psygnal 0.15.0

Fast python callback/event system modeled after Qt Signals

psygnal

Psygnal (pronounced "signal") is a pure python implementation of the observer pattern, with the API of Qt-style Signals with (optional) signature and type checking, and support for threading. It has no dependencies.

This library does not require or use Qt in any way, It simply implements a similar observer pattern API.

Documentation

https://psygnal.readthedocs.io/

Install

pip install psygnal

conda install -c conda-forge psygnal

Usage

The observer pattern is a software design pattern in which an object maintains a list of its dependents ("observers"), and notifies them of any state changes – usually by calling a callback function provided by the observer.

Here is a simple example of using psygnal:

from psygnal import Signal

class MyObject:
    # define one or more signals as class attributes
    value_changed = Signal(str)

# create an instance
my_obj = MyObject()

# You (or others) can connect callbacks to your signals
@my_obj.value_changed.connect
def on_change(new_value: str):
    print(f"The value changed to {new_value}!")

# The object may now emit signals when appropriate,
# (for example in a setter method)
my_obj.value_changed.emit('hi')  # prints "The value changed to hi!"

Much more detail available in the documentation!

Evented Dataclasses

A particularly nice usage of the signal pattern is to emit signals whenever a field of a dataclass changes. Psygnal provides an @evented decorator that will emit a signal whenever a field changes. It is compatible with dataclasses from the standard library, as well as attrs, and pydantic:

from psygnal import evented
from dataclasses import dataclass

@evented
@dataclass
class Person:
    name: str
    age: int = 0

person = Person('John', age=30)

# connect callbacks
@person.events.age.connect
def _on_age_change(new_age: str):
    print(f"Age changed to {new_age}")

person.age = 31  # prints: Age changed to 31

See the dataclass documentation for more details.

Evented Containers

psygnal.containers provides evented versions of mutable data structures (dict, list, set), for cases when you need to monitor mutation:

from psygnal.containers import EventedList

my_list = EventedList([1, 2, 3, 4, 5])

my_list.events.inserted.connect(lambda i, val: print(f"Inserted {val} at index {i}"))
my_list.events.removed.connect(lambda i, val: print(f"Removed {val} at index {i}"))

my_list.append(6)  # Output: Inserted 6 at index 5
my_list.pop()  # Output: Removed 6 at index 5

See the evented containers documentation for more details.

Benchmark history

https://pyapp-kit.github.io/psygnal/

and

https://codspeed.io/pyapp-kit/psygnal

Developers

Setup

This project uses PEP 735 dependency groups.

After cloning, setup your env with uv sync or pip install -e . --group dev

Compiling

While psygnal is a pure python package, it is compiled with mypyc to increase performance. To test the compiled version locally, you can run:

HATCH_BUILD_HOOKS_ENABLE=1 uv sync --force-reinstall

(which is also available as make build if you have make installed)

Debugging

To disable all compiled files and run the pure python version, you may run:

python -c "import psygnal.utils; psygnal.utils.decompile()"

To return the compiled version, run:

python -c "import psygnal.utils; psygnal.utils.recompile()"

The psygnal._compiled variable will tell you if you're using the compiled version or not.