Skip to main content

TonIO

TonIO is a multi-threaded async runtime for free-threaded Python, built in Rust on top of the mio crate, and inspired by tinyio, trio and tokio.

Warning: TonIO is currently a work in progress. The APIs are subject to breaking changes.

Note: TonIO is available on free-threaded Python only. Windows is supported on a best-effort basis.

TonIO supports both using yield and the more canonical async/await notations, with the latter being available as part of the tonio.colored module. Following code snippets show both the usages.

Warning: despite the fact TonIO supports async and await notations, it's not compatible with any asyncio object like futures and tasks. The TonIO-Monkey project provides patches for some popular asyncio packages.

In a nutshell

yield syntax

import tonio

def wait_and_add(x):
    yield tonio.sleep(1)
    return x + 1

@tonio.main
def main():
    parallel = tonio.spawn(
        wait_and_add(3), 
        wait_and_add(4)
    )
    six = yield wait_and_add(5)
    four, five = yield parallel
    return four, five, six

assert main() == (4, 5, 6)

await syntax

import tonio.colored as tonio

async def wait_and_add(x):
    await tonio.sleep(1)
    return x + 1

@tonio.main
async def main():
    parallel = tonio.spawn(
        wait_and_add(3), 
        wait_and_add(4)
    )
    six = await wait_and_add(5)
    four, five = await parallel
    return four, five, six

assert main() == (4, 5, 6)

Usage

Entrypoint

Every TonIO program consists of an entrypoint, which should be passed to the run method:

yield syntax

import tonio

def main():
    yield
    print("Hello world")

tonio.run(main())

await syntax

import tonio.colored as tonio

async def main():
    await tonio.yield_now()
    print("Hello world")

tonio.run(main())

TonIO also provides a main decorator, thus we can rewrite the previous example as:

yield syntax

import tonio

@tonio.main
def main():
    yield
    print("Hello world")

main()

await syntax

import tonio.colored as tonio

@tonio.main
async def main():
    await tonio.yield_now()
    print("Hello world")

main()

Note: as you can see the colored module provides the additional yield_now coroutine, a quick way to define a suspension point, given you cannot just yield as in the non-colored notation.

Note: both run and main can only be called once per program. To run the runtime multiple times in the same program, follow the section below.

Manually managing the runtime

TonIO also provides the runtime function, to manually manage the runtime lifecycle:

import tonio

def _run1():
    ...

async def _run2():
    ...

def main():
    runtime = tonio.runtime()
    runtime.run_until_complete(_run1())
    runtime.run_until_complete(_run2())

Runtime options

The run, main and runtime methods accept options, specifically:

option name description default
context enable contextvars usage in coroutines False
signals list of signals to listen to
threads Number of runtime threads # of CPU cores
blocking_threadpool_size Maximum number of blocking threads 128
blocking_threadpool_idle_ttl Idle timeout for blocking threads (in seconds) 30

Events

The core object in TonIO is Event. It's basically a wrapper around an atomic boolean flag, initialised with False. Event provides the following methods:

  • is_set(): return the value of the flag
  • set(): set the flag to True
  • clear(): set the flag to False
  • wait(timeout=None): returns a coroutine you can yield on that unblocks when the flag is set to True or the timeout expires. Timeout is in seconds.

yield syntax

import tonio

@tonio.main
def main():
    event = tonio.Event()

    def setter():
        yield tonio.sleep(1)
        event.set()

    tonio.spawn(setter())
    yield event.wait()

await syntax

import tonio.colored as tonio

@tonio.main
async def main():
    event = tonio.Event()

    async def setter():
        await tonio.sleep(1)
        event.set()

    tonio.spawn(setter())
    await event.wait()

Waiters

Event.wait returns a Waiter, the object the runtime actually suspends on. Waiters can be combined: w1 & w2 unblocks when both are set, w1 | w2 when either is. When the operands carry timeouts, & keeps the longest and | the shortest.

Waiters can also be built directly from events with Waiter(ev1, ev2) (all) and Waiter.any(ev1, ev2). Both accept a timeout that applies to the waiter as a whole, and expressed in microseconds.

yield syntax

import tonio

@tonio.main
def main():
    ready, stop = tonio.Event(), tonio.Event()
    yield ready.wait() | stop.wait()
    if stop.is_set():
        return

await syntax

import tonio.colored as tonio

@tonio.main
async def main():
    ready, stop = tonio.Event(), tonio.Event()
    await (ready.wait() | stop.wait())
    if stop.is_set():
        return

Result

Result is a thread-safe slot to hand values across coroutines, typically paired with an Event. store(value) writes it and fetch() reads it back. Built with a size, it holds that many slots, store(value, index) fills one and fetch() returns them as a list.

Spawning tasks

TonIO provides the spawn method to schedule new coroutines onto the runtime:

yield syntax

import tonio

def doubv(v):
    yield
    return v * 2

@tonio.main
def main():
    parallel = tonio.spawn(doubv(2), doubv(3))
    v3 = yield doubv(4)
    v1, v2 = yield parallel
    print([v1, v2, v3])

await syntax

import tonio.colored as tonio

async def doubv(v):
    await tonio.yield_now()
    return v * 2

@tonio.main
async def main():
    parallel = tonio.spawn(doubv(2), doubv(3))
    v3 = await doubv(4)
    v1, v2 = await parallel
    print([v1, v2, v3])

Coroutines passed to spawn get scheduled onto the runtime immediately. Using yield or await on the return value of spawn just waits for the coroutines to complete and retrieve the results.

When results are not needed, spawn.without_results waits for completion without collecting them. spawn.without_tracking schedules the coroutines and returns nothing at all, for fire-and-forget work.

Blocking tasks

TonIO provides the spawn_blocking method to schedule blocking operations onto the runtime:

yield syntax

import tonio

def read_file(path):
    with open(path, "r") as f:
        return f.read()

@tonio.main
def main():
    file_data = yield tonio.spawn_blocking(
        read_file, 
        "sometext.txt"
    )

await syntax

import tonio.colored as tonio

def read_file(path):
    with open(path, "r") as f:
        return f.read()

@tonio.main
async def main():
    file_data = await tonio.spawn_blocking(
        read_file, 
        "sometext.txt"
    )

Running tasks from synchronous contexts

TonIO provides the block_on method to spawn coroutines from a synchronous context. It works the same way of spawn, except it accepts a single coroutine and it blocks the current thread until the coroutine is completed.

Warning: using block_on from within a coroutine might produce a runtime deadlock.

Map utilities

TonIO provides the map and map_blocking utilities to spawn the same operation with an iterable of parameters:

yield syntax

import tonio

accum = []

def task(no):
    yield tonio.sleep(0.5)
    accum.append(no * 2)

@tonio.main
def main():
    yield tonio.map(task, range(4))

await syntax

import tonio.colored as tonio

accum = []

async def task(no):
    await tonio.sleep(0.5)
    accum.append(no * 2)

@tonio.main
async def main():
    await tonio.map(task, range(4))

Completion-based iterators

TonIO provides the as_completed utility to iterate over task results based on completion order:

yield syntax

import tonio

def _sleep(v):
    yield tonio.sleep(v)
    return v

@tonio.main
def main():
    vals = []
    for task in tonio.as_completed(
        _sleep(0.5),
        _sleep(0.1),
        _sleep(0.3),
    ):
        vals.append(yield task)

await syntax

import tonio.colored as tonio

async def _sleep(v):
    await tonio.sleep(v)
    return v

@tonio.main
async def main():
    vals = []
    async for val in tonio.as_completed(
        _sleep(0.5),
        _sleep(0.1),
        _sleep(0.3),
    ):
        vals.append(val)

Scopes and cancellations

Note: cancellations in TonIO are optimistic and terminal; optimistic means that cancelling takes effect only at suspension points, terminal means that once a coroutine is cancelled every later suspension is cancelled as well.

Warning: some TonIO primitives are not cancel-safe; they are marked as such in their sections. An operation is considered cancel-safe when cancelling it leaves the objects it operates on untouched.

TonIO provides a scope context, that lets you cancel work spawned within it:

yield syntax

import tonio

def slow_push(target, sleep):
    yield tonio.sleep(sleep)
    target.append(True)

@tonio.main
def main():
    values = []
    with tonio.scope() as scope:
        scope.spawn(slow_push(values, 0.1))
        scope.spawn(slow_push(values, 2))
        yield tonio.sleep(0.2)
        scope.cancel()
    yield scope()
    assert len(values) == 1

await syntax

import tonio.colored as tonio

async def slow_push(target, sleep):
    await tonio.sleep(sleep)
    target.append(True)

@tonio.main
async def main():
    values = []
    async with tonio.scope() as scope:
        scope.spawn(slow_push(values, 0.1))
        scope.spawn(slow_push(values, 2))
        await tonio.sleep(0.2)
        scope.cancel()
    assert len(values) == 1

When you yield on the scope, it will wait for all the spawned coroutines to end. If the scope was canceled, then all the pending coroutines will be canceled. By default, an exception in the scope context won't cancel the scope itself. If you want those exceptions to cancel the scope, you can pass cancel_on_exc=True to scope.

Note: as you can see, the colored version of scope doesn't require to be awaited, as it will yield when exiting the context.

Select first completing task

TonIO also provides a select utility to cancel remaining work on the first completing task:

yield syntax

import tonio

def slow_push(target, sleep):
    yield tonio.sleep(sleep)
    target.append(True)

@tonio.main
def main():
    values = []
    yield tonio.select(
        slow_push(values, 0.1),
        slow_push(values, 2)
    )
    assert len(values) == 1

await syntax

import tonio.colored as tonio

async def slow_push(target, sleep):
    await tonio.sleep(sleep)
    target.append(True)

@tonio.main
async def main():
    values = []
    await tonio.select(
        slow_push(values, 0.1),
        slow_push(values, 2)
    )
    assert len(values) == 1

Note: select also accepts waiters, so an Event.wait() can race against coroutines.

  • tonio.time.time(): a function returning the runtime's clock (in seconds, microsecond resolution)
  • tonio.time.sleep(delay): a coroutine you can yield on to sleep (delay is in seconds)
  • tonio.time.timeout(coro, timeout): a coroutine you can yield on returning a tuple (output, success). If the coroutine succeeds in the given time then the pair (output, True) is returned. Otherwise this will return (None, False).

Note: time.sleep is also exported to the main tonio module.

Note: all of the above functions are also present in tonio.colored.time module.

Scheduling work

TonIO provides the time.interval function to create interval objects you can yield on a scheduled basis:

yield syntax

import tonio
from tonio import time

def some_task():
    ...

def scheduler():
    interval = time.interval(1)
    while True:
        yield interval.tick()
        tonio.spawn(some_task())

@tonio.main
def main():
    tonio.spawn(scheduler())
    # do some other work

await syntax

import tonio.colored as tonio
from tonio.colored import time

async def some_task():
    ...

async def scheduler():
    interval = time.interval(1)
    while True:
        await interval.tick()
        tonio.spawn(some_task())

@tonio.main
async def main():
    tonio.spawn(scheduler())
    # do some other work

The interval method first argument is the interval in seconds resolution, and the method also accepts an optional at argument, to delay the first execution at a specific time (from the runtime's clock perspective):

from tonio import time

# tick every 500ms, with the first tick happening in 5 seconds from now
interval = time.interval(0.5, time.time() + 5)

Warning: Interval.tick is not cancel-safe.

Synchronization primitives

Synchronization primitives are exposed in the tonio.sync module.

Lock

Implements a classic mutex, or a non-reentrant, single-owner lock for coroutines:

yield syntax

import tonio
from tonio import sync

@tonio.main
def main():
    # counter can't go above 1
    counter = 0

    def _count(lock):
        nonlocal counter
        with (yield lock()):
            counter += 1
            yield
            counter -= 1
    
    lock = sync.Lock()
    yield tonio.spawn(*[
        _count(lock)
        for _ in range(10)
    ])

await syntax

import tonio.colored as tonio
from tonio.colored import sync

@tonio.main
async def main():
    # counter can't go above 1
    counter = 0

    async def _count(lock):
        nonlocal counter
        async with lock:
            counter += 1
            await tonio.yield_now()
            counter -= 1
    
    lock = sync.Lock()
    await tonio.spawn(*[
        _count(lock)
        for _ in range(10)
    ])

The Lock object also implements an or_raise method, that will immediately fail when the lock cannot be acquired:

from tonio.exceptions import WouldBlock

try:
    with lock.or_raise():
        ...
except WouldBlock:
    ...

Semaphore

A semaphore for coroutines:

yield syntax

import tonio
from tonio import sync

@tonio.main
def main():
    # counter can't go above 2
    counter = 0

    def _count(semaphore):
        nonlocal counter
        with (yield semaphore()):
            counter += 1
            yield
            counter -= 1
    
    semaphore = sync.Semaphore(2)
    yield tonio.spawn(*[
        _count(semaphore)
        for _ in range(10)
    ])

await syntax

import tonio.colored as tonio
from tonio.colored import sync

@tonio.main
async def main():
    # counter can't go above 2
    counter = 0

    async def _count(semaphore):
        nonlocal counter
        async with semaphore:
            counter += 1
            await tonio.yield_now()
            counter -= 1
    
    semaphore = sync.Semaphore(2)
    await tonio.spawn(*[
        _count(semaphore)
        for _ in range(10)
    ])

As for locks, the Semaphore object also implements an or_raise method, that will immediately fail when the lock cannot be acquired:

from tonio.exceptions import WouldBlock

try:
    with semaphore.or_raise():
        ...
except WouldBlock:
    ...

The Semaphore object also implements a tokens method, that returns the number of available tokens.

Barrier

A barrier for coroutines:

yield syntax

import tonio
from tonio import sync

@tonio.main
def main():
    barrier = sync.Barrier(3)
    count = 0

    def _start_at_3():
        nonlocal count
        count += 1
        i = yield barrier.wait()
        assert count == 3
        return i

    yield tonio.spawn(*[
        _start_at_3()
        for _ in range(3)
    ])

await syntax

import tonio.colored as tonio
from tonio.colored import sync

@tonio.main
async def main():
    barrier = sync.Barrier(3)
    count = 0

    async def _start_at_3():
        nonlocal count
        count += 1
        i = await barrier.wait()
        assert count == 3
        return i

    await tonio.spawn(*[
        _start_at_3()
        for _ in range(3)
    ])

Warning: Barrier.wait is not cancel-safe.

The Barrier object also implements a value method, which returns the current value of the barrier.

Channels

Multi-producer multi-consumer channels for inter-coroutine communication.

The tonio.sync.channel module provides both a channel and an unbounded constructors.
The main difference between bounded and unbounded channels, as the names suggest, is that while the first will suspend sending messages once the specified length is reached, and it will resume accepting messages once the existing buffer is consumed, the latter will always accept new messages. That's also why, the sender part of a bounded channel is async, while in the unbounded is not.

Bounded channel

yield syntax

import tonio
from tonio import sync
from tonio.sync import channel

def producer(sender, barrier, offset):
    for i in range(20):
        message = offset + i
        yield sender.send(message)
    yield barrier.wait()

def consumer(receiver):
    while True:
        try:
            message = yield receiver.receive()
            print(message)
        except BrokenPipeError:
            break

@tonio.main
def main():
    def close(sender, barrier):
        yield barrier.wait()
        sender.close()

    sender, receiver = channel.channel(2)
    barrier = sync.Barrier(3)
    yield tonio.spawn(*[
        producer(sender, barrier, 100),
        producer(sender, barrier, 200),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        close(sender, barrier),
    ])

await syntax

import tonio.colored as tonio
from tonio.colored import sync
from tonio.colored.sync import channel

async def producer(sender, barrier, offset):
    for i in range(20):
        message = offset + i
        await sender.send(message)
    await barrier.wait()

async def consumer(receiver):
    while True:
        try:
            message = await receiver.receive()
            print(message)
        except BrokenPipeError:
            break

@tonio.main
async def main():
    async def close(sender, barrier):
        await barrier.wait()
        sender.close()

    sender, receiver = channel.channel(2)
    barrier = sync.Barrier(3)
    await tonio.spawn(*[
        producer(sender, barrier, 100),
        producer(sender, barrier, 200),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        close(sender, barrier),
    ])

Warning: receive and send operations are not cancel-safe in bounded channels.

Unbounded channel

yield syntax

import tonio
from tonio import sync
from tonio.sync import channel

def producer(sender, barrier, offset):
    for i in range(20):
        message = offset + i
        sender.send(message)
    yield barrier.wait()

def consumer(receiver):
    while True:
        try:
            message = yield receiver.receive()
            print(message)
        except BrokenPipeError:
            break

@tonio.main
def main():
    def close(sender, barrier):
        yield barrier.wait()
        sender.close()

    sender, receiver = channel.unbounded()
    barrier = sync.Barrier(3)
    yield tonio.spawn(*[
        producer(sender, barrier, 100),
        producer(sender, barrier, 200),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        close(sender, barrier),
    ])

await syntax

import tonio.colored as tonio
from tonio.colored import sync
from tonio.colored.sync import channel

async def producer(sender, barrier, offset):
    for i in range(20):
        message = offset + i
        sender.send(message)
    await barrier.wait()

async def consumer(receiver):
    while True:
        try:
            message = await receiver.receive()
            print(message)
        except BrokenPipeError:
            break

@tonio.main
async def main():
    async def close(sender, barrier):
        await barrier.wait()
        sender.close()

    sender, receiver = channel.unbounded()
    barrier = sync.Barrier(3)
    await tonio.spawn(*[
        producer(sender, barrier, 100),
        producer(sender, barrier, 200),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        consumer(receiver),
        close(sender, barrier),
    ])

Warning: receive is not cancel-safe in unbounded channels.

Non-blocking operations

Receivers of both channel kinds offer receive_nowait, a synchronous variant that never suspends: it returns the message or one of the Empty and Closed sentinels. Bounded senders, the only suspending ones, offer send_nowait in the same way: it returns None on success or one of the Full and Closed sentinels. The sentinels are available as attributes on the objects exposing them:

sender, receiver = channel.channel(8)

if sender.send_nowait(message) is sender.Full:
    ...
if (message := receiver.receive_nowait()) is receiver.Empty:
    ...

The same objects also expose try_send and try_receive, which raise instead: WouldBlock when the channel is full or empty, BrokenPipeError when it's closed.

Markers

The mark module provides decorator shortcuts for spawn_blocking and Semaphore:

yield syntax

import tonio

# run through `spawn_blocking`
@tonio.mark.blocking
def read_file(path):
    ...

# no more than 2 running at the same time
@tonio.mark.max_concurrency(2)
def fetch(url):
    ...

# `blocking` + `max_concurrency`
@tonio.mark.cpu_bound(4)
def resize(image):
    ...

await syntax

import tonio.colored as tonio

# run through `spawn_blocking`
@tonio.mark.blocking
def read_file(path):
    ...

# no more than 2 running at the same time
@tonio.mark.max_concurrency(2)
async def fetch(url):
    ...

# `blocking` + `max_concurrency`
@tonio.mark.cpu_bound(4)
def resize(image):
    ...

Network module

Network primitives are exposed under the tonio.net module.

Streams

The high-level network primitives in TonIO are centered around the SocketStream and SocketListener objects.

The SocketListener object implements an accept coroutine which returns a SocketStream object.
The SocketStream object implements the send_all and receive_some coroutines to send and receive data, and a send_eof method to shutdown the sending side.
Both objects implement a close method to shutdown the underlying socket.

You can create and interact with the above objects using some high-level helpers in the net module, specifically:

  • open_tcp_stream: a coroutine to open a SocketStream connected to a TCP endpoint
  • open_unix_socket: a coroutine to open a SocketStream connected to a Unix socket
  • open_tcp_listeners: a coroutine to initialise SocketListener objects
  • open_unix_listener: a coroutine to initialise a SocketListener on a Unix socket path
  • serve_listeners: a coroutine to spawn SocketListener accept loops targeting a handler
  • serve_tcp: a coroutine that joins open_tcp_listeners and serve_listeners in one call
  • serve_unix: a coroutine that joins open_unix_listener and serve_listeners in one call

yield syntax

from tonio.net import open_tcp_stream, serve_tcp

def server():
    yield serve_tcp(
        server_handle, 
        host='127.0.0.1', 
        port=8000
    )

def server_handle(stream):
    # receive some data
    data = yield stream.receive_some()

def client():
    stream = yield open_tcp_stream(
        host='127.0.0.1', 
        port=8000
    )
    # send some data
    yield stream.send_all(b"message")

await syntax

from tonio.colored.net import open_tcp_stream, serve_tcp

async def server():
    await serve_tcp(
        server_handle, 
        host='127.0.0.1', 
        port=8000
    )

async def server_handle(stream):
    # receive some data
    data = await stream.receive_some()

async def client():
    stream = await open_tcp_stream(
        host='127.0.0.1', 
        port=8000
    )
    # send some data
    await stream.send_all(b"message")

Unix domain sockets use the same objects, with serve_unix and open_unix_socket:

yield syntax

from tonio.net import open_unix_socket, serve_unix

def server():
    yield serve_unix(
        server_handle, 
        '/tmp/app.sock', 
        mode=0o600
    )

def client():
    stream = yield open_unix_socket(
        '/tmp/app.sock'
    )
    yield stream.send_all(b"message")

await syntax

from tonio.colored.net import open_unix_socket, serve_unix

async def server():
    await serve_unix(
        server_handle, 
        '/tmp/app.sock', 
        mode=0o600
    )

async def client():
    stream = await open_unix_socket(
        '/tmp/app.sock'
    )
    await stream.send_all(b"message")
Readiness and non-blocking operations

SocketStream also exposes its readiness state, for code that wants to decide when to read or write rather than just block on it:

  • wait_readable(timeout=None) and wait_writable(timeout=None): coroutines that suspend until the socket is ready, returning False if the timeout (in seconds) expires first
  • receive_some_nowait(max_bytes=None): a synchronous receive, returning the NotReady sentinel (available as stream.NotReady) when no data is available
  • try_receive_some(max_bytes=None): same, but raising WouldBlock instead
  • watch_readable() and watch_writable(): context managers producing a watcher, whose waiter() method returns a Waiter (or None when ready) you can combine with others, and whose ready() method tells whether the socket is ready

yield syntax

def handler(stream, stop):
    while True:
        with stream.watch_readable() as watcher:
            if (waiter := watcher.waiter()) is not None:
                yield waiter | stop.wait()
            if stop.is_set():
                break
        data = stream.receive_some_nowait()
        if data is stream.NotReady:
            continue
        ...

await syntax

async def handler(stream, stop):
    while True:
        with stream.watch_readable() as watcher:
            if (waiter := watcher.waiter()) is not None:
                await (waiter | stop.wait())
            if stop.is_set():
                break
        data = stream.receive_some_nowait()
        if data is stream.NotReady:
            continue
        ...

TLS streams

TonIO implement TLS wrappers around the streaming APIs through primitives in the tonio.net.tls module.

TonIO provides the TLSStream and TLSListener object wrappers and the following high-level helpers:

  • open_tls_over_tcp_stream: a coroutine to open a TLSStream wrapping a TCP SocketStream
  • open_tls_over_tcp_listeners: a coroutine to initialise TLSListener objects
  • serve_tls_over_tcp: a coroutine that joins open_tls_over_tcp_listeners and serve_listeners in one call

Low-level sockets

The tonio.net.socket module provides TonIO's basic low-level networking API.
Generally, the API exposed by this module mirrors the standard library socket module.

TonIO socket objects are overall very similar to the standard library socket objects, with the main difference being that blocking methods become coroutines.

yield syntax

import tonio
from tonio.net import socket

def server():
    sock = socket.socket()
    with sock:
        yield sock.bind(('127.0.0.1', 8000))
        sock.listen()

        while True:
            client, _ = yield sock.accept()
            tonio.spawn(server_handle(client))

def server_handle(connection):
    with connection:
        # receive some data
        data = yield connection.recv(4096)

def client():
    sock = socket.socket()
    with sock:
        yield sock.connect(('127.0.0.1', 8000))
        yield sock.send(b"message")

await syntax

import tonio.colored as tonio
from tonio.colored.net import socket

async def server():
    sock = socket.socket()
    with sock:
        await sock.bind(('127.0.0.1', 8000))
        sock.listen()

        while True:
            client, _ = await sock.accept()
            tonio.spawn(server_handle(client))

async def server_handle(connection):
    with connection:
        # receive some data
        data = await connection.recv(4096)

async def client():
    sock = socket.socket()
    with sock:
        await sock.connect(('127.0.0.1', 8000))
        await sock.send(b"message")

Filesystem module

TonIO's fs module exposes async API for filesystem operations (that are run in the blocking thread-pool). It provides open, a Path class mirroring pathlib.Path, and wrap_file to adopt an already-open file object.

yield syntax

import tonio
import tonio.fs as fs

def main():
    f = yield fs.open('data.txt', 'w')
    yield f.write('hello')
    yield f.close()

    path = fs.Path('data.txt')
    if (yield path.exists()):
        print((yield path.read_text()))

    for entry in (yield fs.Path('.').iterdir()):
        print(entry.name)

await syntax

import tonio.colored as tonio
import tonio.colored.fs as fs

async def main():
    async with await fs.open('data.txt', 'w') as f:
        await f.write('hello')

    path = fs.Path('data.txt')
    if await path.exists():
        print(await path.read_text())

    for entry in await fs.Path('.').iterdir():
        print(entry.name)

Operations that touch the filesystem are asynchronous; everything else stays synchronous.

Note: methods returning several paths (iterdir, glob, rglob, walk) are resolved in a single hop and give back a list. Since walk is fully materialised, mutating its dirnames does not prune the traversal, unlike pathlib.Path.walk.

Reading a file line by line differs between the two flavours. The await syntax supports async for and async with, neither of which the yield syntax can express:

yield syntax

def read_lines(path):
    f = yield fs.open(path, 'r')
    try:
        while True:
            line = yield f.readline()
            if not line:
                break
            print(line)
    finally:
        yield f.close()

await syntax

async def read_lines(path):
    async with await fs.open(path, 'r') as f:
        async for line in f:
            print(line)

Subprocesses

TonIO exposes two coroutines to run child processes: run_process for the common "run it and collect the outcome" case, and open_process for interacting with a process while it runs. Both spawn the process on the blocking thread-pool.

run_process returns a subprocess.CompletedProcess and accepts:

  • stdin: bytes to feed to the child (defaults to b'', meaning "close stdin immediately"), or a file descriptor/subprocess constant
  • capture_stdout / capture_stderr: when true, the relevant stream is collected and available on the result
  • check: when true (the default), a non-zero exit code raises subprocess.CalledProcessError

Any other keyword argument is forwarded to subprocess.Popen.

yield syntax

import tonio

def main():
    result = yield tonio.run_process(
        ['echo', 'hello'],
        capture_stdout=True
    )
    print(result.returncode)
    print(result.stdout)

await syntax

import tonio.colored as tonio

async def main():
    result = await tonio.run_process(
        ['echo', 'hello'],
        capture_stdout=True
    )
    print(result.returncode)
    print(result.stdout)

open_process returns a Process object instead, giving access to the running child. Passing subprocess.PIPE for stdin, stdout or stderr exposes the corresponding pipe as a stream on the process object, implementing the same send_all and receive_some coroutines of network streams:

yield syntax

import subprocess
import tonio

def main():
    proc = yield tonio.open_process(
        ['cat'],
        stdin=subprocess.PIPE,
        stdout=subprocess.PIPE
    )
    yield proc.stdin.send_all(b'hello')
    proc.stdin.close()
    data = yield proc.stdout.receive_some()
    code = yield proc.wait()

await syntax

import subprocess
import tonio.colored as tonio

async def main():
    proc = await tonio.open_process(
        ['cat'],
        stdin=subprocess.PIPE,
        stdout=subprocess.PIPE
    )
    await proc.stdin.send_all(b'hello')
    proc.stdin.close()
    data = await proc.stdout.receive_some()
    code = await proc.wait()

The Process object exposes:

  • args and pid: the command and the process identifier
  • stdin, stdout, stderr: the piped streams, or None when not piped
  • stdio: a (stdin, stdout) tuple, when both are piped
  • returncode and poll(): the exit code, or None while the process is still running
  • wait: a coroutine waiting for the process to exit, returning its exit code
  • send_signal, terminate, kill: synchronous methods to signal the process

Note: unlike run_process, open_process does not reap the child for you: remember to wait on it — possibly after a kill — otherwise the child outlives your task.

Note: processes in TonIO only communicate over unbuffered byte streams: the universal_newlines, text, encoding, errors and bufsize options of subprocess are not supported.

Note: on Windows, due to the platform's lack of features, the subprocess readiness implementation falls back to the blocking thread-pool. Thus, waiting on a process or read/write operations on pipes can't be interrupted while blocked: cancellations take effect only once the OS call returns.

Driving your own I/O

The io module exposes the primitives TonIO's own sockets, pipes and processes are built on, so you can plug any file descriptor the platform poller understands into the runtime, with the same readiness model.

register puts a descriptor under the poller's watch, for both reading and writing, and returns a ScheduledIO object. The registration is edge-triggered and lasts until you close it. The descriptor itself is neither owned nor switched to non-blocking mode: that's up to you.

Readiness is consumed in user space through a small protocol: arm_r (or arm_w) returns None if the descriptor is known to be ready, otherwise a Waiter to suspend on. Once ready, you perform the actual system call. If it would block anyway, clear_r (or clear_w) drops the stale readiness, so the next arm_r suspends again.

yield syntax

import os
import tonio

def read_some(sched, fd, max_bytes=65536):
    while True:
        if (waiter := sched.arm_r()) is not None:
            yield waiter
            continue
        try:
            return os.read(fd, max_bytes)
        except BlockingIOError:
            sched.clear_r()

@tonio.main
def main():
    r, w = os.pipe()
    os.set_blocking(r, False)
    sched = tonio.io.register(r)
    os.write(w, b"hello")
    print((yield read_some(sched, r)))
    sched.close()

await syntax

import os
import tonio.colored as tonio

async def read_some(sched, fd, max_bytes=65536):
    while True:
        if (waiter := sched.arm_r()) is not None:
            await waiter
            continue
        try:
            return os.read(fd, max_bytes)
        except BlockingIOError:
            sched.clear_r()

@tonio.main
async def main():
    r, w = os.pipe()
    os.set_blocking(r, False)
    sched = tonio.io.register(r)
    os.write(w, b"hello")
    print(await read_some(sched, r))
    sched.close()

arm_r and arm_w accept a timeout in seconds. An expired waiter just resumes, so a further arm_* call tells whether the descriptor got ready in the meantime. Readiness also covers hang-ups and errors: you'll be woken up when the peer goes away, and the following system call will report it.

consume_r and consume_w drain a direction's readiness and tell whether it was set, for descriptors signalling through readiness alone.

Descriptor streams

FdStream wraps a pipe-like descriptor into the same stream interface of the network module, taking ownership of it. It provides the send_all and receive_some coroutines, fileno and close, the latter also invoked when leaving a with block. This is what the pipes of Process objects are.

yield syntax

import os
import tonio
from tonio.io import FdStream

@tonio.main
def main():
    r, w = os.pipe()
    with FdStream(r) as reader, FdStream(w) as writer:
        yield writer.send_all(b"hello")
        print((yield reader.receive_some()))

await syntax

import os
import tonio.colored as tonio
from tonio.colored.io import FdStream

@tonio.main
async def main():
    r, w = os.pipe()
    with FdStream(r) as reader, FdStream(w) as writer:
        await writer.send_all(b"hello")
        print(await reader.receive_some())

Concurrent operations on the same direction of a stream raise WouldBlock, and a broken pipe surfaces as ResourceBroken.

Note: on Windows FdStream falls back to the blocking thread-pool, with the same limitations described for subprocesses.

Signals

TonIO provides a context manager to catch signals.

The usage of such context manager requires to first configure the runtime to listen for such signals:

yield syntax

import signal
import tonio
from tonio.time import interval

def sig_handle():
    with tonio.signal_receiver(
        signal.SIGHUP, 
        signal.SIGUSR1
    ) as sigs:
        for ev in sigs:
            sig = yield ev
            if sig == signal.SIGHUP:
                ...

@tonio.main(
    signals=[signal.SIGHUP, signal.SIGUSR1]
)
def main():
    tonio.spawn(sig_handle())
    ticker = interval(1)
    while True:
        yield ticker.tick()

await syntax

import signal
import tonio.colored as tonio
from tonio.colored.time import interval

async def sig_handle():
    with tonio.signal_receiver(
        signal.SIGHUP, 
        signal.SIGUSR1
    ) as sigs:
        async for sig in sigs:
            if sig == signal.SIGHUP:
                ...

@tonio.main(
    signals=[signal.SIGHUP, signal.SIGUSR1]
)
async def main():
    tonio.spawn(sig_handle())
    ticker = interval(1)
    while True:
        await ticker.tick()

Exceptions

The tonio.exceptions module exposes:

  • CancelledError: raised inside a coroutine when it gets cancelled
  • TimeoutError: raised when a timed operation expires
  • WouldBlock: raised by the or_raise and try_* variants when the operation cannot complete immediately
  • ResourceBroken: raised by streams when the underlying transport is unusable

Note: CancelledError and TimeoutError derive from BaseException, so they pass through except Exception clauses.

Testing

TonIO ships with a pytest plugin, which runs tests marked with the tonio marker on the runtime:

yield syntax

import pytest
import tonio

@pytest.mark.tonio
def test_sleep():
    yield tonio.sleep(0.1)

await syntax

import pytest
import tonio.colored as tonio

@pytest.mark.tonio
async def test_sleep():
    await tonio.sleep(0.1)

The marker can also be applied at module level with pytestmark = pytest.mark.tonio, or at class level.

To avoid marking tests entirely, the plugin also provides an auto mode, in which every async test gets run on the TonIO runtime. Auto mode can be enabled with the tonio_mode option in the pytest configuration:

[tool.pytest.ini_options]
tonio_mode = 'auto'

Async fixtures

For asynchronous fixtures where setup and teardown are required, the two syntaxes differ.

await syntax

Async fixtures involved in TonIO tests run on the runtime. Coroutine fixtures simply return their value, while async generator fixtures can yield it and run teardown code after the yield:

import pytest
from tonio.colored.net import open_tcp_stream

@pytest.fixture
async def connection():
    stream = await open_tcp_stream(host='127.0.0.1', port=8000)
    yield stream
    stream.close()
yield syntax

Given non-colored TonIO coroutines are indistinguishable from standard pytest generator fixtures, the plugin never runs generator fixtures on the runtime. This is usually not a limitation: since yield syntax tests already run as coroutines, simple setup can just happen within the test itself:

import pytest
from tonio.net import open_tcp_stream

def _connect():
    stream = yield open_tcp_stream(host='127.0.0.1', port=8000)
    return stream

@pytest.mark.tonio
def test_conn():
    stream = yield _connect()
    yield stream.send_all(b'ping')

When an actual teardown is required, the plugin provides the tonio_run fixture, which runs the given coroutine on the runtime:

@pytest.fixture
def connection(tonio_run):
    stream = tonio_run(_connect())

    def _disconnect():
        yield stream.send_all(b'bye')
        stream.close()

    yield stream
    tonio_run(_disconnect())

Runtime configuration in pytest

The TonIO runtime is always initialised once per test session, with context enabled and 2 threads. Runtime options can be customized overriding the session-scoped tonio_runtime_options fixture, for example in conftest.py:

import pytest

@pytest.fixture(scope='session')
def tonio_runtime_options():
    return {'threads': 4}

The runtime object itself is available to tests and fixtures via the session-scoped tonio_runtime fixture.

Libraries built on TonIO

In addition to the patches provided by the TonIO-Monkey project, the following libraries target TonIO natively:

  • httpunk: a low-level async HTTP library
  • punkreq: a high-level async HTTP client
  • punkasgi: an ASGI server built on TonIO

License

TonIO is released under the BSD License.

Metadata

Release files for tonio 0.10.4

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for tonio 0.10.4
File Size Uploaded
tonio-0.10.4.tar.gz 107.2 kB Details

Built distributions (wheels)

Table of built distributions (wheels) for tonio 0.10.4
File
tonio-0.10.4-cp315-cp315t-musllinux_1_1_x86_64.whl CPython 3.15 CPython 3.15 free-threading Linux musl 1.1+ x86-64 Details
tonio-0.10.4-cp315-cp315t-musllinux_1_1_armv7l.whl CPython 3.15 CPython 3.15 free-threading Linux musl 1.1+ ARMv7l Details
tonio-0.10.4-cp315-cp315t-musllinux_1_1_aarch64.whl CPython 3.15 CPython 3.15 free-threading Linux musl 1.1+ ARM64 Details
tonio-0.10.4-cp315-cp315t-manylinux_2_28_armv7l.whl CPython 3.15 CPython 3.15 free-threading Linux glibc 2.28+ ARMv7l Details
tonio-0.10.4-cp315-cp315t-manylinux_2_28_aarch64.whl CPython 3.15 CPython 3.15 free-threading Linux glibc 2.28+ ARM64 Details
tonio-0.10.4-cp315-cp315t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl CPython 3.15 CPython 3.15 free-threading Linux glibc 2.17+ x86-64 Details
tonio-0.10.4-cp315-cp315t-manylinux_2_12_i686.manylinux2010_i686.whl CPython 3.15 CPython 3.15 free-threading Linux glibc 2.12+ x86-32 Details
tonio-0.10.4-cp315-cp315t-macosx_11_0_arm64.whl CPython 3.15 CPython 3.15 free-threading macOS 11.0+ ARM64 Details
tonio-0.10.4-cp315-cp315t-macosx_10_12_x86_64.whl CPython 3.15 CPython 3.15 free-threading macOS 10.12+ x86-64 Details
tonio-0.10.4-cp314-cp314t-win_amd64.whl CPython 3.14 CPython 3.14 free-threading Windows x86-64 Details
tonio-0.10.4-cp314-cp314t-musllinux_1_1_x86_64.whl CPython 3.14 CPython 3.14 free-threading Linux musl 1.1+ x86-64 Details
tonio-0.10.4-cp314-cp314t-musllinux_1_1_armv7l.whl CPython 3.14 CPython 3.14 free-threading Linux musl 1.1+ ARMv7l Details
tonio-0.10.4-cp314-cp314t-musllinux_1_1_aarch64.whl CPython 3.14 CPython 3.14 free-threading Linux musl 1.1+ ARM64 Details
tonio-0.10.4-cp314-cp314t-manylinux_2_28_armv7l.whl CPython 3.14 CPython 3.14 free-threading Linux glibc 2.28+ ARMv7l Details
tonio-0.10.4-cp314-cp314t-manylinux_2_28_aarch64.whl CPython 3.14 CPython 3.14 free-threading Linux glibc 2.28+ ARM64 Details
tonio-0.10.4-cp314-cp314t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl CPython 3.14 CPython 3.14 free-threading Linux glibc 2.17+ x86-64 Details
tonio-0.10.4-cp314-cp314t-manylinux_2_12_i686.manylinux2010_i686.whl CPython 3.14 CPython 3.14 free-threading Linux glibc 2.12+ x86-32 Details
tonio-0.10.4-cp314-cp314t-macosx_11_0_arm64.whl CPython 3.14 CPython 3.14 free-threading macOS 11.0+ ARM64 Details
tonio-0.10.4-cp314-cp314t-macosx_10_12_x86_64.whl CPython 3.14 CPython 3.14 free-threading macOS 10.12+ x86-64 Details

Total release size: 11.2 MB

Release files / tonio-0.10.4.tar.gz

Download URL tonio-0.10.4.tar.gz
Size 107.2 kB
Tags Source
SHA-256 checksum
How to use checksums
9e14f3dc0d4a0ac943a2453b34a3f10e40343231cce1d440f2744182c9435022
BLAKE2b-256 checksum
How to use checksums
e8b4ef1ffbbab01d2051a557e87f81aac14648eed588f09cf48dcae66b102862
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-musllinux_1_1_x86_64.whl

Download URL tonio-0.10.4-cp315-cp315t-musllinux_1_1_x86_64.whl
Size 711.5 kB
Tags CPython 3.15 CPython 3.15 free-threading Linux musl 1.1+ x86-64
SHA-256 checksum
How to use checksums
ee7faecc732d4266bc1499a08385127168326c38fe8317120d4d246510946e3c
BLAKE2b-256 checksum
How to use checksums
48306ecb7256fd637913d18cc0886a795f9308d30100ea61cf46a821f62ff2da
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-musllinux_1_1_armv7l.whl

Download URL tonio-0.10.4-cp315-cp315t-musllinux_1_1_armv7l.whl
Size 763.7 kB
Tags CPython 3.15 CPython 3.15 free-threading Linux musl 1.1+ ARMv7l
SHA-256 checksum
How to use checksums
fc01214ef4ec09880b7bed07eae54eec75cc1988f1946cfc37b25a7124174105
BLAKE2b-256 checksum
How to use checksums
219bec9cd8a8fcc9a17558edc2f2a7fe8cf32ffb0a945cfbf77704c632002692
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-musllinux_1_1_aarch64.whl

Download URL tonio-0.10.4-cp315-cp315t-musllinux_1_1_aarch64.whl
Size 664.2 kB
Tags CPython 3.15 CPython 3.15 free-threading Linux musl 1.1+ ARM64
SHA-256 checksum
How to use checksums
4e4a4412ce289406892150265d68513b6ccf786fac9f8081c4f5faa1a3c693a2
BLAKE2b-256 checksum
How to use checksums
4176ef2f83170fd5a1c32ecb4015f9fd7afd32eb6b14270f487fae2903a592e4
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-manylinux_2_28_armv7l.whl

Download URL tonio-0.10.4-cp315-cp315t-manylinux_2_28_armv7l.whl
Size 481.1 kB
Tags CPython 3.15 CPython 3.15 free-threading Linux glibc 2.28+ ARMv7l
SHA-256 checksum
How to use checksums
6ecccb4d5afc74c141987f1a260a2f4136b13355067fe78f71ee5c2a4732a86a
BLAKE2b-256 checksum
How to use checksums
c7111126995ed565b5433e4fe49592d0d13b6bdf3d25c1636f81b103a858e36a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-manylinux_2_28_aarch64.whl

Download URL tonio-0.10.4-cp315-cp315t-manylinux_2_28_aarch64.whl
Size 476.8 kB
Tags CPython 3.15 CPython 3.15 free-threading Linux glibc 2.28+ ARM64
SHA-256 checksum
How to use checksums
f73bacd776d325de384d0f8e23f32316dffdbf19b5f9925e67093bfe211c8e97
BLAKE2b-256 checksum
How to use checksums
4312efdfc47139f9e03a48535ae345fa464f241f9cc573e1ba822713c1fbee58
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

Download URL tonio-0.10.4-cp315-cp315t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Size 606.6 kB
Tags CPython 3.15 CPython 3.15 free-threading Linux glibc 2.17+ x86-64
SHA-256 checksum
How to use checksums
8679afc7fc4b6551dc6919ec1d371deb253f38fc3141ce86168a5122a1aed24f
BLAKE2b-256 checksum
How to use checksums
d5173ab29851f8b195b106cf6b15f895ef5384987920011679536df9be0b8fc2
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-manylinux_2_12_i686.manylinux2010_i686.whl

Download URL tonio-0.10.4-cp315-cp315t-manylinux_2_12_i686.manylinux2010_i686.whl
Size 528.1 kB
Tags CPython 3.15 CPython 3.15 free-threading Linux glibc 2.12+ x86-32
SHA-256 checksum
How to use checksums
93453bf1aef82fb9d6d8f6d8f75f0009a8236ca2e1b8e99efa888b837bc4108d
BLAKE2b-256 checksum
How to use checksums
87b777013afb88daf2a42e6a5bab1acfca302cdf571f2e1db749736c9a4ce4eb
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-macosx_11_0_arm64.whl

Download URL tonio-0.10.4-cp315-cp315t-macosx_11_0_arm64.whl
Size 549.2 kB
Tags CPython 3.15 CPython 3.15 free-threading macOS 11.0+ ARM64
SHA-256 checksum
How to use checksums
8c53d8942fbaa187e2248ed28bb326570666a1e032efdc43bfc4b4a87aa18038
BLAKE2b-256 checksum
How to use checksums
d7f96a21e28d76c5aac310a18c5f457fb3930c1752c631adad5c28d056da5902
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp315-cp315t-macosx_10_12_x86_64.whl

Download URL tonio-0.10.4-cp315-cp315t-macosx_10_12_x86_64.whl
Size 588.3 kB
Tags CPython 3.15 CPython 3.15 free-threading macOS 10.12+ x86-64
SHA-256 checksum
How to use checksums
007e33e9f65b14cb72fc58dbc0e8af1349795f9ea2c4b3b8b374c374b305050e
BLAKE2b-256 checksum
How to use checksums
8bd2dfb41b2b6b32e8e9eb8efd9bab3b9cdcc0c8328ccee3f780e36a0f394a28
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-win_amd64.whl

Download URL tonio-0.10.4-cp314-cp314t-win_amd64.whl
Size 395.1 kB
Tags CPython 3.14 CPython 3.14 free-threading Windows x86-64
SHA-256 checksum
How to use checksums
48f068827a64fbbc2018f54adb38b52e0a9e8121ee6a8a62379aa5726282e3f7
BLAKE2b-256 checksum
How to use checksums
c17916a4518dc42b25693da7a4ef418690173567362fed7f6c777034b36158cc
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-musllinux_1_1_x86_64.whl

Download URL tonio-0.10.4-cp314-cp314t-musllinux_1_1_x86_64.whl
Size 711.7 kB
Tags CPython 3.14 CPython 3.14 free-threading Linux musl 1.1+ x86-64
SHA-256 checksum
How to use checksums
84459522f945c49d5ccba1cdc16f062532cea7401ff3d1842ed771768e9f6258
BLAKE2b-256 checksum
How to use checksums
5777a36bcfa743eefe086979d3dfb94d4360c01ebe0b9055be3dddfe2a5cbf4a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-musllinux_1_1_armv7l.whl

Download URL tonio-0.10.4-cp314-cp314t-musllinux_1_1_armv7l.whl
Size 763.6 kB
Tags CPython 3.14 CPython 3.14 free-threading Linux musl 1.1+ ARMv7l
SHA-256 checksum
How to use checksums
6db929138fc2cf86738e3690b0d1f3f43dd24c3c779d50830c88a57ab9ff8dc4
BLAKE2b-256 checksum
How to use checksums
b09ff4f31831fa00ca0720039f82daf58e040bfe89336b83e0604122e0fa0db8
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-musllinux_1_1_aarch64.whl

Download URL tonio-0.10.4-cp314-cp314t-musllinux_1_1_aarch64.whl
Size 664.3 kB
Tags CPython 3.14 CPython 3.14 free-threading Linux musl 1.1+ ARM64
SHA-256 checksum
How to use checksums
1a610080d807c2e39f292c3a24cde8ed54305ba5de0984569aa52c375b56d686
BLAKE2b-256 checksum
How to use checksums
22c477d21e84ab1632e161b85c27b6bc8a70e4fec0bd2997cb14b48b89f87f9d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-manylinux_2_28_armv7l.whl

Download URL tonio-0.10.4-cp314-cp314t-manylinux_2_28_armv7l.whl
Size 481.1 kB
Tags CPython 3.14 CPython 3.14 free-threading Linux glibc 2.28+ ARMv7l
SHA-256 checksum
How to use checksums
71ef0a6b9e0797c2e469fedbf397fc474e29aa5a941c28191fb77db84b0414b3
BLAKE2b-256 checksum
How to use checksums
41d96be4a3b6deabb537967479aa99c95284b33c3da3c6e8ae01103e11a9d42d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-manylinux_2_28_aarch64.whl

Download URL tonio-0.10.4-cp314-cp314t-manylinux_2_28_aarch64.whl
Size 476.7 kB
Tags CPython 3.14 CPython 3.14 free-threading Linux glibc 2.28+ ARM64
SHA-256 checksum
How to use checksums
9670e348a5806cd9114b8b2b8c4b0105538368e52ee49dd340bc228a9e88224e
BLAKE2b-256 checksum
How to use checksums
633eeb4122ed49ec7bd62b6bee6fc5338c60413fe3b86e34afbafb8e4b7ce5a3
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

Download URL tonio-0.10.4-cp314-cp314t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Size 608.6 kB
Tags CPython 3.14 CPython 3.14 free-threading Linux glibc 2.17+ x86-64
SHA-256 checksum
How to use checksums
b2d4cf38f6b4094fc99c6de909bc6f706d299985cbd43bd54488a3741b0ac4b6
BLAKE2b-256 checksum
How to use checksums
2e127f4a615ae04d53a9cb5182dda52d202b32c6180a9197dd780c5627f9031c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-manylinux_2_12_i686.manylinux2010_i686.whl

Download URL tonio-0.10.4-cp314-cp314t-manylinux_2_12_i686.manylinux2010_i686.whl
Size 527.9 kB
Tags CPython 3.14 CPython 3.14 free-threading Linux glibc 2.12+ x86-32
SHA-256 checksum
How to use checksums
669e00cc12e8ad9948be08d734c0afa6ee6a9d2f98119e7286573deb9daa5668
BLAKE2b-256 checksum
How to use checksums
90d5c087ecde8ed05768679a9a072ff4bc0a75964e6ffbb077f38ef0fd28d0d3
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-macosx_11_0_arm64.whl

Download URL tonio-0.10.4-cp314-cp314t-macosx_11_0_arm64.whl
Size 549.7 kB
Tags CPython 3.14 CPython 3.14 free-threading macOS 11.0+ ARM64
SHA-256 checksum
How to use checksums
c7926257125e2fb91a18251886cd928b332ae16e795db09dee041914860fbe3a
BLAKE2b-256 checksum
How to use checksums
2f9f7eccfa6d49eaaa45deb7b48cc881ce726ad24d05b9b6d913890a575337df
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release files / tonio-0.10.4-cp314-cp314t-macosx_10_12_x86_64.whl

Download URL tonio-0.10.4-cp314-cp314t-macosx_10_12_x86_64.whl
Size 588.1 kB
Tags CPython 3.14 CPython 3.14 free-threading macOS 10.12+ x86-64
SHA-256 checksum
How to use checksums
4a75277221de3c9443d8eea9022d4678c303fcbde8b6b7c75798351581f33a54
BLAKE2b-256 checksum
How to use checksums
391471afcad6fe0aa3c57749c7ecf677d31f63512d0036b0dd8a81f754a6f2ac
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 2, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.10.4 This release

20 release files

0.9.9

19 release files

0.9.8

19 release files

0.9.7

19 release files

0.9.2

19 release files

0.9.1

19 release files

0.9.0

19 release files

0.8.3

19 release files

0.8.2

19 release files

0.8.1

19 release files

0.8.0

19 release files

0.7.2

19 release files

0.7.0

19 release files

0.3.0

19 release files

0.2.1

19 release files

0.1.0

19 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page