Merge pull request #1960 from bagerard/bump_version_0_16_2

bump version to 0.16.2

.install_mongodb_on_travis.sh

@@ -1,23 +1,28 @@
 #!/bin/bash
 
+sudo apt-get remove mongodb-org-server
 sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv 7F0CEB10
 
-if [ "$MONGODB" = "2.4" ]; then
-    echo "deb http://downloads-distro.mongodb.org/repo/ubuntu-upstart dist 10gen" | sudo tee /etc/apt/sources.list.d/mongodb.list
-    sudo apt-get update
-    sudo apt-get install mongodb-10gen=2.4.14
-    sudo service mongodb start
-elif [ "$MONGODB" = "2.6" ]; then
+if [ "$MONGODB" = "2.6" ]; then
     echo "deb http://downloads-distro.mongodb.org/repo/ubuntu-upstart dist 10gen" | sudo tee /etc/apt/sources.list.d/mongodb.list
     sudo apt-get update
     sudo apt-get install mongodb-org-server=2.6.12
     # service should be started automatically
 elif [ "$MONGODB" = "3.0" ]; then
-    echo "deb http://repo.mongodb.org/apt/ubuntu precise/mongodb-org/3.0 multiverse" | sudo tee /etc/apt/sources.list.d/mongodb.list
+    echo "deb http://repo.mongodb.org/apt/ubuntu trusty/mongodb-org/3.0 multiverse" | sudo tee /etc/apt/sources.list.d/mongodb.list
     sudo apt-get update
     sudo apt-get install mongodb-org-server=3.0.14
     # service should be started automatically
+elif [ "$MONGODB" = "3.2" ]; then
+    sudo apt-key adv --keyserver keyserver.ubuntu.com --recv EA312927
+    echo "deb http://repo.mongodb.org/apt/ubuntu trusty/mongodb-org/3.2 multiverse" | sudo tee /etc/apt/sources.list.d/mongodb-org-3.2.list
+    sudo apt-get update
+    sudo apt-get install mongodb-org-server=3.2.20
+    # service should be started automatically
 else
-    echo "Invalid MongoDB version, expected 2.4, 2.6, or 3.0."
+    echo "Invalid MongoDB version, expected 2.6, 3.0, or 3.2"
     exit 1
 fi;
+
+mkdir db
+1>db/logs mongod --dbpath=db &

.travis.yml

@@ -2,12 +2,10 @@
 # PyMongo combinations. However, that would result in an overly long build
 # with a very large number of jobs, hence we only test a subset of all the
 # combinations:
-# * MongoDB v2.4 & v3.0 are only tested against Python v2.7 & v3.5.
-# * MongoDB v2.4 is tested against PyMongo v2.7 & v3.x.
-# * MongoDB v3.0 is tested against PyMongo v3.x.
 # * MongoDB v2.6 is currently the "main" version tested against Python v2.7,
-#   v3.5, PyPy & PyPy3, and PyMongo v2.7, v2.8 & v3.x.
-#
+#   v3.5, v3.6, PyPy, and PyMongo v3.x.
+# * MongoDB v3.0 & v3.2 are tested against Python v2.7, v3.5 & v3.6
+#   and Pymongo v3.5 & v3.x
 # Reminder: Update README.rst if you change MongoDB versions we test.
 
 language: python
@@ -15,12 +13,11 @@ language: python
 python:
 - 2.7
 - 3.5
+- 3.6
 - pypy
 
 env:
-- MONGODB=2.6 PYMONGO=2.7
-- MONGODB=2.6 PYMONGO=2.8
-- MONGODB=2.6 PYMONGO=3.0
+- MONGODB=2.6 PYMONGO=3.x
 
 matrix:
   # Finish the build as soon as one job fails
@@ -28,20 +25,22 @@ matrix:
 
   include:
   - python: 2.7
-    env: MONGODB=2.4 PYMONGO=2.7
+    env: MONGODB=3.0 PYMONGO=3.5
   - python: 2.7
-    env: MONGODB=2.4 PYMONGO=3.0
-  - python: 2.7
-    env: MONGODB=3.0 PYMONGO=3.0
+    env: MONGODB=3.2 PYMONGO=3.x
   - python: 3.5
-    env: MONGODB=2.4 PYMONGO=2.7
+    env: MONGODB=3.0 PYMONGO=3.5
   - python: 3.5
-    env: MONGODB=2.4 PYMONGO=3.0
-  - python: 3.5
-    env: MONGODB=3.0 PYMONGO=3.0
+    env: MONGODB=3.2 PYMONGO=3.x
+  - python: 3.6
+    env: MONGODB=3.0 PYMONGO=3.5
+  - python: 3.6
+    env: MONGODB=3.2 PYMONGO=3.x
 
 before_install:
 - bash .install_mongodb_on_travis.sh
+- sleep 15  # https://docs.travis-ci.com/user/database-setup/#MongoDB-does-not-immediately-accept-connections
+- mongo --eval 'db.version();'
 
 install:
 - sudo apt-get install python-dev python3-dev libopenjpeg-dev zlib1g-dev libjpeg-turbo8-dev
@@ -90,11 +89,11 @@ deploy:
   distributions: "sdist bdist_wheel"
 
   # only deploy on tagged commits (aka GitHub releases) and only for the
-  # parent repo's builds running Python 2.7 along with dev PyMongo (we run
+  # parent repo's builds running Python 2.7 along with PyMongo v3.x (we run
   # Travis against many different Python and PyMongo versions and we don't
   # want the deploy to occur multiple times).
   on:
     tags: true
     repo: MongoEngine/mongoengine
-    condition: "$PYMONGO = 3.0"
+    condition: "$PYMONGO = 3.x"
     python: 2.7

AUTHORS

@@ -243,3 +243,7 @@ that much better:
  * Victor Varvaryuk
  * Stanislav Kaledin (https://github.com/sallyruthstruik)
  * Dmitry Yantsen (https://github.com/mrTable)
+ * Renjianxin (https://github.com/Davidrjx)
+ * Erdenezul Batmunkh (https://github.com/erdenezul)
+ * Andy Yankovsky (https://github.com/werat)
+ * Bastien Gérard (https://github.com/bagerard)

CONTRIBUTING.rst

@@ -22,8 +22,11 @@ Supported Interpreters
 
 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.
+The codebase is written in python 2 so you must be using python 2
+when developing new features. Compatibility of the library with Python 3
+relies on the 2to3 package that gets executed as part of the installation
+build. You should ensure that your code is properly converted by
+`2to3 <http://docs.python.org/library/2to3.html>`_.
 
 Style Guide
 -----------

README.rst

@@ -26,19 +26,21 @@ an `API reference <https://mongoengine-odm.readthedocs.io/apireference.html>`_.
 
 Supported MongoDB Versions
 ==========================
-MongoEngine is currently tested against MongoDB v2.4, v2.6, and v3.0. Future
+MongoEngine is currently tested against MongoDB v2.6, v3.0 and v3.2. Future
 versions should be supported as well, but aren't actively tested at the moment.
 Make sure to open an issue or submit a pull request if you experience any
-problems with MongoDB v3.2+.
+problems with MongoDB v3.4+.
 
 Installation
 ============
 We recommend the use of `virtualenv <https://virtualenv.pypa.io/>`_ and of
 `pip <https://pip.pypa.io/>`_. You can then use ``pip install -U mongoengine``.
 You may also have `setuptools <http://peak.telecommunity.com/DevCenter/setuptools>`_
-and thus you can use ``easy_install -U mongoengine``. Otherwise, you can download the
-source from `GitHub <http://github.com/MongoEngine/mongoengine>`_ and run ``python
-setup.py install``.
+and thus you can use ``easy_install -U mongoengine``. Another option is 
+`pipenv <https://docs.pipenv.org/>`_. You can then use ``pipenv install mongoengine``
+to both create the virtual environment and install the package. Otherwise, you can
+download the source from `GitHub <http://github.com/MongoEngine/mongoengine>`_ and
+run ``python setup.py install``.
 
 Dependencies
 ============

docs/apireference.rst

@@ -87,7 +87,9 @@ Fields
 .. autoclass:: mongoengine.fields.DictField
 .. autoclass:: mongoengine.fields.MapField
 .. autoclass:: mongoengine.fields.ReferenceField
+.. autoclass:: mongoengine.fields.LazyReferenceField
 .. autoclass:: mongoengine.fields.GenericReferenceField
+.. autoclass:: mongoengine.fields.GenericLazyReferenceField
 .. autoclass:: mongoengine.fields.CachedReferenceField
 .. autoclass:: mongoengine.fields.BinaryField
 .. autoclass:: mongoengine.fields.FileField

docs/changelog.rst

@@ -6,6 +6,81 @@ Development
 ===========
 - (Fill this out as you fix issues and develop your features).
 
+=================
+Changes in 0.16.2
+=================
+- Fix .save() that fails when called with write_concern=None (regression of 0.16.1) #1958
+
+=================
+Changes in 0.16.1
+=================
+- Fix `_cls` that is not set properly in Document constructor (regression) #1950
+- Fix bug in _delta method - Update of a ListField depends on an unrelated dynamic field update #1733
+- Remove deprecated `save()` method and used `insert_one()` #1899
+
+=================
+Changes in 0.16.0
+=================
+- Various improvements to the doc
+- Improvement to code quality
+- POTENTIAL BREAKING CHANGES:
+    - EmbeddedDocumentField will no longer accept references to Document classes in its constructor #1661
+    - Get rid of the `basecls` parameter from the DictField constructor (dead code) #1876
+    - default value of ComplexDateTime is now None (and no longer the current datetime) #1368
+- Fix unhashable TypeError when referencing a Document with a compound key in an EmbeddedDocument #1685
+- Fix bug where an EmbeddedDocument with the same id as its parent would not be tracked for changes #1768
+- Fix the fact that bulk `insert()` was not setting primary keys of inserted documents instances #1919
+- Fix bug when referencing the abstract class in a ReferenceField #1920
+- Allow modification to the document made in pre_save_post_validation to be taken into account #1202
+- Replaced MongoDB 2.4 tests in CI by MongoDB 3.2 #1903
+- Fix side effects of using queryset.`no_dereference` on other documents #1677
+- Fix TypeError when using lazy django translation objects as translated choices #1879
+- Improve 2-3 codebase compatibility #1889
+- Fix the support for changing the default value of ComplexDateTime #1368
+- Improves error message in case an EmbeddedDocumentListField receives an EmbeddedDocument instance
+    instead of a list #1877
+- Fix the Decimal operator inc/dec #1517 #1320
+- Ignore killcursors queries in `query_counter` context manager #1869
+- Fix the fact that `query_counter` was modifying the initial profiling_level in case it was != 0 #1870
+- Repaired the `no_sub_classes` context manager + fix the fact that it was swallowing exceptions #1865
+- Fix index creation error that was swallowed by hasattr under python2 #1688
+- QuerySet limit function behaviour: Passing 0 as parameter will return all the documents in the cursor #1611
+- bulk insert updates the ids of the input documents instances #1919
+- Fix an harmless bug related to GenericReferenceField where modifications in the generic-referenced document
+    were tracked in the parent #1934
+- Improve validator of BinaryField #273
+- Implemented lazy regex compiling in Field classes to improve 'import mongoengine' performance #1806
+- Updated GridFSProxy.__str__  so that it would always print both the filename and grid_id #710
+- Add __repr__ to Q and QCombination #1843
+- fix bug in BaseList.__iter__ operator (was occuring when modifying a BaseList while iterating over it) #1676
+- Added field `DateField`#513
+
+Changes in 0.15.3
+=================
+-  Subfield resolve error in generic_emdedded_document query #1651 #1652
+-  use each modifier only with $position #1673 #1675
+-  Improve LazyReferenceField and GenericLazyReferenceField with nested fields #1704
+-  Fix validation error instance in GenericEmbeddedDocumentField #1067
+-  Update cached fields when fields argument is given #1712
+-  Add a db parameter to register_connection for compatibility with connect
+-  Use insert_one, insert_many in Document.insert #1491
+-  Use new update_one, update_many on document/queryset update #1491
+-  Use insert_one, insert_many in Document.insert #1491
+-  Fix reload(fields) affect changed fields #1371
+-  Fix Read-only access to database fails when trying to create indexes #1338
+
+Changes in 0.15.0
+=================
+- Add LazyReferenceField and GenericLazyReferenceField to address #1230
+
+Changes in 0.14.1
+=================
+- Removed SemiStrictDict and started using a regular dict for `BaseDocument._data` #1630
+- Added support for the `$position` param in the `$push` operator #1566
+- Fixed `DateTimeField` interpreting an empty string as today #1533
+- Added a missing `__ne__` method to the `GridFSProxy` class #1632
+- Fixed `BaseQuerySet._fields_to_db_fields` #1553
+
 Changes in 0.14.0
 =================
 - BREAKING CHANGE: Removed the `coerce_types` param from `QuerySet.as_pymongo` #1549

docs/code/tumblelog.py

@@ -45,27 +45,27 @@ post2.link_url = 'http://tractiondigital.com/labs/mongoengine/docs'
 post2.tags = ['mongoengine']
 post2.save()
 
-print 'ALL POSTS'
-print
+print('ALL POSTS')
+print()
 for post in Post.objects:
-    print post.title
+    print(post.title)
     #print '=' * post.title.count()
-    print "=" * 20
+    print("=" * 20)
 
     if isinstance(post, TextPost):
-        print post.content
+        print(post.content)
 
     if isinstance(post, LinkPost):
-        print 'Link:', post.link_url
+        print('Link:', post.link_url)
 
-    print
-print
+    print()
+print()
 
-print 'POSTS TAGGED \'MONGODB\''
-print
+print('POSTS TAGGED \'MONGODB\'')
+print()
 for post in Post.objects(tags='mongodb'):
-    print post.title
-print
+    print(post.title)
+print()
 
 num_posts = Post.objects(tags='mongodb').count()
-print 'Found %d posts with tag "mongodb"' % num_posts
+print('Found %d posts with tag "mongodb"' % num_posts)

docs/guide/connecting.rst

@@ -18,10 +18,10 @@ provide the :attr:`host` and :attr:`port` arguments to
 
     connect('project1', host='192.168.1.35', port=12345)
 
-If the database requires authentication, :attr:`username` and :attr:`password`
-arguments should be provided::
+If the database requires authentication, :attr:`username`, :attr:`password`
+and :attr:`authentication_source` arguments should be provided::
 
-    connect('project1', username='webapp', password='pwd123')
+    connect('project1', username='webapp', password='pwd123', authentication_source='admin')
 
 URI style connections are also supported -- just supply the URI as
 the :attr:`host` to

docs/guide/defining-documents.rst

@@ -22,7 +22,7 @@ objects** as class attributes to the document class::
 
     class Page(Document):
         title = StringField(max_length=200, required=True)
-        date_modified = DateTimeField(default=datetime.datetime.now)
+        date_modified = DateTimeField(default=datetime.datetime.utcnow)
 
 As BSON (the binary format for storing data in mongodb) is order dependent,
 documents are serialized based on their field order.
@@ -80,6 +80,7 @@ are as follows:
 * :class:`~mongoengine.fields.FloatField`
 * :class:`~mongoengine.fields.GenericEmbeddedDocumentField`
 * :class:`~mongoengine.fields.GenericReferenceField`
+* :class:`~mongoengine.fields.GenericLazyReferenceField`
 * :class:`~mongoengine.fields.GeoPointField`
 * :class:`~mongoengine.fields.ImageField`
 * :class:`~mongoengine.fields.IntField`
@@ -87,6 +88,7 @@ are as follows:
 * :class:`~mongoengine.fields.MapField`
 * :class:`~mongoengine.fields.ObjectIdField`
 * :class:`~mongoengine.fields.ReferenceField`
+* :class:`~mongoengine.fields.LazyReferenceField`
 * :class:`~mongoengine.fields.SequenceField`
 * :class:`~mongoengine.fields.SortedListField`
 * :class:`~mongoengine.fields.StringField`
@@ -224,7 +226,7 @@ store; in this situation a :class:`~mongoengine.fields.DictField` is appropriate
         user = ReferenceField(User)
         answers = DictField()
 
-    survey_response = SurveyResponse(date=datetime.now(), user=request.user)
+    survey_response = SurveyResponse(date=datetime.utcnow(), user=request.user)
     response_form = ResponseForm(request.POST)
     survey_response.answers = response_form.cleaned_data()
     survey_response.save()
@@ -490,7 +492,9 @@ the field name with a **#**::
             ]
         }
 
-If a dictionary is passed then the following options are available:
+If a dictionary is passed then additional options become available. Valid options include,
+but are not limited to:
+
 
 :attr:`fields` (Default: None)
     The fields to index. Specified in the same format as described above.
@@ -511,8 +515,15 @@ If a dictionary is passed then the following options are available:
     Allows you to automatically expire data from a collection by setting the
     time in seconds to expire the a field.
 
+:attr:`name` (Optional)
+    Allows you to specify a name for the index
+
+:attr:`collation` (Optional)
+    Allows to create case insensitive indexes (MongoDB v3.4+ only)
+
 .. note::
 
+    Additional options are forwarded as **kwargs to pymongo's create_index method.
     Inheritance adds extra fields indices see: :ref:`document-inheritance`.
 
 Global index default options
@@ -524,15 +535,16 @@ There are a few top level defaults for all indexes that can be set::
         title = StringField()
         rating = StringField()
         meta = {
-            'index_options': {},
+            'index_opts': {},
             'index_background': True,
+            'index_cls': False,
+            'auto_create_index': True,
             'index_drop_dups': True,
-            'index_cls': False
         }
 
 
-:attr:`index_options` (Optional)
-    Set any default index options - see the `full options list <http://docs.mongodb.org/manual/reference/method/db.collection.ensureIndex/#db.collection.ensureIndex>`_
+:attr:`index_opts` (Optional)
+    Set any default index options - see the `full options list <https://docs.mongodb.com/manual/reference/method/db.collection.createIndex/#db.collection.createIndex>`_
 
 :attr:`index_background` (Optional)
     Set the default value for if an index should be indexed in the background
@@ -540,10 +552,15 @@ There are a few top level defaults for all indexes that can be set::
 :attr:`index_cls` (Optional)
     A way to turn off a specific index for _cls.
 
+:attr:`auto_create_index` (Optional)
+    When this is True (default), MongoEngine will ensure that the correct
+    indexes exist in MongoDB each time a command is run. This can be disabled
+    in systems where indexes are managed separately. Disabling this will improve
+    performance.
+
 :attr:`index_drop_dups` (Optional)
     Set the default value for if an index should drop duplicates
-
-.. note:: Since MongoDB 3.0 drop_dups is not supported anymore. Raises a Warning
+    Since MongoDB 3.0 drop_dups is not supported anymore. Raises a Warning
     and has no effect
 
 
@@ -618,7 +635,7 @@ collection after a given period. See the official
 documentation for more information.  A common usecase might be session data::
 
     class Session(Document):
-        created = DateTimeField(default=datetime.now)
+        created = DateTimeField(default=datetime.utcnow)
         meta = {
             'indexes': [
                 {'fields': ['created'], 'expireAfterSeconds': 3600}
@@ -725,6 +742,9 @@ document.::
 .. note:: From 0.8 onwards :attr:`allow_inheritance` defaults
           to False, meaning you must set it to True to use inheritance.
 
+          Setting :attr:`allow_inheritance` to True should also be used in
+          :class:`~mongoengine.EmbeddedDocument` class in case you need to subclass it
+
 Working with existing data
 --------------------------
 As MongoEngine no longer defaults to needing :attr:`_cls`, you can quickly and

docs/guide/document-instances.rst

@@ -57,7 +57,8 @@ document values for example::
 
         def clean(self):
             """Ensures that only published essays have a `pub_date` and
-            automatically sets the pub_date if published and not set"""
+            automatically sets `pub_date` if essay is published and `pub_date`
+            is not set"""
             if self.status == 'Draft' and self.pub_date is not None:
                 msg = 'Draft entries should not have a publication date.'
                 raise ValidationError(msg)

docs/guide/gridfs.rst

@@ -53,7 +53,8 @@ Deletion
 
 Deleting stored files is achieved with the :func:`delete` method::
 
-    marmot.photo.delete()
+    marmot.photo.delete()    # Deletes the GridFS document
+    marmot.save()            # Saves the GridFS reference (being None) contained in the marmot instance
 
 .. warning::
 
@@ -71,4 +72,5 @@ Files can be replaced with the :func:`replace` method. This works just like
 the :func:`put` method so even metadata can (and should) be replaced::
 
     another_marmot = open('another_marmot.png', 'rb')
-    marmot.photo.replace(another_marmot, content_type='image/png')
+    marmot.photo.replace(another_marmot, content_type='image/png')  # Replaces the GridFS document
+    marmot.save()                                                   # Replaces the GridFS reference contained in marmot instance

docs/guide/querying.rst

@@ -456,14 +456,14 @@ data. To turn off dereferencing of the results of a query use
 :func:`~mongoengine.queryset.QuerySet.no_dereference` on the queryset like so::
 
     post = Post.objects.no_dereference().first()
-    assert(isinstance(post.author, ObjectId))
+    assert(isinstance(post.author, DBRef))
 
 You can also turn off all dereferencing for a fixed period by using the
 :class:`~mongoengine.context_managers.no_dereference` context manager::
 
     with no_dereference(Post) as Post:
         post = Post.objects.first()
-        assert(isinstance(post.author, ObjectId))
+        assert(isinstance(post.author, DBRef))
 
     # Outside the context manager dereferencing occurs.
     assert(isinstance(post.author, User))
@@ -565,6 +565,15 @@ cannot use the `$` syntax in keyword arguments it has been mapped to `S`::
     >>> post.tags
     ['database', 'mongodb']
 
+From MongoDB version 2.6, push operator supports $position value which allows
+to push values with index.
+    >>> post = BlogPost(title="Test", tags=["mongo"])
+    >>> post.save()
+    >>> post.update(push__tags__0=["database", "code"])
+    >>> post.reload()
+    >>> post.tags
+    ['database', 'code', 'mongo']
+
 .. note::
     Currently only top level lists are handled, future versions of mongodb /
     pymongo plan to support nested positional operators.  See `The $ positional

docs/guide/signals.rst

@@ -43,10 +43,10 @@ Available signals include:
   has taken place but before saving.
 
 `post_save`
-  Called within :meth:`~mongoengine.Document.save` after all actions
-  (validation, insert/update, cascades, clearing dirty flags) have completed
-  successfully.  Passed the additional boolean keyword argument `created` to
-  indicate if the save was an insert or an update.
+  Called within :meth:`~mongoengine.Document.save` after most actions
+  (validation, insert/update, and cascades, but not clearing dirty flags) have 
+  completed successfully.  Passed the additional boolean keyword argument 
+  `created` to indicate if the save was an insert or an update.
 
 `pre_delete`
   Called within :meth:`~mongoengine.Document.delete` prior to
@@ -113,6 +113,10 @@ handlers within your subclass::
     signals.pre_save.connect(Author.pre_save, sender=Author)
     signals.post_save.connect(Author.post_save, sender=Author)
 
+.. warning::
+
+    Note that EmbeddedDocument only supports pre/post_init signals. pre/post_save, etc should be attached to Document's class only. Attaching pre_save to an EmbeddedDocument is ignored silently.
+
 Finally, you can also use this small decorator to quickly create a number of
 signals and attach them to your :class:`~mongoengine.Document` or
 :class:`~mongoengine.EmbeddedDocument` subclasses as class decorators::

docs/guide/text-indexes.rst

@@ -48,4 +48,4 @@ Ordering by text score
 
 ::
 
-  objects = News.objects.search('mongo').order_by('$text_score')
+  objects = News.objects.search_text('mongo').order_by('$text_score')

docs/tutorial.rst

@@ -86,7 +86,7 @@ of them stand out as particularly intuitive solutions.
 Posts
 ^^^^^
 
-Happily mongoDB *isn't* a relational database, so we're not going to do it that
+Happily MongoDB *isn't* a relational database, so we're not going to do it that
 way. As it turns out, we can use MongoDB's schemaless nature to provide us with
 a much nicer solution. We will store all of the posts in *one collection* and
 each post type will only store the fields it needs. If we later want to add
@@ -153,7 +153,7 @@ post. This works, but there is no real reason to be storing the comments
 separately from their associated posts, other than to work around the
 relational model. Using MongoDB we can store the comments as a list of
 *embedded documents* directly on a post document. An embedded document should
-be treated no differently that a regular document; it just doesn't have its own
+be treated no differently than a regular document; it just doesn't have its own
 collection in the database. Using MongoEngine, we can define the structure of
 embedded documents, along with utility methods, in exactly the same way we do
 with regular documents::

docs/upgrade.rst

@@ -6,6 +6,11 @@ Development
 ***********
 (Fill this out whenever you introduce breaking changes to MongoEngine)
 
+URLField's constructor no longer takes `verify_exists`
+
+0.15.0
+******
+
 0.14.0
 ******
 This release includes a few bug fixes and a significant code cleanup. The most

mongoengine/__init__.py

@@ -23,7 +23,7 @@ __all__ = (list(document.__all__) + list(fields.__all__) +
            list(signals.__all__) + list(errors.__all__))
 
 
-VERSION = (0, 14, 0)
+VERSION = (0, 16, 2)
 
 
 def get_version():

mongoengine/base/__init__.py

@@ -15,7 +15,7 @@ __all__ = (
     'UPDATE_OPERATORS', '_document_registry', 'get_document',
 
     # datastructures
-    'BaseDict', 'BaseList', 'EmbeddedDocumentList',
+    'BaseDict', 'BaseList', 'EmbeddedDocumentList', 'LazyReference',
 
     # document
     'BaseDocument',

mongoengine/base/common.py

@@ -3,9 +3,10 @@ from mongoengine.errors import NotRegistered
 __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'])
+UPDATE_OPERATORS = {'set', 'unset', 'inc', 'dec', 'mul',
+                    'pop', 'push', 'push_all', 'pull',
+                    'pull_all', 'add_to_set', 'set_on_insert',
+                    'min', 'max', 'rename'}
 
 
 _document_registry = {}
@@ -18,7 +19,7 @@ def get_document(name):
         # Possible old style name
         single_end = name.split('.')[-1]
         compound_end = '.%s' % single_end
-        possible_match = [k for k in _document_registry.keys()
+        possible_match = [k for k in _document_registry
                           if k.endswith(compound_end) or k == single_end]
         if len(possible_match) == 1:
             doc = _document_registry.get(possible_match.pop(), None)

mongoengine/base/datastructures.py

@@ -1,12 +1,12 @@
-import itertools
 import weakref
 
+from bson import DBRef
 import six
 
 from mongoengine.common import _import_class
 from mongoengine.errors import DoesNotExist, MultipleObjectsReturned
 
-__all__ = ('BaseDict', 'BaseList', 'EmbeddedDocumentList')
+__all__ = ('BaseDict', 'BaseList', 'EmbeddedDocumentList', 'LazyReference')
 
 
 class BaseDict(dict):
@@ -17,10 +17,9 @@ class BaseDict(dict):
     _name = None
 
     def __init__(self, dict_items, instance, name):
-        Document = _import_class('Document')
-        EmbeddedDocument = _import_class('EmbeddedDocument')
+        BaseDocument = _import_class('BaseDocument')
 
-        if isinstance(instance, (Document, EmbeddedDocument)):
+        if isinstance(instance, BaseDocument):
             self._instance = weakref.proxy(instance)
         self._name = name
         super(BaseDict, self).__init__(dict_items)
@@ -31,11 +30,11 @@ class BaseDict(dict):
         EmbeddedDocument = _import_class('EmbeddedDocument')
         if isinstance(value, EmbeddedDocument) and value._instance is None:
             value._instance = self._instance
-        elif not isinstance(value, BaseDict) and isinstance(value, dict):
+        elif isinstance(value, dict) and not isinstance(value, BaseDict):
             value = BaseDict(value, None, '%s.%s' % (self._name, key))
             super(BaseDict, self).__setitem__(key, value)
             value._instance = self._instance
-        elif not isinstance(value, BaseList) and isinstance(value, list):
+        elif isinstance(value, list) and not isinstance(value, BaseList):
             value = BaseList(value, None, '%s.%s' % (self._name, key))
             super(BaseDict, self).__setitem__(key, value)
             value._instance = self._instance
@@ -102,10 +101,9 @@ class BaseList(list):
     _name = None
 
     def __init__(self, list_items, instance, name):
-        Document = _import_class('Document')
-        EmbeddedDocument = _import_class('EmbeddedDocument')
+        BaseDocument = _import_class('BaseDocument')
 
-        if isinstance(instance, (Document, EmbeddedDocument)):
+        if isinstance(instance, BaseDocument):
             self._instance = weakref.proxy(instance)
         self._name = name
         super(BaseList, self).__init__(list_items)
@@ -116,19 +114,19 @@ class BaseList(list):
         EmbeddedDocument = _import_class('EmbeddedDocument')
         if isinstance(value, EmbeddedDocument) and value._instance is None:
             value._instance = self._instance
-        elif not isinstance(value, BaseDict) and isinstance(value, dict):
+        elif isinstance(value, dict) and not isinstance(value, BaseDict):
             value = BaseDict(value, None, '%s.%s' % (self._name, key))
             super(BaseList, self).__setitem__(key, value)
             value._instance = self._instance
-        elif not isinstance(value, BaseList) and isinstance(value, list):
+        elif isinstance(value, list) and not isinstance(value, BaseList):
             value = BaseList(value, None, '%s.%s' % (self._name, key))
             super(BaseList, self).__setitem__(key, value)
             value._instance = self._instance
         return value
 
     def __iter__(self):
-        for i in xrange(self.__len__()):
-            yield self[i]
+        for v in super(BaseList, self).__iter__():
+            yield v
 
     def __setitem__(self, key, value, *args, **kwargs):
         if isinstance(key, slice):
@@ -137,7 +135,7 @@ class BaseList(list):
             self._mark_as_changed(key)
         return super(BaseList, self).__setitem__(key, value)
 
-    def __delitem__(self, key, *args, **kwargs):
+    def __delitem__(self, key):
         self._mark_as_changed()
         return super(BaseList, self).__delitem__(key)
 
@@ -186,7 +184,7 @@ class BaseList(list):
         self._mark_as_changed()
         return super(BaseList, self).remove(*args, **kwargs)
 
-    def reverse(self, *args, **kwargs):
+    def reverse(self):
         self._mark_as_changed()
         return super(BaseList, self).reverse()
 
@@ -233,6 +231,9 @@ class EmbeddedDocumentList(BaseList):
         Filters the list by only including embedded documents with the
         given keyword arguments.
 
+        This method only supports simple comparison (e.g: .filter(name='John Doe'))
+        and does not support operators like __gte, __lte, __icontains like queryset.filter does
+
         :param kwargs: The keyword arguments corresponding to the fields to
          filter on. *Multiple arguments are treated as if they are ANDed
          together.*
@@ -350,7 +351,8 @@ class EmbeddedDocumentList(BaseList):
 
     def update(self, **update):
         """
-        Updates the embedded documents with the given update values.
+        Updates the embedded documents with the given replacement values. This
+        function does not support mongoDB update operators such as ``inc__``.
 
         .. note::
             The embedded document changes are not automatically saved
@@ -372,7 +374,7 @@ class EmbeddedDocumentList(BaseList):
 
 class StrictDict(object):
     __slots__ = ()
-    _special_fields = set(['get', 'pop', 'iteritems', 'items', 'keys', 'create'])
+    _special_fields = {'get', 'pop', 'iteritems', 'items', 'keys', 'create'}
     _classes = {}
 
     def __init__(self, **kwargs):
@@ -447,40 +449,40 @@ class StrictDict(object):
         return cls._classes[allowed_keys]
 
 
-class SemiStrictDict(StrictDict):
-    __slots__ = ('_extras', )
-    _classes = {}
+class LazyReference(DBRef):
+    __slots__ = ('_cached_doc', 'passthrough', 'document_type')
 
-    def __getattr__(self, attr):
-        try:
-            super(SemiStrictDict, self).__getattr__(attr)
-        except AttributeError:
-            try:
-                return self.__getattribute__('_extras')[attr]
-            except KeyError as e:
-                raise AttributeError(e)
+    def fetch(self, force=False):
+        if not self._cached_doc or force:
+            self._cached_doc = self.document_type.objects.get(pk=self.pk)
+            if not self._cached_doc:
+                raise DoesNotExist('Trying to dereference unknown document %s' % (self))
+        return self._cached_doc
 
-    def __setattr__(self, attr, value):
-        try:
-            super(SemiStrictDict, self).__setattr__(attr, value)
-        except AttributeError:
-            try:
-                self._extras[attr] = value
-            except AttributeError:
-                self._extras = {attr: value}
+    @property
+    def pk(self):
+        return self.id
 
-    def __delattr__(self, attr):
-        try:
-            super(SemiStrictDict, self).__delattr__(attr)
-        except AttributeError:
-            try:
-                del self._extras[attr]
-            except KeyError as e:
-                raise AttributeError(e)
+    def __init__(self, document_type, pk, cached_doc=None, passthrough=False):
+        self.document_type = document_type
+        self._cached_doc = cached_doc
+        self.passthrough = passthrough
+        super(LazyReference, self).__init__(self.document_type._get_collection_name(), pk)
 
-    def __iter__(self):
+    def __getitem__(self, name):
+        if not self.passthrough:
+            raise KeyError()
+        document = self.fetch()
+        return document[name]
+
+    def __getattr__(self, name):
+        if not object.__getattribute__(self, 'passthrough'):
+            raise AttributeError()
+        document = self.fetch()
         try:
-            extras_iter = iter(self.__getattribute__('_extras'))
-        except AttributeError:
-            extras_iter = ()
-        return itertools.chain(super(SemiStrictDict, self).__iter__(), extras_iter)
+            return document[name]
+        except KeyError:
+            raise AttributeError()
+
+    def __repr__(self):
+        return "<LazyReference(%s, %r)>" % (self.document_type, self.pk)

mongoengine/base/document.py

@@ -1,6 +1,5 @@
 import copy
 import numbers
-from collections import Hashable
 from functools import partial
 
 from bson import ObjectId, json_util
@@ -13,13 +12,15 @@ from mongoengine import signals
 from mongoengine.base.common import get_document
 from mongoengine.base.datastructures import (BaseDict, BaseList,
                                              EmbeddedDocumentList,
-                                             SemiStrictDict, StrictDict)
+                                             LazyReference,
+                                             StrictDict)
 from mongoengine.base.fields import ComplexBaseField
 from mongoengine.common import _import_class
 from mongoengine.errors import (FieldDoesNotExist, InvalidDocumentError,
                                 LookUpError, OperationError, ValidationError)
+from mongoengine.python_support import Hashable
 
-__all__ = ('BaseDocument',)
+__all__ = ('BaseDocument', 'NON_FIELD_ERRORS')
 
 NON_FIELD_ERRORS = '__all__'
 
@@ -79,8 +80,7 @@ class BaseDocument(object):
         if self.STRICT and not self._dynamic:
             self._data = StrictDict.create(allowed_keys=self._fields_ordered)()
         else:
-            self._data = SemiStrictDict.create(
-                allowed_keys=self._fields_ordered)()
+            self._data = {}
 
         self._dynamic_fields = SON()
 
@@ -100,13 +100,11 @@ class BaseDocument(object):
             for key, value in values.iteritems():
                 if key in self._fields or key == '_id':
                     setattr(self, key, value)
-                elif self._dynamic:
+                else:
                     dynamic_data[key] = value
         else:
             FileField = _import_class('FileField')
             for key, value in values.iteritems():
-                if key == '__auto_convert':
-                    continue
                 key = self._reverse_db_field_map.get(key, key)
                 if key in self._fields or key in ('id', 'pk', '_cls'):
                     if __auto_convert and value is not None:
@@ -147,7 +145,7 @@ class BaseDocument(object):
 
             if not hasattr(self, name) and not name.startswith('_'):
                 DynamicField = _import_class('DynamicField')
-                field = DynamicField(db_field=name)
+                field = DynamicField(db_field=name, null=True)
                 field.name = name
                 self._dynamic_fields[name] = field
                 self._fields_ordered += (name,)
@@ -304,7 +302,7 @@ class BaseDocument(object):
         data['_cls'] = self._class_name
 
         # only root fields ['test1.a', 'test2'] => ['test1', 'test2']
-        root_fields = set([f.split('.')[0] for f in fields])
+        root_fields = {f.split('.')[0] for f in fields}
 
         for field_name in self:
             if root_fields and field_name not in root_fields:
@@ -337,7 +335,7 @@ class BaseDocument(object):
                 value = field.generate()
                 self._data[field_name] = value
 
-            if value is not None:
+            if (value is not None) or (field.null):
                 if use_db_field:
                     data[field.db_field] = value
                 else:
@@ -406,7 +404,15 @@ class BaseDocument(object):
 
     @classmethod
     def from_json(cls, json_data, created=False):
-        """Converts json data to an unsaved document instance"""
+        """Converts json data to a Document instance
+
+        :param json_data: The json data to load into the Document
+        :param created: If True, the document will be considered as a brand new document
+                        If False and an id is provided, it will consider that the data being
+                        loaded corresponds to what's already in the database (This has an impact of subsequent call to .save())
+                        If False and no id is provided, it will consider the data as a new document
+                        (default ``False``)
+        """
         return cls._from_son(json_util.loads(json_data), created=created)
 
     def __expand_dynamic_values(self, name, value):
@@ -489,7 +495,7 @@ class BaseDocument(object):
                 else:
                     data = getattr(data, part, None)
 
-                if hasattr(data, '_changed_fields'):
+                if not isinstance(data, LazyReference) and hasattr(data, '_changed_fields'):
                     if getattr(data, '_is_document', False):
                         continue
 
@@ -497,7 +503,13 @@ class BaseDocument(object):
 
         self._changed_fields = []
 
-    def _nestable_types_changed_fields(self, changed_fields, key, data, inspected):
+    def _nestable_types_changed_fields(self, changed_fields, base_key, data):
+        """Inspect nested data for changed fields
+
+        :param changed_fields: Previously collected changed fields
+        :param base_key: The base key that must be used to prepend changes to this data
+        :param data: data to inspect for changes
+        """
         # Loop list / dict fields as they contain documents
         # Determine the iterator to use
         if not hasattr(data, 'items'):
@@ -505,68 +517,60 @@ class BaseDocument(object):
         else:
             iterator = data.iteritems()
 
-        for index, value in iterator:
-            list_key = '%s%s.' % (key, index)
+        for index_or_key, value in iterator:
+            item_key = '%s%s.' % (base_key, index_or_key)
             # don't check anything lower if this key is already marked
             # as changed.
-            if list_key[:-1] in changed_fields:
+            if item_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)
-                                   for k in changed if k]
+                changed = value._get_changed_fields()
+                changed_fields += ['%s%s' % (item_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)
+                    changed_fields, item_key, value)
 
-    def _get_changed_fields(self, inspected=None):
+    def _get_changed_fields(self):
         """Return a list of all fields that have explicitly been changed.
         """
         EmbeddedDocument = _import_class('EmbeddedDocument')
-        DynamicEmbeddedDocument = _import_class('DynamicEmbeddedDocument')
         ReferenceField = _import_class('ReferenceField')
+        GenericReferenceField = _import_class('GenericReferenceField')
         SortedListField = _import_class('SortedListField')
 
         changed_fields = []
         changed_fields += getattr(self, '_changed_fields', [])
 
-        inspected = inspected or set()
-        if hasattr(self, 'id') and isinstance(self.id, Hashable):
-            if self.id in inspected:
-                return changed_fields
-            inspected.add(self.id)
-
         for field_name in self._fields_ordered:
             db_field_name = self._db_field_map.get(field_name, field_name)
             key = '%s.' % db_field_name
             data = self._data.get(field_name, None)
             field = self._fields.get(field_name)
 
-            if hasattr(data, 'id'):
-                if data.id in inspected:
-                    continue
-            if isinstance(field, ReferenceField):
+            if db_field_name in changed_fields:
+                # Whole field already marked as changed, no need to go further
                 continue
-            elif (
-                isinstance(data, (EmbeddedDocument, DynamicEmbeddedDocument)) and
-                db_field_name not in changed_fields
-            ):
+
+            if isinstance(field, ReferenceField):   # Don't follow referenced documents
+                continue
+
+            if isinstance(data, EmbeddedDocument):
                 # Find all embedded fields that have been changed
-                changed = data._get_changed_fields(inspected)
+                changed = data._get_changed_fields()
                 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):
+            elif isinstance(data, (list, tuple, dict)):
                 if (hasattr(field, 'field') and
-                        isinstance(field.field, ReferenceField)):
+                        isinstance(field.field, (ReferenceField, GenericReferenceField))):
                     continue
                 elif isinstance(field, SortedListField) and field._ordering:
                     # if ordering is affected whole list is changed
-                    if any(map(lambda d: field._ordering in d._changed_fields, data)):
+                    if any(field._ordering in d._changed_fields for d in data):
                         changed_fields.append(db_field_name)
                         continue
 
                 self._nestable_types_changed_fields(
-                    changed_fields, key, data, inspected)
+                    changed_fields, key, data)
         return changed_fields
 
     def _delta(self):
@@ -578,7 +582,6 @@ class BaseDocument(object):
 
         set_fields = self._get_changed_fields()
         unset_data = {}
-        parts = []
         if hasattr(self, '_changed_fields'):
             set_data = {}
             # Fetch each set item from its path
@@ -588,15 +591,13 @@ class BaseDocument(object):
                 new_path = []
                 for p in parts:
                     if isinstance(d, (ObjectId, DBRef)):
+                        # Don't dig in the references
                         break
-                    elif isinstance(d, list) and p.lstrip('-').isdigit():
-                        if p[0] == '-':
-                            p = str(len(d) + int(p))
-                        try:
-                            d = d[int(p)]
-                        except IndexError:
-                            d = None
+                    elif isinstance(d, list) and p.isdigit():
+                        # An item of a list (identified by its index) is updated
+                        d = d[int(p)]
                     elif hasattr(d, 'get'):
+                        # dict-like (dict, embedded document)
                         d = d.get(p)
                     new_path.append(p)
                 path = '.'.join(new_path)
@@ -608,26 +609,26 @@ class BaseDocument(object):
 
         # Determine if any changed items were actually unset.
         for path, value in set_data.items():
-            if value or isinstance(value, (numbers.Number, bool)):
+            if value or isinstance(value, (numbers.Number, bool)):  # Account for 0 and True that are truthy
                 continue
 
-            # If we've set a value that ain't the default value don't unset it.
-            default = None
+            parts = path.split('.')
+
             if (self._dynamic and len(parts) and parts[0] in
                     self._dynamic_fields):
                 del set_data[path]
                 unset_data[path] = 1
                 continue
-            elif path in self._fields:
+
+            # If we've set a value that ain't the default value don't unset it.
+            default = None
+            if path in self._fields:
                 default = self._fields[path].default
             else:  # Perform a full lookup for lists / embedded lookups
                 d = self
-                parts = path.split('.')
                 db_field_name = parts.pop()
                 for p in parts:
-                    if isinstance(d, list) and p.lstrip('-').isdigit():
-                        if p[0] == '-':
-                            p = str(len(d) + int(p))
+                    if isinstance(d, list) and p.isdigit():
                         d = d[int(p)]
                     elif (hasattr(d, '__getattribute__') and
                           not isinstance(d, dict)):
@@ -645,10 +646,9 @@ class BaseDocument(object):
                         default = None
 
             if default is not None:
-                if callable(default):
-                    default = default()
+                default = default() if callable(default) else default
 
-            if default != value:
+            if value != default:
                 continue
 
             del set_data[path]
@@ -694,7 +694,7 @@ class BaseDocument(object):
 
         fields = cls._fields
         if not _auto_dereference:
-            fields = copy.copy(fields)
+            fields = copy.deepcopy(fields)
 
         for field_name, field in fields.iteritems():
             field._auto_dereference = _auto_dereference
@@ -1080,5 +1080,11 @@ class BaseDocument(object):
         """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)
+            if value is None:
+                return None
+            sep = getattr(field, 'display_sep', ' ')
+            values = value if field.__class__.__name__ in ('ListField', 'SortedListField') else [value]
+            return sep.join([
+                six.text_type(dict(field.choices).get(val, val))
+                for val in values or []])
         return value

mongoengine/base/fields.py

@@ -55,7 +55,7 @@ class BaseField(object):
             field.  Generally this is deprecated in favour of the
             `FIELD.validate` method
         :param choices: (optional) The valid choices
-        :param null: (optional) Is the field value can be null. If no and there is a default value
+        :param null: (optional) If the field value can be null. If no and there is a default value
             then the default value is set
         :param sparse: (optional) `sparse=True` combined with `unique=True` and `required=False`
             means that uniqueness won't be enforced for `None` values
@@ -130,7 +130,6 @@ class BaseField(object):
     def __set__(self, instance, value):
         """Descriptor for assigning a value to a field in a document.
         """
-
         # If setting to None and there is a default
         # Then set the value to the default value
         if value is None:
@@ -213,8 +212,10 @@ class BaseField(object):
                     )
                 )
         # Choices which are types other than Documents
-        elif value not in choice_list:
-            self.error('Value must be one of %s' % six.text_type(choice_list))
+        else:
+            values = value if isinstance(value, (list, tuple)) else [value]
+            if len(set(values) - set(choice_list)):
+                self.error('Value must be one of %s' % six.text_type(choice_list))
 
     def _validate(self, value, **kwargs):
         # Check the Choices Constraint
@@ -265,13 +266,15 @@ class ComplexBaseField(BaseField):
         ReferenceField = _import_class('ReferenceField')
         GenericReferenceField = _import_class('GenericReferenceField')
         EmbeddedDocumentListField = _import_class('EmbeddedDocumentListField')
-        dereference = (self._auto_dereference and
+
+        auto_dereference = instance._fields[self.name]._auto_dereference
+
+        dereference = (auto_dereference and
                        (self.field is None or isinstance(self.field,
                                                          (GenericReferenceField, ReferenceField))))
 
         _dereference = _import_class('DeReference')()
 
-        self._auto_dereference = instance._fields[self.name]._auto_dereference
         if instance._initialised and dereference and instance._data.get(self.name):
             instance._data[self.name] = _dereference(
                 instance._data.get(self.name), max_depth=1, instance=instance,
@@ -292,7 +295,7 @@ class ComplexBaseField(BaseField):
             value = BaseDict(value, instance, self.name)
             instance._data[self.name] = value
 
-        if (self._auto_dereference and instance._initialised and
+        if (auto_dereference and instance._initialised and
                 isinstance(value, (BaseList, BaseDict)) and
                 not value._dereferenced):
             value = _dereference(
@@ -311,11 +314,16 @@ class ComplexBaseField(BaseField):
         if hasattr(value, 'to_python'):
             return value.to_python()
 
+        BaseDocument = _import_class('BaseDocument')
+        if isinstance(value, BaseDocument):
+            # Something is wrong, return the value as it is
+            return value
+
         is_list = False
         if not hasattr(value, 'items'):
             try:
                 is_list = True
-                value = {k: v for k, v in enumerate(value)}
+                value = {idx: v for idx, v in enumerate(value)}
             except TypeError:  # Not iterable return the value
                 return value
 
@@ -500,7 +508,7 @@ class GeoJsonBaseField(BaseField):
     def validate(self, value):
         """Validate the GeoJson object based on its type."""
         if isinstance(value, dict):
-            if set(value.keys()) == set(['type', 'coordinates']):
+            if set(value.keys()) == {'type', 'coordinates'}:
                 if value['type'] != self._type:
                     self.error('%s type must be "%s"' %
                                (self._name, self._type))

mongoengine/base/metaclasses.py

@@ -18,14 +18,14 @@ class DocumentMetaclass(type):
     """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__
+    def __new__(mcs, name, bases, attrs):
+        flattened_bases = mcs._get_bases(bases)
+        super_new = super(DocumentMetaclass, mcs).__new__
 
         # If a base class just call super
         metaclass = attrs.get('my_metaclass')
         if metaclass and issubclass(metaclass, DocumentMetaclass):
-            return super_new(cls, name, bases, attrs)
+            return super_new(mcs, name, bases, attrs)
 
         attrs['_is_document'] = attrs.get('_is_document', False)
         attrs['_cached_reference_fields'] = []
@@ -121,7 +121,8 @@ class DocumentMetaclass(type):
                 # inheritance of classes where inheritance is set to False
                 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' %
+                    raise ValueError('Document %s may not be subclassed. '
+                                     'To enable inheritance, use the "allow_inheritance" meta attribute.' %
                                      base.__name__)
 
         # Get superclasses from last base superclass
@@ -138,7 +139,7 @@ class DocumentMetaclass(type):
         attrs['_types'] = attrs['_subclasses']  # TODO depreciate _types
 
         # Create the new_class
-        new_class = super_new(cls, name, bases, attrs)
+        new_class = super_new(mcs, name, bases, attrs)
 
         # Set _subclasses
         for base in document_bases:
@@ -147,7 +148,7 @@ class DocumentMetaclass(type):
             base._types = base._subclasses  # TODO depreciate _types
 
         (Document, EmbeddedDocument, DictField,
-         CachedReferenceField) = cls._import_classes()
+         CachedReferenceField) = mcs._import_classes()
 
         if issubclass(new_class, Document):
             new_class._collection = None
@@ -219,29 +220,26 @@ class DocumentMetaclass(type):
 
         return new_class
 
-    def add_to_class(self, name, value):
-        setattr(self, name, value)
-
     @classmethod
-    def _get_bases(cls, bases):
+    def _get_bases(mcs, bases):
         if isinstance(bases, BasesTuple):
             return bases
         seen = []
-        bases = cls.__get_bases(bases)
+        bases = mcs.__get_bases(bases)
         unique_bases = (b for b in bases if not (b in seen or seen.append(b)))
         return BasesTuple(unique_bases)
 
     @classmethod
-    def __get_bases(cls, bases):
+    def __get_bases(mcs, bases):
         for base in bases:
             if base is object:
                 continue
             yield base
-            for child_base in cls.__get_bases(base.__bases__):
+            for child_base in mcs.__get_bases(base.__bases__):
                 yield child_base
 
     @classmethod
-    def _import_classes(cls):
+    def _import_classes(mcs):
         Document = _import_class('Document')
         EmbeddedDocument = _import_class('EmbeddedDocument')
         DictField = _import_class('DictField')
@@ -254,9 +252,9 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
     collection in the database.
     """
 
-    def __new__(cls, name, bases, attrs):
-        flattened_bases = cls._get_bases(bases)
-        super_new = super(TopLevelDocumentMetaclass, cls).__new__
+    def __new__(mcs, name, bases, attrs):
+        flattened_bases = mcs._get_bases(bases)
+        super_new = super(TopLevelDocumentMetaclass, mcs).__new__
 
         # Set default _meta data if base class, otherwise get user defined meta
         if attrs.get('my_metaclass') == TopLevelDocumentMetaclass:
@@ -319,7 +317,7 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
                     not parent_doc_cls._meta.get('abstract', False)):
                 msg = 'Abstract document cannot have non-abstract base'
                 raise ValueError(msg)
-            return super_new(cls, name, bases, attrs)
+            return super_new(mcs, name, bases, attrs)
 
         # Merge base class metas.
         # Uses a special MetaDict that handles various merging rules
@@ -360,7 +358,7 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
         attrs['_meta'] = meta
 
         # Call super and get the new class
-        new_class = super_new(cls, name, bases, attrs)
+        new_class = super_new(mcs, name, bases, attrs)
 
         meta = new_class._meta
 
@@ -394,7 +392,7 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
                                            '_auto_id_field', False)
         if not new_class._meta.get('id_field'):
             # After 0.10, find not existing names, instead of overwriting
-            id_name, id_db_name = cls.get_auto_id_names(new_class)
+            id_name, id_db_name = mcs.get_auto_id_names(new_class)
             new_class._auto_id_field = True
             new_class._meta['id_field'] = id_name
             new_class._fields[id_name] = ObjectIdField(db_field=id_db_name)
@@ -419,7 +417,7 @@ class TopLevelDocumentMetaclass(DocumentMetaclass):
         return new_class
 
     @classmethod
-    def get_auto_id_names(cls, new_class):
+    def get_auto_id_names(mcs, new_class):
         id_name, id_db_name = ('id', '_id')
         if id_name not in new_class._fields and \
                 id_db_name not in (v.db_field for v in new_class._fields.values()):

mongoengine/base/utils.py

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

mongoengine/connection.py

@@ -28,7 +28,7 @@ _connections = {}
 _dbs = {}
 
 
-def register_connection(alias, name=None, host=None, port=None,
+def register_connection(alias, db=None, name=None, host=None, port=None,
                         read_preference=READ_PREFERENCE,
                         username=None, password=None,
                         authentication_source=None,
@@ -39,6 +39,7 @@ def register_connection(alias, name=None, host=None, port=None,
     :param alias: the name that will be used to refer to this connection
         throughout MongoEngine
     :param name: the name of the specific database to use
+    :param db: the name of the database to use, for compatibility with connect
     :param host: the host name of the :program:`mongod` instance to connect to
     :param port: the port that the :program:`mongod` instance is running on
     :param read_preference: The read preference for the collection
@@ -58,7 +59,7 @@ def register_connection(alias, name=None, host=None, port=None,
     .. versionchanged:: 0.10.6 - added mongomock support
     """
     conn_settings = {
-        'name': name or 'test',
+        'name': name or db or 'test',
         'host': host or 'localhost',
         'port': port or 27017,
         'read_preference': read_preference,
@@ -103,6 +104,18 @@ def register_connection(alias, name=None, host=None, port=None,
                 conn_settings['authentication_source'] = uri_options['authsource']
             if 'authmechanism' in uri_options:
                 conn_settings['authentication_mechanism'] = uri_options['authmechanism']
+            if IS_PYMONGO_3 and 'readpreference' in uri_options:
+                read_preferences = (
+                    ReadPreference.NEAREST,
+                    ReadPreference.PRIMARY,
+                    ReadPreference.PRIMARY_PREFERRED,
+                    ReadPreference.SECONDARY,
+                    ReadPreference.SECONDARY_PREFERRED)
+                read_pf_mode = uri_options['readpreference'].lower()
+                for preference in read_preferences:
+                    if preference.name.lower() == read_pf_mode:
+                        conn_settings['read_preference'] = preference
+                        break
         else:
             resolved_hosts.append(entity)
     conn_settings['host'] = resolved_hosts
@@ -146,13 +159,14 @@ def get_connection(alias=DEFAULT_CONNECTION_NAME, reconnect=False):
         raise MongoEngineConnectionError(msg)
 
     def _clean_settings(settings_dict):
-        irrelevant_fields = set([
-            'name', 'username', 'password', 'authentication_source',
-            'authentication_mechanism'
-        ])
+        # set literal more efficient than calling set function
+        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
+            if k not in irrelevant_fields_set
         }
 
     # Retrieve a copy of the connection settings associated with the requested

mongoengine/context_managers.py

@@ -1,9 +1,11 @@
+from contextlib import contextmanager
+from pymongo.write_concern import WriteConcern
 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')
+           'no_sub_classes', 'query_counter', 'set_write_concern')
 
 
 class switch_db(object):
@@ -143,66 +145,85 @@ class no_sub_classes(object):
         :param cls: the class to turn querying sub classes on
         """
         self.cls = cls
+        self.cls_initial_subclasses = None
 
     def __enter__(self):
         """Change the objects default and _auto_dereference values."""
-        self.cls._all_subclasses = self.cls._subclasses
-        self.cls._subclasses = (self.cls,)
+        self.cls_initial_subclasses = self.cls._subclasses
+        self.cls._subclasses = (self.cls._class_name,)
         return self.cls
 
     def __exit__(self, t, value, traceback):
         """Reset the default and _auto_dereference values."""
-        self.cls._subclasses = self.cls._all_subclasses
-        delattr(self.cls, '_all_subclasses')
-        return self.cls
+        self.cls._subclasses = self.cls_initial_subclasses
 
 
 class query_counter(object):
-    """Query_counter context manager to get the number of queries."""
+    """Query_counter context manager to get the number of queries.
+    This works by updating the `profiling_level` of the database so that all queries get logged,
+    resetting the db.system.profile collection at the beginnig of the context and counting the new entries.
+
+    This was designed for debugging purpose. In fact it is a global counter so queries issued by other threads/processes
+    can interfere with it
+
+    Be aware that:
+    - Iterating over large amount of documents (>101) makes pymongo issue `getmore` queries to fetch the next batch of
+        documents (https://docs.mongodb.com/manual/tutorial/iterate-a-cursor/#cursor-batches)
+    - Some queries are ignored by default by the counter (killcursors, db.system.indexes)
+    """
 
     def __init__(self):
-        """Construct the query_counter."""
-        self.counter = 0
+        """Construct the query_counter
+        """
         self.db = get_db()
+        self.initial_profiling_level = None
+        self._ctx_query_counter = 0             # number of queries issued by the context
 
-    def __enter__(self):
-        """On every with block we need to drop the profile collection."""
+        self._ignored_query = {
+            'ns':
+                {'$ne': '%s.system.indexes' % self.db.name},
+            'op':                       # MONGODB < 3.2
+                {'$ne': 'killcursors'},
+            'command.killCursors':      # MONGODB >= 3.2
+                {'$exists': False}
+        }
+
+    def _turn_on_profiling(self):
+        self.initial_profiling_level = self.db.profiling_level()
         self.db.set_profiling_level(0)
         self.db.system.profile.drop()
         self.db.set_profiling_level(2)
+
+    def _resets_profiling(self):
+        self.db.set_profiling_level(self.initial_profiling_level)
+
+    def __enter__(self):
+        self._turn_on_profiling()
         return self
 
     def __exit__(self, t, value, traceback):
-        """Reset the profiling level."""
-        self.db.set_profiling_level(0)
+        self._resets_profiling()
 
     def __eq__(self, value):
-        """== Compare querycounter."""
         counter = self._get_count()
         return value == counter
 
     def __ne__(self, value):
-        """!= Compare querycounter."""
         return not self.__eq__(value)
 
     def __lt__(self, value):
-        """< Compare querycounter."""
         return self._get_count() < value
 
     def __le__(self, value):
-        """<= Compare querycounter."""
         return self._get_count() <= value
 
     def __gt__(self, value):
-        """> Compare querycounter."""
         return self._get_count() > value
 
     def __ge__(self, value):
-        """>= Compare querycounter."""
         return self._get_count() >= value
 
     def __int__(self):
-        """int representation."""
         return self._get_count()
 
     def __repr__(self):
@@ -210,8 +231,17 @@ class query_counter(object):
         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}}
-        count = self.db.system.profile.find(ignore_query).count() - self.counter
-        self.counter += 1
+        """Get the number of queries by counting the current number of entries in db.system.profile
+        and substracting the queries issued by this context. In fact everytime this is called, 1 query is
+        issued so we need to balance that
+        """
+        count = self.db.system.profile.find(self._ignored_query).count() - self._ctx_query_counter
+        self._ctx_query_counter += 1    # Account for the query we just issued to gather the information
         return count
+
+
+@contextmanager
+def set_write_concern(collection, write_concerns):
+    combined_concerns = dict(collection.write_concern.document.items())
+    combined_concerns.update(write_concerns)
+    yield collection.with_options(write_concern=WriteConcern(**combined_concerns))

mongoengine/dereference.py

@@ -3,6 +3,7 @@ import six
 
 from mongoengine.base import (BaseDict, BaseList, EmbeddedDocumentList,
                               TopLevelDocumentMetaclass, get_document)
+from mongoengine.base.datastructures import LazyReference
 from mongoengine.connection import get_db
 from mongoengine.document import Document, EmbeddedDocument
 from mongoengine.fields import DictField, ListField, MapField, ReferenceField
@@ -99,7 +100,10 @@ class DeReference(object):
             if isinstance(item, (Document, EmbeddedDocument)):
                 for field_name, field in item._fields.iteritems():
                     v = item._data.get(field_name, None)
-                    if isinstance(v, DBRef):
+                    if isinstance(v, LazyReference):
+                        # LazyReference inherits DBRef but should not be dereferenced here !
+                        continue
+                    elif isinstance(v, DBRef):
                         reference_map.setdefault(field.document_type, set()).add(v.id)
                     elif isinstance(v, (dict, SON)) and '_ref' in v:
                         reference_map.setdefault(get_document(v['_cls']), set()).add(v['_ref'].id)
@@ -110,6 +114,9 @@ class DeReference(object):
                             if isinstance(field_cls, (Document, TopLevelDocumentMetaclass)):
                                 key = field_cls
                             reference_map.setdefault(key, set()).update(refs)
+            elif isinstance(item, LazyReference):
+                # LazyReference inherits DBRef but should not be dereferenced here !
+                continue
             elif isinstance(item, DBRef):
                 reference_map.setdefault(item.collection, set()).add(item.id)
             elif isinstance(item, (dict, SON)) and '_ref' in item:
@@ -126,7 +133,12 @@ class DeReference(object):
         """
         object_map = {}
         for collection, dbrefs in self.reference_map.iteritems():
-            if hasattr(collection, 'objects'):  # We have a document class for the refs
+
+            # we use getattr instead of hasattr because hasattr swallows any exception under python2
+            # so it could hide nasty things without raising exceptions (cfr bug #1688))
+            ref_document_cls_exists = (getattr(collection, 'objects', None) is not None)
+
+            if ref_document_cls_exists:
                 col_name = collection._get_collection_name()
                 refs = [dbref for dbref in dbrefs
                         if (col_name, dbref) not in object_map]
@@ -134,7 +146,7 @@ class DeReference(object):
                 for key, doc in references.iteritems():
                     object_map[(col_name, key)] = doc
             else:  # Generic reference: use the refs data to convert to document
-                if isinstance(doc_type, (ListField, DictField, MapField,)):
+                if isinstance(doc_type, (ListField, DictField, MapField)):
                     continue
 
                 refs = [dbref for dbref in dbrefs
@@ -230,7 +242,7 @@ class DeReference(object):
             elif isinstance(v, (dict, list, tuple)) and depth <= self.max_depth:
                 item_name = '%s.%s' % (name, k) if name else name
                 data[k] = self._attach_objects(v, depth - 1, instance=instance, name=item_name)
-            elif hasattr(v, 'id'):
+            elif isinstance(v, DBRef) and hasattr(v, 'id'):
                 data[k] = self.object_map.get((v.collection, v.id), v)
 
         if instance and name:

mongoengine/document.py

@@ -12,7 +12,9 @@ from mongoengine.base import (BaseDict, BaseDocument, BaseList,
                               TopLevelDocumentMetaclass, get_document)
 from mongoengine.common import _import_class
 from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
-from mongoengine.context_managers import switch_collection, switch_db
+from mongoengine.context_managers import (set_write_concern,
+                                          switch_collection,
+                                          switch_db)
 from mongoengine.errors import (InvalidDocumentError, InvalidQueryError,
                                 SaveConditionError)
 from mongoengine.python_support import IS_PYMONGO_3
@@ -39,7 +41,7 @@ class InvalidCollectionError(Exception):
     pass
 
 
-class EmbeddedDocument(BaseDocument):
+class EmbeddedDocument(six.with_metaclass(DocumentMetaclass, BaseDocument)):
     """A :class:`~mongoengine.Document` that isn't stored in its own
     collection.  :class:`~mongoengine.EmbeddedDocument`\ s should be used as
     fields on :class:`~mongoengine.Document`\ s through the
@@ -58,7 +60,6 @@ class EmbeddedDocument(BaseDocument):
     # The __metaclass__ attribute is removed by 2to3 when running with Python3
     # my_metaclass is defined so that metaclass can be queried in Python 2 & 3
     my_metaclass = DocumentMetaclass
-    __metaclass__ = DocumentMetaclass
 
     # A generic embedded document doesn't have any immutable properties
     # that describe it uniquely, hence it shouldn't be hashable. You can
@@ -95,7 +96,7 @@ class EmbeddedDocument(BaseDocument):
         self._instance.reload(*args, **kwargs)
 
 
-class Document(BaseDocument):
+class Document(six.with_metaclass(TopLevelDocumentMetaclass, BaseDocument)):
     """The base class used for defining the structure and properties of
     collections of documents stored in MongoDB. Inherit from this class, and
     add fields as class attributes to define a document's structure.
@@ -150,7 +151,6 @@ class Document(BaseDocument):
     # The __metaclass__ attribute is removed by 2to3 when running with Python3
     # my_metaclass is defined so that metaclass can be queried in Python 2 & 3
     my_metaclass = TopLevelDocumentMetaclass
-    __metaclass__ = TopLevelDocumentMetaclass
 
     __slots__ = ('__objects',)
 
@@ -172,8 +172,8 @@ class Document(BaseDocument):
         """
         if self.pk is None:
             return super(BaseDocument, self).__hash__()
-        else:
-            return hash(self.pk)
+
+        return hash(self.pk)
 
     @classmethod
     def _get_db(cls):
@@ -195,7 +195,10 @@ class Document(BaseDocument):
 
             # Ensure indexes on the collection unless auto_create_index was
             # set to False.
-            if cls._meta.get('auto_create_index', True):
+            # Also there is no need to ensure indexes on slave.
+            db = cls._get_db()
+            if cls._meta.get('auto_create_index', True) and\
+                    db.client.is_primary:
                 cls.ensure_indexes()
 
         return cls._collection
@@ -280,6 +283,9 @@ class Document(BaseDocument):
         elif query[id_field] != self.pk:
             raise InvalidQueryError('Invalid document modify query: it must modify only this document.')
 
+        # Need to add shard key to query, or you get an error
+        query.update(self._object_key)
+
         updated = self._qs(**query).modify(new=True, **update)
         if updated is None:
             return False
@@ -320,7 +326,7 @@ 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
+        :param signal_kwargs: (optional) kwargs dictionary to be passed to
             the signal calls.
 
         .. versionchanged:: 0.5
@@ -364,6 +370,8 @@ class Document(BaseDocument):
 
         signals.pre_save_post_validation.send(self.__class__, document=self,
                                               created=created, **signal_kwargs)
+        # it might be refreshed by the pre_save_post_validation hook, e.g., for etag generation
+        doc = self.to_mongo()
 
         if self._meta.get('auto_create_index', True):
             self.ensure_indexes()
@@ -423,11 +431,18 @@ class Document(BaseDocument):
         Helper method, should only be used inside save().
         """
         collection = self._get_collection()
+        with set_write_concern(collection, write_concern) as wc_collection:
+            if force_insert:
+                return wc_collection.insert_one(doc).inserted_id
+            # insert_one will provoke UniqueError alongside save does not
+            # therefore, it need to catch and call replace_one.
+            if '_id' in doc:
+                raw_object = wc_collection.find_one_and_replace(
+                    {'_id': doc['_id']}, doc)
+                if raw_object:
+                    return doc['_id']
 
-        if force_insert:
-            return collection.insert(doc, **write_concern)
-
-        object_id = collection.save(doc, **write_concern)
+            object_id = wc_collection.insert_one(doc).inserted_id
 
         # 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
@@ -576,12 +591,11 @@ class Document(BaseDocument):
         """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
+        :param 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
+            will be used as options for the resultant ``getLastError`` command.
+            For example, ``save(..., w: 2, fsync: True)`` will
             wait until at least two servers have recorded the write and
             will force an fsync on the primary server.
 
@@ -702,7 +716,6 @@ class Document(BaseDocument):
             obj = obj[0]
         else:
             raise self.DoesNotExist('Document does not exist')
-
         for field in obj._data:
             if not fields or field in fields:
                 try:
@@ -710,7 +723,7 @@ class Document(BaseDocument):
                 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
+                        # a 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
@@ -718,7 +731,9 @@ class Document(BaseDocument):
                         # i.e. obj.update(unset__field=1) followed by obj.reload()
                         delattr(self, field)
 
-        self._changed_fields = obj._changed_fields
+        self._changed_fields = list(
+            set(self._changed_fields) - set(fields)
+        ) if fields else obj._changed_fields
         self._created = False
         return self
 
@@ -964,8 +979,16 @@ class Document(BaseDocument):
         """
 
         required = cls.list_indexes()
-        existing = [info['key']
-                    for info in cls._get_collection().index_information().values()]
+
+        existing = []
+        for info in cls._get_collection().index_information().values():
+            if '_fts' in info['key'][0]:
+                index_type = info['key'][0][1]
+                text_index_fields = info.get('weights').keys()
+                existing.append(
+                    [(key, index_type) for key in text_index_fields])
+            else:
+                existing.append(info['key'])
         missing = [index for index in required if index not in existing]
         extra = [index for index in existing if index not in required]
 
@@ -982,10 +1005,10 @@ class Document(BaseDocument):
         return {'missing': missing, 'extra': extra}
 
 
-class DynamicDocument(Document):
+class DynamicDocument(six.with_metaclass(TopLevelDocumentMetaclass, Document)):
     """A Dynamic Document class allowing flexible, expandable and uncontrolled
     schemas.  As a :class:`~mongoengine.Document` subclass, acts in the same
-    way as an ordinary document but has expando style properties.  Any data
+    way as an ordinary document but has expanded style properties.  Any data
     passed or set against the :class:`~mongoengine.DynamicDocument` that is
     not a field is automatically converted into a
     :class:`~mongoengine.fields.DynamicField` and data can be attributed to that
@@ -999,7 +1022,6 @@ class DynamicDocument(Document):
     # The __metaclass__ attribute is removed by 2to3 when running with Python3
     # my_metaclass is defined so that metaclass can be queried in Python 2 & 3
     my_metaclass = TopLevelDocumentMetaclass
-    __metaclass__ = TopLevelDocumentMetaclass
 
     _dynamic = True
 
@@ -1010,11 +1032,12 @@ class DynamicDocument(Document):
         field_name = args[0]
         if field_name in self._dynamic_fields:
             setattr(self, field_name, None)
+            self._dynamic_fields[field_name].null = False
         else:
             super(DynamicDocument, self).__delattr__(*args, **kwargs)
 
 
-class DynamicEmbeddedDocument(EmbeddedDocument):
+class DynamicEmbeddedDocument(six.with_metaclass(DocumentMetaclass, EmbeddedDocument)):
     """A Dynamic Embedded Document class allowing flexible, expandable and
     uncontrolled schemas. See :class:`~mongoengine.DynamicDocument` for more
     information about dynamic documents.
@@ -1023,7 +1046,6 @@ class DynamicEmbeddedDocument(EmbeddedDocument):
     # The __metaclass__ attribute is removed by 2to3 when running with Python3
     # my_metaclass is defined so that metaclass can be queried in Python 2 & 3
     my_metaclass = DocumentMetaclass
-    __metaclass__ = DocumentMetaclass
 
     _dynamic = True
 

mongoengine/errors.py

@@ -71,6 +71,7 @@ class ValidationError(AssertionError):
     _message = None
 
     def __init__(self, message='', **kwargs):
+        super(ValidationError, self).__init__(message)
         self.errors = kwargs.get('errors', {})
         self.field_name = kwargs.get('field_name')
         self.message = message

mongoengine/fields.py

@@ -5,7 +5,6 @@ import re
 import socket
 import time
 import uuid
-import warnings
 from operator import itemgetter
 
 from bson import Binary, DBRef, ObjectId, SON
@@ -25,13 +24,18 @@ try:
 except ImportError:
     Int64 = long
 
+
 from mongoengine.base import (BaseDocument, BaseField, ComplexBaseField,
-                              GeoJsonBaseField, ObjectIdField, get_document)
+                              GeoJsonBaseField, LazyReference, ObjectIdField,
+                              get_document)
+from mongoengine.base.utils import LazyRegexCompiler
+from mongoengine.common import _import_class
 from mongoengine.connection import DEFAULT_CONNECTION_NAME, get_db
 from mongoengine.document import Document, EmbeddedDocument
 from mongoengine.errors import DoesNotExist, InvalidQueryError, ValidationError
 from mongoengine.python_support import StringIO
-from mongoengine.queryset import DO_NOTHING, QuerySet
+from mongoengine.queryset import DO_NOTHING
+from mongoengine.queryset.base import BaseQuerySet
 
 try:
     from PIL import Image, ImageOps
@@ -39,13 +43,20 @@ except ImportError:
     Image = None
     ImageOps = None
 
+if six.PY3:
+    # Useless as long as 2to3 gets executed
+    # as it turns `long` into `int` blindly
+    long = int
+
+
 __all__ = (
     'StringField', 'URLField', 'EmailField', 'IntField', 'LongField',
-    'FloatField', 'DecimalField', 'BooleanField', 'DateTimeField',
+    'FloatField', 'DecimalField', 'BooleanField', 'DateTimeField', 'DateField',
     'ComplexDateTimeField', 'EmbeddedDocumentField', 'ObjectIdField',
     'GenericEmbeddedDocumentField', 'DynamicField', 'ListField',
     'SortedListField', 'EmbeddedDocumentListField', 'DictField',
     'MapField', 'ReferenceField', 'CachedReferenceField',
+    'LazyReferenceField', 'GenericLazyReferenceField',
     'GenericReferenceField', 'BinaryField', 'GridFSError', 'GridFSProxy',
     'FileField', 'ImageGridFsProxy', 'ImproperlyConfigured', 'ImageField',
     'GeoPointField', 'PointField', 'LineStringField', 'PolygonField',
@@ -120,9 +131,9 @@ class URLField(StringField):
     .. versionadded:: 0.3
     """
 
-    _URL_REGEX = re.compile(
+    _URL_REGEX = LazyRegexCompiler(
         r'^(?:[a-z0-9\.\-]*)://'  # scheme is validated separately
-        r'(?:(?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+(?:[A-Z]{2,6}\.?|[A-Z0-9-]{2,}(?<!-)\.?)|'  # domain...
+        r'(?:(?:[A-Z0-9](?:[A-Z0-9-_]{0,61}[A-Z0-9])?\.)+(?:[A-Z]{2,6}\.?|[A-Z0-9-]{2,}(?<!-)\.?)|'  # domain...
         r'localhost|'  # localhost...
         r'\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}|'  # ...or ipv4
         r'\[?[A-F0-9]*:[A-F0-9:]+\]?)'  # ...or ipv6
@@ -130,8 +141,7 @@ class URLField(StringField):
         r'(?:/?|[/?]\S+)$', re.IGNORECASE)
     _URL_SCHEMES = ['http', 'https', 'ftp', 'ftps']
 
-    def __init__(self, verify_exists=False, url_regex=None, schemes=None, **kwargs):
-        self.verify_exists = verify_exists
+    def __init__(self, url_regex=None, schemes=None, **kwargs):
         self.url_regex = url_regex or self._URL_REGEX
         self.schemes = schemes or self._URL_SCHEMES
         super(URLField, self).__init__(**kwargs)
@@ -154,7 +164,7 @@ class EmailField(StringField):
 
     .. versionadded:: 0.4
     """
-    USER_REGEX = re.compile(
+    USER_REGEX = LazyRegexCompiler(
         # `dot-atom` defined in RFC 5322 Section 3.2.3.
         r"(^[-!#$%&'*+/=?^_`{}|~0-9A-Z]+(\.[-!#$%&'*+/=?^_`{}|~0-9A-Z]+)*\Z"
         # `quoted-string` defined in RFC 5322 Section 3.2.4.
@@ -162,7 +172,7 @@ class EmailField(StringField):
         re.IGNORECASE
     )
 
-    UTF8_USER_REGEX = re.compile(
+    UTF8_USER_REGEX = LazyRegexCompiler(
         six.u(
             # RFC 6531 Section 3.3 extends `atext` (used by dot-atom) to
             # include `UTF8-non-ascii`.
@@ -172,7 +182,7 @@ class EmailField(StringField):
         ), re.IGNORECASE | re.UNICODE
     )
 
-    DOMAIN_REGEX = re.compile(
+    DOMAIN_REGEX = LazyRegexCompiler(
         r'((?:[A-Z0-9](?:[A-Z0-9-]{0,61}[A-Z0-9])?\.)+)(?:[A-Z0-9-]{2,63}(?<!-))\Z',
         re.IGNORECASE
     )
@@ -264,14 +274,14 @@ class IntField(BaseField):
     def to_python(self, value):
         try:
             value = int(value)
-        except ValueError:
+        except (TypeError, ValueError):
             pass
         return value
 
     def validate(self, value):
         try:
             value = int(value)
-        except Exception:
+        except (TypeError, ValueError):
             self.error('%s could not be converted to int' % value)
 
         if self.min_value is not None and value < self.min_value:
@@ -297,7 +307,7 @@ class LongField(BaseField):
     def to_python(self, value):
         try:
             value = long(value)
-        except ValueError:
+        except (TypeError, ValueError):
             pass
         return value
 
@@ -307,7 +317,7 @@ class LongField(BaseField):
     def validate(self, value):
         try:
             value = long(value)
-        except Exception:
+        except (TypeError, ValueError):
             self.error('%s could not be converted to long' % value)
 
         if self.min_value is not None and value < self.min_value:
@@ -361,7 +371,8 @@ class FloatField(BaseField):
 
 
 class DecimalField(BaseField):
-    """Fixed-point decimal number field.
+    """Fixed-point decimal number field. Stores the value as a float by default unless `force_string` is used.
+    If using floats, beware of Decimal to float conversion (potential precision loss)
 
     .. versionchanged:: 0.8
     .. versionadded:: 0.3
@@ -372,7 +383,9 @@ class DecimalField(BaseField):
         """
         :param min_value: Validation rule for the minimum acceptable value.
         :param max_value: Validation rule for the maximum acceptable value.
-        :param force_string: Store as a string.
+        :param force_string: Store the value as a string (instead of a float).
+         Be aware that this affects query sorting and operation like lte, gte (as string comparison is applied)
+         and some query operator won't work (e.g: inc, dec)
         :param precision: Number of decimal places to store.
         :param rounding: The rounding rule from the python decimal library:
 
@@ -403,7 +416,7 @@ class DecimalField(BaseField):
         # Convert to string for python 2.6 before casting to Decimal
         try:
             value = decimal.Decimal('%s' % value)
-        except decimal.InvalidOperation:
+        except (TypeError, ValueError, decimal.InvalidOperation):
             return value
         return value.quantize(decimal.Decimal('.%s' % ('0' * self.precision)), rounding=self.rounding)
 
@@ -420,7 +433,7 @@ class DecimalField(BaseField):
                 value = six.text_type(value)
             try:
                 value = decimal.Decimal(value)
-            except Exception as exc:
+            except (TypeError, ValueError, decimal.InvalidOperation) as exc:
                 self.error('Could not convert value to decimal: %s' % exc)
 
         if self.min_value is not None and value < self.min_value:
@@ -459,6 +472,8 @@ class DateTimeField(BaseField):
     installed you can utilise it to convert varying types of date formats into valid
     python datetime objects.
 
+    Note: To default the field to the current datetime, use: DateTimeField(default=datetime.utcnow)
+
     Note: Microseconds are rounded to the nearest millisecond.
       Pre UTC microsecond support is effectively broken.
       Use :class:`~mongoengine.fields.ComplexDateTimeField` if you
@@ -483,6 +498,10 @@ class DateTimeField(BaseField):
         if not isinstance(value, six.string_types):
             return None
 
+        value = value.strip()
+        if not value:
+            return None
+
         # Attempt to parse a datetime:
         if dateutil:
             try:
@@ -518,6 +537,22 @@ class DateTimeField(BaseField):
         return super(DateTimeField, self).prepare_query_value(op, self.to_mongo(value))
 
 
+class DateField(DateTimeField):
+    def to_mongo(self, value):
+        value = super(DateField, self).to_mongo(value)
+        # drop hours, minutes, seconds
+        if isinstance(value, datetime.datetime):
+            value = datetime.datetime(value.year, value.month, value.day)
+        return value
+
+    def to_python(self, value):
+        value = super(DateField, self).to_python(value)
+        # convert datetime to date
+        if isinstance(value, datetime.datetime):
+            value = datetime.date(value.year, value.month, value.day)
+        return value
+
+
 class ComplexDateTimeField(StringField):
     """
     ComplexDateTimeField handles microseconds exactly instead of rounding
@@ -534,11 +569,15 @@ class ComplexDateTimeField(StringField):
     The `,` as the separator can be easily modified by passing the `separator`
     keyword when initializing the field.
 
+    Note: To default the field to the current datetime, use: DateTimeField(default=datetime.utcnow)
+
     .. versionadded:: 0.5
     """
 
     def __init__(self, separator=',', **kwargs):
-        self.names = ['year', 'month', 'day', 'hour', 'minute', 'second', 'microsecond']
+        """
+        :param separator: Allows to customize the separator used for storage (default ``,``)
+        """
         self.separator = separator
         self.format = separator.join(['%Y', '%m', '%d', '%H', '%M', '%S', '%f'])
         super(ComplexDateTimeField, self).__init__(**kwargs)
@@ -565,20 +604,24 @@ class ComplexDateTimeField(StringField):
         >>> ComplexDateTimeField()._convert_from_string(a)
         datetime.datetime(2011, 6, 8, 20, 26, 24, 92284)
         """
-        values = map(int, data.split(self.separator))
+        values = [int(d) for d in data.split(self.separator)]
         return datetime.datetime(*values)
 
     def __get__(self, instance, owner):
+        if instance is None:
+            return self
+
         data = super(ComplexDateTimeField, self).__get__(instance, owner)
-        if data is None:
-            return None if self.null else datetime.datetime.now()
-        if isinstance(data, datetime.datetime):
+
+        if isinstance(data, datetime.datetime) or data is None:
             return data
         return self._convert_from_string(data)
 
     def __set__(self, instance, value):
-        value = self._convert_from_datetime(value) if value else value
-        return super(ComplexDateTimeField, self).__set__(instance, value)
+        super(ComplexDateTimeField, self).__set__(instance, value)
+        value = instance._data[self.name]
+        if value is not None:
+            instance._data[self.name] = self._convert_from_datetime(value)
 
     def validate(self, value):
         value = self.to_python(value)
@@ -607,6 +650,7 @@ class EmbeddedDocumentField(BaseField):
     """
 
     def __init__(self, document_type, **kwargs):
+        # XXX ValidationError raised outside of the "validate" method.
         if not (
             isinstance(document_type, six.string_types) or
             issubclass(document_type, EmbeddedDocument)
@@ -621,9 +665,17 @@ class EmbeddedDocumentField(BaseField):
     def document_type(self):
         if isinstance(self.document_type_obj, six.string_types):
             if self.document_type_obj == RECURSIVE_REFERENCE_CONSTANT:
-                self.document_type_obj = self.owner_document
+                resolved_document_type = self.owner_document
             else:
-                self.document_type_obj = get_document(self.document_type_obj)
+                resolved_document_type = get_document(self.document_type_obj)
+
+            if not issubclass(resolved_document_type, EmbeddedDocument):
+                # Due to the late resolution of the document_type
+                # There is a chance that it won't be an EmbeddedDocument (#1661)
+                self.error('Invalid embedded document class provided to an '
+                           'EmbeddedDocumentField')
+            self.document_type_obj = resolved_document_type
+
         return self.document_type_obj
 
     def to_python(self, value):
@@ -682,16 +734,28 @@ class GenericEmbeddedDocumentField(BaseField):
         return value
 
     def validate(self, value, clean=True):
+        if self.choices and isinstance(value, SON):
+            for choice in self.choices:
+                if value['_cls'] == choice._class_name:
+                    return True
+
         if not isinstance(value, EmbeddedDocument):
             self.error('Invalid embedded document instance provided to an '
                        'GenericEmbeddedDocumentField')
 
         value.validate(clean=clean)
 
+    def lookup_member(self, member_name):
+        if self.choices:
+            for choice in self.choices:
+                field = choice._fields.get(member_name)
+                if field:
+                    return field
+        return None
+
     def to_mongo(self, document, use_db_field=True, fields=None):
         if document is None:
             return None
-
         data = document.to_mongo(use_db_field, fields)
         if '_cls' not in data:
             data['_cls'] = document._class_name
@@ -775,10 +839,20 @@ class ListField(ComplexBaseField):
         kwargs.setdefault('default', lambda: [])
         super(ListField, self).__init__(**kwargs)
 
+    def __get__(self, instance, owner):
+        if instance is None:
+            # Document class being used rather than a document object
+            return self
+        value = instance._data.get(self.name)
+        LazyReferenceField = _import_class('LazyReferenceField')
+        GenericLazyReferenceField = _import_class('GenericLazyReferenceField')
+        if isinstance(self.field, (LazyReferenceField, GenericLazyReferenceField)) and value:
+            instance._data[self.name] = [self.field.build_lazyref(x) for x in value]
+        return super(ListField, self).__get__(instance, owner)
+
     def validate(self, value):
         """Make sure that a list of valid fields is being used."""
-        if (not isinstance(value, (list, tuple, QuerySet)) or
-                isinstance(value, six.string_types)):
+        if not isinstance(value, (list, tuple, BaseQuerySet)):
             self.error('Only lists and tuples may be used in a list field')
         super(ListField, self).validate(value)
 
@@ -885,12 +959,10 @@ class DictField(ComplexBaseField):
     .. versionchanged:: 0.5 - Can now handle complex / varying types of data
     """
 
-    def __init__(self, basecls=None, field=None, *args, **kwargs):
+    def __init__(self, field=None, *args, **kwargs):
         self.field = field
         self._auto_dereference = False
-        self.basecls = basecls or BaseField
-        if not issubclass(self.basecls, BaseField):
-            self.error('DictField only accepts dict values')
+
         kwargs.setdefault('default', lambda: {})
         super(DictField, self).__init__(*args, **kwargs)
 
@@ -909,7 +981,7 @@ class DictField(ComplexBaseField):
         super(DictField, self).validate(value)
 
     def lookup_member(self, member_name):
-        return DictField(basecls=self.basecls, db_field=member_name)
+        return DictField(db_field=member_name)
 
     def prepare_query_value(self, op, value):
         match_operators = ['contains', 'icontains', 'startswith',
@@ -919,7 +991,7 @@ class DictField(ComplexBaseField):
         if op in match_operators and isinstance(value, six.string_types):
             return StringField().prepare_query_value(op, value)
 
-        if hasattr(self.field, 'field'):
+        if hasattr(self.field, 'field'):    # Used for instance when using DictField(ListField(IntField()))
             if op in ('set', 'unset') and isinstance(value, dict):
                 return {
                     k: self.field.prepare_query_value(op, v)
@@ -939,6 +1011,7 @@ class MapField(DictField):
     """
 
     def __init__(self, field=None, *args, **kwargs):
+        # XXX ValidationError raised outside of the "validate" method.
         if not isinstance(field, BaseField):
             self.error('Argument to MapField constructor must be a valid '
                        'field')
@@ -949,6 +1022,15 @@ class ReferenceField(BaseField):
     """A reference to a document that will be automatically dereferenced on
     access (lazily).
 
+    Note this means you will get a database I/O access everytime you access
+    this field. This is necessary because the field returns a :class:`~mongoengine.Document`
+    which precise type can depend of the value of the `_cls` field present in the
+    document in database.
+    In short, using this type of field can lead to poor performances (especially
+    if you access this field only to retrieve it `pk` field which is already
+    known before dereference). To solve this you should consider using the
+    :class:`~mongoengine.fields.LazyReferenceField`.
+
     Use the `reverse_delete_rule` to handle what should happen if the document
     the field is referencing is deleted.  EmbeddedDocuments, DictFields and
     MapFields does not support reverse_delete_rule and an `InvalidDocumentError`
@@ -967,11 +1049,13 @@ class ReferenceField(BaseField):
 
     .. code-block:: python
 
-        class Bar(Document):
-            content = StringField()
-            foo = ReferenceField('Foo')
+        class Org(Document):
+            owner = ReferenceField('User')
 
-        Foo.register_delete_rule(Bar, 'foo', NULLIFY)
+        class User(Document):
+            org = ReferenceField('Org', reverse_delete_rule=CASCADE)
+
+        User.register_delete_rule(Org, 'owner', DENY)
 
     .. versionchanged:: 0.5 added `reverse_delete_rule`
     """
@@ -989,6 +1073,7 @@ class ReferenceField(BaseField):
             A reference to an abstract document type is always stored as a
             :class:`~pymongo.dbref.DBRef`, regardless of the value of `dbref`.
         """
+        # XXX ValidationError raised outside of the "validate" method.
         if (
             not isinstance(document_type, six.string_types) and
             not issubclass(document_type, Document)
@@ -1018,9 +1103,9 @@ class ReferenceField(BaseField):
 
         # Get value from document instance if available
         value = instance._data.get(self.name)
-        self._auto_dereference = instance._fields[self.name]._auto_dereference
+        auto_dereference = instance._fields[self.name]._auto_dereference
         # Dereference DBRefs
-        if self._auto_dereference and isinstance(value, DBRef):
+        if auto_dereference and isinstance(value, DBRef):
             if hasattr(value, 'cls'):
                 # Dereference using the class type specified in the reference
                 cls = get_document(value.cls)
@@ -1043,6 +1128,8 @@ class ReferenceField(BaseField):
         if isinstance(document, Document):
             # We need the id from the saved object to create the DBRef
             id_ = document.pk
+
+            # XXX ValidationError raised outside of the "validate" method.
             if id_ is None:
                 self.error('You can only reference documents once they have'
                            ' been saved to the database')
@@ -1082,21 +1169,13 @@ class ReferenceField(BaseField):
         return self.to_mongo(value)
 
     def validate(self, value):
-
-        if not isinstance(value, (self.document_type, DBRef, ObjectId)):
-            self.error('A ReferenceField only accepts DBRef, ObjectId or documents')
+        if not isinstance(value, (self.document_type, LazyReference, DBRef, ObjectId)):
+            self.error('A ReferenceField only accepts DBRef, LazyReference, ObjectId or documents')
 
         if isinstance(value, Document) and value.id is None:
             self.error('You can only reference documents once they have been '
                        'saved to the database')
 
-        if self.document_type._meta.get('abstract') and \
-                not isinstance(value, self.document_type):
-            self.error(
-                '%s is not an instance of abstract reference type %s' % (
-                    self.document_type._class_name)
-            )
-
     def lookup_member(self, member_name):
         return self.document_type._fields.get(member_name)
 
@@ -1117,6 +1196,7 @@ class CachedReferenceField(BaseField):
         if fields is None:
             fields = []
 
+        # XXX ValidationError raised outside of the "validate" method.
         if (
             not isinstance(document_type, six.string_types) and
             not issubclass(document_type, Document)
@@ -1176,9 +1256,10 @@ class CachedReferenceField(BaseField):
 
         # Get value from document instance if available
         value = instance._data.get(self.name)
-        self._auto_dereference = instance._fields[self.name]._auto_dereference
+        auto_dereference = instance._fields[self.name]._auto_dereference
+
         # Dereference DBRefs
-        if self._auto_dereference and isinstance(value, DBRef):
+        if auto_dereference and isinstance(value, DBRef):
             dereferenced = self.document_type._get_db().dereference(value)
             if dereferenced is None:
                 raise DoesNotExist('Trying to dereference unknown document %s' % value)
@@ -1191,6 +1272,7 @@ class CachedReferenceField(BaseField):
         id_field_name = self.document_type._meta['id_field']
         id_field = self.document_type._fields[id_field_name]
 
+        # XXX ValidationError raised outside of the "validate" method.
         if isinstance(document, Document):
             # We need the id from the saved object to create the DBRef
             id_ = document.pk
@@ -1199,7 +1281,6 @@ class CachedReferenceField(BaseField):
                            ' been saved to the database')
         else:
             self.error('Only accept a document object')
-            # TODO: should raise here or will fail next statement
 
         value = SON((
             ('_id', id_field.to_mongo(id_)),
@@ -1217,16 +1298,20 @@ class CachedReferenceField(BaseField):
         if value is None:
             return None
 
+        # XXX ValidationError raised outside of the "validate" method.
         if isinstance(value, Document):
             if value.pk is None:
                 self.error('You can only reference documents once they have'
                            ' been saved to the database')
-            return {'_id': value.pk}
+            value_dict = {'_id': value.pk}
+            for field in self.fields:
+                value_dict.update({field: value[field]})
+
+            return value_dict
 
         raise NotImplementedError
 
     def validate(self, value):
-
         if not isinstance(value, self.document_type):
             self.error('A CachedReferenceField only accepts documents')
 
@@ -1259,6 +1344,12 @@ class GenericReferenceField(BaseField):
     """A reference to *any* :class:`~mongoengine.document.Document` subclass
     that will be automatically dereferenced on access (lazily).
 
+    Note this field works the same way as :class:`~mongoengine.document.ReferenceField`,
+    doing database I/O access the first time it is accessed (even if it's to access
+    it ``pk`` or ``id`` field).
+    To solve this you should consider using the
+    :class:`~mongoengine.fields.GenericLazyReferenceField`.
+
     .. note ::
         * Any documents used as a generic reference must be registered in the
           document registry.  Importing the model will automatically register
@@ -1281,6 +1372,8 @@ class GenericReferenceField(BaseField):
                 elif isinstance(choice, type) and issubclass(choice, Document):
                     self.choices.append(choice._class_name)
                 else:
+                    # XXX ValidationError raised outside of the "validate"
+                    # method.
                     self.error('Invalid choices provided: must be a list of'
                                'Document subclasses and/or six.string_typess')
 
@@ -1299,8 +1392,8 @@ class GenericReferenceField(BaseField):
 
         value = instance._data.get(self.name)
 
-        self._auto_dereference = instance._fields[self.name]._auto_dereference
-        if self._auto_dereference and isinstance(value, (dict, SON)):
+        auto_dereference = instance._fields[self.name]._auto_dereference
+        if auto_dereference and isinstance(value, (dict, SON)):
             dereferenced = self.dereference(value)
             if dereferenced is None:
                 raise DoesNotExist('Trying to dereference unknown document %s' % value)
@@ -1344,6 +1437,7 @@ class GenericReferenceField(BaseField):
             # We need the id from the saved object to create the DBRef
             id_ = document.id
             if id_ is None:
+                # XXX ValidationError raised outside of the "validate" method.
                 self.error('You can only reference documents once they have'
                            ' been saved to the database')
         else:
@@ -1381,10 +1475,10 @@ class BinaryField(BaseField):
         return Binary(value)
 
     def validate(self, value):
-        if not isinstance(value, (six.binary_type, six.text_type, Binary)):
+        if not isinstance(value, (six.binary_type, Binary)):
             self.error('BinaryField only accepts instances of '
                        '(%s, %s, Binary)' % (
-                           six.binary_type.__name__, six.text_type.__name__))
+                           six.binary_type.__name__, Binary.__name__))
 
         if self.max_bytes is not None and len(value) > self.max_bytes:
             self.error('Binary value is too long')
@@ -1429,9 +1523,11 @@ class GridFSProxy(object):
     def __get__(self, instance, value):
         return self
 
-    def __nonzero__(self):
+    def __bool__(self):
         return bool(self.grid_id)
 
+    __nonzero__ = __bool__  # For Py2 support
+
     def __getstate__(self):
         self_dict = self.__dict__
         self_dict['_fs'] = None
@@ -1449,9 +1545,9 @@ class GridFSProxy(object):
         return '<%s: %s>' % (self.__class__.__name__, self.grid_id)
 
     def __str__(self):
-        name = getattr(
-            self.get(), 'filename', self.grid_id) if self.get() else '(no file)'
-        return '<%s: %s>' % (self.__class__.__name__, name)
+        gridout = self.get()
+        filename = getattr(gridout, 'filename') if gridout else '<no file>'
+        return '<%s: %s (%s)>' % (self.__class__.__name__, filename, self.grid_id)
 
     def __eq__(self, other):
         if isinstance(other, GridFSProxy):
@@ -1461,6 +1557,9 @@ class GridFSProxy(object):
         else:
             return False
 
+    def __ne__(self, other):
+        return not self == other
+
     @property
     def fs(self):
         if not self._fs:
@@ -1768,12 +1867,9 @@ class ImageField(FileField):
     """
     A Image File storage field.
 
-    @size (width, height, force):
-        max size to store images, if larger will be automatically resized
-        ex: size=(800, 600, True)
-
-    @thumbnail (width, height, force):
-        size to generate a thumbnail
+    :param size: max size to store images, provided as (width, height, force)
+        if larger, it will be automatically resized (ex: size=(800, 600, True))
+    :param thumbnail_size: size to generate a thumbnail, provided as (width, height, force)
 
     .. versionadded:: 0.6
     """
@@ -1844,8 +1940,7 @@ class SequenceField(BaseField):
         self.collection_name = collection_name or self.COLLECTION_NAME
         self.db_alias = db_alias or DEFAULT_CONNECTION_NAME
         self.sequence_name = sequence_name
-        self.value_decorator = (callable(value_decorator) and
-                                value_decorator or self.VALUE_DECORATOR)
+        self.value_decorator = value_decorator if callable(value_decorator) else self.VALUE_DECORATOR
         super(SequenceField, self).__init__(*args, **kwargs)
 
     def generate(self):
@@ -1954,7 +2049,7 @@ class UUIDField(BaseField):
                 if not isinstance(value, six.string_types):
                     value = six.text_type(value)
                 return uuid.UUID(value)
-            except Exception:
+            except (ValueError, TypeError, AttributeError):
                 return original_value
         return value
 
@@ -1976,7 +2071,7 @@ class UUIDField(BaseField):
                 value = str(value)
             try:
                 uuid.UUID(value)
-            except Exception as exc:
+            except (ValueError, TypeError, AttributeError) as exc:
                 self.error('Could not convert to UUID: %s' % exc)
 
 
@@ -2134,3 +2229,201 @@ class MultiPolygonField(GeoJsonBaseField):
     .. versionadded:: 0.9
     """
     _type = 'MultiPolygon'
+
+
+class LazyReferenceField(BaseField):
+    """A really lazy reference to a document.
+    Unlike the :class:`~mongoengine.fields.ReferenceField` it will
+    **not** be automatically (lazily) dereferenced on access.
+    Instead, access will return a :class:`~mongoengine.base.LazyReference` class
+    instance, allowing access to `pk` or manual dereference by using
+    ``fetch()`` method.
+
+    .. versionadded:: 0.15
+    """
+
+    def __init__(self, document_type, passthrough=False, dbref=False,
+                 reverse_delete_rule=DO_NOTHING, **kwargs):
+        """Initialises the Reference Field.
+
+        :param dbref:  Store the reference as :class:`~pymongo.dbref.DBRef`
+          or as the :class:`~pymongo.objectid.ObjectId`.id .
+        :param reverse_delete_rule: Determines what to do when the referring
+          object is deleted
+        :param passthrough: When trying to access unknown fields, the
+        :class:`~mongoengine.base.datastructure.LazyReference` instance will
+        automatically call `fetch()` and try to retrive the field on the fetched
+        document. Note this only work getting field (not setting or deleting).
+        """
+        # XXX ValidationError raised outside of the "validate" method.
+        if (
+            not isinstance(document_type, six.string_types) and
+            not issubclass(document_type, Document)
+        ):
+            self.error('Argument to LazyReferenceField constructor must be a '
+                       'document class or a string')
+
+        self.dbref = dbref
+        self.passthrough = passthrough
+        self.document_type_obj = document_type
+        self.reverse_delete_rule = reverse_delete_rule
+        super(LazyReferenceField, self).__init__(**kwargs)
+
+    @property
+    def document_type(self):
+        if isinstance(self.document_type_obj, six.string_types):
+            if self.document_type_obj == RECURSIVE_REFERENCE_CONSTANT:
+                self.document_type_obj = self.owner_document
+            else:
+                self.document_type_obj = get_document(self.document_type_obj)
+        return self.document_type_obj
+
+    def build_lazyref(self, value):
+        if isinstance(value, LazyReference):
+            if value.passthrough != self.passthrough:
+                value = LazyReference(value.document_type, value.pk, passthrough=self.passthrough)
+        elif value is not None:
+            if isinstance(value, self.document_type):
+                value = LazyReference(self.document_type, value.pk, passthrough=self.passthrough)
+            elif isinstance(value, DBRef):
+                value = LazyReference(self.document_type, value.id, passthrough=self.passthrough)
+            else:
+                # value is the primary key of the referenced document
+                value = LazyReference(self.document_type, value, passthrough=self.passthrough)
+        return value
+
+    def __get__(self, instance, owner):
+        """Descriptor to allow lazy dereferencing."""
+        if instance is None:
+            # Document class being used rather than a document object
+            return self
+
+        value = self.build_lazyref(instance._data.get(self.name))
+        if value:
+            instance._data[self.name] = value
+
+        return super(LazyReferenceField, self).__get__(instance, owner)
+
+    def to_mongo(self, value):
+        if isinstance(value, LazyReference):
+            pk = value.pk
+        elif isinstance(value, self.document_type):
+            pk = value.pk
+        elif isinstance(value, DBRef):
+            pk = value.id
+        else:
+            # value is the primary key of the referenced document
+            pk = value
+        id_field_name = self.document_type._meta['id_field']
+        id_field = self.document_type._fields[id_field_name]
+        pk = id_field.to_mongo(pk)
+        if self.dbref:
+            return DBRef(self.document_type._get_collection_name(), pk)
+        else:
+            return pk
+
+    def validate(self, value):
+        if isinstance(value, LazyReference):
+            if value.collection != self.document_type._get_collection_name():
+                self.error('Reference must be on a `%s` document.' % self.document_type)
+            pk = value.pk
+        elif isinstance(value, self.document_type):
+            pk = value.pk
+        elif isinstance(value, DBRef):
+            # TODO: check collection ?
+            collection = self.document_type._get_collection_name()
+            if value.collection != collection:
+                self.error("DBRef on bad collection (must be on `%s`)" % collection)
+            pk = value.id
+        else:
+            # value is the primary key of the referenced document
+            id_field_name = self.document_type._meta['id_field']
+            id_field = getattr(self.document_type, id_field_name)
+            pk = value
+            try:
+                id_field.validate(pk)
+            except ValidationError:
+                self.error(
+                    "value should be `{0}` document, LazyReference or DBRef on `{0}` "
+                    "or `{0}`'s primary key (i.e. `{1}`)".format(
+                        self.document_type.__name__, type(id_field).__name__))
+
+        if pk is None:
+            self.error('You can only reference documents once they have been '
+                       'saved to the database')
+
+    def prepare_query_value(self, op, value):
+        if value is None:
+            return None
+        super(LazyReferenceField, self).prepare_query_value(op, value)
+        return self.to_mongo(value)
+
+    def lookup_member(self, member_name):
+        return self.document_type._fields.get(member_name)
+
+
+class GenericLazyReferenceField(GenericReferenceField):
+    """A reference to *any* :class:`~mongoengine.document.Document` subclass.
+    Unlike the :class:`~mongoengine.fields.GenericReferenceField` it will
+    **not** be automatically (lazily) dereferenced on access.
+    Instead, access will return a :class:`~mongoengine.base.LazyReference` class
+    instance, allowing access to `pk` or manual dereference by using
+    ``fetch()`` method.
+
+    .. note ::
+        * Any documents used as a generic reference must be registered in the
+          document registry.  Importing the model will automatically register
+          it.
+
+        * You can use the choices param to limit the acceptable Document types
+
+    .. versionadded:: 0.15
+    """
+
+    def __init__(self, *args, **kwargs):
+        self.passthrough = kwargs.pop('passthrough', False)
+        super(GenericLazyReferenceField, self).__init__(*args, **kwargs)
+
+    def _validate_choices(self, value):
+        if isinstance(value, LazyReference):
+            value = value.document_type._class_name
+        super(GenericLazyReferenceField, self)._validate_choices(value)
+
+    def build_lazyref(self, value):
+        if isinstance(value, LazyReference):
+            if value.passthrough != self.passthrough:
+                value = LazyReference(value.document_type, value.pk, passthrough=self.passthrough)
+        elif value is not None:
+            if isinstance(value, (dict, SON)):
+                value = LazyReference(get_document(value['_cls']), value['_ref'].id, passthrough=self.passthrough)
+            elif isinstance(value, Document):
+                value = LazyReference(type(value), value.pk, passthrough=self.passthrough)
+        return value
+
+    def __get__(self, instance, owner):
+        if instance is None:
+            return self
+
+        value = self.build_lazyref(instance._data.get(self.name))
+        if value:
+            instance._data[self.name] = value
+
+        return super(GenericLazyReferenceField, self).__get__(instance, owner)
+
+    def validate(self, value):
+        if isinstance(value, LazyReference) and value.pk is None:
+            self.error('You can only reference documents once they have been'
+                       ' saved to the database')
+        return super(GenericLazyReferenceField, self).validate(value)
+
+    def to_mongo(self, document):
+        if document is None:
+            return None
+
+        if isinstance(document, LazyReference):
+            return SON((
+                ('_cls', document.document_type._class_name),
+                ('_ref', DBRef(document.document_type._get_collection_name(), document.pk))
+            ))
+        else:
+            return super(GenericLazyReferenceField, self).to_mongo(document)

mongoengine/python_support.py

@@ -6,11 +6,7 @@ import pymongo
 import six
 
 
-if pymongo.version_tuple[0] < 3:
-    IS_PYMONGO_3 = False
-else:
-    IS_PYMONGO_3 = True
-
+IS_PYMONGO_3 = pymongo.version_tuple[0] >= 3
 
 # six.BytesIO resolves to StringIO.StringIO in Py2 and io.BytesIO in Py3.
 StringIO = six.BytesIO
@@ -23,3 +19,10 @@ if not six.PY3:
         pass
     else:
         StringIO = cStringIO.StringIO
+
+
+if six.PY3:
+    from collections.abc import Hashable
+else:
+    # raises DeprecationWarnings in Python >=3.7
+    from collections import Hashable

mongoengine/queryset/base.py

@@ -2,7 +2,6 @@ from __future__ import absolute_import
 
 import copy
 import itertools
-import operator
 import pprint
 import re
 import warnings
@@ -18,7 +17,7 @@ from mongoengine import signals
 from mongoengine.base import get_document
 from mongoengine.common import _import_class
 from mongoengine.connection import get_db
-from mongoengine.context_managers import switch_db
+from mongoengine.context_managers import set_write_concern, switch_db
 from mongoengine.errors import (InvalidQueryError, LookUpError,
                                 NotUniqueError, OperationError)
 from mongoengine.python_support import IS_PYMONGO_3
@@ -39,8 +38,6 @@ CASCADE = 2
 DENY = 3
 PULL = 4
 
-RE_TYPE = type(re.compile(''))
-
 
 class BaseQuerySet(object):
     """A set of results returned from a query. Wraps a MongoDB cursor,
@@ -209,14 +206,12 @@ class BaseQuerySet(object):
         queryset = self.order_by()
         return False if queryset.first() is None else True
 
-    def __nonzero__(self):
-        """Avoid to open all records in an if stmt in Py2."""
-        return self._has_data()
-
     def __bool__(self):
         """Avoid to open all records in an if stmt in Py3."""
         return self._has_data()
 
+    __nonzero__ = __bool__  # For Py2 support
+
     # Core functions
 
     def all(self):
@@ -269,13 +264,13 @@ class BaseQuerySet(object):
         queryset = queryset.filter(*q_objs, **query)
 
         try:
-            result = queryset.next()
+            result = six.next(queryset)
         except StopIteration:
             msg = ('%s matching query does not exist.'
                    % queryset._document._class_name)
             raise queryset._document.DoesNotExist(msg)
         try:
-            queryset.next()
+            six.next(queryset)
         except StopIteration:
             return result
 
@@ -350,11 +345,24 @@ class BaseQuerySet(object):
                                      documents=docs, **signal_kwargs)
 
         raw = [doc.to_mongo() for doc in docs]
+
+        with set_write_concern(self._collection, write_concern) as collection:
+            insert_func = collection.insert_many
+            if return_one:
+                raw = raw[0]
+                insert_func = collection.insert_one
+
         try:
-            ids = self._collection.insert(raw, **write_concern)
+            inserted_result = insert_func(raw)
+            ids = [inserted_result.inserted_id] if return_one else inserted_result.inserted_ids
         except pymongo.errors.DuplicateKeyError as err:
             message = 'Could not save document (%s)'
             raise NotUniqueError(message % six.text_type(err))
+        except pymongo.errors.BulkWriteError as err:
+            # inserting documents that already have an _id field will
+            # give huge performance debt or raise
+            message = u'Document must not have _id value before bulk write (%s)'
+            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', six.text_type(err)):
@@ -364,18 +372,20 @@ class BaseQuerySet(object):
                 raise NotUniqueError(message % six.text_type(err))
             raise OperationError(message % six.text_type(err))
 
+        # Apply inserted_ids to documents
+        for doc, doc_id in zip(docs, ids):
+            doc.pk = doc_id
+
         if not load_bulk:
             signals.post_bulk_insert.send(
                 self._document, documents=docs, loaded=False, **signal_kwargs)
-            return return_one and ids[0] or ids
+            return ids[0] if return_one else ids
 
         documents = self.in_bulk(ids)
-        results = []
-        for obj_id in ids:
-            results.append(documents.get(obj_id))
+        results = [documents.get(obj_id) for obj_id in ids]
         signals.post_bulk_insert.send(
             self._document, documents=results, loaded=True, **signal_kwargs)
-        return return_one and results[0] or results
+        return results[0] if return_one else results
 
     def count(self, with_limit_and_skip=False):
         """Count the selected elements in the query.
@@ -384,7 +394,7 @@ class BaseQuerySet(object):
             :meth:`skip` that has been applied to this cursor into account when
             getting the count
         """
-        if self._limit == 0 and with_limit_and_skip or self._none:
+        if self._limit == 0 and with_limit_and_skip is False or self._none:
             return 0
         return self._cursor.count(with_limit_and_skip=with_limit_and_skip)
 
@@ -486,8 +496,9 @@ class BaseQuerySet(object):
             ``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.
-        :param full_result: Return the full result rather than just the number
-            updated.
+        :param full_result: Return the full result dictionary rather than just the number
+            updated, e.g. return
+            ``{'n': 2, 'nModified': 2, 'ok': 1.0, 'updatedExisting': True}``.
         :param update: Django-style update keyword arguments
 
         .. versionadded:: 0.2
@@ -510,12 +521,15 @@ class BaseQuerySet(object):
             else:
                 update['$set'] = {'_cls': queryset._document._class_name}
         try:
-            result = queryset._collection.update(query, update, multi=multi,
-                                                 upsert=upsert, **write_concern)
+            with set_write_concern(queryset._collection, write_concern) as collection:
+                update_func = collection.update_one
+                if multi:
+                    update_func = collection.update_many
+                result = update_func(query, update, upsert=upsert)
             if full_result:
                 return result
-            elif result:
-                return result['n']
+            elif result.raw_result:
+                return result.raw_result['n']
         except pymongo.errors.DuplicateKeyError as err:
             raise NotUniqueError(u'Update failed (%s)' % six.text_type(err))
         except pymongo.errors.OperationFailure as err:
@@ -544,10 +558,10 @@ class BaseQuerySet(object):
                                     write_concern=write_concern,
                                     full_result=True, **update)
 
-        if atomic_update['updatedExisting']:
+        if atomic_update.raw_result['updatedExisting']:
             document = self.get()
         else:
-            document = self._document.objects.with_id(atomic_update['upserted'])
+            document = self._document.objects.with_id(atomic_update.upserted_id)
         return document
 
     def update_one(self, upsert=False, write_concern=None, **update):
@@ -759,10 +773,11 @@ class BaseQuerySet(object):
         """Limit the number of returned documents to `n`. This may also be
         achieved using array-slicing syntax (e.g. ``User.objects[:5]``).
 
-        :param n: the maximum number of objects to return
+        :param n: the maximum number of objects to return if n is greater than 0.
+        When 0 is passed, returns all the documents in the cursor
         """
         queryset = self.clone()
-        queryset._limit = n if n != 0 else 1
+        queryset._limit = n
 
         # If a cursor object has already been created, apply the limit to it.
         if queryset._cursor_obj:
@@ -960,11 +975,10 @@ class BaseQuerySet(object):
         # explicitly included, and then more complicated operators such as
         # $slice.
         def _sort_key(field_tuple):
-            key, value = field_tuple
-            if isinstance(value, (int)):
+            _, value = field_tuple
+            if isinstance(value, int):
                 return value  # 0 for exclusion, 1 for inclusion
-            else:
-                return 2  # so that complex values appear last
+            return 2  # so that complex values appear last
 
         fields = sorted(cleaned_fields, key=_sort_key)
 
@@ -1182,6 +1196,10 @@ class BaseQuerySet(object):
 
         pipeline = initial_pipeline + list(pipeline)
 
+        if IS_PYMONGO_3 and self._read_preference is not None:
+            return self._collection.with_options(read_preference=self._read_preference) \
+                       .aggregate(pipeline, cursor={}, **kwargs)
+
         return self._collection.aggregate(pipeline, cursor={}, **kwargs)
 
     # JS functionality
@@ -1457,13 +1475,13 @@ class BaseQuerySet(object):
 
     # Iterator helpers
 
-    def next(self):
+    def __next__(self):
         """Wrap the result in a :class:`~mongoengine.Document` object.
         """
         if self._limit == 0 or self._none:
             raise StopIteration
 
-        raw_doc = self._cursor.next()
+        raw_doc = six.next(self._cursor)
 
         if self._as_pymongo:
             return self._get_as_pymongo(raw_doc)
@@ -1477,6 +1495,8 @@ class BaseQuerySet(object):
 
         return doc
 
+    next = __next__     # For Python2 support
+
     def rewind(self):
         """Rewind the cursor to its unevaluated state.
 
@@ -1578,6 +1598,9 @@ class BaseQuerySet(object):
         if self._batch_size is not None:
             self._cursor_obj.batch_size(self._batch_size)
 
+        if self._comment is not None:
+            self._cursor_obj.comment(self._comment)
+
         return self._cursor_obj
 
     def __deepcopy__(self, memo):
@@ -1722,25 +1745,33 @@ class BaseQuerySet(object):
         return frequencies
 
     def _fields_to_dbfields(self, fields):
-        """Translate fields paths to its db equivalents"""
-        ret = []
+        """Translate fields' paths to their db equivalents."""
         subclasses = []
-        document = self._document
-        if document._meta['allow_inheritance']:
+        if self._document._meta['allow_inheritance']:
             subclasses = [get_document(x)
-                          for x in document._subclasses][1:]
+                          for x in self._document._subclasses][1:]
+
+        db_field_paths = []
         for field in fields:
+            field_parts = field.split('.')
             try:
-                field = '.'.join(f.db_field for f in
-                                 document._lookup_field(field.split('.')))
-                ret.append(field)
+                field = '.'.join(
+                    f if isinstance(f, six.string_types) else f.db_field
+                    for f in self._document._lookup_field(field_parts)
+                )
+                db_field_paths.append(field)
             except LookUpError as err:
                 found = False
+
+                # If a field path wasn't found on the main document, go
+                # through its subclasses and see if it exists on any of them.
                 for subdoc in subclasses:
                     try:
-                        subfield = '.'.join(f.db_field for f in
-                                            subdoc._lookup_field(field.split('.')))
-                        ret.append(subfield)
+                        subfield = '.'.join(
+                            f if isinstance(f, six.string_types) else f.db_field
+                            for f in subdoc._lookup_field(field_parts)
+                        )
+                        db_field_paths.append(subfield)
                         found = True
                         break
                     except LookUpError:
@@ -1748,7 +1779,8 @@ class BaseQuerySet(object):
 
                 if not found:
                     raise err
-        return ret
+
+        return db_field_paths
 
     def _get_order_by(self, keys):
         """Given a list of MongoEngine-style sort keys, return a list
@@ -1840,8 +1872,8 @@ class BaseQuerySet(object):
             # Substitute the correct name for the field into the javascript
             return '.'.join([f.db_field for f in fields])
 
-        code = re.sub(u'\[\s*~([A-z_][A-z_0-9.]+?)\s*\]', field_sub, code)
-        code = re.sub(u'\{\{\s*~([A-z_][A-z_0-9.]+?)\s*\}\}', field_path_sub,
+        code = re.sub(r'\[\s*~([A-z_][A-z_0-9.]+?)\s*\]', field_sub, code)
+        code = re.sub(r'\{\{\s*~([A-z_][A-z_0-9.]+?)\s*\}\}', field_path_sub,
                       code)
         return code
 

mongoengine/queryset/field_list.py

@@ -63,9 +63,11 @@ class QueryFieldList(object):
             self._only_called = True
         return self
 
-    def __nonzero__(self):
+    def __bool__(self):
         return bool(self.fields)
 
+    __nonzero__ = __bool__  # For Py2 support
+
     def as_dict(self):
         field_list = {field: self.value for field in self.fields}
         if self.slice:

mongoengine/queryset/manager.py

@@ -36,7 +36,7 @@ class QuerySetManager(object):
         queryset_class = owner._meta.get('queryset_class', self.default)
         queryset = queryset_class(owner, owner._get_collection())
         if self.get_queryset:
-            arg_count = self.get_queryset.func_code.co_argcount
+            arg_count = self.get_queryset.__code__.co_argcount
             if arg_count == 1:
                 queryset = self.get_queryset(queryset)
             elif arg_count == 2:

mongoengine/queryset/queryset.py

@@ -1,3 +1,5 @@
+import six
+
 from mongoengine.errors import OperationError
 from mongoengine.queryset.base import (BaseQuerySet, CASCADE, DENY, DO_NOTHING,
                                        NULLIFY, PULL)
@@ -87,10 +89,10 @@ class QuerySet(BaseQuerySet):
                 yield self._result_cache[pos]
                 pos += 1
 
-            # Raise StopIteration if we already established there were no more
+            # return if we already established there were no more
             # docs in the db cursor.
             if not self._has_more:
-                raise StopIteration
+                return
 
             # Otherwise, populate more of the cache and repeat.
             if len(self._result_cache) <= pos:
@@ -112,8 +114,8 @@ class QuerySet(BaseQuerySet):
         # 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())
+            for _ in six.moves.range(ITER_CHUNK_SIZE):
+                self._result_cache.append(six.next(self))
         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
@@ -166,9 +168,9 @@ class QuerySetNoCache(BaseQuerySet):
             return '.. queryset mid-iteration ..'
 
         data = []
-        for _ in xrange(REPR_OUTPUT_SIZE + 1):
+        for _ in six.moves.range(REPR_OUTPUT_SIZE + 1):
             try:
-                data.append(self.next())
+                data.append(six.next(self))
             except StopIteration:
                 break
 
@@ -184,10 +186,3 @@ class QuerySetNoCache(BaseQuerySet):
             queryset = self.clone()
         queryset.rewind()
         return queryset
-
-
-class QuerySetNoDeRef(QuerySet):
-    """Special no_dereference QuerySet"""
-
-    def __dereference(items, max_depth=1, instance=None, name=None):
-        return items

mongoengine/queryset/transform.py

@@ -101,21 +101,8 @@ def query(_doc_cls=None, **kwargs):
                         value = value['_id']
 
             elif op in ('in', 'nin', 'all', 'near') and not isinstance(value, dict):
-                # 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]
+                # Raise an error if the in/nin/all/near param is not iterable.
+                value = _prepare_query_for_iterable(field, op, value)
 
             # If we're querying a GenericReferenceField, we need to alter the
             # key depending on the value:
@@ -160,7 +147,7 @@ def query(_doc_cls=None, **kwargs):
         if op is None or key not in mongo_query:
             mongo_query[key] = value
         elif key in mongo_query:
-            if isinstance(mongo_query[key], dict):
+            if isinstance(mongo_query[key], dict) and isinstance(value, dict):
                 mongo_query[key].update(value)
                 # $max/minDistance needs to come last - convert to SON
                 value_dict = mongo_query[key]
@@ -214,30 +201,37 @@ def update(_doc_cls=None, **update):
     format.
     """
     mongo_update = {}
+
     for key, value in update.items():
         if key == '__raw__':
             mongo_update.update(value)
             continue
+
         parts = key.split('__')
+
         # 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
         op = None
         if parts[0] in UPDATE_OPERATORS:
             op = parts.pop(0)
             # Convert Pythonic names to Mongo equivalents
-            if op in ('push_all', 'pull_all'):
-                op = op.replace('_all', 'All')
-            elif op == 'dec':
+            operator_map = {
+                'push_all': 'pushAll',
+                'pull_all': 'pullAll',
+                'dec': 'inc',
+                'add_to_set': 'addToSet',
+                'set_on_insert': 'setOnInsert'
+            }
+            if op == 'dec':
                 # Support decrement by flipping a positive value's sign
                 # and using 'inc'
-                op = 'inc'
                 value = -value
-            elif op == 'add_to_set':
-                op = 'addToSet'
-            elif op == 'set_on_insert':
-                op = 'setOnInsert'
+            # If the operator doesn't found from operator map, the op value
+            # will stay unchanged
+            op = operator_map.get(op, op)
 
         match = None
         if parts[-1] in COMPARISON_OPERATORS:
@@ -284,7 +278,15 @@ def update(_doc_cls=None, **update):
             if isinstance(field, GeoJsonBaseField):
                 value = field.to_mongo(value)
 
-            if op in (None, 'set', 'push', 'pull'):
+            if op == 'pull':
+                if field.required or value is not None:
+                    if match == 'in' and not isinstance(value, dict):
+                        value = _prepare_query_for_iterable(field, op, value)
+                    else:
+                        value = field.prepare_query_value(op, value)
+            elif op == 'push' and isinstance(value, (list, tuple, set)):
+                value = [field.prepare_query_value(op, v) for v in value]
+            elif op in (None, 'set', 'push'):
                 if field.required or value is not None:
                     value = field.prepare_query_value(op, value)
             elif op in ('pushAll', 'pullAll'):
@@ -296,6 +298,8 @@ def update(_doc_cls=None, **update):
                     value = field.prepare_query_value(op, value)
             elif op == 'unset':
                 value = 1
+            elif op == 'inc':
+                value = field.prepare_query_value(op, value)
 
         if match:
             match = '$' + match
@@ -319,11 +323,17 @@ def update(_doc_cls=None, **update):
             field_classes = [c.__class__ for c in cleaned_fields]
             field_classes.reverse()
             ListField = _import_class('ListField')
-            if ListField in field_classes:
-                # Join all fields via dot notation to the last ListField
+            EmbeddedDocumentListField = _import_class('EmbeddedDocumentListField')
+            if ListField in field_classes or EmbeddedDocumentListField in field_classes:
+                # Join all fields via dot notation to the last ListField or EmbeddedDocumentListField
                 # Then process as normal
+                if ListField in field_classes:
+                    _check_field = ListField
+                else:
+                    _check_field = EmbeddedDocumentListField
+
                 last_listField = len(
-                    cleaned_fields) - field_classes.index(ListField)
+                    cleaned_fields) - field_classes.index(_check_field)
                 key = '.'.join(parts[:last_listField])
                 parts = parts[last_listField:]
                 parts.insert(0, key)
@@ -333,10 +343,26 @@ def update(_doc_cls=None, **update):
                 value = {key: value}
         elif op == 'addToSet' and isinstance(value, list):
             value = {key: {'$each': value}}
+        elif op in ('push', 'pushAll'):
+            if parts[-1].isdigit():
+                key = parts[0]
+                position = int(parts[-1])
+                # $position expects an iterable. If pushing a single value,
+                # wrap it in a list.
+                if not isinstance(value, (set, tuple, list)):
+                    value = [value]
+                value = {key: {'$each': value, '$position': position}}
+            else:
+                if op == 'pushAll':
+                    op = 'push'  # convert to non-deprecated keyword
+                    if not isinstance(value, (set, tuple, list)):
+                        value = [value]
+                    value = {key: {'$each': value}}
+                else:
+                    value = {key: value}
         else:
             value = {key: value}
         key = '$' + op
-
         if key not in mongo_update:
             mongo_update[key] = value
         elif key in mongo_update and isinstance(mongo_update[key], dict):
@@ -403,7 +429,6 @@ def _infer_geometry(value):
                                 '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]
@@ -425,3 +450,22 @@ def _infer_geometry(value):
 
     raise InvalidQueryError('Invalid $geometry data. Can be either a '
                             'dictionary or (nested) lists of coordinate(s)')
+
+
+def _prepare_query_for_iterable(field, op, value):
+    # 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]).")
+
+    if not hasattr(value, '__iter__'):
+        raise TypeError("The `in`, `nin`, `all`, or "
+                        "`near`-operators must be applied to an "
+                        "iterable (e.g. a list).")
+
+    return [field.prepare_query_value(op, v) for v in value]

mongoengine/queryset/visitor.py

@@ -3,7 +3,7 @@ import copy
 from mongoengine.errors import InvalidQueryError
 from mongoengine.queryset import transform
 
-__all__ = ('Q',)
+__all__ = ('Q', 'QNode')
 
 
 class QNodeVisitor(object):
@@ -131,6 +131,10 @@ class QCombination(QNode):
             else:
                 self.children.append(node)
 
+    def __repr__(self):
+        op = ' & ' if self.operation is self.AND else ' | '
+        return '(%s)' % op.join([repr(node) for node in self.children])
+
     def accept(self, visitor):
         for i in range(len(self.children)):
             if isinstance(self.children[i], QNode):
@@ -151,6 +155,9 @@ class Q(QNode):
     def __init__(self, **query):
         self.query = query
 
+    def __repr__(self):
+        return 'Q(**%s)' % repr(self.query)
+
     def accept(self, visitor):
         return visitor.visit_query(self)
 

setup.cfg

@@ -1,11 +1,11 @@
 [nosetests]
 verbosity=2
 detailed-errors=1
-tests=tests
+#tests=tests
 cover-package=mongoengine
 
 [flake8]
-ignore=E501,F401,F403,F405,I201
+ignore=E501,F401,F403,F405,I201,I202,W504, W605
 exclude=build,dist,docs,venv,venv3,.tox,.eggs,tests
 max-complexity=47
 application-import-names=mongoengine,tests

setup.py

@@ -44,9 +44,8 @@ CLASSIFIERS = [
     "Programming Language :: Python :: 2",
     "Programming Language :: Python :: 2.7",
     "Programming Language :: Python :: 3",
-    "Programming Language :: Python :: 3.3",
-    "Programming Language :: Python :: 3.4",
     "Programming Language :: Python :: 3.5",
+    "Programming Language :: Python :: 3.6",
     "Programming Language :: Python :: Implementation :: CPython",
     "Programming Language :: Python :: Implementation :: PyPy",
     'Topic :: Database',
@@ -70,9 +69,9 @@ 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",
+    author_email='harry.marr@gmail.com',
+    maintainer="Stefan Wojcik",
+    maintainer_email="wojcikstefan@gmail.com",
     url='http://mongoengine.org/',
     download_url='https://github.com/MongoEngine/mongoengine/tarball/master',
     license='MIT',

tests/__init__.py

@@ -1,4 +1,4 @@
-from all_warnings import AllWarnings
-from document import *
-from queryset import *
-from fields import *
+from .all_warnings import AllWarnings
+from .document import *
+from .queryset import *
+from .fields import *

tests/document/__init__.py

@@ -1,13 +1,13 @@
 import unittest
 
-from class_methods import *
-from delta import *
-from dynamic import *
-from indexes import *
-from inheritance import *
-from instance import *
-from json_serialisation import *
-from validation import *
+from .class_methods import *
+from .delta import *
+from .dynamic import *
+from .indexes import *
+from .inheritance import *
+from .instance import *
+from .json_serialisation import *
+from .validation import *
 
 if __name__ == '__main__':
     unittest.main()

tests/document/class_methods.py

@@ -5,6 +5,7 @@ from mongoengine import *
 
 from mongoengine.queryset import NULLIFY, PULL
 from mongoengine.connection import get_db
+from tests.utils import requires_mongodb_gte_26
 
 __all__ = ("ClassMethodsTest", )
 
@@ -65,10 +66,10 @@ class ClassMethodsTest(unittest.TestCase):
         """
         collection_name = 'person'
         self.Person(name='Test').save()
-        self.assertTrue(collection_name in self.db.collection_names())
+        self.assertIn(collection_name,  self.db.collection_names())
 
         self.Person.drop_collection()
-        self.assertFalse(collection_name in self.db.collection_names())
+        self.assertNotIn(collection_name, self.db.collection_names())
 
     def test_register_delete_rule(self):
         """Ensure that register delete rule adds a delete rule to the document
@@ -187,6 +188,26 @@ class ClassMethodsTest(unittest.TestCase):
         self.assertEqual(BlogPostWithTags.compare_indexes(), { 'missing': [], 'extra': [] })
         self.assertEqual(BlogPostWithCustomField.compare_indexes(), { 'missing': [], 'extra': [] })
 
+    @requires_mongodb_gte_26
+    def test_compare_indexes_for_text_indexes(self):
+        """ Ensure that compare_indexes behaves correctly for text indexes """
+
+        class Doc(Document):
+            a = StringField()
+            b = StringField()
+            meta = {'indexes': [
+                {'fields': ['$a', "$b"],
+                 'default_language': 'english',
+                 'weights': {'a': 10, 'b': 2}
+                }
+            ]}
+
+        Doc.drop_collection()
+        Doc.ensure_indexes()
+        actual = Doc.compare_indexes()
+        expected = {'missing': [], 'extra': []}
+        self.assertEqual(actual, expected)
+
     def test_list_indexes_inheritance(self):
         """ ensure that all of the indexes are listed regardless of the super-
         or sub-class that we call it from
@@ -319,7 +340,7 @@ class ClassMethodsTest(unittest.TestCase):
             meta = {'collection': collection_name}
 
         Person(name="Test User").save()
-        self.assertTrue(collection_name in self.db.collection_names())
+        self.assertIn(collection_name, self.db.collection_names())
 
         user_obj = self.db[collection_name].find_one()
         self.assertEqual(user_obj['name'], "Test User")
@@ -328,7 +349,7 @@ class ClassMethodsTest(unittest.TestCase):
         self.assertEqual(user_obj.name, "Test User")
 
         Person.drop_collection()
-        self.assertFalse(collection_name in self.db.collection_names())
+        self.assertNotIn(collection_name, self.db.collection_names())
 
     def test_collection_name_and_primary(self):
         """Ensure that a collection with a specified name may be used.

tests/document/delta.py

@@ -694,7 +694,7 @@ class DeltaTest(unittest.TestCase):
         organization.employees.append(person)
         updates, removals = organization._delta()
         self.assertEqual({}, removals)
-        self.assertTrue('employees' in updates)
+        self.assertIn('employees', updates)
 
     def test_delta_with_dbref_false(self):
         person, organization, employee = self.circular_reference_deltas_2(Document, Document, False)
@@ -709,7 +709,7 @@ class DeltaTest(unittest.TestCase):
         organization.employees.append(person)
         updates, removals = organization._delta()
         self.assertEqual({}, removals)
-        self.assertTrue('employees' in updates)
+        self.assertIn('employees', updates)
 
     def test_nested_nested_fields_mark_as_changed(self):
         class EmbeddedDoc(EmbeddedDocument):

tests/document/dynamic.py

@@ -174,8 +174,8 @@ class DynamicTest(unittest.TestCase):
 
         Employee.drop_collection()
 
-        self.assertTrue('name' in Employee._fields)
-        self.assertTrue('salary' in Employee._fields)
+        self.assertIn('name', Employee._fields)
+        self.assertIn('salary', Employee._fields)
         self.assertEqual(Employee._get_collection_name(),
                          self.Person._get_collection_name())
 
@@ -189,7 +189,7 @@ class DynamicTest(unittest.TestCase):
         self.assertEqual(1, Employee.objects(age=20).count())
 
         joe_bloggs = self.Person.objects.first()
-        self.assertTrue(isinstance(joe_bloggs, Employee))
+        self.assertIsInstance(joe_bloggs, Employee)
 
     def test_embedded_dynamic_document(self):
         """Test dynamic embedded documents"""

tests/document/indexes.py

@@ -1,15 +1,14 @@
 # -*- coding: utf-8 -*-
 import unittest
-import sys
+from datetime import datetime
 
 from nose.plugins.skip import SkipTest
-from datetime import datetime
+from pymongo.errors import OperationFailure
 import pymongo
 
 from mongoengine import *
 from mongoengine.connection import get_db
-
-from tests.utils import get_mongodb_version, needs_mongodb_v26
+from tests.utils import get_mongodb_version, requires_mongodb_gte_26, MONGODB_32, MONGODB_3
 
 __all__ = ("IndexesTest", )
 
@@ -19,6 +18,7 @@ class IndexesTest(unittest.TestCase):
     def setUp(self):
         self.connection = connect(db='mongoenginetest')
         self.db = get_db()
+        self.mongodb_version = get_mongodb_version()
 
         class Person(Document):
             name = StringField()
@@ -70,7 +70,7 @@ class IndexesTest(unittest.TestCase):
         self.assertEqual(len(info), 4)
         info = [value['key'] for key, value in info.iteritems()]
         for expected in expected_specs:
-            self.assertTrue(expected['fields'] in info)
+            self.assertIn(expected['fields'], info)
 
     def _index_test_inheritance(self, InheritFrom):
 
@@ -102,7 +102,7 @@ class IndexesTest(unittest.TestCase):
         self.assertEqual(len(info), 4)
         info = [value['key'] for key, value in info.iteritems()]
         for expected in expected_specs:
-            self.assertTrue(expected['fields'] in info)
+            self.assertIn(expected['fields'], info)
 
         class ExtendedBlogPost(BlogPost):
             title = StringField()
@@ -117,7 +117,7 @@ class IndexesTest(unittest.TestCase):
         info = ExtendedBlogPost.objects._collection.index_information()
         info = [value['key'] for key, value in info.iteritems()]
         for expected in expected_specs:
-            self.assertTrue(expected['fields'] in info)
+            self.assertIn(expected['fields'], info)
 
     def test_indexes_document_inheritance(self):
         """Ensure that indexes are used when meta[indexes] is specified for
@@ -226,7 +226,7 @@ class IndexesTest(unittest.TestCase):
         list(Person.objects)
         info = Person.objects._collection.index_information()
         info = [value['key'] for key, value in info.iteritems()]
-        self.assertTrue([('rank.title', 1)] in info)
+        self.assertIn([('rank.title', 1)], info)
 
     def test_explicit_geo2d_index(self):
         """Ensure that geo2d indexes work when created via meta[indexes]
@@ -246,7 +246,7 @@ class IndexesTest(unittest.TestCase):
         Place.ensure_indexes()
         info = Place._get_collection().index_information()
         info = [value['key'] for key, value in info.iteritems()]
-        self.assertTrue([('location.point', '2d')] in info)
+        self.assertIn([('location.point', '2d')], info)
 
     def test_explicit_geo2d_index_embedded(self):
         """Ensure that geo2d indexes work when created via meta[indexes]
@@ -269,7 +269,7 @@ class IndexesTest(unittest.TestCase):
         Place.ensure_indexes()
         info = Place._get_collection().index_information()
         info = [value['key'] for key, value in info.iteritems()]
-        self.assertTrue([('current.location.point', '2d')] in info)
+        self.assertIn([('current.location.point', '2d')], info)
 
     def test_explicit_geosphere_index(self):
         """Ensure that geosphere indexes work when created via meta[indexes]
@@ -289,7 +289,7 @@ class IndexesTest(unittest.TestCase):
         Place.ensure_indexes()
         info = Place._get_collection().index_information()
         info = [value['key'] for key, value in info.iteritems()]
-        self.assertTrue([('location.point', '2dsphere')] in info)
+        self.assertIn([('location.point', '2dsphere')], info)
 
     def test_explicit_geohaystack_index(self):
         """Ensure that geohaystack indexes work when created via meta[indexes]
@@ -311,7 +311,7 @@ class IndexesTest(unittest.TestCase):
         Place.ensure_indexes()
         info = Place._get_collection().index_information()
         info = [value['key'] for key, value in info.iteritems()]
-        self.assertTrue([('location.point', 'geoHaystack')] in info)
+        self.assertIn([('location.point', 'geoHaystack')], info)
 
     def test_create_geohaystack_index(self):
         """Ensure that geohaystack indexes can be created
@@ -323,7 +323,7 @@ class IndexesTest(unittest.TestCase):
         Place.create_index({'fields': (')location.point', 'name')}, bucketSize=10)
         info = Place._get_collection().index_information()
         info = [value['key'] for key, value in info.iteritems()]
-        self.assertTrue([('location.point', 'geoHaystack'), ('name', 1)] in info)
+        self.assertIn([('location.point', 'geoHaystack'), ('name', 1)], info)
 
     def test_dictionary_indexes(self):
         """Ensure that indexes are used when meta[indexes] contains
@@ -356,7 +356,7 @@ class IndexesTest(unittest.TestCase):
                  value.get('unique', False),
                  value.get('sparse', False))
                 for key, value in info.iteritems()]
-        self.assertTrue(([('addDate', -1)], True, True) in info)
+        self.assertIn(([('addDate', -1)], True, True), info)
 
         BlogPost.drop_collection()
 
@@ -491,7 +491,7 @@ class IndexesTest(unittest.TestCase):
         obj = Test(a=1)
         obj.save()
 
-        IS_MONGODB_3 = get_mongodb_version()[0] >= 3
+        IS_MONGODB_3 = get_mongodb_version() >= MONGODB_3
 
         # Need to be explicit about covered indexes as mongoDB doesn't know if
         # the documents returned might have more keys in that here.
@@ -541,19 +541,24 @@ class IndexesTest(unittest.TestCase):
                                  [('categories', 1), ('_id', 1)])
 
     def test_hint(self):
+        MONGO_VER = self.mongodb_version
 
+        TAGS_INDEX_NAME = 'tags_1'
         class BlogPost(Document):
             tags = ListField(StringField())
             meta = {
                 'indexes': [
-                    'tags',
+                    {
+                        'fields': ['tags'],
+                        'name': TAGS_INDEX_NAME
+                    }
                 ],
             }
 
         BlogPost.drop_collection()
 
-        for i in range(0, 10):
-            tags = [("tag %i" % n) for n in range(0, i % 2)]
+        for i in range(10):
+            tags = [("tag %i" % n) for n in range(i % 2)]
             BlogPost(tags=tags).save()
 
         self.assertEqual(BlogPost.objects.count(), 10)
@@ -563,18 +568,18 @@ class IndexesTest(unittest.TestCase):
         if pymongo.version != '3.0':
             self.assertEqual(BlogPost.objects.hint([('tags', 1)]).count(), 10)
 
+        if MONGO_VER == MONGODB_32:
+            # Mongo32 throws an error if an index exists (i.e `tags` in our case)
+            # and you use hint on an index name that does not exist
+            with self.assertRaises(OperationFailure):
+                BlogPost.objects.hint([('ZZ', 1)]).count()
+        else:
             self.assertEqual(BlogPost.objects.hint([('ZZ', 1)]).count(), 10)
 
-        if pymongo.version >= '2.8':
-            self.assertEqual(BlogPost.objects.hint('tags').count(), 10)
-        else:
-            def invalid_index():
-                BlogPost.objects.hint('tags').next()
-            self.assertRaises(TypeError, invalid_index)
+        self.assertEqual(BlogPost.objects.hint(TAGS_INDEX_NAME ).count(), 10)
 
-        def invalid_index_2():
-            return BlogPost.objects.hint(('tags', 1)).next()
-        self.assertRaises(Exception, invalid_index_2)
+        with self.assertRaises(Exception):
+            BlogPost.objects.hint(('tags', 1)).next()
 
     def test_unique(self):
         """Ensure that uniqueness constraints are applied to fields.
@@ -749,7 +754,7 @@ class IndexesTest(unittest.TestCase):
         except NotUniqueError:
             pass
 
-    def test_unique_and_primary(self):
+    def test_primary_save_duplicate_update_existing_object(self):
         """If you set a field as primary, then unexpected behaviour can occur.
         You won't create a duplicate but you will update an existing document.
         """
@@ -803,7 +808,7 @@ class IndexesTest(unittest.TestCase):
         info = BlogPost.objects._collection.index_information()
         info = [value['key'] for key, value in info.iteritems()]
         index_item = [('_id', 1), ('comments.comment_id', 1)]
-        self.assertTrue(index_item in info)
+        self.assertIn(index_item, info)
 
     def test_compound_key_embedded(self):
 
@@ -850,8 +855,8 @@ class IndexesTest(unittest.TestCase):
 
         info = MyDoc.objects._collection.index_information()
         info = [value['key'] for key, value in info.iteritems()]
-        self.assertTrue([('provider_ids.foo', 1)] in info)
-        self.assertTrue([('provider_ids.bar', 1)] in info)
+        self.assertIn([('provider_ids.foo', 1)], info)
+        self.assertIn([('provider_ids.bar', 1)], info)
 
     def test_sparse_compound_indexes(self):
 
@@ -867,7 +872,7 @@ class IndexesTest(unittest.TestCase):
                          info['provider_ids.foo_1_provider_ids.bar_1']['key'])
         self.assertTrue(info['provider_ids.foo_1_provider_ids.bar_1']['sparse'])
 
-    @needs_mongodb_v26
+    @requires_mongodb_gte_26
     def test_text_indexes(self):
         class Book(Document):
             title = DictField()
@@ -876,9 +881,9 @@ class IndexesTest(unittest.TestCase):
             }
 
         indexes = Book.objects._collection.index_information()
-        self.assertTrue("title_text" in indexes)
+        self.assertIn("title_text", indexes)
         key = indexes["title_text"]["key"]
-        self.assertTrue(('_fts', 'text') in key)
+        self.assertIn(('_fts', 'text'), key)
 
     def test_hashed_indexes(self):
 
@@ -889,8 +894,8 @@ class IndexesTest(unittest.TestCase):
             }
 
         indexes = Book.objects._collection.index_information()
-        self.assertTrue("ref_id_hashed" in indexes)
-        self.assertTrue(('ref_id', 'hashed') in indexes["ref_id_hashed"]["key"])
+        self.assertIn("ref_id_hashed", indexes)
+        self.assertIn(('ref_id', 'hashed'), indexes["ref_id_hashed"]["key"])
 
     def test_indexes_after_database_drop(self):
         """
@@ -1013,7 +1018,7 @@ class IndexesTest(unittest.TestCase):
         TestDoc.ensure_indexes()
 
         index_info = TestDoc._get_collection().index_information()
-        self.assertTrue('shard_1_1__cls_1_txt_1_1' in index_info)
+        self.assertIn('shard_1_1__cls_1_txt_1_1', index_info)
 
 
 if __name__ == '__main__':

tests/document/inheritance.py

@@ -2,14 +2,11 @@
 import unittest
 import warnings
 
-from datetime import datetime
-
-from tests.fixtures import Base
-
-from mongoengine import Document, EmbeddedDocument, connect
+from mongoengine import (BooleanField, Document, EmbeddedDocument,
+                         EmbeddedDocumentField, GenericReferenceField,
+                         IntField, ReferenceField, StringField, connect)
 from mongoengine.connection import get_db
-from mongoengine.fields import (BooleanField, GenericReferenceField,
-                                IntField, StringField)
+from tests.fixtures import Base
 
 __all__ = ('InheritanceTest', )
 
@@ -26,6 +23,27 @@ class InheritanceTest(unittest.TestCase):
                 continue
             self.db.drop_collection(collection)
 
+    def test_constructor_cls(self):
+        # Ensures _cls is properly set during construction
+        # and when object gets reloaded (prevent regression of #1950)
+        class EmbedData(EmbeddedDocument):
+            data = StringField()
+            meta = {'allow_inheritance': True}
+
+        class DataDoc(Document):
+            name = StringField()
+            embed = EmbeddedDocumentField(EmbedData)
+            meta = {'allow_inheritance': True}
+
+        test_doc = DataDoc(name='test', embed=EmbedData(data='data'))
+        assert test_doc._cls == 'DataDoc'
+        assert test_doc.embed._cls == 'EmbedData'
+        test_doc.save()
+        saved_doc = DataDoc.objects.with_id(test_doc.id)
+        assert test_doc._cls == saved_doc._cls
+        assert test_doc.embed._cls == saved_doc.embed._cls
+        test_doc.delete()
+
     def test_superclasses(self):
         """Ensure that the correct list of superclasses is assembled.
         """
@@ -258,9 +276,10 @@ class InheritanceTest(unittest.TestCase):
             name = StringField()
 
         # can't inherit because Animal didn't explicitly allow inheritance
-        with self.assertRaises(ValueError):
+        with self.assertRaises(ValueError) as cm:
             class Dog(Animal):
                 pass
+        self.assertIn("Document Animal may not be subclassed", str(cm.exception))
 
         # Check that _cls etc aren't present on simple documents
         dog = Animal(name='dog').save()
@@ -268,7 +287,7 @@ class InheritanceTest(unittest.TestCase):
 
         collection = self.db[Animal._get_collection_name()]
         obj = collection.find_one()
-        self.assertFalse('_cls' in obj)
+        self.assertNotIn('_cls', obj)
 
     def test_cant_turn_off_inheritance_on_subclass(self):
         """Ensure if inheritance is on in a subclass you cant turn it off.
@@ -277,9 +296,10 @@ class InheritanceTest(unittest.TestCase):
             name = StringField()
             meta = {'allow_inheritance': True}
 
-        with self.assertRaises(ValueError):
+        with self.assertRaises(ValueError) as cm:
             class Mammal(Animal):
                 meta = {'allow_inheritance': False}
+        self.assertEqual(str(cm.exception), 'Only direct subclasses of Document may set "allow_inheritance" to False')
 
     def test_allow_inheritance_abstract_document(self):
         """Ensure that abstract documents can set inheritance rules and that
@@ -292,13 +312,48 @@ class InheritanceTest(unittest.TestCase):
         class Animal(FinalDocument):
             name = StringField()
 
-        with self.assertRaises(ValueError):
+        with self.assertRaises(ValueError) as cm:
             class Mammal(Animal):
                 pass
 
         # Check that _cls isn't present in simple documents
         doc = Animal(name='dog')
-        self.assertFalse('_cls' in doc.to_mongo())
+        self.assertNotIn('_cls', doc.to_mongo())
+
+    def test_using_abstract_class_in_reference_field(self):
+        # Ensures no regression of #1920
+        class AbstractHuman(Document):
+            meta = {'abstract': True}
+
+        class Dad(AbstractHuman):
+            name = StringField()
+
+        class Home(Document):
+            dad = ReferenceField(AbstractHuman)  # Referencing the abstract class
+            address = StringField()
+
+        dad = Dad(name='5').save()
+        Home(dad=dad, address='street').save()
+
+        home = Home.objects.first()
+        home.address = 'garbage'
+        home.save()     # Was failing with ValidationError
+
+    def test_abstract_class_referencing_self(self):
+        # Ensures no regression of #1920
+        class Human(Document):
+            meta = {'abstract': True}
+            creator = ReferenceField('self', dbref=True)
+
+        class User(Human):
+            name = StringField()
+
+        user = User(name='John').save()
+        user2 = User(name='Foo', creator=user).save()
+
+        user2 = User.objects.with_id(user2.id)
+        user2.name = 'Bar'
+        user2.save()    # Was failing with ValidationError
 
     def test_abstract_handle_ids_in_metaclass_properly(self):
 
@@ -358,11 +413,11 @@ class InheritanceTest(unittest.TestCase):
             meta = {'abstract': True,
                     'allow_inheritance': False}
 
-        bkk = City(continent='asia')
-        self.assertEqual(None, bkk.pk)
+        city = City(continent='asia')
+        self.assertEqual(None, city.pk)
         # TODO: expected error? Shouldn't we create a new error type?
         with self.assertRaises(KeyError):
-            setattr(bkk, 'pk', 1)
+            setattr(city, 'pk', 1)
 
     def test_allow_inheritance_embedded_document(self):
         """Ensure embedded documents respect inheritance."""
@@ -374,14 +429,14 @@ class InheritanceTest(unittest.TestCase):
                 pass
 
         doc = Comment(content='test')
-        self.assertFalse('_cls' in doc.to_mongo())
+        self.assertNotIn('_cls', doc.to_mongo())
 
         class Comment(EmbeddedDocument):
             content = StringField()
             meta = {'allow_inheritance': True}
 
         doc = Comment(content='test')
-        self.assertTrue('_cls' in doc.to_mongo())
+        self.assertIn('_cls', doc.to_mongo())
 
     def test_document_inheritance(self):
         """Ensure mutliple inheritance of abstract documents
@@ -434,8 +489,8 @@ class InheritanceTest(unittest.TestCase):
             for cls in [Animal, Fish, Guppy]:
                 self.assertEqual(cls._meta[k], v)
 
-        self.assertFalse('collection' in Animal._meta)
-        self.assertFalse('collection' in Mammal._meta)
+        self.assertNotIn('collection', Animal._meta)
+        self.assertNotIn('collection', Mammal._meta)
 
         self.assertEqual(Animal._get_collection_name(), None)
         self.assertEqual(Mammal._get_collection_name(), None)

tests/document/instance.py

@@ -8,9 +8,12 @@ import weakref
 
 from datetime import datetime
 from bson import DBRef, ObjectId
+from pymongo.errors import DuplicateKeyError
+
 from tests import fixtures
 from tests.fixtures import (PickleEmbedded, PickleTest, PickleSignalsTest,
                             PickleDynamicEmbedded, PickleDynamicTest)
+from tests.utils import MongoDBTestCase
 
 from mongoengine import *
 from mongoengine.base import get_document, _document_registry
@@ -22,18 +25,17 @@ from mongoengine.queryset import NULLIFY, Q
 from mongoengine.context_managers import switch_db, query_counter
 from mongoengine import signals
 
+from tests.utils import requires_mongodb_gte_26
+
 TEST_IMAGE_PATH = os.path.join(os.path.dirname(__file__),
                                '../fields/mongoengine.png')
 
 __all__ = ("InstanceTest",)
 
 
-class InstanceTest(unittest.TestCase):
+class InstanceTest(MongoDBTestCase):
 
     def setUp(self):
-        connect(db='mongoenginetest')
-        self.db = get_db()
-
         class Job(EmbeddedDocument):
             name = StringField()
             years = IntField()
@@ -355,7 +357,7 @@ class InstanceTest(unittest.TestCase):
 
         user_son = User.objects._collection.find_one()
         self.assertEqual(user_son['_id'], 'test')
-        self.assertTrue('username' not in user_son['_id'])
+        self.assertNotIn('username', user_son['_id'])
 
         User.drop_collection()
 
@@ -368,7 +370,7 @@ class InstanceTest(unittest.TestCase):
 
         user_son = User.objects._collection.find_one()
         self.assertEqual(user_son['_id'], 'mongo')
-        self.assertTrue('username' not in user_son['_id'])
+        self.assertNotIn('username', user_son['_id'])
 
     def test_document_not_registered(self):
         class Place(Document):
@@ -474,6 +476,24 @@ class InstanceTest(unittest.TestCase):
         doc.save()
         doc.reload()
 
+    def test_reload_with_changed_fields(self):
+        """Ensures reloading will not affect changed fields"""
+        class User(Document):
+            name = StringField()
+            number = IntField()
+        User.drop_collection()
+
+        user = User(name="Bob", number=1).save()
+        user.name = "John"
+        user.number = 2
+
+        self.assertEqual(user._get_changed_fields(), ['name', 'number'])
+        user.reload('number')
+        self.assertEqual(user._get_changed_fields(), ['name'])
+        user.save()
+        user.reload()
+        self.assertEqual(user.name, "John")
+
     def test_reload_referencing(self):
         """Ensures reloading updates weakrefs correctly."""
         class Embedded(EmbeddedDocument):
@@ -519,7 +539,7 @@ class InstanceTest(unittest.TestCase):
         doc.save()
         doc.dict_field['extra'] = 1
         doc = doc.reload(10, 'list_field')
-        self.assertEqual(doc._get_changed_fields(), [])
+        self.assertEqual(doc._get_changed_fields(), ['dict_field.extra'])
         self.assertEqual(len(doc.list_field), 5)
         self.assertEqual(len(doc.dict_field), 3)
         self.assertEqual(len(doc.embedded_field.list_field), 4)
@@ -530,21 +550,14 @@ class InstanceTest(unittest.TestCase):
             pass
 
         f = Foo()
-        try:
+        with self.assertRaises(Foo.DoesNotExist):
             f.reload()
-        except Foo.DoesNotExist:
-            pass
-        except Exception:
-            self.assertFalse("Threw wrong exception")
 
         f.save()
         f.delete()
-        try:
+
+        with self.assertRaises(Foo.DoesNotExist):
             f.reload()
-        except Foo.DoesNotExist:
-            pass
-        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."""
@@ -581,10 +594,10 @@ class InstanceTest(unittest.TestCase):
         # Length = length(assigned fields + id)
         self.assertEqual(len(person), 5)
 
-        self.assertTrue('age' in person)
+        self.assertIn('age', person)
         person.age = None
-        self.assertFalse('age' in person)
-        self.assertFalse('nationality' in person)
+        self.assertNotIn('age', person)
+        self.assertNotIn('nationality', person)
 
     def test_embedded_document_to_mongo(self):
         class Person(EmbeddedDocument):
@@ -614,8 +627,8 @@ class InstanceTest(unittest.TestCase):
         class Comment(EmbeddedDocument):
             content = StringField()
 
-        self.assertTrue('content' in Comment._fields)
-        self.assertFalse('id' in Comment._fields)
+        self.assertIn('content', Comment._fields)
+        self.assertNotIn('id', Comment._fields)
 
     def test_embedded_document_instance(self):
         """Ensure that embedded documents can reference parent instance."""
@@ -714,12 +727,12 @@ class InstanceTest(unittest.TestCase):
 
         t = TestDocument(status="draft", pub_date=datetime.now())
 
-        try:
+        with self.assertRaises(ValidationError) as cm:
             t.save()
-        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})
+
+        expected_msg = "Draft entries may not have a publication date."
+        self.assertIn(expected_msg, cm.exception.message)
+        self.assertEqual(cm.exception.to_dict(), {'__all__': expected_msg})
 
         t = TestDocument(status="published")
         t.save(clean=False)
@@ -753,12 +766,13 @@ class InstanceTest(unittest.TestCase):
         TestDocument.drop_collection()
 
         t = TestDocument(doc=TestEmbeddedDocument(x=10, y=25, z=15))
-        try:
+
+        with self.assertRaises(ValidationError) as cm:
             t.save()
-        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}})
+
+        expected_msg = "Value of z != x + y"
+        self.assertIn(expected_msg, cm.exception.message)
+        self.assertEqual(cm.exception.to_dict(), {'doc': {'__all__': expected_msg}})
 
         t = TestDocument(doc=TestEmbeddedDocument(x=10, y=25)).save()
         self.assertEqual(t.doc.z, 35)
@@ -826,6 +840,22 @@ class InstanceTest(unittest.TestCase):
 
         self.assertDbEqual([dict(other_doc.to_mongo()), dict(doc.to_mongo())])
 
+    @requires_mongodb_gte_26
+    def test_modify_with_positional_push(self):
+        class BlogPost(Document):
+            tags = ListField(StringField())
+
+        post = BlogPost.objects.create(tags=['python'])
+        self.assertEqual(post.tags, ['python'])
+        post.modify(push__tags__0=['code', 'mongo'])
+        self.assertEqual(post.tags, ['code', 'mongo', 'python'])
+
+        # Assert same order of the list items is maintained in the db
+        self.assertEqual(
+            BlogPost._get_collection().find_one({'_id': post.pk})['tags'],
+            ['code', 'mongo', 'python']
+        )
+
     def test_save(self):
         """Ensure that a document may be saved in the database."""
 
@@ -1323,6 +1353,23 @@ class InstanceTest(unittest.TestCase):
         site = Site.objects.first()
         self.assertEqual(site.page.log_message, "Error: Dummy message")
 
+    def test_update_list_field(self):
+        """Test update on `ListField` with $pull + $in.
+        """
+        class Doc(Document):
+            foo = ListField(StringField())
+
+        Doc.drop_collection()
+        doc = Doc(foo=['a', 'b', 'c'])
+        doc.save()
+
+        # Update
+        doc = Doc.objects.first()
+        doc.update(pull__foo__in=['a', 'c'])
+
+        doc = Doc.objects.first()
+        self.assertEqual(doc.foo, ['b'])
+
     def test_embedded_update_db_field(self):
         """Test update on `EmbeddedDocumentField` fields when db_field
         is other than default.
@@ -1375,6 +1422,60 @@ class InstanceTest(unittest.TestCase):
         self.assertEqual(person.age, 21)
         self.assertEqual(person.active, False)
 
+    def test__get_changed_fields_same_ids_reference_field_does_not_enters_infinite_loop(self):
+        # Refers to Issue #1685
+        class EmbeddedChildModel(EmbeddedDocument):
+            id = DictField(primary_key=True)
+
+        class ParentModel(Document):
+            child = EmbeddedDocumentField(
+                EmbeddedChildModel)
+
+        emb = EmbeddedChildModel(id={'1': [1]})
+        ParentModel(children=emb)._get_changed_fields()
+
+    def test__get_changed_fields_same_ids_reference_field_does_not_enters_infinite_loop(self):
+        class User(Document):
+            id = IntField(primary_key=True)
+            name = StringField()
+
+        class Message(Document):
+            id = IntField(primary_key=True)
+            author = ReferenceField(User)
+
+        Message.drop_collection()
+
+        # All objects share the same id, but each in a different collection
+        user = User(id=1, name='user-name').save()
+        message = Message(id=1, author=user).save()
+
+        message.author.name = 'tutu'
+        self.assertEqual(message._get_changed_fields(), [])
+        self.assertEqual(user._get_changed_fields(), ['name'])
+
+    def test__get_changed_fields_same_ids_embedded(self):
+        # Refers to Issue #1768
+        class User(EmbeddedDocument):
+            id = IntField()
+            name = StringField()
+
+        class Message(Document):
+            id = IntField(primary_key=True)
+            author = EmbeddedDocumentField(User)
+
+        Message.drop_collection()
+
+        # All objects share the same id, but each in a different collection
+        user = User(id=1, name='user-name')#.save()
+        message = Message(id=1, author=user).save()
+
+        message.author.name = 'tutu'
+        self.assertEqual(message._get_changed_fields(), ['author.name'])
+        message.save()
+
+        message_fetched = Message.objects.with_id(message.id)
+        self.assertEqual(message_fetched.author.name, 'tutu')
+
     def test_query_count_when_saving(self):
         """Ensure references don't cause extra fetches when saving"""
         class Organization(Document):
@@ -1408,9 +1509,9 @@ class InstanceTest(unittest.TestCase):
         user = User.objects.first()
         # Even if stored as ObjectId's internally mongoengine uses DBRefs
         # As ObjectId's aren't automatically derefenced
-        self.assertTrue(isinstance(user._data['orgs'][0], DBRef))
-        self.assertTrue(isinstance(user.orgs[0], Organization))
-        self.assertTrue(isinstance(user._data['orgs'][0], Organization))
+        self.assertIsInstance(user._data['orgs'][0], DBRef)
+        self.assertIsInstance(user.orgs[0], Organization)
+        self.assertIsInstance(user._data['orgs'][0], Organization)
 
         # Changing a value
         with query_counter() as q:
@@ -1790,9 +1891,8 @@ class InstanceTest(unittest.TestCase):
         post_obj = BlogPost.objects.first()
 
         # Test laziness
-        self.assertTrue(isinstance(post_obj._data['author'],
-                                   bson.DBRef))
-        self.assertTrue(isinstance(post_obj.author, self.Person))
+        self.assertIsInstance(post_obj._data['author'], bson.DBRef)
+        self.assertIsInstance(post_obj.author, self.Person)
         self.assertEqual(post_obj.author.name, 'Test User')
 
         # Ensure that the dereferenced object may be changed and saved
@@ -1866,6 +1966,25 @@ class InstanceTest(unittest.TestCase):
         author.delete()
         self.assertEqual(BlogPost.objects.count(), 0)
 
+    def test_reverse_delete_rule_pull(self):
+        """Ensure that a referenced document is also deleted with
+        pull.
+        """
+        class Record(Document):
+            name = StringField()
+            children = ListField(ReferenceField('self', reverse_delete_rule=PULL))
+
+        Record.drop_collection()
+
+        parent_record = Record(name='parent').save()
+        child_record = Record(name='child').save()
+        parent_record.children.append(child_record)
+        parent_record.save()
+
+        child_record.delete()
+        self.assertEqual(Record.objects(name='parent').get().children, [])
+
+
     def test_reverse_delete_rule_with_custom_id_field(self):
         """Ensure that a referenced document with custom primary key
         is also deleted upon deletion.
@@ -2179,12 +2298,12 @@ class InstanceTest(unittest.TestCase):
         # Make sure docs are properly identified in a list (__eq__ is used
         # for the comparison).
         all_user_list = list(User.objects.all())
-        self.assertTrue(u1 in all_user_list)
-        self.assertTrue(u2 in all_user_list)
-        self.assertTrue(u3 in all_user_list)
-        self.assertTrue(u4 not in all_user_list)  # New object
-        self.assertTrue(b1 not in all_user_list)  # Other object
-        self.assertTrue(b2 not in all_user_list)  # Other object
+        self.assertIn(u1, all_user_list)
+        self.assertIn(u2, all_user_list)
+        self.assertIn(u3, all_user_list)
+        self.assertNotIn(u4, all_user_list)  # New object
+        self.assertNotIn(b1, all_user_list)  # Other object
+        self.assertNotIn(b2, all_user_list)  # Other object
 
         # Make sure docs can be used as keys in a dict (__hash__ is used
         # for hashing the docs).
@@ -2202,10 +2321,10 @@ class InstanceTest(unittest.TestCase):
         # Make sure docs are properly identified in a set (__hash__ is used
         # for hashing the docs).
         all_user_set = set(User.objects.all())
-        self.assertTrue(u1 in all_user_set)
-        self.assertTrue(u4 not in all_user_set)
-        self.assertTrue(b1 not in all_user_list)
-        self.assertTrue(b2 not in all_user_list)
+        self.assertIn(u1, all_user_set)
+        self.assertNotIn(u4, all_user_set)
+        self.assertNotIn(b1, all_user_list)
+        self.assertNotIn(b2, all_user_list)
 
         # Make sure duplicate docs aren't accepted in the set
         self.assertEqual(len(all_user_set), 3)
@@ -2906,7 +3025,7 @@ class InstanceTest(unittest.TestCase):
         Person(name="Harry Potter").save()
 
         person = Person.objects.first()
-        self.assertTrue('id' in person._data.keys())
+        self.assertIn('id', person._data.keys())
         self.assertEqual(person._data.get('id'), person.id)
 
     def test_complex_nesting_document_and_embedded_document(self):
@@ -2998,36 +3117,36 @@ class InstanceTest(unittest.TestCase):
 
         dbref2 = f._data['test2']
         obj2 = f.test2
-        self.assertTrue(isinstance(dbref2, DBRef))
-        self.assertTrue(isinstance(obj2, Test2))
-        self.assertTrue(obj2.id == dbref2.id)
-        self.assertTrue(obj2 == dbref2)
-        self.assertTrue(dbref2 == obj2)
+        self.assertIsInstance(dbref2, DBRef)
+        self.assertIsInstance(obj2, Test2)
+        self.assertEqual(obj2.id, dbref2.id)
+        self.assertEqual(obj2, dbref2)
+        self.assertEqual(dbref2, obj2)
 
         dbref3 = f._data['test3']
         obj3 = f.test3
-        self.assertTrue(isinstance(dbref3, DBRef))
-        self.assertTrue(isinstance(obj3, Test3))
-        self.assertTrue(obj3.id == dbref3.id)
-        self.assertTrue(obj3 == dbref3)
-        self.assertTrue(dbref3 == obj3)
+        self.assertIsInstance(dbref3, DBRef)
+        self.assertIsInstance(obj3, Test3)
+        self.assertEqual(obj3.id, dbref3.id)
+        self.assertEqual(obj3, dbref3)
+        self.assertEqual(dbref3, obj3)
 
-        self.assertTrue(obj2.id == obj3.id)
-        self.assertTrue(dbref2.id == dbref3.id)
-        self.assertFalse(dbref2 == dbref3)
-        self.assertFalse(dbref3 == dbref2)
-        self.assertTrue(dbref2 != dbref3)
-        self.assertTrue(dbref3 != dbref2)
+        self.assertEqual(obj2.id, obj3.id)
+        self.assertEqual(dbref2.id, dbref3.id)
+        self.assertNotEqual(dbref2, dbref3)
+        self.assertNotEqual(dbref3, dbref2)
+        self.assertNotEqual(dbref2, dbref3)
+        self.assertNotEqual(dbref3, dbref2)
 
-        self.assertFalse(obj2 == dbref3)
-        self.assertFalse(dbref3 == obj2)
-        self.assertTrue(obj2 != dbref3)
-        self.assertTrue(dbref3 != obj2)
+        self.assertNotEqual(obj2, dbref3)
+        self.assertNotEqual(dbref3, obj2)
+        self.assertNotEqual(obj2, dbref3)
+        self.assertNotEqual(dbref3, obj2)
 
-        self.assertFalse(obj3 == dbref2)
-        self.assertFalse(dbref2 == obj3)
-        self.assertTrue(obj3 != dbref2)
-        self.assertTrue(dbref2 != obj3)
+        self.assertNotEqual(obj3, dbref2)
+        self.assertNotEqual(dbref2, obj3)
+        self.assertNotEqual(obj3, dbref2)
+        self.assertNotEqual(dbref2, obj3)
 
     def test_default_values(self):
         class Person(Document):
@@ -3076,6 +3195,64 @@ class InstanceTest(unittest.TestCase):
         self.assertEquals(p.id, None)
         p.id = "12345"  # in case it is not working: "OperationError: Shard Keys are immutable..." will be raised here
 
+    def test_from_son_created_False_without_id(self):
+        class MyPerson(Document):
+            name = StringField()
+
+        MyPerson.objects.delete()
+
+        p = MyPerson.from_json('{"name": "a_fancy_name"}', created=False)
+        self.assertFalse(p._created)
+        self.assertIsNone(p.id)
+        p.save()
+        self.assertIsNotNone(p.id)
+        saved_p = MyPerson.objects.get(id=p.id)
+        self.assertEqual(saved_p.name, 'a_fancy_name')
+
+    def test_from_son_created_False_with_id(self):
+        # 1854
+        class MyPerson(Document):
+            name = StringField()
+
+        MyPerson.objects.delete()
+
+        p = MyPerson.from_json('{"_id": "5b85a8b04ec5dc2da388296e", "name": "a_fancy_name"}', created=False)
+        self.assertFalse(p._created)
+        self.assertEqual(p._changed_fields, [])
+        self.assertEqual(p.name, 'a_fancy_name')
+        self.assertEqual(p.id, ObjectId('5b85a8b04ec5dc2da388296e'))
+        p.save()
+
+        with self.assertRaises(DoesNotExist):
+            # Since created=False and we gave an id in the json and _changed_fields is empty
+            # mongoengine assumes that the document exits with that structure already
+            # and calling .save() didn't save anything
+            MyPerson.objects.get(id=p.id)
+
+        self.assertFalse(p._created)
+        p.name = 'a new fancy name'
+        self.assertEqual(p._changed_fields, ['name'])
+        p.save()
+        saved_p = MyPerson.objects.get(id=p.id)
+        self.assertEqual(saved_p.name, p.name)
+
+    def test_from_son_created_True_with_an_id(self):
+        class MyPerson(Document):
+            name = StringField()
+
+        MyPerson.objects.delete()
+
+        p = MyPerson.from_json('{"_id": "5b85a8b04ec5dc2da388296e", "name": "a_fancy_name"}', created=True)
+        self.assertTrue(p._created)
+        self.assertEqual(p._changed_fields, [])
+        self.assertEqual(p.name, 'a_fancy_name')
+        self.assertEqual(p.id, ObjectId('5b85a8b04ec5dc2da388296e'))
+        p.save()
+
+        saved_p = MyPerson.objects.get(id=p.id)
+        self.assertEqual(saved_p, p)
+        self.assertEqual(p.name, 'a_fancy_name')
+
     def test_null_field(self):
         # 734
         class User(Document):
@@ -3149,6 +3326,50 @@ class InstanceTest(unittest.TestCase):
 
         person.update(set__height=2.0)
 
+    @requires_mongodb_gte_26
+    def test_push_with_position(self):
+        """Ensure that push with position works properly for an instance."""
+        class BlogPost(Document):
+            slug = StringField()
+            tags = ListField(StringField())
+
+        blog = BlogPost()
+        blog.slug = "ABC"
+        blog.tags = ["python"]
+        blog.save()
+
+        blog.update(push__tags__0=["mongodb", "code"])
+        blog.reload()
+        self.assertEqual(blog.tags, ['mongodb', 'code', 'python'])
+
+    def test_push_nested_list(self):
+        """Ensure that push update works in nested list"""
+        class BlogPost(Document):
+            slug = StringField()
+            tags = ListField()
+
+        blog = BlogPost(slug="test").save()
+        blog.update(push__tags=["value1", 123])
+        blog.reload()
+        self.assertEqual(blog.tags, [["value1", 123]])
+
+    def test_accessing_objects_with_indexes_error(self):
+        insert_result = self.db.company.insert_many([{'name': 'Foo'},
+                                                     {'name': 'Foo'}])  # Force 2 doc with same name
+        REF_OID = insert_result.inserted_ids[0]
+        self.db.user.insert_one({'company': REF_OID})   # Force 2 doc with same name
+
+        class Company(Document):
+            name = StringField(unique=True)
+
+        class User(Document):
+            company = ReferenceField(Company)
+
+
+        # Ensure index creation exception aren't swallowed (#1688)
+        with self.assertRaises(DuplicateKeyError):
+            User.objects().select_related()
+
 
 if __name__ == '__main__':
     unittest.main()

tests/document/validation.py

@@ -20,16 +20,16 @@ class ValidatorErrorTest(unittest.TestCase):
 
         # 1st level error schema
         error.errors = {'1st': ValidationError('bad 1st'), }
-        self.assertTrue('1st' in error.to_dict())
+        self.assertIn('1st', error.to_dict())
         self.assertEqual(error.to_dict()['1st'], 'bad 1st')
 
         # 2nd level error schema
         error.errors = {'1st': ValidationError('bad 1st', errors={
             '2nd': ValidationError('bad 2nd'),
         })}
-        self.assertTrue('1st' in error.to_dict())
-        self.assertTrue(isinstance(error.to_dict()['1st'], dict))
-        self.assertTrue('2nd' in error.to_dict()['1st'])
+        self.assertIn('1st', error.to_dict())
+        self.assertIsInstance(error.to_dict()['1st'], dict)
+        self.assertIn('2nd', error.to_dict()['1st'])
         self.assertEqual(error.to_dict()['1st']['2nd'], 'bad 2nd')
 
         # moar levels
@@ -40,10 +40,10 @@ class ValidatorErrorTest(unittest.TestCase):
                 }),
             }),
         })}
-        self.assertTrue('1st' in error.to_dict())
-        self.assertTrue('2nd' in error.to_dict()['1st'])
-        self.assertTrue('3rd' in error.to_dict()['1st']['2nd'])
-        self.assertTrue('4th' in error.to_dict()['1st']['2nd']['3rd'])
+        self.assertIn('1st', error.to_dict())
+        self.assertIn('2nd', error.to_dict()['1st'])
+        self.assertIn('3rd', error.to_dict()['1st']['2nd'])
+        self.assertIn('4th', error.to_dict()['1st']['2nd']['3rd'])
         self.assertEqual(error.to_dict()['1st']['2nd']['3rd']['4th'],
                          'Inception')
 
@@ -58,7 +58,7 @@ class ValidatorErrorTest(unittest.TestCase):
         try:
             User().validate()
         except ValidationError as e:
-            self.assertTrue("User:None" in e.message)
+            self.assertIn("User:None", e.message)
             self.assertEqual(e.to_dict(), {
                 'username': 'Field is required',
                 'name': 'Field is required'})
@@ -68,7 +68,7 @@ class ValidatorErrorTest(unittest.TestCase):
         try:
             user.save()
         except ValidationError as e:
-            self.assertTrue("User:RossC0" in e.message)
+            self.assertIn("User:RossC0", e.message)
             self.assertEqual(e.to_dict(), {
                 'name': 'Field is required'})
 
@@ -116,7 +116,7 @@ class ValidatorErrorTest(unittest.TestCase):
         try:
             Doc(id="bad").validate()
         except ValidationError as e:
-            self.assertTrue("SubDoc:None" in e.message)
+            self.assertIn("SubDoc:None", e.message)
             self.assertEqual(e.to_dict(), {
                 "e": {'val': 'OK could not be converted to int'}})
 
@@ -127,14 +127,14 @@ class ValidatorErrorTest(unittest.TestCase):
         doc = Doc.objects.first()
         keys = doc._data.keys()
         self.assertEqual(2, len(keys))
-        self.assertTrue('e' in keys)
-        self.assertTrue('id' in keys)
+        self.assertIn('e', keys)
+        self.assertIn('id', keys)
 
         doc.e.val = "OK"
         try:
             doc.save()
         except ValidationError as e:
-            self.assertTrue("Doc:test" in e.message)
+            self.assertIn("Doc:test", e.message)
             self.assertEqual(e.to_dict(), {
                 "e": {'val': 'OK could not be converted to int'}})
 

tests/fields/__init__.py

@@ -1,3 +1,3 @@
-from fields import *
-from file_tests import *
-from geo import *
+from .fields import *
+from .file_tests import *
+from .geo import *

tests/fields/file_tests.py

@@ -53,8 +53,8 @@ class FileTest(MongoDBTestCase):
         putfile.save()
 
         result = PutFile.objects.first()
-        self.assertTrue(putfile == result)
-        self.assertEqual("%s" % result.the_file, "<GridFSProxy: hello>")
+        self.assertEqual(putfile, result)
+        self.assertEqual("%s" % result.the_file, "<GridFSProxy: hello (%s)>" % result.the_file.grid_id)
         self.assertEqual(result.the_file.read(), text)
         self.assertEqual(result.the_file.content_type, content_type)
         result.the_file.delete()  # Remove file from GridFS
@@ -71,7 +71,7 @@ class FileTest(MongoDBTestCase):
         putfile.save()
 
         result = PutFile.objects.first()
-        self.assertTrue(putfile == result)
+        self.assertEqual(putfile, result)
         self.assertEqual(result.the_file.read(), text)
         self.assertEqual(result.the_file.content_type, content_type)
         result.the_file.delete()
@@ -96,7 +96,7 @@ class FileTest(MongoDBTestCase):
         streamfile.save()
 
         result = StreamFile.objects.first()
-        self.assertTrue(streamfile == result)
+        self.assertEqual(streamfile, result)
         self.assertEqual(result.the_file.read(), text + more_text)
         self.assertEqual(result.the_file.content_type, content_type)
         result.the_file.seek(0)
@@ -132,7 +132,7 @@ class FileTest(MongoDBTestCase):
         streamfile.save()
 
         result = StreamFile.objects.first()
-        self.assertTrue(streamfile == result)
+        self.assertEqual(streamfile, result)
         self.assertEqual(result.the_file.read(), text + more_text)
         # self.assertEqual(result.the_file.content_type, content_type)
         result.the_file.seek(0)
@@ -161,7 +161,7 @@ class FileTest(MongoDBTestCase):
         setfile.save()
 
         result = SetFile.objects.first()
-        self.assertTrue(setfile == result)
+        self.assertEqual(setfile, result)
         self.assertEqual(result.the_file.read(), text)
 
         # Try replacing file with new one
@@ -169,7 +169,7 @@ class FileTest(MongoDBTestCase):
         result.save()
 
         result = SetFile.objects.first()
-        self.assertTrue(setfile == result)
+        self.assertEqual(setfile, result)
         self.assertEqual(result.the_file.read(), more_text)
         result.the_file.delete()
 
@@ -231,8 +231,8 @@ class FileTest(MongoDBTestCase):
         test_file_dupe = TestFile()
         data = test_file_dupe.the_file.read()  # Should be None
 
-        self.assertTrue(test_file.name != test_file_dupe.name)
-        self.assertTrue(test_file.the_file.read() != data)
+        self.assertNotEqual(test_file.name, test_file_dupe.name)
+        self.assertNotEqual(test_file.the_file.read(), data)
 
         TestFile.drop_collection()
 
@@ -291,7 +291,7 @@ class FileTest(MongoDBTestCase):
             the_file = FileField()
 
         test_file = TestFile()
-        self.assertFalse(test_file.the_file in [{"test": 1}])
+        self.assertNotIn(test_file.the_file, [{"test": 1}])
 
     def test_file_disk_space(self):
         """ Test disk space usage when we delete/replace a file """

tests/fields/geo.py

@@ -298,9 +298,9 @@ class GeoFieldTest(unittest.TestCase):
             polygon = PolygonField()
 
         geo_indicies = Event._geo_indices()
-        self.assertTrue({'fields': [('line', '2dsphere')]} in geo_indicies)
-        self.assertTrue({'fields': [('polygon', '2dsphere')]} in geo_indicies)
-        self.assertTrue({'fields': [('point', '2dsphere')]} in geo_indicies)
+        self.assertIn({'fields': [('line', '2dsphere')]}, geo_indicies)
+        self.assertIn({'fields': [('polygon', '2dsphere')]}, geo_indicies)
+        self.assertIn({'fields': [('point', '2dsphere')]}, geo_indicies)
 
     def test_indexes_2dsphere_embedded(self):
         """Ensure that indexes are created automatically for GeoPointFields.
@@ -316,9 +316,9 @@ class GeoFieldTest(unittest.TestCase):
             venue = EmbeddedDocumentField(Venue)
 
         geo_indicies = Event._geo_indices()
-        self.assertTrue({'fields': [('venue.line', '2dsphere')]} in geo_indicies)
-        self.assertTrue({'fields': [('venue.polygon', '2dsphere')]} in geo_indicies)
-        self.assertTrue({'fields': [('venue.point', '2dsphere')]} in geo_indicies)
+        self.assertIn({'fields': [('venue.line', '2dsphere')]}, geo_indicies)
+        self.assertIn({'fields': [('venue.polygon', '2dsphere')]}, geo_indicies)
+        self.assertIn({'fields': [('venue.point', '2dsphere')]}, geo_indicies)
 
     def test_geo_indexes_recursion(self):
 
@@ -335,9 +335,9 @@ class GeoFieldTest(unittest.TestCase):
 
         Parent(name='Berlin').save()
         info = Parent._get_collection().index_information()
-        self.assertFalse('location_2d' in info)
+        self.assertNotIn('location_2d', info)
         info = Location._get_collection().index_information()
-        self.assertTrue('location_2d' in info)
+        self.assertIn('location_2d', info)
 
         self.assertEqual(len(Parent._geo_indices()), 0)
         self.assertEqual(len(Location._geo_indices()), 1)

tests/queryset/__init__.py

@@ -1,6 +1,6 @@
-from transform import *
-from field_list import *
-from queryset import *
-from visitor import *
-from geo import *
-from modify import *
+from .transform import *
+from .field_list import *
+from .queryset import *
+from .visitor import *
+from .geo import *
+from .modify import *

tests/queryset/field_list.py

@@ -181,7 +181,7 @@ class OnlyExcludeAllTest(unittest.TestCase):
         employee.save()
 
         obj = self.Person.objects(id=employee.id).only('age').get()
-        self.assertTrue(isinstance(obj, Employee))
+        self.assertIsInstance(obj, Employee)
 
         # Check field names are looked up properly
         obj = Employee.objects(id=employee.id).only('salary').get()
@@ -197,14 +197,18 @@ class OnlyExcludeAllTest(unittest.TestCase):
             title = StringField()
             text = StringField()
 
+        class VariousData(EmbeddedDocument):
+            some = BooleanField()
+
         class BlogPost(Document):
             content = StringField()
             author = EmbeddedDocumentField(User)
             comments = ListField(EmbeddedDocumentField(Comment))
+            various = MapField(field=EmbeddedDocumentField(VariousData))
 
         BlogPost.drop_collection()
 
-        post = BlogPost(content='Had a good coffee today...')
+        post = BlogPost(content='Had a good coffee today...', various={'test_dynamic':{'some': True}})
         post.author = User(name='Test User')
         post.comments = [Comment(title='I aggree', text='Great post!'), Comment(title='Coffee', text='I hate coffee')]
         post.save()
@@ -215,6 +219,9 @@ class OnlyExcludeAllTest(unittest.TestCase):
         self.assertEqual(obj.author.name, 'Test User')
         self.assertEqual(obj.comments, [])
 
+        obj = BlogPost.objects.only('various.test_dynamic.some').get()
+        self.assertEqual(obj.various["test_dynamic"].some, True)
+
         obj = BlogPost.objects.only('content', 'comments.title',).get()
         self.assertEqual(obj.content, 'Had a good coffee today...')
         self.assertEqual(obj.author, None)

tests/queryset/geo.py

@@ -3,7 +3,7 @@ import unittest
 
 from mongoengine import *
 
-from tests.utils import MongoDBTestCase, needs_mongodb_v3
+from tests.utils import MongoDBTestCase, requires_mongodb_gte_3
 
 
 __all__ = ("GeoQueriesTest",)
@@ -72,7 +72,7 @@ class GeoQueriesTest(MongoDBTestCase):
 
     # $minDistance was added in MongoDB v2.6, but continued being buggy
     # until v3.0; skip for older versions
-    @needs_mongodb_v3
+    @requires_mongodb_gte_3
     def test_near_and_min_distance(self):
         """Ensure the "min_distance" operator works alongside the "near"
         operator.
@@ -95,9 +95,9 @@ class GeoQueriesTest(MongoDBTestCase):
             location__within_distance=point_and_distance)
         self.assertEqual(events.count(), 2)
         events = list(events)
-        self.assertTrue(event2 not in events)
-        self.assertTrue(event1 in events)
-        self.assertTrue(event3 in events)
+        self.assertNotIn(event2, events)
+        self.assertIn(event1, events)
+        self.assertIn(event3, events)
 
         # find events within 10 degrees of san francisco
         point_and_distance = [[-122.415579, 37.7566023], 10]
@@ -245,7 +245,7 @@ class GeoQueriesTest(MongoDBTestCase):
 
     # $minDistance was added in MongoDB v2.6, but continued being buggy
     # until v3.0; skip for older versions
-    @needs_mongodb_v3
+    @requires_mongodb_gte_3
     def test_2dsphere_near_and_min_max_distance(self):
         """Ensure "min_distace" and "max_distance" operators work well
         together with the "near" operator in a 2dsphere index.
@@ -285,9 +285,9 @@ class GeoQueriesTest(MongoDBTestCase):
             location__geo_within_center=point_and_distance)
         self.assertEqual(events.count(), 2)
         events = list(events)
-        self.assertTrue(event2 not in events)
-        self.assertTrue(event1 in events)
-        self.assertTrue(event3 in events)
+        self.assertNotIn(event2, events)
+        self.assertIn(event1, events)
+        self.assertIn(event3, events)
 
     def _test_embedded(self, point_field_class):
         """Helper test method ensuring given point field class works
@@ -329,7 +329,7 @@ class GeoQueriesTest(MongoDBTestCase):
         self._test_embedded(point_field_class=PointField)
 
     # Needs MongoDB > 2.6.4 https://jira.mongodb.org/browse/SERVER-14039
-    @needs_mongodb_v3
+    @requires_mongodb_gte_3
     def test_spherical_geospatial_operators(self):
         """Ensure that spherical geospatial queries are working."""
         class Point(Document):
@@ -510,6 +510,24 @@ class GeoQueriesTest(MongoDBTestCase):
         roads = Road.objects.filter(poly__geo_intersects={"$geometry": polygon}).count()
         self.assertEqual(1, roads)
 
+    def test_aspymongo_with_only(self):
+        """Ensure as_pymongo works with only"""
+        class Place(Document):
+            location = PointField()
+
+        Place.drop_collection()
+        p = Place(location=[24.946861267089844, 60.16311983618494])
+        p.save()
+        qs = Place.objects().only('location')
+        self.assertDictEqual(
+            qs.as_pymongo()[0]['location'],
+            {u'type': u'Point',
+             u'coordinates': [
+                24.946861267089844,
+                60.16311983618494]
+            }
+        )
+
     def test_2dsphere_point_sets_correctly(self):
         class Location(Document):
             loc = PointField()

tests/queryset/modify.py

@@ -1,6 +1,8 @@
 import unittest
 
-from mongoengine import connect, Document, IntField
+from mongoengine import connect, Document, IntField, StringField, ListField
+
+from tests.utils import requires_mongodb_gte_26
 
 __all__ = ("FindAndModifyTest",)
 
@@ -94,6 +96,37 @@ class FindAndModifyTest(unittest.TestCase):
         self.assertEqual(old_doc.to_mongo(), {"_id": 1})
         self.assertDbEqual([{"_id": 0, "value": 0}, {"_id": 1, "value": -1}])
 
+    @requires_mongodb_gte_26
+    def test_modify_with_push(self):
+        class BlogPost(Document):
+            tags = ListField(StringField())
+
+        BlogPost.drop_collection()
+
+        blog = BlogPost.objects.create()
+
+        # Push a new tag via modify with new=False (default).
+        BlogPost(id=blog.id).modify(push__tags='code')
+        self.assertEqual(blog.tags, [])
+        blog.reload()
+        self.assertEqual(blog.tags, ['code'])
+
+        # Push a new tag via modify with new=True.
+        blog = BlogPost.objects(id=blog.id).modify(push__tags='java', new=True)
+        self.assertEqual(blog.tags, ['code', 'java'])
+
+        # Push a new tag with a positional argument.
+        blog = BlogPost.objects(id=blog.id).modify(
+            push__tags__0='python',
+            new=True)
+        self.assertEqual(blog.tags, ['python', 'code', 'java'])
+
+        # Push multiple new tags with a positional argument.
+        blog = BlogPost.objects(id=blog.id).modify(
+            push__tags__1=['go', 'rust'],
+            new=True)
+        self.assertEqual(blog.tags, ['python', 'go', 'rust', 'code', 'java'])
+
 
 if __name__ == '__main__':
     unittest.main()

tests/queryset/transform.py

@@ -1,5 +1,7 @@
 import unittest
 
+from bson.son import SON
+
 from mongoengine import *
 from mongoengine.queryset import Q, transform
 
@@ -28,12 +30,16 @@ class TransformTest(unittest.TestCase):
                          {'name': {'$exists': True}})
 
     def test_transform_update(self):
+        class LisDoc(Document):
+            foo = ListField(StringField())
+
         class DicDoc(Document):
             dictField = DictField()
 
         class Doc(Document):
             pass
 
+        LisDoc.drop_collection()
         DicDoc.drop_collection()
         Doc.drop_collection()
 
@@ -42,14 +48,28 @@ class TransformTest(unittest.TestCase):
 
         for k, v in (("set", "$set"), ("set_on_insert", "$setOnInsert"), ("push", "$push")):
             update = transform.update(DicDoc, **{"%s__dictField__test" % k: doc})
-            self.assertTrue(isinstance(update[v]["dictField.test"], dict))
+            self.assertIsInstance(update[v]["dictField.test"], dict)
 
         # Update special cases
         update = transform.update(DicDoc, unset__dictField__test=doc)
         self.assertEqual(update["$unset"]["dictField.test"], 1)
 
         update = transform.update(DicDoc, pull__dictField__test=doc)
-        self.assertTrue(isinstance(update["$pull"]["dictField"]["test"], dict))
+        self.assertIsInstance(update["$pull"]["dictField"]["test"], dict)
+
+        update = transform.update(LisDoc, pull__foo__in=['a'])
+        self.assertEqual(update, {'$pull': {'foo': {'$in': ['a']}}})
+
+    def test_transform_update_push(self):
+        """Ensure the differences in behvaior between 'push' and 'push_all'"""
+        class BlogPost(Document):
+            tags = ListField(StringField())
+
+        update = transform.update(BlogPost, push__tags=['mongo', 'db'])
+        self.assertEqual(update, {'$push': {'tags': ['mongo', 'db']}})
+
+        update = transform.update(BlogPost, push_all__tags=['mongo', 'db'])
+        self.assertEqual(update, {'$push': {'tags': {'$each': ['mongo', 'db']}}})
 
     def test_query_field_name(self):
         """Ensure that the correct field name is used when querying.
@@ -68,17 +88,15 @@ class TransformTest(unittest.TestCase):
         post = BlogPost(**data)
         post.save()
 
-        self.assertTrue('postTitle' in
-                        BlogPost.objects(title=data['title'])._query)
+        self.assertIn('postTitle', BlogPost.objects(title=data['title'])._query)
         self.assertFalse('title' in
                          BlogPost.objects(title=data['title'])._query)
         self.assertEqual(BlogPost.objects(title=data['title']).count(), 1)
 
-        self.assertTrue('_id' in BlogPost.objects(pk=post.id)._query)
+        self.assertIn('_id', BlogPost.objects(pk=post.id)._query)
         self.assertEqual(BlogPost.objects(pk=post.id).count(), 1)
 
-        self.assertTrue('postComments.commentContent' in
-                        BlogPost.objects(comments__content='test')._query)
+        self.assertIn('postComments.commentContent', BlogPost.objects(comments__content='test')._query)
         self.assertEqual(BlogPost.objects(comments__content='test').count(), 1)
 
         BlogPost.drop_collection()
@@ -96,8 +114,8 @@ class TransformTest(unittest.TestCase):
         post = BlogPost(**data)
         post.save()
 
-        self.assertTrue('_id' in BlogPost.objects(pk=data['title'])._query)
-        self.assertTrue('_id' in BlogPost.objects(title=data['title'])._query)
+        self.assertIn('_id', BlogPost.objects(pk=data['title'])._query)
+        self.assertIn('_id', BlogPost.objects(title=data['title'])._query)
         self.assertEqual(BlogPost.objects(pk=data['title']).count(), 1)
 
         BlogPost.drop_collection()
@@ -241,6 +259,30 @@ class TransformTest(unittest.TestCase):
         with self.assertRaises(InvalidQueryError):
             events.count()
 
+    def test_update_pull_for_list_fields(self):
+        """
+        Test added to check pull operation in update for
+        EmbeddedDocumentListField which is inside a EmbeddedDocumentField
+        """
+        class Word(EmbeddedDocument):
+            word = StringField()
+            index = IntField()
+
+        class SubDoc(EmbeddedDocument):
+            heading = ListField(StringField())
+            text = EmbeddedDocumentListField(Word)
+
+        class MainDoc(Document):
+            title = StringField()
+            content = EmbeddedDocumentField(SubDoc)
+
+        word = Word(word='abc', index=1)
+        update = transform.update(MainDoc, pull__content__text=word)
+        self.assertEqual(update, {'$pull': {'content.text': SON([('word', u'abc'), ('index', 1)])}})
+
+        update = transform.update(MainDoc, pull__content__heading='xyz')
+        self.assertEqual(update, {'$pull': {'content.heading': 'xyz'}})
+
 
 if __name__ == '__main__':
     unittest.main()

tests/queryset/visitor.py

@@ -196,7 +196,7 @@ class QTest(unittest.TestCase):
 
         test2 = test.clone()
         self.assertEqual(test2.count(), 3)
-        self.assertFalse(test2 == test)
+        self.assertNotEqual(test2, test)
 
         test3 = test2.filter(x=6)
         self.assertEqual(test3.count(), 1)
@@ -296,6 +296,18 @@ class QTest(unittest.TestCase):
         obj = self.Person.objects(Q(name__not=re.compile('^Gui'))).first()
         self.assertEqual(obj, None)
 
+    def test_q_repr(self):
+        self.assertEqual(repr(Q()), 'Q(**{})')
+        self.assertEqual(repr(Q(name='test')), "Q(**{'name': 'test'})")
+
+        self.assertEqual(
+            repr(Q(name='test') & Q(age__gte=18)),
+            "(Q(**{'name': 'test'}) & Q(**{'age__gte': 18}))")
+
+        self.assertEqual(
+            repr(Q(name='test') | Q(age__gte=18)),
+            "(Q(**{'name': 'test'}) | Q(**{'age__gte': 18}))")
+
     def test_q_lists(self):
         """Ensure that Q objects query ListFields correctly.
         """

tests/test_connection.py

@@ -39,15 +39,15 @@ class ConnectionTest(unittest.TestCase):
         connect('mongoenginetest')
 
         conn = get_connection()
-        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+        self.assertIsInstance(conn, pymongo.mongo_client.MongoClient)
 
         db = get_db()
-        self.assertTrue(isinstance(db, pymongo.database.Database))
+        self.assertIsInstance(db, pymongo.database.Database)
         self.assertEqual(db.name, 'mongoenginetest')
 
         connect('mongoenginetest2', alias='testdb')
         conn = get_connection('testdb')
-        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+        self.assertIsInstance(conn, pymongo.mongo_client.MongoClient)
 
     def test_connect_in_mocking(self):
         """Ensure that the connect() method works properly in mocking.
@@ -59,31 +59,31 @@ class ConnectionTest(unittest.TestCase):
 
         connect('mongoenginetest', host='mongomock://localhost')
         conn = get_connection()
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect('mongoenginetest2', host='mongomock://localhost', alias='testdb2')
         conn = get_connection('testdb2')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect('mongoenginetest3', host='mongodb://localhost', is_mock=True, alias='testdb3')
         conn = get_connection('testdb3')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect('mongoenginetest4', is_mock=True, alias='testdb4')
         conn = get_connection('testdb4')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect(host='mongodb://localhost:27017/mongoenginetest5', is_mock=True, alias='testdb5')
         conn = get_connection('testdb5')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect(host='mongomock://localhost:27017/mongoenginetest6', alias='testdb6')
         conn = get_connection('testdb6')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect(host='mongomock://localhost:27017/mongoenginetest7', is_mock=True, alias='testdb7')
         conn = get_connection('testdb7')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
     def test_connect_with_host_list(self):
         """Ensure that the connect() method works when host is a list
@@ -97,27 +97,27 @@ class ConnectionTest(unittest.TestCase):
 
         connect(host=['mongomock://localhost'])
         conn = get_connection()
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect(host=['mongodb://localhost'], is_mock=True,  alias='testdb2')
         conn = get_connection('testdb2')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect(host=['localhost'], is_mock=True,  alias='testdb3')
         conn = get_connection('testdb3')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect(host=['mongomock://localhost:27017', 'mongomock://localhost:27018'], alias='testdb4')
         conn = get_connection('testdb4')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(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))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
         connect(host=['localhost:27017', 'localhost:27018'], is_mock=True,  alias='testdb6')
         conn = get_connection('testdb6')
-        self.assertTrue(isinstance(conn, mongomock.MongoClient))
+        self.assertIsInstance(conn, mongomock.MongoClient)
 
     def test_disconnect(self):
         """Ensure that the disconnect() method works properly
@@ -163,10 +163,10 @@ class ConnectionTest(unittest.TestCase):
         connect("testdb_uri", host='mongodb://username:password@localhost/mongoenginetest')
 
         conn = get_connection()
-        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+        self.assertIsInstance(conn, pymongo.mongo_client.MongoClient)
 
         db = get_db()
-        self.assertTrue(isinstance(db, pymongo.database.Database))
+        self.assertIsInstance(db, pymongo.database.Database)
         self.assertEqual(db.name, 'mongoenginetest')
 
         c.admin.system.users.remove({})
@@ -179,10 +179,10 @@ class ConnectionTest(unittest.TestCase):
         connect("mongoenginetest", host='mongodb://localhost/')
 
         conn = get_connection()
-        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+        self.assertIsInstance(conn, pymongo.mongo_client.MongoClient)
 
         db = get_db()
-        self.assertTrue(isinstance(db, pymongo.database.Database))
+        self.assertIsInstance(db, pymongo.database.Database)
         self.assertEqual(db.name, 'mongoenginetest')
 
     def test_connect_uri_default_db(self):
@@ -192,10 +192,10 @@ class ConnectionTest(unittest.TestCase):
         connect(host='mongodb://localhost/')
 
         conn = get_connection()
-        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+        self.assertIsInstance(conn, pymongo.mongo_client.MongoClient)
 
         db = get_db()
-        self.assertTrue(isinstance(db, pymongo.database.Database))
+        self.assertIsInstance(db, pymongo.database.Database)
         self.assertEqual(db.name, 'test')
 
     def test_uri_without_credentials_doesnt_override_conn_settings(self):
@@ -242,7 +242,7 @@ class ConnectionTest(unittest.TestCase):
                   'mongoenginetest?authSource=admin')
         )
         db = get_db('test2')
-        self.assertTrue(isinstance(db, pymongo.database.Database))
+        self.assertIsInstance(db, pymongo.database.Database)
         self.assertEqual(db.name, 'mongoenginetest')
 
         # Clear all users
@@ -255,10 +255,10 @@ class ConnectionTest(unittest.TestCase):
 
         self.assertRaises(MongoEngineConnectionError, get_connection)
         conn = get_connection('testdb')
-        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+        self.assertIsInstance(conn, pymongo.mongo_client.MongoClient)
 
         db = get_db('testdb')
-        self.assertTrue(isinstance(db, pymongo.database.Database))
+        self.assertIsInstance(db, pymongo.database.Database)
         self.assertEqual(db.name, 'mongoenginetest2')
 
     def test_register_connection_defaults(self):
@@ -267,7 +267,7 @@ class ConnectionTest(unittest.TestCase):
         register_connection('testdb', 'mongoenginetest', host=None, port=None)
 
         conn = get_connection('testdb')
-        self.assertTrue(isinstance(conn, pymongo.mongo_client.MongoClient))
+        self.assertIsInstance(conn, pymongo.mongo_client.MongoClient)
 
     def test_connection_kwargs(self):
         """Ensure that connection kwargs get passed to pymongo."""
@@ -326,7 +326,7 @@ class ConnectionTest(unittest.TestCase):
         if IS_PYMONGO_3:
             c = connect(host='mongodb://localhost/test?replicaSet=local-rs')
             db = get_db()
-            self.assertTrue(isinstance(db, pymongo.database.Database))
+            self.assertIsInstance(db, pymongo.database.Database)
             self.assertEqual(db.name, 'test')
         else:
             # PyMongo < v3.x raises an exception:
@@ -343,7 +343,7 @@ class ConnectionTest(unittest.TestCase):
             self.assertEqual(c._MongoClient__options.replica_set_name,
                              'local-rs')
             db = get_db()
-            self.assertTrue(isinstance(db, pymongo.database.Database))
+            self.assertIsInstance(db, pymongo.database.Database)
             self.assertEqual(db.name, 'test')
         else:
             # PyMongo < v3.x raises an exception:
@@ -364,6 +364,12 @@ class ConnectionTest(unittest.TestCase):
         date_doc = DateDoc.objects.first()
         self.assertEqual(d, date_doc.the_date)
 
+    def test_read_preference_from_parse(self):
+        if IS_PYMONGO_3:
+            from pymongo import ReadPreference
+            conn = connect(host="mongodb://a1.vpc,a2.vpc,a3.vpc/prod?readPreference=secondaryPreferred")
+            self.assertEqual(conn.read_preference, ReadPreference.SECONDARY_PREFERRED)
+
     def test_multiple_connection_settings(self):
         connect('mongoenginetest', alias='t1', host="localhost")
 
@@ -371,8 +377,8 @@ class ConnectionTest(unittest.TestCase):
 
         mongo_connections = mongoengine.connection._connections
         self.assertEqual(len(mongo_connections.items()), 2)
-        self.assertTrue('t1' in mongo_connections.keys())
-        self.assertTrue('t2' in mongo_connections.keys())
+        self.assertIn('t1', mongo_connections.keys())
+        self.assertIn('t2', mongo_connections.keys())
         if not IS_PYMONGO_3:
             self.assertEqual(mongo_connections['t1'].host, 'localhost')
             self.assertEqual(mongo_connections['t2'].host, '127.0.0.1')

tests/test_context_managers.py

@@ -89,15 +89,15 @@ class ContextManagersTest(unittest.TestCase):
 
         with no_dereference(Group) as Group:
             group = Group.objects.first()
-            self.assertTrue(all([not isinstance(m, User)
-                                for m in group.members]))
-            self.assertFalse(isinstance(group.ref, User))
-            self.assertFalse(isinstance(group.generic, User))
+            for m in group.members:
+                self.assertNotIsInstance(m, User)
+            self.assertNotIsInstance(group.ref, User)
+            self.assertNotIsInstance(group.generic, User)
 
-        self.assertTrue(all([isinstance(m, User)
-                             for m in group.members]))
-        self.assertTrue(isinstance(group.ref, User))
-        self.assertTrue(isinstance(group.generic, User))
+        for m in group.members:
+            self.assertIsInstance(m, User)
+        self.assertIsInstance(group.ref, User)
+        self.assertIsInstance(group.generic, User)
 
     def test_no_dereference_context_manager_dbref(self):
         """Ensure that DBRef items in ListFields aren't dereferenced.
@@ -129,19 +129,17 @@ class ContextManagersTest(unittest.TestCase):
             group = Group.objects.first()
             self.assertTrue(all([not isinstance(m, User)
                                 for m in group.members]))
-            self.assertFalse(isinstance(group.ref, User))
-            self.assertFalse(isinstance(group.generic, User))
+            self.assertNotIsInstance(group.ref, User)
+            self.assertNotIsInstance(group.generic, User)
 
         self.assertTrue(all([isinstance(m, User)
                              for m in group.members]))
-        self.assertTrue(isinstance(group.ref, User))
-        self.assertTrue(isinstance(group.generic, User))
+        self.assertIsInstance(group.ref, User)
+        self.assertIsInstance(group.generic, User)
 
     def test_no_sub_classes(self):
         class A(Document):
             x = IntField()
-            y = IntField()
-
             meta = {'allow_inheritance': True}
 
         class B(A):
@@ -152,29 +150,29 @@ class ContextManagersTest(unittest.TestCase):
 
         A.drop_collection()
 
-        A(x=10, y=20).save()
-        A(x=15, y=30).save()
-        B(x=20, y=40).save()
-        B(x=30, y=50).save()
-        C(x=40, y=60).save()
+        A(x=10).save()
+        A(x=15).save()
+        B(x=20).save()
+        B(x=30).save()
+        C(x=40).save()
 
         self.assertEqual(A.objects.count(), 5)
         self.assertEqual(B.objects.count(), 3)
         self.assertEqual(C.objects.count(), 1)
 
-        with no_sub_classes(A) as A:
+        with no_sub_classes(A):
             self.assertEqual(A.objects.count(), 2)
 
             for obj in A.objects:
                 self.assertEqual(obj.__class__, A)
 
-        with no_sub_classes(B) as B:
+        with no_sub_classes(B):
             self.assertEqual(B.objects.count(), 2)
 
             for obj in B.objects:
                 self.assertEqual(obj.__class__, B)
 
-        with no_sub_classes(C) as C:
+        with no_sub_classes(C):
             self.assertEqual(C.objects.count(), 1)
 
             for obj in C.objects:
@@ -185,18 +183,124 @@ class ContextManagersTest(unittest.TestCase):
         self.assertEqual(B.objects.count(), 3)
         self.assertEqual(C.objects.count(), 1)
 
+    def test_no_sub_classes_modification_to_document_class_are_temporary(self):
+        class A(Document):
+            x = IntField()
+            meta = {'allow_inheritance': True}
+
+        class B(A):
+            z = IntField()
+
+        self.assertEqual(A._subclasses, ('A', 'A.B'))
+        with no_sub_classes(A):
+            self.assertEqual(A._subclasses, ('A',))
+        self.assertEqual(A._subclasses, ('A', 'A.B'))
+
+        self.assertEqual(B._subclasses, ('A.B',))
+        with no_sub_classes(B):
+            self.assertEqual(B._subclasses, ('A.B',))
+        self.assertEqual(B._subclasses, ('A.B',))
+
+    def test_no_subclass_context_manager_does_not_swallow_exception(self):
+        class User(Document):
+            name = StringField()
+
+        with self.assertRaises(TypeError):
+            with no_sub_classes(User):
+                raise TypeError()
+
+    def test_query_counter_does_not_swallow_exception(self):
+
+        with self.assertRaises(TypeError):
+            with query_counter() as q:
+                raise TypeError()
+
+    def test_query_counter_temporarily_modifies_profiling_level(self):
+        connect('mongoenginetest')
+        db = get_db()
+
+        initial_profiling_level = db.profiling_level()
+
+        try:
+            NEW_LEVEL = 1
+            db.set_profiling_level(NEW_LEVEL)
+            self.assertEqual(db.profiling_level(), NEW_LEVEL)
+            with query_counter() as q:
+                self.assertEqual(db.profiling_level(), 2)
+            self.assertEqual(db.profiling_level(), NEW_LEVEL)
+        except Exception:
+            db.set_profiling_level(initial_profiling_level)    # Ensures it gets reseted no matter the outcome of the test
+            raise
+
     def test_query_counter(self):
         connect('mongoenginetest')
         db = get_db()
-        db.test.find({})
+
+        collection = db.query_counter
+        collection.drop()
+
+        def issue_1_count_query():
+            collection.find({}).count()
+
+        def issue_1_insert_query():
+            collection.insert_one({'test': 'garbage'})
+
+        def issue_1_find_query():
+            collection.find_one()
+
+        counter = 0
+        with query_counter() as q:
+            self.assertEqual(q, counter)
+            self.assertEqual(q, counter)    # Ensures previous count query did not get counted
+
+            for _ in range(10):
+                issue_1_insert_query()
+                counter += 1
+            self.assertEqual(q, counter)
+
+            for _ in range(4):
+                issue_1_find_query()
+                counter += 1
+            self.assertEqual(q, counter)
+
+            for _ in range(3):
+                issue_1_count_query()
+                counter += 1
+            self.assertEqual(q, counter)
+
+    def test_query_counter_counts_getmore_queries(self):
+        connect('mongoenginetest')
+        db = get_db()
+
+        collection = db.query_counter
+        collection.drop()
+
+        many_docs = [{'test': 'garbage %s' % i} for i in range(150)]
+        collection.insert_many(many_docs)   # first batch of documents contains 101 documents
 
         with query_counter() as q:
-            self.assertEqual(0, q)
+            self.assertEqual(q, 0)
+            list(collection.find())
+            self.assertEqual(q, 2)  # 1st select + 1 getmore
 
-            for i in range(1, 51):
-                db.test.find({}).count()
+    def test_query_counter_ignores_particular_queries(self):
+        connect('mongoenginetest')
+        db = get_db()
 
-            self.assertEqual(50, q)
+        collection = db.query_counter
+        collection.insert_many([{'test': 'garbage %s' % i} for i in range(10)])
+
+        with query_counter() as q:
+            self.assertEqual(q, 0)
+            cursor = collection.find()
+            self.assertEqual(q, 0)      # cursor wasn't opened yet
+            _ = next(cursor)            # opens the cursor and fires the find query
+            self.assertEqual(q, 1)
+
+            cursor.close()              # issues a `killcursors` query that is ignored by the context
+            self.assertEqual(q, 1)
+            _ = db.system.indexes.find_one()    # queries on db.system.indexes are ignored as well
+            self.assertEqual(q, 1)
 
 if __name__ == '__main__':
     unittest.main()

tests/test_datastructures.py

@@ -1,6 +1,21 @@
 import unittest
 
-from mongoengine.base.datastructures import StrictDict, SemiStrictDict
+from mongoengine.base.datastructures import StrictDict, BaseList
+
+
+class TestBaseList(unittest.TestCase):
+
+    def test_iter_simple(self):
+        values = [True, False, True, False]
+        base_list = BaseList(values, instance=None, name='my_name')
+        self.assertEqual(values, list(base_list))
+
+    def test_iter_allow_modification_while_iterating_withou_error(self):
+        # regular list allows for this, thus this subclass must comply to that
+        base_list = BaseList([True, False, True, False], instance=None, name='my_name')
+        for idx, val in enumerate(base_list):
+            if val:
+                base_list.pop(idx)
 
 
 class TestStrictDict(unittest.TestCase):
@@ -76,44 +91,5 @@ class TestStrictDict(unittest.TestCase):
         assert dict(**d) == {'a': 1, 'b': 2}
 
 
-class TestSemiSrictDict(TestStrictDict):
-    def strict_dict_class(self, *args, **kwargs):
-        return SemiStrictDict.create(*args, **kwargs)
-
-    def test_init_fails_on_nonexisting_attrs(self):
-        # disable irrelevant test
-        pass
-
-    def test_setattr_raises_on_nonexisting_attr(self):
-        # disable irrelevant test
-        pass
-
-    def test_setattr_getattr_nonexisting_attr_succeeds(self):
-        d = self.dtype()
-        d.x = 1
-        self.assertEqual(d.x, 1)
-
-    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'])
-
-    def test_iteritems_with_nonexisting_attrs(self):
-        d = self.dtype(a=1, b=1, c=1, x=2)
-        self.assertEqual(list(d.iteritems()), [('a', 1), ('b', 1), ('c', 1), ('x', 2)])
-
-    def tets_cmp_with_strict_dicts(self):
-        d = self.dtype(a=1, b=1, c=1)
-        dd = StrictDict.create(("a", "b", "c"))(a=1, b=1, c=1)
-        self.assertEqual(d, dd)
-
-    def test_cmp_with_strict_dict_with_nonexisting_attrs(self):
-        d = self.dtype(a=1, b=1, c=1, x=2)
-        dd = StrictDict.create(("a", "b", "c", "x"))(a=1, b=1, c=1, x=2)
-        self.assertEqual(d, dd)
-
 if __name__ == '__main__':
     unittest.main()

tests/test_dereference.py

@@ -200,8 +200,8 @@ class FieldTest(unittest.TestCase):
         group = Group(author=user, members=[user]).save()
 
         raw_data = Group._get_collection().find_one()
-        self.assertTrue(isinstance(raw_data['author'], DBRef))
-        self.assertTrue(isinstance(raw_data['members'][0], DBRef))
+        self.assertIsInstance(raw_data['author'], DBRef)
+        self.assertIsInstance(raw_data['members'][0], DBRef)
         group = Group.objects.first()
 
         self.assertEqual(group.author, user)
@@ -224,8 +224,8 @@ class FieldTest(unittest.TestCase):
         self.assertEqual(group.members, [user])
 
         raw_data = Group._get_collection().find_one()
-        self.assertTrue(isinstance(raw_data['author'], ObjectId))
-        self.assertTrue(isinstance(raw_data['members'][0], ObjectId))
+        self.assertIsInstance(raw_data['author'], ObjectId)
+        self.assertIsInstance(raw_data['members'][0], ObjectId)
 
     def test_recursive_reference(self):
         """Ensure that ReferenceFields can reference their own documents.
@@ -469,7 +469,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for m in group_obj.members:
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Document select_related
         with query_counter() as q:
@@ -485,7 +485,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for m in group_obj.members:
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Queryset select_related
         with query_counter() as q:
@@ -502,7 +502,7 @@ class FieldTest(unittest.TestCase):
                 self.assertEqual(q, 4)
 
                 for m in group_obj.members:
-                    self.assertTrue('User' in m.__class__.__name__)
+                    self.assertIn('User', m.__class__.__name__)
 
         UserA.drop_collection()
         UserB.drop_collection()
@@ -560,7 +560,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for m in group_obj.members:
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Document select_related
         with query_counter() as q:
@@ -576,7 +576,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for m in group_obj.members:
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Queryset select_related
         with query_counter() as q:
@@ -593,7 +593,7 @@ class FieldTest(unittest.TestCase):
                 self.assertEqual(q, 4)
 
                 for m in group_obj.members:
-                    self.assertTrue('User' in m.__class__.__name__)
+                    self.assertIn('User', m.__class__.__name__)
 
         UserA.drop_collection()
         UserB.drop_collection()
@@ -633,7 +633,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 2)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue(isinstance(m, User))
+                self.assertIsInstance(m, User)
 
         # Document select_related
         with query_counter() as q:
@@ -646,7 +646,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 2)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue(isinstance(m, User))
+                self.assertIsInstance(m, User)
 
        # Queryset select_related
         with query_counter() as q:
@@ -660,7 +660,7 @@ class FieldTest(unittest.TestCase):
                 self.assertEqual(q, 2)
 
                 for k, m in group_obj.members.iteritems():
-                    self.assertTrue(isinstance(m, User))
+                    self.assertIsInstance(m, User)
 
         User.drop_collection()
         Group.drop_collection()
@@ -715,7 +715,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Document select_related
         with query_counter() as q:
@@ -731,7 +731,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Queryset select_related
         with query_counter() as q:
@@ -748,7 +748,7 @@ class FieldTest(unittest.TestCase):
                 self.assertEqual(q, 4)
 
                 for k, m in group_obj.members.iteritems():
-                    self.assertTrue('User' in m.__class__.__name__)
+                    self.assertIn('User', m.__class__.__name__)
 
         Group.objects.delete()
         Group().save()
@@ -806,7 +806,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 2)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue(isinstance(m, UserA))
+                self.assertIsInstance(m, UserA)
 
         # Document select_related
         with query_counter() as q:
@@ -822,7 +822,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 2)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue(isinstance(m, UserA))
+                self.assertIsInstance(m, UserA)
 
         # Queryset select_related
         with query_counter() as q:
@@ -839,7 +839,7 @@ class FieldTest(unittest.TestCase):
                 self.assertEqual(q, 2)
 
                 for k, m in group_obj.members.iteritems():
-                    self.assertTrue(isinstance(m, UserA))
+                    self.assertIsInstance(m, UserA)
 
         UserA.drop_collection()
         Group.drop_collection()
@@ -894,7 +894,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Document select_related
         with query_counter() as q:
@@ -910,7 +910,7 @@ class FieldTest(unittest.TestCase):
             self.assertEqual(q, 4)
 
             for k, m in group_obj.members.iteritems():
-                self.assertTrue('User' in m.__class__.__name__)
+                self.assertIn('User', m.__class__.__name__)
 
         # Queryset select_related
         with query_counter() as q:
@@ -927,7 +927,7 @@ class FieldTest(unittest.TestCase):
                 self.assertEqual(q, 4)
 
                 for k, m in group_obj.members.iteritems():
-                    self.assertTrue('User' in m.__class__.__name__)
+                    self.assertIn('User', m.__class__.__name__)
 
         Group.objects.delete()
         Group().save()
@@ -1029,7 +1029,6 @@ class FieldTest(unittest.TestCase):
         self.assertEqual(type(foo.bar), Bar)
         self.assertEqual(type(foo.baz), Baz)
 
-
     def test_document_reload_reference_integrity(self):
         """
         Ensure reloading a document with multiple similar id
@@ -1209,10 +1208,10 @@ class FieldTest(unittest.TestCase):
 
         # Can't use query_counter across databases - so test the _data object
         book = Book.objects.first()
-        self.assertFalse(isinstance(book._data['author'], User))
+        self.assertNotIsInstance(book._data['author'], User)
 
         book.select_related()
-        self.assertTrue(isinstance(book._data['author'], User))
+        self.assertIsInstance(book._data['author'], User)
 
     def test_non_ascii_pk(self):
         """

tests/test_utils.py

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

tests/utils.py

@@ -7,12 +7,19 @@ from mongoengine.connection import get_db, get_connection
 from mongoengine.python_support import IS_PYMONGO_3
 
 
-MONGO_TEST_DB = 'mongoenginetest'
+MONGO_TEST_DB = 'mongoenginetest'   # standard name for the test database
+
+
+# Constant that can be used to compare the version retrieved with
+# get_mongodb_version()
+MONGODB_26 = (2, 6)
+MONGODB_3 = (3,0)
+MONGODB_32 = (3, 2)
 
 
 class MongoDBTestCase(unittest.TestCase):
     """Base class for tests that need a mongodb connection
-    db is being dropped automatically
+    It ensures that the db is clean at the beginning and dropped at the end automatically
     """
 
     @classmethod
@@ -27,40 +34,46 @@ class MongoDBTestCase(unittest.TestCase):
 
 
 def get_mongodb_version():
-    """Return the version tuple of the MongoDB server that the default
-    connection is connected to.
-    """
-    return tuple(get_connection().server_info()['versionArray'])
+    """Return the version of the connected mongoDB (first 2 digits)
 
-def _decorated_with_ver_requirement(func, ver_tuple):
+    :return: tuple(int, int)
+    """
+    version_list = get_connection().server_info()['versionArray'][:2]     # e.g: (3, 2)
+    return tuple(version_list)
+
+
+def _decorated_with_ver_requirement(func, version):
     """Return a given function decorated with the version requirement
     for a particular MongoDB version tuple.
+
+    :param version: The version required (tuple(int, int))
     """
     def _inner(*args, **kwargs):
-        mongodb_ver = get_mongodb_version()
-        if mongodb_ver >= ver_tuple:
+        MONGODB_V = get_mongodb_version()
+        if MONGODB_V >= version:
             return func(*args, **kwargs)
 
-        raise SkipTest('Needs MongoDB v{}+'.format(
-            '.'.join([str(v) for v in ver_tuple])
-        ))
+        raise SkipTest('Needs MongoDB v{}+'.format('.'.join(str(n) for n in version)))
 
     _inner.__name__ = func.__name__
     _inner.__doc__ = func.__doc__
 
     return _inner
 
-def needs_mongodb_v26(func):
+
+def requires_mongodb_gte_26(func):
     """Raise a SkipTest exception if we're working with MongoDB version
     lower than v2.6.
     """
-    return _decorated_with_ver_requirement(func, (2, 6))
+    return _decorated_with_ver_requirement(func, MONGODB_26)
 
-def needs_mongodb_v3(func):
+
+def requires_mongodb_gte_3(func):
     """Raise a SkipTest exception if we're working with MongoDB version
     lower than v3.0.
     """
-    return _decorated_with_ver_requirement(func, (3, 0))
+    return _decorated_with_ver_requirement(func, MONGODB_3)
+
 
 def skip_pymongo3(f):
     """Raise a SkipTest exception if we're running a test against

tox.ini

@@ -1,13 +1,12 @@
 [tox]
-envlist = {py27,py35,pypy,pypy3}-{mg27,mg28,mg30}
+envlist = {py27,py35,pypy,pypy3}-{mg35,mg3x}
 
 [testenv]
 commands =
     python setup.py nosetests {posargs}
 deps =
     nose
-    mg27: PyMongo<2.8
-    mg28: PyMongo>=2.8,<2.9
-    mg30: PyMongo>=3.0
+    mg35: PyMongo==3.5
+    mg3x: PyMongo>=3.0,<3.7
 setenv =
     PYTHON_EGG_CACHE = {envdir}/python-eggs