Following system colour scheme Selected dark colour scheme Selected light colour scheme

Python Enhancement Proposals

PEP 696 – Type Defaults for Type Parameters

Author:
James Hilton-Balfe <gobot1234yt at gmail.com>
Sponsor:
Jelle Zijlstra <jelle.zijlstra at gmail.com>
Discussions-To:
Discourse thread
Status:
Draft
Type:
Standards Track
Topic:
Typing
Created:
14-Jul-2022
Python-Version:
3.12
Post-History:
22-Mar-2022, 08-Jan-2023

Table of Contents

Abstract

This PEP introduces the concept of type defaults for type parameters, including TypeVar, ParamSpec, and TypeVarTuple, which act as defaults for type parameters for which no type is specified.

Default type argument support is available in some popular languages such as C++, TypeScript, and Rust. A survey of type parameter syntax in some common languages has been conducted by the author of PEP 695 and can be found in its Appendix A.

Motivation

T = TypeVar("T", default=int)  # This means that if no type is specified T = int

@dataclass
class Box(Generic[T]):
    value: T | None = None

reveal_type(Box())                      # type is Box[int]
reveal_type(Box(value="Hello World!"))  # type is Box[str]

One place this regularly comes up is Generator. I propose changing the stub definition to something like:

YieldT = TypeVar("YieldT")
SendT = TypeVar("SendT", default=None)
ReturnT = TypeVar("ReturnT", default=None)

class Generator(Generic[YieldT, SendT, ReturnT]): ...

Generator[int] == Generator[int, None] == Generator[int, None, None]

This is also useful for a Generic that is commonly over one type.

class Bot: ...

BotT = TypeVar("BotT", bound=Bot, default=Bot)

class Context(Generic[BotT]):
    bot: BotT

class MyBot(Bot): ...

reveal_type(Context().bot)         # type is Bot  # notice this is not Any which is what it would be currently
reveal_type(Context[MyBot]().bot)  # type is MyBot

Not only does this improve typing for those who explicitly use it, it also helps non-typing users who rely on auto-complete to speed up their development.

This design pattern is common in projects like:

  • discord.py — where the example above was taken from.
  • NumPy — the default for types like ndarray’s dtype would be float64. Currently it’s Unknown or Any.
  • TensorFlow — this could be used for Tensor similarly to numpy.ndarray and would be useful to simplify the definition of Layer.

Specification

Default Ordering and Subscription Rules

The order for defaults should follow the standard function parameter rules, so a type parameter with no default cannot follow one with a default value. Doing so should ideally raise a TypeError in typing._GenericAlias/types.GenericAlias, and a type checker should flag this as an error.

DefaultStrT = TypeVar("DefaultStrT", default=str)
DefaultIntT = TypeVar("DefaultIntT", default=int)
DefaultBoolT = TypeVar("DefaultBoolT", default=bool)
T = TypeVar("T")
T2 = TypeVar("T2")

class NonDefaultFollowsDefault(Generic[DefaultStrT, T]): ...  # Invalid: non-default TypeVars cannot follow ones with defaults


class NoNonDefaults(Generic[DefaultStrT, DefaultIntT]): ...

(
    NoNoneDefaults ==
    NoNoneDefaults[str] ==
    NoNoneDefaults[str, int]
)  # All valid


class OneDefault(Generic[T, DefaultBoolT]): ...

OneDefault[float] == OneDefault[float, bool]  # Valid
reveal_type(OneDefault)          # type is type[OneDefault[T, DefaultBoolT = bool]]
reveal_type(OneDefault[float]()) # type is OneDefault[float, bool]


class AllTheDefaults(Generic[T1, T2, DefaultStrT, DefaultIntT, DefaultBoolT]): ...

reveal_type(AllTheDefaults)                  # type is type[AllTheDefaults[T1, T2, DefaultStrT = str, DefaultIntT = int, DefaultBoolT = bool]]
reveal_type(AllTheDefaults[int, complex]())  # type is AllTheDefaults[int, complex, str, int, bool]
AllTheDefaults[int]  # Invalid: expected 2 arguments to AllTheDefaults
(
    AllTheDefaults[int, complex] ==
    AllTheDefaults[int, complex, str] ==
    AllTheDefaults[int, complex, str, int] ==
    AllTheDefaults[int, complex, str, int, bool]
)  # All valid

With the new Python 3.12 syntax for generics (introduced by PEP 695), this can be enforced at compile time:

type Alias[DefaultT = int, T] = tuple[DefaultT, T]  # SyntaxError: non-default TypeVars cannot follow ones with defaults

def generic_func[DefaultT = int, T](x: DefaultT, y: T) -> None: ...  # SyntaxError: non-default TypeVars cannot follow ones with defaults

class GenericClass[DefaultT = int, T]: ...  # SyntaxError: non-default TypeVars cannot follow ones with defaults

ParamSpec Defaults

ParamSpec defaults are defined using the same syntax as TypeVar s but use a list of types or an ellipsis literal “...” or another in-scope ParamSpec (see Scoping Rules).

DefaultP = ParamSpec("DefaultP", default=[str, int])

class Foo(Generic[DefaultP]): ...

reveal_type(Foo)                  # type is type[Foo[DefaultP = [str, int]]]
reveal_type(Foo())                # type is Foo[[str, int]]
reveal_type(Foo[[bool, bool]]())  # type is Foo[[bool, bool]]

TypeVarTuple Defaults

TypeVarTuple defaults are defined using the same syntax as TypeVar s but use an unpacked tuple of types instead of a single type or another in-scope TypeVarTuple (see Scoping Rules).

DefaultTs = TypeVarTuple("DefaultTs", default=Unpack[tuple[str, int]])

class Foo(Generic[*DefaultTs]): ...

reveal_type(Foo)               # type is type[Foo[DefaultTs = *tuple[str, int]]]
reveal_type(Foo())             # type is Foo[str, int]
reveal_type(Foo[int, bool]())  # type is Foo[int, bool]

Using Another Type Parameter as default

This allows for a value to be used again when the type parameter to a generic is missing but another type parameter is specified.

To use another type parameter as a default the default and the type parameter must be the same type (a TypeVar’s default must be a TypeVar, etc.).

This could be used on builtins.slice where the start parameter should default to int, stop default to the type of start and step default to int | None.

StartT = TypeVar("StartT", default=int)
StopT = TypeVar("StopT", default=StartT)
StepT = TypeVar("StepT", default=int | None)

class slice(Generic[StartT, StopT, StepT]): ...

reveal_type(slice)  # type is type[slice[StartT = int, StopT = StartT, StepT = int | None]]
reveal_type(slice())                        # type is slice[int, int, int | None]
reveal_type(slice[str]())                   # type is slice[str, str, int | None]
reveal_type(slice[str, bool, timedelta]())  # type is slice[str, bool, timedelta]

T2 = TypeVar("T2", default=DefaultStrT)

class Foo(Generic[DefaultStrT, T2]):
    def __init__(self, a: DefaultStrT, b: T2) -> None: ...

reveal_type(Foo(1, ""))  # type is Foo[int, str]
Foo[int](1, "")          # Invalid: Foo[int, str] cannot be assigned to self: Foo[int, int] in Foo.__init__
Foo[int]("", 1)          # Invalid: Foo[str, int] cannot be assigned to self: Foo[int, int] in Foo.__init__

When using a type parameter as the default to another type parameter, the following rules apply, where T1 is the default for T2.

Scoping Rules

T1 must be used before T2 in the parameter list of the generic.

T2 = TypeVar("T2", default=T1)

class Foo(Generic[T1, T2]): ...   # Valid
class Foo(Generic[T1]):
    class Bar(Generic[T2]): ...   # Valid

StartT = TypeVar("StartT", default="StopT")  # Swapped defaults around from previous example
StopT = TypeVar("StopT", default=int)
class slice(Generic[StartT, StopT, StepT]): ...
                  # ^^^^^^ Invalid: ordering does not allow StopT to be bound

Using a type parameter from an outer scope as a default is not supported.

Bound Rules

T1’s bound must be a subtype of T2’s bound.

T1 = TypeVar("T1", bound=int)
TypeVar("Ok", default=T1, bound=float)     # Valid
TypeVar("AlsoOk", default=T1, bound=int)   # Valid
TypeVar("Invalid", default=T1, bound=str)  # Invalid: int is not a subtype of str

Constraint Rules

The constraints of T2 must be a superset of the constraints of T1.

T1 = TypeVar("T1", bound=int)
TypeVar("Invalid", float, str, default=T1)         # Invalid: upper bound int is incompatible with constraints float or str

T1 = TypeVar("T1", int, str)
TypeVar("AlsoOk", int, str, bool, default=T1)      # Valid
TypeVar("AlsoInvalid", bool, complex, default=T1)  # Invalid: {bool, complex} is not a superset of {int, str}

Type Parameters as Parameters to Generics

Type parameters are valid as parameters to generics inside of a default when the first parameter is in scope as determined by the previous section.

T = TypeVar("T")
ListDefaultT = TypeVar("ListDefaultT", default=list[T])

class Bar(Generic[T, ListDefaultT]):
    def __init__(self, x: T, y: ListDefaultT): ...

reveal_type(Bar)                    # type is type[Bar[T, ListDefaultT = list[T]]]
reveal_type(Bar[int])               # type is type[Bar[int, list[int]]]
reveal_type(Bar[int]())             # type is Bar[int, list[int]]
reveal_type(Bar[int, list[str]]())  # type is Bar[int, list[str]]
reveal_type(Bar[int, str]())        # type is Bar[int, str]

Specialisation Rules

Type parameters currently cannot be further subscripted. This might change if Higher Kinded TypeVars are implemented.

Generic TypeAliases

Generic TypeAliases should be able to be further subscripted following normal subscription rules. If a type parameter has a default that hasn’t been overridden it should be treated like it was substituted into the TypeAlias. However, it can be specialised further down the line.

class SomethingWithNoDefaults(Generic[T, T2]): ...

MyAlias: TypeAlias = SomethingWithNoDefaults[int, DefaultStrT]  # Valid
reveal_type(MyAlias)          # type is type[SomethingWithNoDefaults[int, DefaultStrT]]
reveal_type(MyAlias[bool]())  # type is SomethingWithNoDefaults[int, bool]

MyAlias[bool, int]  # Invalid: too many arguments passed to MyAlias

Subclassing

Subclasses of Generics with type parameters that have defaults behave similarly to Generic TypeAliases. That is, subclasses can be further subscripted following normal subscription rules, non-overridden defaults should be substituted in, and type parameters with such defaults can be further specialised down the line.

class SubclassMe(Generic[T, DefaultStrT]):
    x: DefaultStrT

class Bar(SubclassMe[int, DefaultStrT]): ...
reveal_type(Bar)          # type is type[Bar[DefaultStrT = str]]
reveal_type(Bar())        # type is Bar[str]
reveal_type(Bar[bool]())  # type is Bar[bool]

class Foo(SubclassMe[float]): ...

reveal_type(Foo().x)  # type is str

Foo[str]  # Invalid: Foo cannot be further subscripted

class Baz(Generic[DefaultIntT, DefaultStrT]): ...

class Spam(Baz): ...
reveal_type(Spam())  # type is <subclass of Baz[int, str]>

Using bound and default

If both bound and default are passed default must be a subtype of bound. Otherwise the type checker should generate an error.

TypeVar("Ok", bound=float, default=int)     # Valid
TypeVar("Invalid", bound=str, default=int)  # Invalid: the bound and default are incompatible

Constraints

For constrained TypeVars, the default needs to be one of the constraints. A type checker should generate an error even if it is a subtype of one of the constraints.

TypeVar("Ok", float, str, default=float)     # Valid
TypeVar("Invalid", float, str, default=int)  # Invalid: expected one of float or str got int

Function Defaults

In generic functions, type checkers may use a type parameter’s default when the type parameter cannot be solved to anything. We leave the semantics of this usage unspecified, as ensuring the default is returned in every code path where the type parameter can go unsolved may be too hard to implement. Type checkers are free to either disallow this case or experiment with implementing support.

T = TypeVar('T', default=int)
def func(x: int | set[T]) -> T: ...
reveal_type(func(0))  # a type checker may reveal T's default of int here

Defaults following TypeVarTuple

A TypeVar that immediately follows a TypeVarTuple is not allowed to have a default, because it would be ambiguous whether a type argument should be bound to the TypeVarTuple or the defaulted TypeVar.

Ts = TypeVarTuple("Ts")
T = TypeVar("T", default=bool)

class Foo(Generic[Ts, T]): ...  # Type checker error

# Could be reasonably interpreted as either Ts = (int, str, float), T = bool
# or Ts = (int, str), T = float
Foo[int, str, float]

With the Python 3.12 built-in generic syntax, this case should raise a SyntaxError.

However, it is allowed to have a ParamSpec with a default following a TypeVarTuple with a default, as there can be no ambiguity between a type argument for the ParamSpec and one for the TypeVarTuple.

Ts = TypeVarTuple("Ts")
P = ParamSpec("P", default=[float, bool])

class Foo(Generic[Ts, P]): ...  # Valid

Foo[int, str]  # Ts = (int, str), P = [float, bool]
Foo[int, str, [bytes]]  # Ts = (int, str), P = [bytes]

Subtyping

Type parameter defaults do not affect the subtyping rules for generic classes. In particular, defaults can be ignored when considering whether a class is compatible with a generic protocol.

TypeVarTuples as Defaults

Using a TypeVarTuple as a default is not supported because:

  • Scoping Rules does not allow usage of type parameters from outer scopes.
  • Multiple TypeVarTuples cannot appear in the type parameter list for a single object, as specified in PEP 646.

These reasons leave no current valid location where a TypeVarTuple could be used as the default of another TypeVarTuple.

Binding rules

Type parameter defaults should be bound by attribute access (including call and subscript).

class Foo[T = int]:
    def meth(self) -> Self:
        return self

reveal_type(Foo.meth)  # type is (self: Foo[int]) -> Foo[int]

Implementation

At runtime, this would involve the following changes to the typing module.

  • The classes TypeVar, ParamSpec, and TypeVarTuple should expose the type passed to default. This would be available as a __default__ attribute, which would be None if no argument is passed and NoneType if default=None.

The following changes would be required to both GenericAliases:

  • logic to determine the defaults required for a subscription.
  • ideally, logic to determine if subscription (like Generic[T, DefaultT]) would be valid.

The grammar for type parameter lists would need to be updated to allow defaults; see below.

A reference implementation of the runtime changes can be found at https://github.com/Gobot1234/cpython/tree/pep-696

A reference implementation of the type checker can be found at https://github.com/Gobot1234/mypy/tree/TypeVar-defaults

Pyright currently supports this functionality.

Grammar changes

The syntax added in PEP 695 will be extended to introduce a way to specify defaults for type parameters using the “=” operator inside of the square brackets like so:

# TypeVars
class Foo[T = str]: ...

# ParamSpecs
class Baz[**P = [int, str]]: ...

# TypeVarTuples
class Qux[*Ts = *tuple[int, bool]]: ...

# TypeAliases
type Foo[T, U = str] = Bar[T, U]
type Baz[**P = [int, str]] = Spam[**P]
type Qux[*Ts = *tuple[str]] = Ham[*Ts]
type Rab[U, T = str] = Bar[T, U]

Similarly to the bound for a type parameter, defaults should be lazily evaluated, with the same scoping rules to avoid the unnecessary usage of quotes around them.

This functionality was included in the initial draft of PEP 695 but was removed due to scope creep.

The following changes would be made to the grammar:

type_param:
    | a=NAME b=[type_param_bound] d=[type_param_default]
    | a=NAME c=[type_param_constraint] d=[type_param_default]
    | '*' a=NAME d=[type_param_default]
    | '**' a=NAME d=[type_param_default]

type_param_default:
    | '=' e=expression
    | '=' e=starred_expression

The compiler would enforce that type parameters without defaults cannot follow type parameters with defaults and that TypeVars with defaults cannot immediately follow TypeVarTuples.

Rejected Alternatives

Allowing the Type Parameters Defaults to Be Passed to type.__new__’s **kwargs

T = TypeVar("T")

@dataclass
class Box(Generic[T], T=int):
    value: T | None = None

While this is much easier to read and follows a similar rationale to the TypeVar unary syntax, it would not be backwards compatible as T might already be passed to a metaclass/superclass or support classes that don’t subclass Generic at runtime.

Ideally, if PEP 637 wasn’t rejected, the following would be acceptable:

T = TypeVar("T")

@dataclass
class Box(Generic[T = int]):
    value: T | None = None

Allowing Non-defaults to Follow Defaults

YieldT = TypeVar("YieldT", default=Any)
SendT = TypeVar("SendT", default=Any)
ReturnT = TypeVar("ReturnT")

class Coroutine(Generic[YieldT, SendT, ReturnT]): ...

Coroutine[int] == Coroutine[Any, Any, int]

Allowing non-defaults to follow defaults would alleviate the issues with returning types like Coroutine from functions where the most used type argument is the last (the return). Allowing non-defaults to follow defaults is too confusing and potentially ambiguous, even if only the above two forms were valid. Changing the argument order now would also break a lot of codebases. This is also solvable in most cases using a TypeAlias.

Coro: TypeAlias = Coroutine[Any, Any, T]
Coro[int] == Coroutine[Any, Any, int]

Having default Implicitly Be bound

In an earlier version of this PEP, the default was implicitly set to bound if no value was passed for default. This while convenient, could have a type parameter with no default follow a type parameter with a default. Consider:

T = TypeVar("T", bound=int)  # default is implicitly int
U = TypeVar("U")

class Foo(Generic[T, U]):
    ...

# would expand to

T = TypeVar("T", bound=int, default=int)
U = TypeVar("U")

class Foo(Generic[T, U]):
    ...

This would have also been a breaking change for a small number of cases where the code relied on Any being the implicit default.

Allowing Type Parameters With Defaults To Be Used in Function Signatures

A previous version of this PEP allowed TypeVarLikes with defaults to be used in function signatures. This was removed for the reasons described in Function Defaults. Hopefully, this can be added in the future if a way to get the runtime value of a type parameter is added.

Allowing Type Parameters from Outer Scopes in default

This was deemed too niche a feature to be worth the added complexity. If any cases arise where this is needed, it can be added in a future PEP.

Acknowledgements

Thanks to the following people for their feedback on the PEP:

Eric Traut, Jelle Zijlstra, Joshua Butt, Danny Yamamoto, Kaylynn Morgan and Jakub Kuczys


Source: https://github.com/python/peps/blob/main/peps/pep-0696.rst

Last modified: 2024-02-08 00:44:55 GMT