pytau 0.14.0

A Python functional reactive framework.

A library for composing asynchronous and event-based programs in Python

About Tau

Tau is a functional reactive programming framework designed for Python from the ground up.

import asyncio

from tau.core import RealtimeNetworkScheduler
from tau.event import Lambda
from tau.signal import From


async def main():
    scheduler = RealtimeNetworkScheduler()
    signal = From(scheduler, ["world"])
    Lambda(scheduler.get_network(), signal, lambda x: print(f"Hello, {x[0].get_value()}!"))

asyncio.run(main())

Installation

Tau runs on Python 3.6 or above. To install Tau:

pip3 install pytau

Credits

Thanks to Bjorn Madsen for adding a feature to graph-theory to make Tau’s event propagation work.

Release History

0.14.0 (2020-12-31)

• Lift get_start_time() / get_end_time() to NetworkScheduler
• Add get_clock() helper method to NetworkScheduler
• Add NullEvent

0.13.0 (2020-12-28)

• Add Pipe, Alarm and RepeatingTimer signal types
• Remove websocket dependency & example
• fixed defaulting of time_millis parameter in schedule_xxx_at() methods

0.12.1 (2020-10-24)

• Bug fix: allow scheduling of a lambda as a value update in realtime as well as historical

0.12.0 (2020-10-24)

• Allow scheduling of a lambda as a value update

0.11.0 (2020-10-10)

• Add SignalGenerator API for historical mode

0.10.0 (2020-10-10)

• Add ability to schedule bare lambdas

0.9.4 (2020-10-05)

• Caught all exceptions in XQ-Thread to keep it running in live mode
• Cleaned up some unreliable unit tests

0.9.3 (2020-10-04)

• Third attempt to fix thread starvation in recursive event handlers

0.9.2 (2020-10-03)

• Another attempt to fix thread starvation in recursive event handlers

0.9.1 (2020-10-03)

• Fixed critical event stall

0.9.0 (2020-10-03)

• Switched to producer-consumer queue model in RealtimeNetworkScheduler

0.8.0 (2020-09-26)

• Ensured stable insertion-based sort for events entered with same time

0.7.3 (2020-09-19)

• Updated examples

0.7.2 (2020-09-19)

• Added unit test for BufferWithTime
• Fixed bugs in BufferWithTime

0.7.1 (2020-09-19)

• Fixed BufferWithTime to work in historic mode

0.7.0 (2020-09-19)

• Moved setting of start/end time to HistoricNetworkScheduler constructor

0.6.1 (2020-09-18)

• Added get_start_time() and get_end_time() to HistoricNetworkScheduler

0.6.0 (2020-09-18)

• Split NetworkScheduler into Realtime and Historical variants

0.5.1 (2020-05-24)

• Added MANIFEST.in to fix broken builds

0.5.0 (2020-05-12)

• Added WindowWithCount operator

0.4.3 (2020-05-02)

• Added more unit tests
• Fixed build badge target URL

0.4.2 (2020-05-02)

• Added more unit tests
• Removed declared support for Python 3.6; missing asyncio functions

0.4.1 (2020-05-02)

• Added unit tests
• Integrated Azure DevOps build pipeline

0.4.0 (2020-04-30)

• Rewrote core graph functions using graph-theory; removed networkx dependency
• Properly fixed case where next sibling node activation skipped
• Added Network#attach() method to explicitly add a node to the graph without connecting it

0.3.1 (2020-04-26)

• Fixed case where next sibling node activation skipped

0.3.0 (2020-04-25)

• Added FlatMap operator

0.2.0 (2020-04-13)

• Switched back to Python 3.7.x

0.1.1 (2020-04-04)

• Critical fix to setup.py to pick up package source
• Switch to using Do operator in hello_world.py example
• Improve the subscribe_trades.py example

0.1.0 (2020-04-04)

• Remove dependency on APScheduler
• Rewrite to use asyncio internally
• Added websocket example
• Switched to require Python version >= 3.8

0.0.2 (2020-03-28)

• Renamed OneShot to From and ForEach to Do
• Added BufferWithCount, BufferWithTime, Interval, Just and Scan operators
• Improved documentation

0.0.1 (2020-03-28)

• Initial implementation