streamish

Iterator and async iterator utilities for Python 3.12+.

Features

• Hybrid API: Fluent chains or standalone functions
• Unified sync/async: Same functions work with both iterators and async iterators
• Type safe: Full pyright strict mode support
• Zero dependencies: stdlib only

Installation

pip install streamish

Quick Start

import streamish as st

# Fluent API
result = list(
    st.stream([1, 2, 3, 4, 5])
    .map(lambda x: x * 2)
    .filter(lambda x: x > 4)
    .take(2)
)
# [6, 8]

# Standalone functions
result = list(st.take(2, st.filter(lambda x: x > 4, st.map(lambda x: x * 2, [1, 2, 3, 4, 5]))))
# [6, 8]

Async Support

Functions automatically detect async iterables and async functions:

import streamish as st

async def fetch(url: str) -> Response:
    ...

# Async source
async def urls():
    yield "https://example.com/1"
    yield "https://example.com/2"

# Automatically async
async for response in st.stream(urls()).map(fetch):
    print(response)

# Concurrent execution with map_async
async for response in st.map_async(fetch, urls(), concurrency=10):
    print(response)

Operations

Transform

Filter

Group

Combine

Examples

Processing a file line by line

import streamish as st

with open("data.txt") as f:
    result = list(
        st.stream(f)
        .map(str.strip)
        .filter(bool)  # skip empty lines
        .distinct()
        .take(100)
    )

Batching API requests

import streamish as st

async def send_batch(items: list[Item]) -> None:
    ...

async def process(items: AsyncIterable[Item]) -> None:
    async for batch in st.stream(items).batch(100, timeout=5.0):
        await send_batch(batch)

Concurrent HTTP requests

import httpx
import streamish as st

async def fetch(client: httpx.AsyncClient, url: str) -> Response:
    return await client.get(url)

async def main():
    urls = ["https://example.com/1", "https://example.com/2", ...]

    async with httpx.AsyncClient() as client:
        async for response in st.map_async(
            lambda url: fetch(client, url),
            urls,
            concurrency=10,
        ):
            print(response.status_code)

Windowed statistics

import streamish as st

# Moving average over last 5 values
values = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

averages = list(
    st.stream(values)
    .window(5)
    .map(lambda w: sum(w) / len(w))
)
# [3.0, 4.0, 5.0, 6.0, 7.0, 8.0]

Merging async streams

import streamish as st

async def stream_a():
    for i in range(3):
        await asyncio.sleep(0.1)
        yield f"a{i}"

async def stream_b():
    for i in range(3):
        await asyncio.sleep(0.15)
        yield f"b{i}"

# Items emitted as they arrive
async for item in st.merge(stream_a(), stream_b()):
    print(item)
# a0, b0, a1, a2, b1, b2 (order depends on timing)

License

MIT