Reactive Signals for Python with first-class async support, inspired by Angular's reactivity model.
Full documentation is available at https://reaktiv.readthedocs.io/.
If you've worked with modern frontend frameworks like React, Vue, or Angular, you're familiar with the power of reactive state management. The idea is simple but transformative: when data changes, everything that depends on it updates automatically. This is the magic behind dynamic UIs and real-time systems.
But why should Python miss out on the benefits of reactivity? reaktiv brings these reactive programming advantages to your Python projects:
- Automatic state propagation: Change a value once, and all dependent computations update automatically.
- Efficient updates: Only the necessary parts are recomputed.
- Async-friendly: Seamlessly integrates with Python's
asynciofor managing real-time data flows. - Zero external dependencies: Lightweight and easy to incorporate into any project.
- Type-safe: Fully annotated for clarity and maintainability.
Reactive programming isn't just a frontend paradigm. In Python, it can simplify complex backend scenarios such as:
- Real-time data streams: Stock prices, sensor readings, or live updates.
- User session management: Keep track of state changes without heavy manual subscription management.
- Complex API workflows: Automatically cascade changes across related computations.
By combining these features, reaktiv provides a robust foundation for building reactive, real-time systems - whether for data streaming, live monitoring, or even Python-powered UI frameworks.
reaktiv provides three core primitives:
- Signals: Store a value and notify dependents when it changes.
- Computed Signals: Derive values that automatically update when dependencies change.
- Effects: Run side effects when signals or computed signals change.
graph LR
A[Signal] -->|Value| B[Computed Signal]
A -->|Change| C[Effect]
B -->|Value| C
B -->|Change| C
C -->|Update| D[External System]
classDef signal fill:#4CAF50,color:white;
classDef computed fill:#2196F3,color:white;
classDef effect fill:#FF9800,color:white;
class A,B signal;
class B computed;
class C effect;
pip install reaktiv
# or with uv
uv pip install reaktivimport asyncio
from reaktiv import signal, effect
async def main():
name = signal("Alice")
async def greet():
print(f"Hello, {name()}!")
# Create and schedule effect
# IMPORTANT: Assign the effect to a variable to ensure it is not garbage collected.
greeter = effect(greet)
name.set("Bob") # Prints: "Hello, Bob!"
await asyncio.sleep(0) # Process effects
asyncio.run(main())Instead of calling set(new_value), update() lets you modify a signal based on its current value.
from reaktiv import signal
counter = signal(0)
# Standard way
counter.set(counter() + 1)
# Using update() for cleaner syntax
counter.update(lambda x: x + 1)
print(counter()) # 2from reaktiv import signal, computed
# Synchronous context example
price = signal(100)
tax_rate = signal(0.2)
total = computed(lambda: price() * (1 + tax_rate()))
print(total()) # 120.0
tax_rate.set(0.25)
print(total()) # 125.0import asyncio
from reaktiv import signal, effect
async def main():
counter = signal(0)
async def print_counter():
print(f"Counter value is: {counter()}")
# IMPORTANT: Assign the effect to a variable to prevent it from being garbage collected.
counter_effect = effect(print_counter)
for i in range(1, 4):
await asyncio.sleep(1) # Simulate an asynchronous operation or delay.
counter.set(i)
# Wait a bit to allow the last effect to process.
await asyncio.sleep(1)
asyncio.run(main())effect works with both async and sync functions. If you don't need async functionality, you can use it without asyncio:
from reaktiv import signal, effect
# Create signals
counter = signal(0)
# Define a synchronous effect function
def log_counter():
print(f"Counter value: {counter()}")
# Create and schedule the effect
counter_effect = effect(log_counter) # Prints: "Counter value: 0"
# Update the signal
counter.set(1) # Immediately prints: "Counter value: 1"
# No asyncio needed for synchronous effects!By default, when you access a signal inside a computed function or an effect, it will subscribe to that signal. However, sometimes you may want to access a signal without tracking it as a dependency.
import asyncio
from reaktiv import signal, effect, untracked
async def main():
count = signal(10)
message = signal("Hello")
async def log_message():
tracked_count = count()
untracked_msg = untracked(lambda: message()) # Not tracked as a dependency
print(f"Count: {tracked_count}, Message: {untracked_msg}")
log_effect = effect(log_message)
count.set(20) # Effect runs (count is tracked)
await asyncio.sleep(1)
message.set("New Message") # Effect does NOT run (message is untracked)
await asyncio.sleep(1)
asyncio.run(main())Sometimes, you need to clean up resources (e.g., cancel timers, close files, reset state) when an effect re-runs or is disposed.
import asyncio
from reaktiv import signal, effect
async def main():
active = signal(False)
async def monitor_status(on_cleanup):
print("Monitoring started")
active()
def cleanup():
print("Cleaning up before next run or disposal")
on_cleanup(cleanup)
monitor_effect = effect(monitor_status)
await asyncio.sleep(1)
active.set(True) # Cleanup runs before the effect runs again
await asyncio.sleep(1)
monitor_effect.dispose() # Cleanup runs before the effect is disposed
asyncio.run(main())
# Output:
# Monitoring started
# Cleaning up before next run or disposal
# Monitoring started
# Cleaning up before next run or disposalWhen creating a Signal or ComputeSignal, you can provide a custom equality function to control when updates are triggered. This is useful for comparing objects by value rather than identity.
from reaktiv import signal
# Simple example: compare numbers with tolerance
num = signal(10.0, equal=lambda a, b: abs(a - b) < 0.5)
# This won't trigger updates since 10.0 and 10.3 are within 0.5 of each other
num.set(10.3)
# This will trigger updates since 10.0 and 10.6 differ by more than 0.5
num.set(10.6)
# Custom class comparison example
class User:
def __init__(self, name, role):
self.name = name
self.role = role
# By default, different User instances are considered different even with same data
# Let's create a signal with custom equality that only cares about the role
user = signal(User("Alice", "admin"),
equal=lambda a, b: a.role == b.role)
# This won't trigger updates because the role is still "admin"
# even though it's a different User instance with a different name
user.set(User("Bob", "admin"))
# This will trigger updates because the role changed
user.set(User("Charlie", "user"))
# For deep equality with nested structures, a simple JSON-based approach works well:
import json
def json_equal(a, b):
return json.dumps(a, sort_keys=True) == json.dumps(b, sort_keys=True)
# This signal will only update when the content actually changes,
# even for complex nested structures
user_profile = signal({
'name': 'Alice',
'preferences': {'theme': 'dark', 'notifications': True}
}, equal=json_equal)
changed = effect(lambda: print("User profile updated:", user_profile.get()))
# This won't trigger updates (same content in a new object)
user_profile.set({
'name': 'Alice',
'preferences': {'theme': 'dark', 'notifications': True}
})
# This will trigger updates (content changed)
user_profile.set({
'name': 'Alice',
'preferences': {'theme': 'light', 'notifications': True}
})import asyncio
from reaktiv import signal, computed, effect
async def main():
candidate_a = signal(100)
candidate_b = signal(100)
total_votes = computed(lambda: candidate_a() + candidate_b())
percent_a = computed(lambda: (candidate_a() / total_votes()) * 100)
percent_b = computed(lambda: (candidate_b() / total_votes()) * 100)
async def display_results():
print(f"Total: {total_votes()} | A: {candidate_a()} ({percent_a():.1f}%) | B: {candidate_b()} ({percent_b():.1f}%)")
async def check_dominance():
if percent_a() > 60:
print("Alert: Candidate A is dominating!")
elif percent_b() > 60:
print("Alert: Candidate B is dominating!")
display_effect = effect(display_results)
alert_effect = effect(check_dominance)
for _ in range(3):
await asyncio.sleep(1)
candidate_a.set(candidate_a() + 40)
candidate_b.set(candidate_b() + 10)
await asyncio.sleep(1)
asyncio.run(main())Sample Output:
Total: 200 | A: 100 (50.0%) | B: 100 (50.0%)
Total: 250 | A: 140 (56.0%) | B: 110 (44.0%)
Total: 300 | A: 180 (60.0%) | B: 120 (40.0%)
Total: 350 | A: 220 (62.9%) | B: 130 (37.1%)
Alert: Candidate A is dominating!
This example demonstrates how reaktiv simplifies real-time data processing in web applications by automatically updating computed values and sending changes to clients:
from fastapi import FastAPI, WebSocket
from reaktiv import signal, computed, effect
import asyncio
import uvicorn
app = FastAPI()
measurements = signal([])
average = computed(lambda: sum(measurements()) / len(measurements()) if measurements() else 0)
above_threshold = computed(lambda: [m for m in measurements() if m > average() * 1.1])
@app.websocket("/ws")
async def websocket_endpoint(websocket: WebSocket):
await websocket.accept()
async def send_updates():
await websocket.send_json({
"average": average(),
"above_threshold": above_threshold()
})
update_client = effect(send_updates)
async def add_data():
while True:
await asyncio.sleep(1)
user_input = await websocket.receive_text()
new_data = measurements() + [float(user_input)]
measurements.set(new_data)
await add_data()
if __name__ == "__main__":
uvicorn.run("app:app", host="0.0.0.0", port=8000)This example shows how:
- Signals store measurement data that can be updated in real-time
- Computed values automatically recalculate when measurements change
- Effects automatically send updates to connected clients whenever values change
- The entire reactive system operates with minimal boilerplate code
You can find example scripts in the examples folder to help you get started with using this project.
Inspired by Angular Signals • Built for Python's async-first world • Made in Hamburg