Compare commits
1 Commits
v0.21.0
...
bagerard-p
Author | SHA1 | Date | |
---|---|---|---|
|
d73ca6f90d |
33
.github/workflows/main.yml
vendored
Normal file
33
.github/workflows/main.yml
vendored
Normal file
@@ -0,0 +1,33 @@
|
||||
# This is a basic workflow to help you get started with Actions
|
||||
|
||||
name: CI
|
||||
|
||||
# Controls when the action will run. Triggers the workflow on push or pull request
|
||||
# events but only for the master branch
|
||||
on:
|
||||
push:
|
||||
branches: [ master ]
|
||||
pull_request:
|
||||
branches: [ master ]
|
||||
|
||||
# A workflow run is made up of one or more jobs that can run sequentially or in parallel
|
||||
jobs:
|
||||
# This workflow contains a single job called "build"
|
||||
build:
|
||||
# The type of runner that the job will run on
|
||||
runs-on: ubuntu-latest
|
||||
|
||||
# Steps represent a sequence of tasks that will be executed as part of the job
|
||||
steps:
|
||||
# Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
|
||||
- uses: actions/checkout@v2
|
||||
|
||||
# Runs a single command using the runners shell
|
||||
- name: Run a one-line script
|
||||
run: echo Hello, world!
|
||||
|
||||
# Runs a set of commands using the runners shell
|
||||
- name: Run a multi-line script
|
||||
run: |
|
||||
echo Add other actions to build,
|
||||
echo test, and deploy your project.
|
@@ -33,7 +33,7 @@ clean:
|
||||
html:
|
||||
$(SPHINXBUILD) -b html $(ALLSPHINXOPTS) $(BUILDDIR)/html
|
||||
@echo
|
||||
@echo "Build finished. Check $(BUILDDIR)/html/index.html"
|
||||
@echo "Build finished. The HTML pages are in $(BUILDDIR)/html."
|
||||
|
||||
dirhtml:
|
||||
$(SPHINXBUILD) -b dirhtml $(ALLSPHINXOPTS) $(BUILDDIR)/dirhtml
|
||||
|
@@ -6,23 +6,17 @@ Changelog
|
||||
Development
|
||||
===========
|
||||
- (Fill this out as you fix issues and develop your features).
|
||||
|
||||
Changes in 0.21.0
|
||||
=================
|
||||
- Bug fix in DynamicDocument which is not parsing known fields in constructor like Document do #2412
|
||||
- When using pymongo >= 3.7, make use of Collection.count_documents instead of Collection.count
|
||||
and Cursor.count that got deprecated in pymongo >= 3.7.
|
||||
This should have a negative impact on performance of count see Issue #2219
|
||||
- Fix a bug that made the queryset drop the read_preference after clone().
|
||||
- Remove Py3.5 from CI as it reached EOL and add Python 3.9
|
||||
- Fix some issues related with db_field/field conflict in constructor #2414
|
||||
- BREAKING CHANGE: Fix the behavior of Doc.objects.limit(0) which should return all documents (similar to mongodb) #2311
|
||||
- Fix the behavior of Doc.objects.limit(0) which should return all documents (similar to mongodb) #2311
|
||||
- Bug fix in ListField when updating the first item, it was saving the whole list, instead of
|
||||
just replacing the first item (as usually done when updating 1 item of the list) #2392
|
||||
just replacing the first item (as it's usually done) #2392
|
||||
- Add EnumField: ``mongoengine.fields.EnumField``
|
||||
- Refactoring - Remove useless code related to Document.__only_fields and Queryset.only_fields
|
||||
- Fix query transformation regarding special operators #2365
|
||||
- Bug Fix: Document.save() fails when shard_key is not _id #2154
|
||||
|
||||
Changes in 0.20.0
|
||||
=================
|
||||
|
@@ -14,6 +14,5 @@ User Guide
|
||||
gridfs
|
||||
signals
|
||||
text-indexes
|
||||
migration
|
||||
logging-monitoring
|
||||
mongomock
|
||||
|
@@ -1,267 +0,0 @@
|
||||
===================
|
||||
Documents migration
|
||||
===================
|
||||
|
||||
The structure of your documents and their associated mongoengine schemas are likely
|
||||
to change over the lifetime of an application. This section provides guidance and
|
||||
recommendations on how to deal with migrations.
|
||||
|
||||
Due to the very flexible nature of mongodb, migrations of models aren't trivial and
|
||||
for people that know about `alembic` for `sqlalchemy`, there is unfortunately no equivalent
|
||||
library that will manage the migration in an automatic fashion for mongoengine.
|
||||
|
||||
Example 1: Addition of a field
|
||||
==============================
|
||||
|
||||
Let's start by taking a simple example of a model change and review the different option you
|
||||
have to deal with the migration.
|
||||
|
||||
Let's assume we start with the following schema and save an instance:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
class User(Document):
|
||||
name = StringField()
|
||||
|
||||
User(name="John Doe").save()
|
||||
|
||||
# print the objects as they exist in mongodb
|
||||
print(User.objects().as_pymongo()) # [{u'_id': ObjectId('5d06b9c3d7c1f18db3e7c874'), u'name': u'John Doe'}]
|
||||
|
||||
On the next version of your application, let's now assume that a new field `enabled` gets added to the
|
||||
existing ``User`` model with a `default=True`. Thus you simply update the ``User`` class to the following:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
class User(Document):
|
||||
name = StringField(required=True)
|
||||
enabled = BooleaField(default=True)
|
||||
|
||||
Without applying any migration, we now reload an object from the database into the ``User`` class
|
||||
and checks its `enabled` attribute:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
assert User.objects.count() == 1
|
||||
user = User.objects().first()
|
||||
assert user.enabled is True
|
||||
assert User.objects(enabled=True).count() == 0 # uh?
|
||||
assert User.objects(enabled=False).count() == 0 # uh?
|
||||
|
||||
# this is consistent with what we have in the database
|
||||
# in fact, 'enabled' does not exist
|
||||
print(User.objects().as_pymongo().first()) # {u'_id': ObjectId('5d06b9c3d7c1f18db3e7c874'), u'name': u'John'}
|
||||
assert User.objects(enabled=None).count() == 1
|
||||
|
||||
As you can see, even if the document wasn't updated, mongoengine applies the default value seamlessly when it
|
||||
loads the pymongo dict into a ``User`` instance. At first sight it looks like you don't need to migrate the
|
||||
existing documents when adding new fields but this actually leads to inconsistencies when it comes to querying.
|
||||
|
||||
In fact, when querying, mongoengine isn't trying to account for the default value of the new field and so
|
||||
if you don't actually migrate the existing documents, you are taking a risk that querying/updating
|
||||
will be missing relevant record.
|
||||
|
||||
When adding fields/modifying default values, you can use any of the following to do the migration
|
||||
as a standalone script:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
# Use mongoengine to set a default value for a given field
|
||||
User.objects().update(enabled=True)
|
||||
# or use pymongo
|
||||
user_coll = User._get_collection()
|
||||
user_coll.update_many({}, {'$set': {'enabled': True}})
|
||||
|
||||
|
||||
Example 2: Inheritance change
|
||||
=============================
|
||||
|
||||
Let's consider the following example:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
class Human(Document):
|
||||
name = StringField()
|
||||
meta = {"allow_inheritance": True}
|
||||
|
||||
class Jedi(Human):
|
||||
dark_side = BooleanField()
|
||||
light_saber_color = StringField()
|
||||
|
||||
Jedi(name="Darth Vader", dark_side=True, light_saber_color="red").save()
|
||||
Jedi(name="Obi Wan Kenobi", dark_side=False, light_saber_color="blue").save()
|
||||
|
||||
assert Human.objects.count() == 2
|
||||
assert Jedi.objects.count() == 2
|
||||
|
||||
# Let's check how these documents got stored in mongodb
|
||||
print(Jedi.objects.as_pymongo())
|
||||
# [
|
||||
# {'_id': ObjectId('5fac4aaaf61d7fb06046e0f9'), '_cls': 'Human.Jedi', 'name': 'Darth Vader', 'dark_side': True, 'light_saber_color': 'red'},
|
||||
# {'_id': ObjectId('5fac4ac4f61d7fb06046e0fa'), '_cls': 'Human.Jedi', 'name': 'Obi Wan Kenobi', 'dark_side': False, 'light_saber_color': 'blue'}
|
||||
# ]
|
||||
|
||||
As you can observe, when you use inheritance, MongoEngine stores a field named '_cls' behind the scene to keep
|
||||
track of the Document class.
|
||||
|
||||
Let's now take the scenario that you want to refactor the inheritance schema and:
|
||||
- Have the Jedi's with dark_side=True/False become GoodJedi's/DarkSith
|
||||
- get rid of the 'dark_side' field
|
||||
|
||||
move to the following schemas:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
# unchanged
|
||||
class Human(Document):
|
||||
name = StringField()
|
||||
meta = {"allow_inheritance": True}
|
||||
|
||||
# attribute 'dark_side' removed
|
||||
class GoodJedi(Human):
|
||||
light_saber_color = StringField()
|
||||
|
||||
# new class
|
||||
class BadSith(Human):
|
||||
light_saber_color = StringField()
|
||||
|
||||
MongoEngine doesn't know about the change or how to map them with the existing data
|
||||
so if you don't apply any migration, you will observe a strange behavior, as if the collection was suddenly
|
||||
empty.
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
# As a reminder, the documents that we inserted
|
||||
# have the _cls field = 'Human.Jedi'
|
||||
|
||||
# Following has no match
|
||||
# because the query that is used behind the scene is
|
||||
# filtering on {'_cls': 'Human.GoodJedi'}
|
||||
assert GoodJedi.objects().count() == 0
|
||||
|
||||
# Following has also no match
|
||||
# because it is filtering on {'_cls': {'$in': ('Human', 'Human.GoodJedi', 'Human.BadSith')}}
|
||||
# which has no match
|
||||
assert Human.objects.count() == 0
|
||||
assert Human.objects.first() is None
|
||||
|
||||
# If we bypass MongoEngine and make use of underlying driver (PyMongo)
|
||||
# we can see that the documents are there
|
||||
humans_coll = Human._get_collection()
|
||||
assert humans_coll.count_documents({}) == 2
|
||||
# print first document
|
||||
print(humans_coll.find_one())
|
||||
# {'_id': ObjectId('5fac4aaaf61d7fb06046e0f9'), '_cls': 'Human.Jedi', 'name': 'Darth Vader', 'dark_side': True, 'light_saber_color': 'red'}
|
||||
|
||||
As you can see, first obvious problem is that we need to modify '_cls' values based on existing values of
|
||||
'dark_side' documents.
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
humans_coll = Human._get_collection()
|
||||
old_class = 'Human.Jedi'
|
||||
good_jedi_class = 'Human.GoodJedi'
|
||||
bad_sith_class = 'Human.BadSith'
|
||||
humans_coll.update_many({'_cls': old_class, 'dark_side': False}, {'$set': {'_cls': good_jedi_class}})
|
||||
humans_coll.update_many({'_cls': old_class, 'dark_side': True}, {'$set': {'_cls': bad_sith_class}})
|
||||
|
||||
Let's now check if querying improved in MongoEngine:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
assert GoodJedi.objects().count() == 1 # Hoorah!
|
||||
assert BadSith.objects().count() == 1 # Hoorah!
|
||||
assert Human.objects.count() == 2 # Hoorah!
|
||||
|
||||
# let's now check that documents load correctly
|
||||
jedi = GoodJedi.objects().first()
|
||||
# raises FieldDoesNotExist: The fields "{'dark_side'}" do not exist on the document "Human.GoodJedi"
|
||||
|
||||
In fact we only took care of renaming the _cls values but we havn't removed the 'dark_side' fields
|
||||
which does not exist anymore on the GoodJedi's and BadSith's models.
|
||||
Let's remove the field from the collections:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
humans_coll = Human._get_collection()
|
||||
humans_coll.update_many({}, {'$unset': {'dark_side': 1}})
|
||||
|
||||
.. note:: We did this migration in 2 different steps for the sake of example but it could have been combined
|
||||
with the migration of the _cls fields: ::
|
||||
|
||||
humans_coll.update_many(
|
||||
{'_cls': old_class, 'dark_side': False},
|
||||
{
|
||||
'$set': {'_cls': good_jedi_class},
|
||||
'$unset': {'dark_side': 1}
|
||||
}
|
||||
)
|
||||
|
||||
|
||||
And verify that the documents now load correctly:
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
jedi = GoodJedi.objects().first()
|
||||
assert jedi.name == "Obi Wan Kenobi"
|
||||
|
||||
sith = BadSith.objects().first()
|
||||
assert sith.name == "Darth Vader"
|
||||
|
||||
|
||||
An other way of dealing with this migration is to iterate over
|
||||
the documents and update/replace them one by one. This is way slower but
|
||||
it is often useful for complex migrations of Document models.
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
for doc in humans_coll.find():
|
||||
if doc['_cls'] == 'Human.Jedi':
|
||||
doc['_cls'] = 'Human.BadSith' if doc['dark_side'] else 'Human.GoodJedi'
|
||||
doc.pop('dark_side')
|
||||
humans_coll.replace_one({'_id': doc['_id']}, doc)
|
||||
|
||||
.. warning:: Be aware of this `flaw <https://groups.google.com/g/mongodb-user/c/AFC1ia7MHzk>`_ if you modify documents while iterating
|
||||
|
||||
Recommendations
|
||||
===============
|
||||
|
||||
- Write migration scripts whenever you do changes to the model schemas
|
||||
- Using :class:`~mongoengine.DynamicDocument` or ``meta = {"strict": False}`` may help to avoid some migrations or to have the 2 versions of your application to co-exist.
|
||||
- Write post-processing checks to verify that migrations script worked. See below
|
||||
|
||||
Post-processing checks
|
||||
======================
|
||||
|
||||
The following recipe can be used to sanity check a Document collection after you applied migration.
|
||||
It does not make any assumption on what was migrated, it will fetch 1000 objects randomly and
|
||||
run some quick checks on the documents to make sure the document looks OK. As it is, it will fail
|
||||
on the first occurrence of an error but this is something that can be adapted based on your needs.
|
||||
|
||||
.. code-block:: python
|
||||
|
||||
def get_random_oids(collection, sample_size):
|
||||
pipeline = [{"$project": {'_id': 1}}, {"$sample": {"size": sample_size}}]
|
||||
return [s['_id'] for s in collection.aggregate(pipeline)]
|
||||
|
||||
def get_random_documents(DocCls, sample_size):
|
||||
doc_collection = DocCls._get_collection()
|
||||
random_oids = get_random_oids(doc_collection, sample_size)
|
||||
return DocCls.objects(id__in=random_oids)
|
||||
|
||||
def check_documents(DocCls, sample_size):
|
||||
for doc in get_random_documents(DocCls, sample_size):
|
||||
# general validation (types and values)
|
||||
doc.validate()
|
||||
|
||||
# load all subfields,
|
||||
# this may trigger additional queries if you have ReferenceFields
|
||||
# so it may be slow
|
||||
for field in doc._fields:
|
||||
try:
|
||||
getattr(doc, field)
|
||||
except Exception:
|
||||
LOG.warning(f"Could not load field {field} in Document {doc.id}")
|
||||
raise
|
||||
|
||||
check_documents(Human, sample_size=1000)
|
@@ -28,7 +28,7 @@ __all__ = (
|
||||
)
|
||||
|
||||
|
||||
VERSION = (0, 21, 0)
|
||||
VERSION = (0, 20, 0)
|
||||
|
||||
|
||||
def get_version():
|
||||
|
@@ -101,13 +101,12 @@ class BaseDocument:
|
||||
|
||||
self._dynamic_fields = SON()
|
||||
|
||||
# Assign default values for fields
|
||||
# not set in the constructor
|
||||
for field_name in self._fields:
|
||||
if field_name in values:
|
||||
# Assign default values to the instance.
|
||||
for key, field in self._fields.items():
|
||||
if self._db_field_map.get(key, key) in values:
|
||||
continue
|
||||
value = getattr(self, field_name, None)
|
||||
setattr(self, field_name, value)
|
||||
value = getattr(self, key, None)
|
||||
setattr(self, key, value)
|
||||
|
||||
if "_cls" not in values:
|
||||
self._cls = self._class_name
|
||||
@@ -116,6 +115,7 @@ class BaseDocument:
|
||||
dynamic_data = {}
|
||||
FileField = _import_class("FileField")
|
||||
for key, value in values.items():
|
||||
key = self._reverse_db_field_map.get(key, key)
|
||||
field = self._fields.get(key)
|
||||
if field or key in ("id", "pk", "_cls"):
|
||||
if __auto_convert and value is not None:
|
||||
@@ -750,8 +750,7 @@ class BaseDocument:
|
||||
|
||||
@classmethod
|
||||
def _from_son(cls, son, _auto_dereference=True, created=False):
|
||||
"""Create an instance of a Document (subclass) from a PyMongo SON (dict)
|
||||
"""
|
||||
"""Create an instance of a Document (subclass) from a PyMongo SON."""
|
||||
if son and not isinstance(son, dict):
|
||||
raise ValueError(
|
||||
"The source SON object needs to be of type 'dict' but a '%s' was found"
|
||||
@@ -764,8 +763,6 @@ class BaseDocument:
|
||||
|
||||
# Convert SON to a data dict, making sure each key is a string and
|
||||
# corresponds to the right db field.
|
||||
# This is needed as _from_son is currently called both from BaseDocument.__init__
|
||||
# and from EmbeddedDocumentField.to_python
|
||||
data = {}
|
||||
for key, value in son.items():
|
||||
key = str(key)
|
||||
|
@@ -3822,95 +3822,5 @@ class ObjectKeyTestCase(MongoDBTestCase):
|
||||
assert book._object_key == {"pk": book.pk, "author__name": "Author"}
|
||||
|
||||
|
||||
class DBFieldMappingTest(MongoDBTestCase):
|
||||
def setUp(self):
|
||||
class Fields(object):
|
||||
w1 = BooleanField(db_field="w2")
|
||||
|
||||
x1 = BooleanField(db_field="x2")
|
||||
x2 = BooleanField(db_field="x3")
|
||||
|
||||
y1 = BooleanField(db_field="y0")
|
||||
y2 = BooleanField(db_field="y1")
|
||||
|
||||
z1 = BooleanField(db_field="z2")
|
||||
z2 = BooleanField(db_field="z1")
|
||||
|
||||
class Doc(Fields, Document):
|
||||
pass
|
||||
|
||||
class DynDoc(Fields, DynamicDocument):
|
||||
pass
|
||||
|
||||
self.Doc = Doc
|
||||
self.DynDoc = DynDoc
|
||||
|
||||
def tearDown(self):
|
||||
for collection in list_collection_names(self.db):
|
||||
self.db.drop_collection(collection)
|
||||
|
||||
def test_setting_fields_in_constructor_of_strict_doc_uses_model_names(self):
|
||||
doc = self.Doc(z1=True, z2=False)
|
||||
assert doc.z1 is True
|
||||
assert doc.z2 is False
|
||||
|
||||
def test_setting_fields_in_constructor_of_dyn_doc_uses_model_names(self):
|
||||
doc = self.DynDoc(z1=True, z2=False)
|
||||
assert doc.z1 is True
|
||||
assert doc.z2 is False
|
||||
|
||||
def test_setting_unknown_field_in_constructor_of_dyn_doc_does_not_overwrite_model_fields(
|
||||
self,
|
||||
):
|
||||
doc = self.DynDoc(w2=True)
|
||||
assert doc.w1 is None
|
||||
assert doc.w2 is True
|
||||
|
||||
def test_unknown_fields_of_strict_doc_do_not_overwrite_dbfields_1(self):
|
||||
doc = self.Doc()
|
||||
doc.w2 = True
|
||||
doc.x3 = True
|
||||
doc.y0 = True
|
||||
doc.save()
|
||||
reloaded = self.Doc.objects.get(id=doc.id)
|
||||
assert reloaded.w1 is None
|
||||
assert reloaded.x1 is None
|
||||
assert reloaded.x2 is None
|
||||
assert reloaded.y1 is None
|
||||
assert reloaded.y2 is None
|
||||
|
||||
def test_dbfields_are_loaded_to_the_right_modelfield_for_strict_doc_2(self):
|
||||
doc = self.Doc()
|
||||
doc.x2 = True
|
||||
doc.y2 = True
|
||||
doc.z2 = True
|
||||
doc.save()
|
||||
reloaded = self.Doc.objects.get(id=doc.id)
|
||||
assert (
|
||||
reloaded.x1,
|
||||
reloaded.x2,
|
||||
reloaded.y1,
|
||||
reloaded.y2,
|
||||
reloaded.z1,
|
||||
reloaded.z2,
|
||||
) == (doc.x1, doc.x2, doc.y1, doc.y2, doc.z1, doc.z2)
|
||||
|
||||
def test_dbfields_are_loaded_to_the_right_modelfield_for_dyn_doc_2(self):
|
||||
doc = self.DynDoc()
|
||||
doc.x2 = True
|
||||
doc.y2 = True
|
||||
doc.z2 = True
|
||||
doc.save()
|
||||
reloaded = self.DynDoc.objects.get(id=doc.id)
|
||||
assert (
|
||||
reloaded.x1,
|
||||
reloaded.x2,
|
||||
reloaded.y1,
|
||||
reloaded.y2,
|
||||
reloaded.z1,
|
||||
reloaded.z2,
|
||||
) == (doc.x1, doc.x2, doc.y1, doc.y2, doc.z1, doc.z2)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
unittest.main()
|
||||
|
@@ -2272,13 +2272,6 @@ class TestField(MongoDBTestCase):
|
||||
with pytest.raises(FieldDoesNotExist):
|
||||
Doc(bar="test")
|
||||
|
||||
def test_undefined_field_works_no_confusion_with_db_field(self):
|
||||
class Doc(Document):
|
||||
foo = StringField(db_field="bar")
|
||||
|
||||
with pytest.raises(FieldDoesNotExist):
|
||||
Doc(bar="test")
|
||||
|
||||
|
||||
class TestEmbeddedDocumentListField(MongoDBTestCase):
|
||||
def setUp(self):
|
||||
|
Reference in New Issue
Block a user