# asyncio-pool-ng
[](https://pypi.org/project/asyncio-pool-ng/)
[](https://pypi.org/project/asyncio-pool-ng/)
[](https://opensource.org/licenses/MIT)
[](https://github.com/smithk86/asyncio-pool-ng/actions?query=workflow%3Aci)
[](https://github.com/psf/black)
## About
**AsyncioPoolNG** takes the ideas used in [asyncio-pool](https://github.com/gistart/asyncio-pool) and wraps them around a [TaskGroup](https://anyio.readthedocs.io/en/stable/tasks.html) from [anyio](https://anyio.readthedocs.io/en/stable/index.html).
`AsyncioPool` has three main functions `spawn`, `map`, and `itermap`.
1. `spawn`: Schedule an async function on the pool and get a future back which will eventually have either the result or the exception from the function.
2. `map`: Spawn an async function for each item in an iterable object, and return a set containing a future for each item.
- `asyncio.wait()` can be used to wait for the set of futures to complete.
- When the `AsyncioPool` closes, it will wait for all tasks to complete. All pending futures will be complete once it is closed.
3. `itermap`: Works similarly to `map` but returns an [Async Generator](https://docs.python.org/3/library/typing.html#typing.AsyncGenerator "Async Generator") which yields each future as it completes.
## Differences from asyncio-pool
1. `asyncio-pool-ng` implements [Python typing](https://typing.readthedocs.io/en/latest/) and passes validation checks with [mypy](http://mypy-lang.org/)'s strict mode. This helps IDEs and static type checkers know what type of result to expect when getting data from a completed future.
2. `asyncio-pool` uses callbacks to process data before returning it; `asyncio-pool-ng` only returns [Future](https://docs.python.org/3.10/library/asyncio-future.html#asyncio.Future) instances directly. The future will contain either a result or an exception which can then be handled as needed.
3. While `asyncio-pool` schedules [Coroutine](https://docs.python.org/3/library/typing.html#typing.Coroutine) instances directly, `asyncio-pool-ng` takes the callable and arguments, and creates the Coroutine instance at execution time.
## Example
```python title="example.py"
import asyncio
import logging
from random import random
from asyncio_pool import AsyncioPool
logging.basicConfig(level=logging.INFO)
async def worker(number: int) -> int:
await asyncio.sleep(random() / 2)
return number * 2
async def main() -> None:
result: int = 0
results: list[int] = []
async with AsyncioPool(2) as pool:
"""spawn task and wait for the results"""
result = await pool.spawn(worker, 5)
assert result == 10
logging.info(f"results for pool.spawn(worker, 5): {result}")
"""spawn task and get results later"""
future: asyncio.Future[int] = pool.spawn(worker, 5)
# do other stuff
result = await future
assert result == 10
"""map an async function to a set of values"""
futures: set[asyncio.Future[int]] = pool.map(worker, range(10))
await asyncio.wait(futures)
results = [x.result() for x in futures]
logging.info(f"results for pool.map(worker, range(10)): {results}")
results.sort()
assert results == [0, 2, 4, 6, 8, 10, 12, 14, 16, 18]
"""iterate futures as they complete"""
logging.info("results for pool.itermap(worker, range(10)):")
results = []
async for future in pool.itermap(worker, range(10)):
results.append(future.result())
logging.info(f"> {future.result()}")
results.sort()
assert results == [0, 2, 4, 6, 8, 10, 12, 14, 16, 18]
asyncio.run(main())
```