Python MongoDB light toolkit
Switch branches/tags
Nothing to show
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
doc first release of MongoLite (beta) Aug 5, 2011
mongolite support keywords in Collection Jan 30, 2013
THANKS first release of MongoLite (beta) Aug 5, 2011



MongoLite is a fork of MongoKit by the same author. It aims to come back to simplicity by stripping a lot of features and replacing them with best practices.

The goal of MongoLite is to stick as much as possible to the pymongo api. MongoLite always choose speed over syntaxic sugar, this is why you won't see validation or dot notation features in this project.

MongoLite is perfect for who wants a thin layer on top of pymongo and don't care about validation stuff.

A mongolite is a beautiful stone

Your data is clean:

"Tools change, not data". In order to follow this "credo", just like MongoKit, MongoLite won't add any information into your data saved into the database. So if you need to use other mongo tools or ODMs in other languages, your data won't be polluted by MongoLite's stuff.


  • schema less feature
  • inheritance and polymorphisme support
  • skeleton generation (your object is automaticaly filled by the correct fields)
  • nested and complex schema declaration
  • default values features
  • random query support (which returns a random document from the database)
  • json helpers
  • GridFS support

A quick example

A Document declaration look like that:

>>> from mongolite import *
>>> import datetime

>>> connection = Connection()

>>> @connection.register
... class BlogPost(Document):
...     __database__ = 'test'
...     __collection__ = 'example'
...     skeleton = {
...             'title':unicode,
...             'body':unicode,
...             'author':unicode,
...             'date_creation':datetime.datetime,
...             'rank':int
...     }
...     optional = {
...             'tags': [unicode],
...     }
...     default_values = {'rank':0, 'date_creation':datetime.datetime.utcnow}

>>> blogpost = con.BlogPost() # this use the db "test" from `__database__` and the collection "example" from `__collection__`
>>> blogpost # the skeleton is automatically generated (based on the skeleton attribute)
{'body': None, 'title': None, 'date_creation': datetime.datetime(...), 'rank': 0, 'author': None}
>>> blogpost['title'] = u'my title'
>>> blogpost['body'] = u'a body'
>>> blogpost['author'] = u'me'
>>> blogpost['tags'] = ['about me', 'first post']
>>> blogpost
{'body': u'a body', 'title': u'my title', 'date_creation': datetime.datetime(...), 'rank': 0, 'author': u'me', 'tags': [u'about me', u'first post']}

Note that, while fields in skeleton should be present in the document, fields in optional attribute are not generated by default. It aims to be for documentation only...

To access those fields, use the following convention:

for fields in skeleton:

>>> title = blogpost['title']

for fields in optional:

>>> tags = blogpost.get('tags', [])

MongoLite is written on top of pymongo. All the pymongo's API is accessible and the results are wrapped into Document objects:

>>> blogpost = con.BlogPost.find_one() # this is a blogpost object

However, if you need more performances, you can use the pymongo layer directly:

>>> blogpost = con.test.example.find_one() # this is a dict

Suggestion and patches are really welcome. If you find mistakes in the documentation (english is not my primary langage) feel free to contact me. You can find me (namlook) on twitter.

Recent Change Log


  • support pymongo 2.4


  • support pymongo 2.3
  • added fix for unconditional access to __wrap on cursors
  • add __getitem__ suport on Cursor


  • change pymongo.objectid -> bson.objectid as this alias was deleted in pymongo 2.2
  • add exception in case of bad behavior with inherited queries


  • fix generate_indexes() and slightly api change


  • add inherited queries support
  • add serialize support
  • fix pymongo's version


  • fork from MongoKit, strip all unwanted features