This repository provides:
- PEP 561 typing stubs packages for the Trio project packages:
- trio (
trio-stubs
) - outcome (
outcome-stubs
) - async_generator
(
async_generator-stubs
)
- trio (
- A package
trio_typing
containing types that Trio programs often want to refer to (AsyncGenerator[Y, S]
andTaskStatus[T]
) (Nursery
is exported publicly by mainline Trio as of version 0.12.0.) - A mypy plugin that smooths over some limitations in the basic type hints.
Like Trio, we require Python 3.7 or later. Both PyPy and CPython are supported at runtime, but type checking must occur on CPython (due to limitations of mypy). We test on Linux using the latest releases from the 3.7, 3.8, 3.9, and 3.10 branches. We're not knowingly doing anything OS-specific, so other OSes should work too.
Type checkers other than mypy are not supported, but might work. Experience reports and patches to add support are welcome.
Install with:
pip install trio-typing
Then import some typing names from trio_typing
, like TaskStatus
;
see below for more details.
Install trio-typing with mypy extras:
pip install trio-typing[mypy]
Note that due to recent plugin API changes, trio-typing 0.10.0+ requires mypy 1.0+.
Enable the plugin in your mypy.ini
:
[mypy] plugins = trio_typing.plugin
Start running mypy on your Trio code!
- NOTE:
trio-typing
is the correct dependency to list in the requirements - of your own library using the
trio_typing
module. Don't usetrio-typing[mypy]
, since that would needlessly add a mypy dependency to every app having a transitive dependency on your library.
The stubs packages provide types for all public non-deprecated APIs of
trio
, outcome
, and async_generator
, as of the release date
of the corresponding trio-typing
distribution. You don't need to
explicitly configure these; just say import trio
(for example)
and mypy will know to look in trio-stubs
for the type information.
The trio_typing
package provides:
TaskStatus[T]
, the type of the object passed as thetask_status
argument to a task started withnursery.start()
. The type parameterT
is the type of the value the task provides to be returned fromnursery.start()
. This is implemented as an ABC, and the actual private types inside Trio are registered as virtual subclasses of it. So, you can't instantiatetrio_typing.TaskStatus
, butisinstance(task_status, trio_typing.TaskStatus)
wheretask_status
is a Trio task status object does return True.- (Previous versions of
trio_typing
provided an analogous ABC forNursery
, but the actual class is available astrio.Nursery
as of Trio 0.12.0; you should use that instead.) - A backport of
typing.AsyncGenerator[YieldT, SendT]
to Python 3.5. (YieldT
is the type of values yielded by the generator, andSendT
is the type of values it accepts as an argument toasend()
.) This is an abstract class describing the async generator interface:AsyncIterator
plusasend
,athrow
,aclose
, and theag_*
introspection attributes. On 3.6+,trio_typing.AsyncGenerator
is just a reexport oftyping.AsyncGenerator
. CompatAsyncGenerator[YieldT, SendT, ReturnT]
, a name for the otherwise-anonymous concrete async generator type returned by@async_generator
functions. It is a subtype ofAsyncGenerator[YieldT, SendT]
and provides the same methods. (Native async generators don't have aReturnT
; it is only relevant in determining the return type ofawait async_generator.yield_from_()
.)- A few types that are only useful with the mypy plugin:
YieldType[T]
,SendType[T]
, and the decorator@takes_callable_and_args
.
The trio_typing.plugin
mypy plugin provides:
Inference of more specific
trio.open_file()
andtrio.Path.open()
return types based on constantmode
andbuffering
arguments, soawait trio.open_file("foo", "rb", 0)
returns an unbuffered async file object in binary mode andawait trio.open_file("bar")
returns an async file object in text modeBoilerplate reduction for functions that take parameters
(fn, *args)
and ultimately invokefn(*args)
: just write:from mypy_extensions import VarArg @trio_typing.takes_callable_and_args def start_soon( async_fn: Callable[[VarArg()], Awaitable[T]], *args: Any, other_keywords: str = are_ok_too, ): # your implementation here
start_soon(async_fn, *args)
will raise an error ifasync_fn(*args)
would do so. You can also make the callable take some non-splatted arguments; the*args
get inserted at whatever position in the argument list you writeVarArg()
.The above example will always fail when the plugin is not being used. If you want to always pass in such cases, you can use a union:
@trio_typing.takes_callable_and_args def start_soon( async_fn: Union[ Callable[..., Awaitable[T]], Callable[[VarArg()], Awaitable[T]], ], *args: Any, other_keywords: str = are_ok_too, ): # your implementation here
Without the plugin, this type-checks fine (and allows inference of
T
), since any callable will match theCallable[..., Awaitable[T]]
option. With the plugin, the entire union will be replaced with specific argument types.Note: due to mypy limitations, we only support a maximum of 5 positional arguments, and keyword arguments can't be passed in this way;
nursery.start_soon(functools.partial(...))
will pass the type checker but won't be able to actually check the argument types.Mostly-full support for type checking
@async_generator
functions. You write the decorated function as if it returned a union of its actual return type, its yield type wrapped inYieldType[]
, and its send type wrapped inSendType[]
:from trio_typing import YieldType, SendType @async_generator async def sleep_and_sqrt() -> Union[None, SendType[int], YieldType[float]]: next_yield = 0.0 while True: amount = await yield_(next_yield) # amount is an int if amount < 0: return None await trio.sleep(amount) next_yield = math.sqrt(amount) # prints: CompatAsyncGenerator[float, int, None] reveal_type(sleep_and_sqrt())
Calls to
yield_
andyield_from_
inside an@async_generator
function are type-checked based on these declarations. If you leave off either the yield type or send type, the missing one is assumed to beNone
; if you leave off both (writing justasync def sleep_and_sqrt() -> None:
, like you would if you weren't using the plugin), they're both assumed to beAny
.Note the explicit
return None
; mypy won't acceptreturn
or falling off the end of the function, unless you run it with--no-warn-no-return
.
- Calls to variadic Trio functions like
trio.run()
,nursery.start_soon()
, and so on, only can type-check up to five positional arguments. (This number could be increased easily, but only at the cost of slower typechecking for everyone; mypy's current architecture requires that we generate overload sets initially for every arity we want to be able to use.) You can work around this with a# type: ignore
comment. outcome.capture()
andoutcome.acapture()
currently don't typecheck their arguments at all.
trio-typing
comes with a fairly extensive testsuite; it doesn't test all
the mechanical parts of the stubs, but does exercise most of the interesting
plugin behavior. You can run it after installing, with:
pytest -p trio_typing._tests.datadriven --pyargs trio_typing
Your choice of MIT or Apache 2.0.