Formal!

A Python library for object document mapping with MongoDB and SQL dialects(WiP) using JSON schema.

Description

This is a package for generating classes from a JSON-schema that are to be saved in MongoDB or SQL and (un)pickled via Python's builtin json module or others like simplejson or ujson.

This extends the JSON schema by supporting extra BSON types:

• ObjectId - use the "object_id" type in your JSON schema to validate that a field is a valid ObjectId.
• datetime - use the "date" type in your JSON schema to validate that a field is a valid datetime

Usage

1. Build your schema

>>> schema = {
    'name': 'Country',
    'id': '#country',
    'properties': {
        'name': {'type': 'string'},
        'abbreviation': {'type': 'string'},
    },
    'additionalProperties': False,
}

2. Connect to your database

>>> import formal
>>> formal.connect("test")

3. Create a model

>>> Country = formal.model_factory(schema)

4. Create an object using your model

>>> sweden = Country({"name": 'Sweden', "abbreviation": 'SE'})
>>> sweden.save()
>>> sweden._id
ObjectId('50b506916ee7d81d42ca2190')

5. Let the object validate itself!

>>> sweden = Country.find_one({"name" : "Sweden"})
>>> sweden.name = 5
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "formal/model.py", line 254, in __setattr__
    self.validate_field(attr, self._schema["properties"][attr], value)
  File "formal/model.py", line 189, in validate_field
    self.validate_simple(key, value_schema, value)
  File "formal/model.py", line 236, in validate_simple
    (key, value_type, str(value), type(value)))
formal.exceptions.ValidationError: Field 'name' is of type 'string', received '5' (<type 'int'>)

>>> sweden.overlord = 'Bears'
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "formal/model.py", line 257, in __setattr__
    raise ValidationError("Additional property '%s' not allowed!" % attr)
formal.exceptions.ValidationError: Additional property 'overlord' not allowed!

6. You can also update objects from dictionaries:

>>> sweden.update({"name": "Sverige"})
>>> sweden.save()

7. To get them to a browser or other similar things, serialize them:

>>> sweden.serializablefields()
{'_id': '50b506916ee7d81d42ca2190', 'name': 'Sverige', 'abbreviation': 'SE', 'id': '#country'}

Choosing a collection

By default Formal will use the pluralized version of the model's name. If

you want to use something else, put it in the JSON-schema:

{
    "name": "MyModel",
    ...
    "collectionName": "some_collection",
    ...
}

Multiple Databases

To use multiple databases, simply call connect() multiple times:

>>> import formal
>>> formal.connect("test")
>>> formal.connect("other_db")

By default all models will use the first database specified. If you want to use

a different one, put it in the JSON-schema:

{
    "name": "MyModel",
    ...
    "databaseName": "other_db",
    ...
}

SQL Operation

..is still work in progress.

Roadmap

We have many plans for the future:

• Complete SQL support including:
  • JSON document storage
  • GeoJSON queries
• Time series support
• Support for consensus algorithms like Paxos or Raft
• Automatic data migration (up and down)
• Improved testing
  • multiple interpreters via tox
  • complete coverage
  • more tests

Ping us, if you'd like to contribute!

History

Formal is a fork of warmongo, originally written by Rob Britton.

Things that have changed:

• jsonschema is now truly used to validate objects (it validates far more than just basetypes)
• we do ignore mongo's object_id - not sure if this is a good thing, but it helps with the schemata
• we require (by spec) an 'id' field that lists a uri for the schema
• the resulting field is enforced on instantiated objects, too, so clients can validate by schema-id

Work in progress:

• Migration of versioned object models
• SQL integration
• Deep dot notation
• Delta operation for concurrent editing and object history

Licence

Apache Version 2.0

Change notice

This file has been changed by the Hackerfleet Community and a change notice has

been added to all modified files in accordance to the Apache License 2.0

Production Examples

The Isomer framework uses Formal as object document mapping system to deal with data objects in a developer and

enduser friendly way.

See it in action on http://github.com/isomeric/isomer

The original author uses Warmongo every day at his startup http://www.sweetiq.com/ to share data

definitions between their Python and Node.js applications. It has been running in

production for some time now, so it has been reasonably tested for robustness and performance.