⚡ Power Dataclass ⚡

Installation

pip install powerdataclass

Usage

Python 3.7 have introduced a spiritual successor of NamedTuple: the dataclass. While being nice, the dataclass type hinting is only, well, hinting.

This library gives you an ability to create dataclasses with field values automatically cast to the types defined in the dataclass's type hints:

Typecasting

from powerdataclass import *


class Coordinates(PowerDataclass):
    x: int
    y: int


c1 = Coordinates(1, 2)
c2 = Coordinates('1', '2')
c3 = Coordinates(**{'x': 1.1, 'y': 2.2})

# >>> c1
Coordinates(x=1, y=2)
# >>> c1 == c2 == c3 
True

This also works with every generic type that has a Python primitive type as its origin. This applies to subscriptable types of any level of nestedness as well:

class Vector(PowerDataclass):
    items: List[int]


v1 = Vector(['1', '2', '3'])
v2 = Vector({1.1, 2.2, 3.3})
v3 = Vector(range(1, 4))

# >>> v1
Vector(items=[1, 2, 3])
# >>> v1 == v2 == v3 
True

The typecasting also respects other dataclasses (and Power Dataclasses) declared in type hints. If you pass a mapping or an iterable in place of actual dataclass instance, Power Dataclass will attempt to unpack it to a corresponding dataclass:

class Vector(PowerDataclass):
    items: List[int]


class Tensor(PowerDataclass):
    vectors: List[Vector]


t1 = Tensor(**{
    'vectors': [
        {'items': [1, 2, 3]},
        {'items': [4, 5, 6]},
        ([7, 8, 9],),
    ]
}
            )

# >>> t1
Tensor(vectors=[Vector(items=[1, 2, 3]), Vector(items=[4, 5, 6]), Vector(items=[7, 8, 9])])

If a value type is defined as a dataclass and that dataclass can be instantiated with a sole argument, it will be cast as well

class TimestampedIntValue(PowerDataclass):
    value: int
    timestamp: int = time.time()


class SensorReadings(PowerDataclass):
    moon_phase_angle: TimestampedIntValue
    mars_surface_temperature: TimestampedIntValue


readings = SensorReadings(122, -70)

# >>> readings
SensorReadings(moon_phase_angle=TimestampedIntValue(value=122, timestamp=1570898094),
               mars_surface_temperature=TimestampedIntValue(value=-70, timestamp=1570898094)
               )

Custom typecasting

You can modify the behaviour of type casting by registering two types of handlers on your fancy PowerDataclass:

• type handlers: a unary method marked as a type handler will be applied to any value that has a matching type declared in your dataclass typehints.
• field handlers: a unary method marked as a field handler will be applied to a value of a specific PDC field.

Those functions must always return a value.

You can do this by marking your methods with special decorators:

class CoolBool(PowerDataclass):
    string_bool: bool
    negated_bool: bool

    @type_handler(bool)
    def handle_bools(self, v):
        if type(v) is str:
            return v.lower() in ['y', 'yes', '1', 'True']
        else:
            return bool(v)

    @field_handler('negated_bool')
    def handle_negated_bools(self, v):
        return not self.handle_bools(v)


# >>> CoolBool('yes', 'no')
CoolBool(string_bool=True, negated_bool=True)

Field handlers take precedence over the type handlers. Field handlers and type handlers are scoped to a particular Power Dataclass. Inheritance is respected.

Field Metadata

The behaviour of fields can be modified by providing corresponding flags in a field's metadata dictionary, provided by base Python dataclasses.

Nullability

Fields are considered non-nullable by default. This means that if, during instantiation, the value of a field will be equal to None, a ValueError exception will occur. Type casting will be performed on non-null values, except for non-typecast fields (see below)

If a field has a default value, and it is None, it will be considered nullable. Also, if you want to accept None but you either don't want to provide defaults at all, provide a non-null default or provide a default factory, you can mark your field as nullable by either setting the flag or using a pre-made partial:

class Nihilus(PowerDataclass):
    x: int = field(metadata={FieldMeta.NULLABLE: True})
    y: int = None
    z: list = nullable_field(default_factory=list)


# >>> Nihilus()

# ! TypeError: __init__() missing 1 required positional argument: 'x'

# >>> Nihilus(1)
Nihilus(x=1, y=None, z=[])

# >>> Nihilus('1', '1', (1,))
Nihilus(x=1, y=1, z=[1])

# >>> Nihilus('1', None, None)
Nihilus(x=1, y=1, z=None)

Skipping typecasting (and null checking)

If you want to disable type checking for a specific field you can mark your field as nullable by either setting the corresponding flag in the fields' metadata dictionary or using a pre-made partial:

class Noncasted(PowerDataclass):
    x: int = field(metadata={FieldMeta.SKIP_TYPECASTING: True})
    y: int = noncasted_field()


# >>> Noncasted('1', 2.2)
Noncasted(x='1', y=2.2)

Dependent and calculated fields

If some of your field processing requires other fields typecast before you can declare this field dependencies by name by setting the corresponding value in the fields' metadata:

class Dependent(PowerDataclass):
    a: int
    b: int = field(metadata={FieldMeta.DEPENDS_ON_FIELDS: ['a']})
    c: int = field(metadata={FieldMeta.DEPENDS_ON_FIELDS: ['d', 'b']})
    d: int = field(metadata={FieldMeta.DEPENDS_ON_FIELDS: ['a']})

Fields will be topologically sorted by their dependencies and type casting will be done in this order. For this example, the order will be:

1. a
2. b
3. d
4. c

You can use a combination of field handlers and dependent fields to declare calculated fields:

class CubeSquarer(PowerDataclass):
    n: int
    n_square: int = field(default=None, metadata={FieldMeta.DEPENDS_ON_FIELDS: ['n']})
    n_cube: int = calculated_field(depends_on=['n'])

    @field_handler('n_square')
    def handle_n_square(self, v):
        return self.n ** 2

    @field_handler('n_cube')
    def handle_n_cube(self, v):
        return self.n ** 3


# >>> CubeSquarer(4)
CubeSquarer(n=4, n_square=16, n_cube=256)

It is an error to declare a field as calculatable without registering a corresponding field_handler

Modification of Power Dataclass behaviour

You can modify the behaviour of Power Dataclass by editing the Meta nested class' attributes. All Power Dataclasses have a default value for this Meta nested class equal to powerdataclass.PowerDataclassDefaultMeta This Meta subclass will emulate the behaviour of class variable inheritance, making every attribute of Meta default to powerdataclass.PowerDataclassDefaultMeta

Currently, the following values are now supported:

Name	Default value	Description
dataclass_init	True	passed to the dataclasses.dataclass constructor. See docs
dataclass_repr	True	passed to the dataclasses.dataclass constructor.
dataclass_eq	True	passed to the dataclasses.dataclass constructor.
dataclass_order	False	passed to the dataclasses.dataclass constructor.
dataclass_unsafe_hash	False	passed to the dataclasses.dataclass constructor.
dataclass_frozen	False	passed to the dataclasses.dataclass constructor.
singleton	False	If True enables the Singleton Mode.
json_encoder	None	If set, this class will be used as a cls param to json.dumps in PowerDataclass().to_json() See docs.
json_decoder	None	If set, this class will be used as a cls param to json.loads in PowerDataclass.from_json() See docs.
as_dict_ignored_when_nested	False	If set to True, this PDC won't be converted when this PDC is nested and wrapping PDC's .as_dict() is called. Can be further ignored if as_dict(force=True) was called.

Example of setting the Meta of a PowerDataclass:

class PowerDataclassWithNewBehaviour(PowerDataclass):
    class Meta:
        dataclass_frozen = True
        singleton = True

Singleton Mode

If you set the Meta.singleton value to True, your PowerDataclass will turn into a Singleton.

This means that this PowerDataclass can be instantiated only once, and all further attempts to instantiate this PDC will return that instance instead:

class PDCSingleton(PowerDataclass):
    a: int

    class Meta:
        singleton = True


singleton1 = PDCSingleton(1)
singleton2 = PDCSingleton(2)

# >>> id(singleton1) == id(singleton2)
True

You can test whether a Singleton has been instantiated by calling the class method .get_instance() on your Singleton Mode class. If there is an instance, it will be returned. Otherwise, None will lbe returned.

Other features

• Automatic recursive conversion to dict with the .as_dict() method.
• Automatic recursive conversion to and from JSON strings with the .as_json() and .from_json() methods.

PowerDataclass merging

The PowerDataclass.merge(other) allows you to merge two PowerDataclasses, rewriting the fields' values of the first PDC with the corresponding values of the second PDCs, while retaining the memory address of the first PDC.

class PDC(PowerDataclass):
    x: int
    y: int
    z: int


a = PDC(1, 2, 3)
b = PDC(3, 4, 5)
a.merge(b)
# >>> id(a) != id(b)
True
# >>> a.as_dict() == b.as_dict()
True

PowerDataclass diff

A simple dictionary diff can be calculated between two instances of a same PowerDataclass by using the .diff() method

class DiffPDC(PowerDataclass):
    x: int
    y: int
    z: int


a = DiffPDC(1, 2, 3)
b = DiffPDC(3, 4, 5)

# >>> a.diff(b)
{'x': (1, 3), 'y': (2, 4), 'z': (3, 5)}

Note that an attempt to compare PowerDataclasses of different type will result a DiffImpossible exception

PowerConfig

The powerdataclass.powerconfig package contains two pre-made classes suitable for simple configuration management in your services. Those classes are: the PowerConfig and it's singleton mode subclass, the GlobalPowerConfig

Both of those share two extensions over regular PowerDataclass:

• A PowerConfig can be instantiated from the os environment. PowerDataclass.Meta.envvar_prefix will be prepended to capitalized names of PowerConfig's fields' names.

  class Config(PowerConfig):
      a: int
  
      class Meta:
          envvar_prefix = "CNF"
          
  # >>> Config.from_environ()
  Config(a=5)

  This class method will read the OS environment variable CNF_A. In this example. it ts assumed that this variable is present and is equal to 5.

• there is a predefined type_handler for the bool type, which casts string values in (y, yes, 1, True) to True.

---

Made with ⚡ by Arish Pyne (https://github.com/arishpyne/powerdataclass)