 +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()
 
@@ -86,7 +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()
 
@@ -158,7 +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()
 
@@ -304,6 +306,7 @@ class FieldTest(unittest.TestCase):
 
         User.drop_collection()
         Post.drop_collection()
+        SimpleList.drop_collection()
 
         u1 = User.objects.create(name='u1')
         u2 = User.objects.create(name='u2')
@@ -435,7 +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()
 
@@ -526,7 +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()
 
@@ -609,15 +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:
@@ -682,7 +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()
 
@@ -694,9 +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:
@@ -778,16 +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:
@@ -861,7 +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()
 
@@ -873,9 +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:
@@ -1098,7 +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()
@@ -1127,7 +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()
@@ -1164,7 +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 *
@@ -25,6 +23,8 @@ class SignalTests(unittest.TestCase):
         connect(db='mongoenginetest')
 
         class Author(Document):
+            # Make the id deterministic for easier testing
+            id = SequenceField(primary_key=True)
             name = StringField()
 
             def __unicode__(self):
@@ -33,7 +33,7 @@ class SignalTests(unittest.TestCase):
             @classmethod
             def pre_init(cls, sender, document, *args, **kwargs):
                 signal_output.append('pre_init signal, %s' % cls.__name__)
-                signal_output.append(str(kwargs['values']))
+                signal_output.append(kwargs['values'])
 
             @classmethod
             def post_init(cls, sender, document, **kwargs):
@@ -43,48 +43,55 @@ class SignalTests(unittest.TestCase):
             @classmethod
             def pre_save(cls, sender, document, **kwargs):
                 signal_output.append('pre_save signal, %s' % document)
+                signal_output.append(kwargs)
 
             @classmethod
             def pre_save_post_validation(cls, sender, document, **kwargs):
                 signal_output.append('pre_save_post_validation signal, %s' % document)
-                if 'created' in kwargs:
-                    if kwargs['created']:
-                        signal_output.append('Is created')
-                    else:
-                        signal_output.append('Is updated')
+                if kwargs.pop('created', False):
+                    signal_output.append('Is created')
+                else:
+                    signal_output.append('Is updated')
+                signal_output.append(kwargs)
 
             @classmethod
             def post_save(cls, sender, document, **kwargs):
                 dirty_keys = document._delta()[0].keys() + document._delta()[1].keys()
                 signal_output.append('post_save signal, %s' % document)
                 signal_output.append('post_save dirty keys, %s' % dirty_keys)
-                if 'created' in kwargs:
-                    if kwargs['created']:
-                        signal_output.append('Is created')
-                    else:
-                        signal_output.append('Is updated')
+                if kwargs.pop('created', False):
+                    signal_output.append('Is created')
+                else:
+                    signal_output.append('Is updated')
+                signal_output.append(kwargs)
 
             @classmethod
             def pre_delete(cls, sender, document, **kwargs):
                 signal_output.append('pre_delete signal, %s' % document)
+                signal_output.append(kwargs)
 
             @classmethod
             def post_delete(cls, sender, document, **kwargs):
                 signal_output.append('post_delete signal, %s' % document)
+                signal_output.append(kwargs)
 
             @classmethod
             def pre_bulk_insert(cls, sender, documents, **kwargs):
                 signal_output.append('pre_bulk_insert signal, %s' % documents)
+                signal_output.append(kwargs)
 
             @classmethod
             def post_bulk_insert(cls, sender, documents, **kwargs):
                 signal_output.append('post_bulk_insert signal, %s' % documents)
-                if kwargs.get('loaded', False):
+                if kwargs.pop('loaded', False):
                     signal_output.append('Is loaded')
                 else:
                     signal_output.append('Not loaded')
+                signal_output.append(kwargs)
+
         self.Author = Author
         Author.drop_collection()
+        Author.id.set_next_value(0)
 
         class Another(Document):
 
@@ -96,10 +103,12 @@ class SignalTests(unittest.TestCase):
             @classmethod
             def pre_delete(cls, sender, document, **kwargs):
                 signal_output.append('pre_delete signal, %s' % document)
+                signal_output.append(kwargs)
 
             @classmethod
             def post_delete(cls, sender, document, **kwargs):
                 signal_output.append('post_delete signal, %s' % document)
+                signal_output.append(kwargs)
 
         self.Another = Another
         Another.drop_collection()
@@ -118,6 +127,41 @@ class SignalTests(unittest.TestCase):
         self.ExplicitId = ExplicitId
         ExplicitId.drop_collection()
 
+        class Post(Document):
+            title = StringField()
+            content = StringField()
+            active = BooleanField(default=False)
+
+            def __unicode__(self):
+                return self.title
+
+            @classmethod
+            def pre_bulk_insert(cls, sender, documents, **kwargs):
+                signal_output.append('pre_bulk_insert signal, %s' %
+                                     [(doc, {'active': documents[n].active})
+                                      for n, doc in enumerate(documents)])
+
+                # make changes here, this is just an example -
+                # it could be anything that needs pre-validation or looks-ups before bulk bulk inserting
+                for document in documents:
+                    if not document.active:
+                        document.active = True
+                signal_output.append(kwargs)
+
+            @classmethod
+            def post_bulk_insert(cls, sender, documents, **kwargs):
+                signal_output.append('post_bulk_insert signal, %s' %
+                                     [(doc, {'active': documents[n].active})
+                                      for n, doc in enumerate(documents)])
+                if kwargs.pop('loaded', False):
+                    signal_output.append('Is loaded')
+                else:
+                    signal_output.append('Not loaded')
+                signal_output.append(kwargs)
+
+        self.Post = Post
+        Post.drop_collection()
+
         # Save up the number of connected signals so that we can check at the
         # end that all the signals we register get properly unregistered
         self.pre_signals = (
@@ -147,6 +191,9 @@ class SignalTests(unittest.TestCase):
 
         signals.post_save.connect(ExplicitId.post_save, sender=ExplicitId)
 
+        signals.pre_bulk_insert.connect(Post.pre_bulk_insert, sender=Post)
+        signals.post_bulk_insert.connect(Post.post_bulk_insert, sender=Post)
+
     def tearDown(self):
         signals.pre_init.disconnect(self.Author.pre_init)
         signals.post_init.disconnect(self.Author.post_init)
@@ -163,6 +210,9 @@ class SignalTests(unittest.TestCase):
 
         signals.post_save.disconnect(self.ExplicitId.post_save)
 
+        signals.pre_bulk_insert.disconnect(self.Post.pre_bulk_insert)
+        signals.post_bulk_insert.disconnect(self.Post.post_bulk_insert)
+
         # Check that all our signals got disconnected properly.
         post_signals = (
             len(signals.pre_init.receivers),
@@ -199,66 +249,121 @@ class SignalTests(unittest.TestCase):
             a.save()
             self.get_signal_output(lambda: None) # eliminate signal output
             a1 = self.Author.objects(name='Bill Shakespeare')[0]
-        
+
         self.assertEqual(self.get_signal_output(create_author), [
             "pre_init signal, Author",
-            "{'name': 'Bill Shakespeare'}",
+            {'name': 'Bill Shakespeare'},
             "post_init signal, Bill Shakespeare, document._created = True",
         ])
 
         a1 = self.Author(name='Bill Shakespeare')
         self.assertEqual(self.get_signal_output(a1.save), [
             "pre_save signal, Bill Shakespeare",
+            {},
             "pre_save_post_validation signal, Bill Shakespeare",
             "Is created",
+            {},
             "post_save signal, Bill Shakespeare",
             "post_save dirty keys, ['name']",
-            "Is created"
+            "Is created",
+            {}
         ])
 
         a1.reload()
         a1.name = 'William Shakespeare'
         self.assertEqual(self.get_signal_output(a1.save), [
             "pre_save signal, William Shakespeare",
+            {},
             "pre_save_post_validation signal, William Shakespeare",
             "Is updated",
+            {},
             "post_save signal, William Shakespeare",
             "post_save dirty keys, ['name']",
-            "Is updated"
+            "Is updated",
+            {}
         ])
 
         self.assertEqual(self.get_signal_output(a1.delete), [
             'pre_delete signal, William Shakespeare',
+            {},
             'post_delete signal, William Shakespeare',
+            {}
         ])
 
-        signal_output = self.get_signal_output(load_existing_author)
-        # test signal_output lines separately, because of random ObjectID after object load
-        self.assertEqual(signal_output[0],
+        self.assertEqual(self.get_signal_output(load_existing_author), [
             "pre_init signal, Author",
-        )
-        self.assertEqual(signal_output[2],
-            "post_init signal, Bill Shakespeare, document._created = False",
-        )
+            {'id': 2, 'name': 'Bill Shakespeare'},
+            "post_init signal, Bill Shakespeare, document._created = False"
+        ])
 
-
-        signal_output = self.get_signal_output(bulk_create_author_with_load)
-
-        # The output of this signal is not entirely deterministic. The reloaded
-        # object will have an object ID. Hence, we only check part of the output
-        self.assertEqual(signal_output[3], "pre_bulk_insert signal, [<Author: Bill Shakespeare>]"
-        )
-        self.assertEqual(signal_output[-2:],
-            ["post_bulk_insert signal, [<Author: Bill Shakespeare>]",
-             "Is loaded",])
+        self.assertEqual(self.get_signal_output(bulk_create_author_with_load), [
+            'pre_init signal, Author',
+            {'name': 'Bill Shakespeare'},
+            'post_init signal, Bill Shakespeare, document._created = True',
+            'pre_bulk_insert signal, [<Author: Bill Shakespeare>]',
+            {},
+            'pre_init signal, Author',
+            {'id': 3, 'name': 'Bill Shakespeare'},
+            'post_init signal, Bill Shakespeare, document._created = False',
+            'post_bulk_insert signal, [<Author: Bill Shakespeare>]',
+            'Is loaded',
+            {}
+        ])
 
         self.assertEqual(self.get_signal_output(bulk_create_author_without_load), [
             "pre_init signal, Author",
-            "{'name': 'Bill Shakespeare'}",
+            {'name': 'Bill Shakespeare'},
             "post_init signal, Bill Shakespeare, document._created = True",
             "pre_bulk_insert signal, [<Author: Bill Shakespeare>]",
+            {},
             "post_bulk_insert signal, [<Author: Bill Shakespeare>]",
             "Not loaded",
+            {}
+        ])
+
+    def test_signal_kwargs(self):
+        """ Make sure signal_kwargs is passed to signals calls. """
+
+        def live_and_let_die():
+            a = self.Author(name='Bill Shakespeare')
+            a.save(signal_kwargs={'live': True, 'die': False})
+            a.delete(signal_kwargs={'live': False, 'die': True})
+
+        self.assertEqual(self.get_signal_output(live_and_let_die), [
+            "pre_init signal, Author",
+            {'name': 'Bill Shakespeare'},
+            "post_init signal, Bill Shakespeare, document._created = True",
+            "pre_save signal, Bill Shakespeare",
+            {'die': False, 'live': True},
+            "pre_save_post_validation signal, Bill Shakespeare",
+            "Is created",
+            {'die': False, 'live': True},
+            "post_save signal, Bill Shakespeare",
+            "post_save dirty keys, ['name']",
+            "Is created",
+            {'die': False, 'live': True},
+            'pre_delete signal, Bill Shakespeare',
+            {'die': True, 'live': False},
+            'post_delete signal, Bill Shakespeare',
+            {'die': True, 'live': False}
+        ])
+
+        def bulk_create_author():
+            a1 = self.Author(name='Bill Shakespeare')
+            self.Author.objects.insert([a1], signal_kwargs={'key': True})
+
+        self.assertEqual(self.get_signal_output(bulk_create_author), [
+            'pre_init signal, Author',
+            {'name': 'Bill Shakespeare'},
+            'post_init signal, Bill Shakespeare, document._created = True',
+            'pre_bulk_insert signal, [<Author: Bill Shakespeare>]',
+            {'key': True},
+            'pre_init signal, Author',
+            {'id': 2, 'name': 'Bill Shakespeare'},
+            'post_init signal, Bill Shakespeare, document._created = False',
+            'post_bulk_insert signal, [<Author: Bill Shakespeare>]',
+            'Is loaded',
+            {'key': True}
         ])
 
     def test_queryset_delete_signals(self):
@@ -267,7 +372,9 @@ class SignalTests(unittest.TestCase):
         self.Another(name='Bill Shakespeare').save()
         self.assertEqual(self.get_signal_output(self.Another.objects.delete), [
             'pre_delete signal, Bill Shakespeare',
+            {},
             'post_delete signal, Bill Shakespeare',
+            {}
         ])
 
     def test_signals_with_explicit_doc_ids(self):
@@ -306,6 +413,23 @@ class SignalTests(unittest.TestCase):
         ei.switch_db("testdb-1", keep_created=False)
         self.assertEqual(self.get_signal_output(ei.save), ['Is created'])
 
+    def test_signals_bulk_insert(self):
+        def bulk_set_active_post():
+            posts = [
+                self.Post(title='Post 1'),
+                self.Post(title='Post 2'),
+                self.Post(title='Post 3')
+            ]
+            self.Post.objects.insert(posts)
+
+        results = self.get_signal_output(bulk_set_active_post)
+        self.assertEqual(results, [
+            "pre_bulk_insert signal, [(<Post: Post 1>, {'active': False}), (<Post: Post 2>, {'active': False}), (<Post: Post 3>, {'active': False})]",
+            {},
+            "post_bulk_insert signal, [(<Post: Post 1>, {'active': True}), (<Post: Post 2>, {'active': True}), (<Post: Post 3>, {'active': True})]",
+            'Is loaded',
+            {}
+        ])
 
 if __name__ == '__main__':
     unittest.main()

tests/utils.py

@@ -0,0 +1,22 @@
+import unittest
+
+from mongoengine import connect
+from mongoengine.connection import get_db
+
+MONGO_TEST_DB = 'mongoenginetest'
+
+
+class MongoDBTestCase(unittest.TestCase):
+    """Base class for tests that need a mongodb connection
+    db is being dropped automatically
+    """
+
+    @classmethod
+    def setUpClass(cls):
+        cls._connection = connect(db=MONGO_TEST_DB)
+        cls._connection.drop_database(MONGO_TEST_DB)
+        cls.db = get_db()
+
+    @classmethod
+    def tearDownClass(cls):
+        cls._connection.drop_database(MONGO_TEST_DB)

tox.ini

@@ -1,14 +1,22 @@
 [tox]
-envlist = {py26,py27,py32,py33,py34,py35,pypy,pypy3}-{mg27,mg28}
-#envlist = {py26,py27,py32,py33,py34,pypy,pypy3}-{mg27,mg28,mg30,mgdev}
+envlist = {py26,py27,py33,py34,py35,pypy,pypy3}-{mg27,mg28},flake8
 
 [testenv]
 commands =
     python setup.py nosetests {posargs}
 deps =
     nose
-    rednose
     mg27: PyMongo<2.8
     mg28: PyMongo>=2.8,<3.0
     mg30: PyMongo>=3.0
     mgdev: https://github.com/mongodb/mongo-python-driver/tarball/master
+setenv =
+    PYTHON_EGG_CACHE = {envdir}/python-eggs
+passenv = windir
+
+[testenv:flake8]
+deps =
+    flake8
+    flake8-import-order
+commands =
+   flake8