DictObject

Subclassing this type you can declare your object model. DictObject itself is dict subclass so you can access object properties as attribute or as an item:

class Foo(dicty.DictObject):
    foo = dicty.Field()

obj = Foo(foo='bar')
obj.foo     # 'bar'
obj['foo']  # 'bar'

Object constructor accepts properties as keyword arguments, or you can create instance with fromjson() classmethod that takes object dictionary (e.g. result json.loads()) as an argument:

obj = Foo(foo='bar')
obj = Foo.fromjson({'foo': 'bar'})

You can pass DictObject instance directly to json library or call jsonize() method first that will return plain dict version of your object with only declared fields left:

obj = Foo.fromjson({'foo': 123, 'bar': 123})
obj == {'foo': 123, 'bar': 123}
obj.jsonize() == {'foo': 123}

Name aliasing

If you don't like naming scheme in JSON objects, API and so on. Dicty allows to choose whatever python name you like while manually specify dictionary key. For instance map camel-case keys to their underscore counterparts:

class Foo(dicty.DictObject):
    prop_foo = dicty.Field('propFoo')

obj = Foo(prop_foo=123)
obj = Foo.fromjson({'propFoo': 123})
obj.prop_foo
obj['propFoo']

Subclassing

DictObject supports subclassing:

class Foo(dicty.DictObject):
    foo = dicty.Field()


class Bar(Foo):
    bar = dicty.Field()


obj = Bar.fromjson({'foo': 1, 'bar': 2})
print obj.jsonize()  # {'foo': 1, 'bar': 2}

Mixins are supported as well:

class FooMixIn(object):
    foo = dicty.Field()


class Bar(dicty.DictObject, FooMixIn):
    bar = dicty.Field()


obj = Bar.fromjson({'foo': 1, 'bar': 2})
print obj.jsonize()  # {'foo': 1, 'bar': 2}

Fields

dicty.Field is baseclass for all dicty fields. You can use itself directly to declare a field with no special type info.

Optional fields and default values

Accessing field that is not set will lead to AttributeError: You can specify default value for your field:

class Foo(dicty.DictObject):
    foo = dicty.Field()

obj = Foo()
obj.foo  # raises AttributeError

You can mark field as optional, in this case None will be returned if it was not set before:

class Foo(dicty.DictObject):
    foo = dicty.Field(optional=True)

obj = Foo()
obj.foo  # None

For optional fields you can specify default value other than None with default argument:

class Foo(dicty.DictObject):
    foo = dicty.Field(optional=True, default=123)

obj = Foo()
obj.foo  # 123
obj == {}

Please note that default value does not affect internal dictionary. But if default value is NOT hashable dict key will be set on getattr access.

There is also an option to suply default_func it's get default value for object's field. It takes object instance as an argument. Value returned by default_func is always stored in dict:

class Foo(dicty.DictObject):
    id = dicty.Field(optional=True, default_func=lambda obj: uuid.uuid4().hex)

obj = Foo()
obj == {}
obj.id  # Would be populated with newly generated UUID
obj == {'id': '07d0af8affaf46c885cc251e17dbc37a'}

Available Fields

Dicty is shipped with the follwing:

BooleanField

DateField

DatetimeField

DictField

FloatField

IntegerField

ListField

NativeDateField

NativeDatetimeField

NumericField

RegexpStringField

StringField

TypedDictField

TypedListField

TypedObjectField

Sample usage

With dicty you can easily describe your data model and then use it to encode/decode JSON objects. It supports data validataion, optional parameters, default values, nested objects and so on.

import dicty


class MyDoc(dicty.DictObject):
    prop1 = dicty.StringField()
    prop2 = dicty.IntegerField()

# Regular constructor
doc = MyDoc(prop1='foo', prop2=123)
print doc.prop1     # you can access values as attributes
print doc['prop2']  # as well as dictionary items

print json.dumps(doc)
print json.dumps(doc.jsonify()) # Jsonify will clean and validate output data

# Create instance from dictionary
doc = MyDoc.fromjson({'prop1': 'foo', 'prop2': 123})

# would raise dicty.FieldError here
doc = MyDoc.fromjson({'prop1': 123, 'prop2': 123})

Nested Objects

import dicty


class Foo(dicty.DictObject):
    class Bar(dicty.DictObject):
        prop = dicty.StringField()

    bar = dicty.TypedObjectField(Bar)

obj = Foo()
obj.bar.prop = 123
print obj # {'bar': {'prop': 123}}

Mongo-style key pathes

Dicty allows to build key pathes that can be used to create mongo query:

class Foo(dicty.DictObject):
    bar = dicty.Field('myBar')

print Foo.bar         # 'myBar' full path to the item
print Foo.bar.key     # 'myBar' only leaf key
print Foo.bar.attname # 'bar' python attribute name

Nested object:

class Bar(dicty.DictObject):
    foo = dicty.TypedObjectField(Foo)

print Bar.foo            # 'foo'
print Bar.foo.bar        # 'foo.myBar'

List of objects:

class Bar(dicty.DictObject):
    items = dicty.TypedListField(Foo)

print Bar.items.foo        # 'items.myBar' without index
print Bar.items[0].foo     # 'items.0.myBar' indexed path

Dict of objects:

class Bar(dicty.DictObject):
    items = dicty.TypedDictField(Foo)

# With index
print Bar.items['maurice'].bar  # 'items.maurice.myBar'

# Would raise IndexError
print Bar.items['x.y'].bar