RxIter

RxIter brings observables to python in a bare bones way by using async generators and the async await syntax. In this paradigm observables are analogous to async iterables, and observers analogous to async iterators.

It implements 2 fundamental observable operations, which may be familar to those who know rxpy.

• share
• repeat

Operations

Share

share allows multiple "observers" to subscribe the same observable

import asyncio
from rxiter import share

async def main():

    @share
    async def count():   # a counting "observable"
        v = 0
        while True:
            print(f"returning value {v}")
            yield v
            await asyncio.sleep(1)
            v += 1

    async def count_squared(obs):  # a counting "observer"
        async for v in obs: 
            print(f"{v} squared is {v**2}")

    square_task_subscription = asyncio.Task(count_squared(count()))  # subscribe

    async def count_cubed(obs):  # another counting "observer
        async for v in obs:
            print(f"{v} cubed is {v**3}")

    cube_task_subscription = asyncio.Task(count_cubed(count())) # subscribe

    await asyncio.gather(square_task_subscription, cube_task_subscription)

asyncio.run(main())

The output on this code would be:

returning value 0
0 squared is 0
0 cubed is 0
returning value 1
1 squared is 1
1 cubed is 1
returning value 2
2 squared is 4
2 cubed is 8
etc...

Repeat

repeat takes a iterator, and "records" it's outputed values so that it is turned into an iterable, and can be "listened" back multiple times.

Example

Polling an API

Suppose we have a API endpoint that we would like to poll to get the most up to date weather in Toronto. We could set up an observable as follows:

async get_toronto_weather():
  while True:
    yield await poll_my_api("api_enpoint")
    await asyncio.sleep(60 * 30)  # wait 30 minutes

If you want to "pipe" this to do further operations, like extract some specific content from the dict returned by get_toronto_weather()

async get_temperature():
  async for v in poll_api():
    yield v["temperature"]

Now if we want to have multiple listeners, that is where the share comes into the picture. We can do

@share
async get_toronto_weather():
  while True:
    yield await poll_my_api("api_enpoint")
    await asyncio.sleep(60 * 30)  # wait 30 minutes

async get_temperature():
  async for v in get_toronto_weather():
    yield v["temperature"]

async get_humidity():
  async for v in get_toronto_weather():
    yield v["humidity"]

asyncio.Task(get_temperature())
asyncio.Task(get_humidity())

and get_toronto_weather() will only run once for both get_temperature() and get_humidity()

Realtime stdout on python subprocess

Installation

pip install rxiter