Mongothin

Mongothin is a small wrapper around pymongo. It allows you to do this:

result = UserResource.find_one('XXXX')

Instead of:

client = MongoClient()
result = client['database']['users'].find_one(ObjectId('XXXX'))

Mongothin also handles retries and exponential backoff.

Connections

Mongothin handles clients configuration the way MongoEngine does. Actually the code is taken from MongoEngine.

Resource

A resource wraps metadata such as client alias, collection, number of retries and proxies pymonogo collection methods to pymongo itself.

This is how we define a resource:

class UserResource(Resource):
    # The connection alias registered via :function:clients.register_connection.
    _alias = 'user'
    # The _id field type. Whatever is your doc_id input it will be coerced to this type.
    # Default is ObjectId. This callable must accept a single argument to whatever is used in the method
    # calls.
    _id_type = ObjectId
    # Number of retries
    _retries = 2
    # Exponential backoff base
    _delay = 0.01
    # Collection. If not specified the name of the class is used.
    _collection = 'user'
    # Shard. Make sure the shard info is in the specs
    _shard = (shard_function, shard_key,)

A Resource is heavily oriented to work with _id fields.

Data mapping

Mongothin does not do data mapping. pymongo's input and output are dictionaries, so are Mongothin's. If you need input validation you can use another library such as:

• onctuous
• validino
• FormEncode
• Schematics