marcboeker/mongodb-object

Python object wrapper for MongoDB

MongoDB object wrapper for python

About

mongodb-object is a cocktail of the Django ORM mixed with JavaScript dot object notation. Every document is returned as object that can be traversed using the objects attributes. It is based on the awesome pymongo lib by mike from mongodb.

A simple JSON document

{
    'me': {'age': 24}, 
    'name': 'john', 
    'friends': ['mike', 'dwight', 'elliot']
}

can now be accessed like this

>>> print doc.name
john
>>> print friends[0]
mike
>>> print doc.me.age
24

For those who are familar with the Django ORM, mongodb-object behaves very similar.

Getting started

Installation

git clone git://github.com/marcboeker/mongodb-object.git
cd mongodb-object
sudo python setup.py install

Do the mambo

Let's start with a simple demonstration in the Python interactive console. Open a MongoDB connection and start with a mongodb-object collection. The collection is the base container for our work with mongodb-object.

>>> from pymongo import Connection
>>> from mongodbobject import Collection
>>> from mongodbobject.nested import Nested

>>> db = Connection().test_db
>>> col = Collection(db, 'test_collection')

Now we create a new document using an empty document template from the collection. This is necessary because the document needs to know to which collection it belongs

>>> doc = col.new()
>>> doc.name = 'john'
>>> doc.person = Nested()
>>> doc.person.age = 24
>>> doc.person.gender = 'male'
>>> doc.save()
>>> doc
Doc(id=04be8d4a41950c76e8000000)

We can also reference documents on the fly. Retrieval is powered by lazy loading. Only when you select an attribute of a referenced document, the document will be retrieved.

>>> doc2 = col.new()
>>> doc2.friends = [doc]
>>> doc2.save()
>>> doc2.friends
[Doc(id=04be8d4a41950c76e8000000)]
>>> doc2.friends[0].person.age
24

Finding documents

As a collection holds all documents, we have to ask the collection when we want to find a special document.

>>> col.find_one(person__age=24)
Doc(id=04be8d4a41950c76e8000000)
>>> col.find_one(person__age__lt=25)
Doc(id=04be8d4a41950c76e8000000)
>>> col.find_one(person__age__lt=24)
None
>>> col.find_one(person__age=24, name='john')
Doc(id=04be8d4a41950c76e8000000)

As you can see, we emulate the dot notation of embedded documents using the __ chars like Django does. We can also append query operators ($lt, $lte, $gt, $gte, $ne, $nin, $in, $all, $size) at the end.

You cannot only query for one document, you can retrieve also a list of documents using find() instead of find_one().

>>> doc_list = col.find()
>>> doc_list
<mongodbobject.doclist.DocList instance at 0x11d5f80>
>>> doc_list.count()
2
>>> for doc in doc_list:
...     print doc, doc.keys()
... 
Doc(id=c8c88d4a41950c42e9000000) ['name', 'person']
Doc(id=c8c88d4a41950c42e9010000) ['friends']

find() accepts the same parameters as find_one() does.

Bulk update and remove

To do a bulk update or remove, we use the query() method.

>>> col.query(name='john').update(set__name='jack')
>>> col.find_one(name='jack').name
jack
>>> col.query(name='jack').remove()
>>> col.find().count()
1

update() makes use of the MongoDB update operators ($set, $inc, $push, $pushAll, $pull, $pullAll). You have to prefix the document key with one of the operators above. Embedded docs can be accessed using the previously mentiond __ chars. For example:

set__person__gender=10

Document list options (count, skip, limit, sort)

Using the find() method, you get a list of documents. This list can be sorted, or paginated using the known methods from pymongo.

>>> col.find(name='jack').count()
1
>>> col.find().skip(1).limit(1).next()
Doc(id=fecc8d4a41950c83e9010000)

>>> for doc in col.find().sort(friends=1):
...    print doc
Doc(id=82cd8d4a41950c8fe9000000)
Doc(id=82cd8d4a41950c8fe9010000)

And now sort on another key.

>>> for doc in col.find().sort(name=1):
...    print doc
Doc(id=82cd8d4a41950c8fe9010000)
Doc(id=82cd8d4a41950c8fe9000000)

Document operations

A single document can be modified and afterwards saved without any restrictions. You can also delete the document, return its keys or get a dict represntation.

>>> col.find_one(name='jack').keys()
['name', 'person']
>>> doc = col.find_one(name='jack')
>>> doc.name = 'harry'
>>> doc.save()
>>> col.find_one(name='harry').name
harry
>>> doc.delete()
>>> col.find_one(name='jack')
None

mongodb-object's Issues

find_one behaviour

Just compare this lib to pymongo on find_one()
Refer here: http://www.mongodb.org/display/DOCS/Python+Tutorial

When call collection.find_one() without any arguments it will return 'to_dict' error.

The error happen because _parse_query in collection.py does not process anything.

Sould has something that return a random record from collection. I'm trying a few quick and dirty fix now. Will post something when it working.

regards,
azraai

Python Support Prior 2.6

Doc class repr function is using str.format() which is not supported until 2.6. Hope can use replacement method prior 2.6

"import mongodbobject" in setup.py

It's not needed, but causes import exception by "from pymongo.dbref import DBRef" in collection.py, because pymongo is not installed yet and gets installed after setup.py finishes.