Shortcuts

osiiso.amap(), osiiso.tmap(), osiiso.pmap() — one-shot mapping helpers. Each builds a queue, maps a callable over an iterable, runs it, and returns the values in input order.

from osiiso import amap, tmap, pmap

Because every helper raises ExecutionError if any task failed or was cancelled, a returned tuple always lines up 1:1 with the input iterable.

---

Signatures

await amap(fn, iterable, *, workers=None, rate=None, timeout=None, **options) -> tuple
tmap(fn, iterable, *, workers=None, rate=None, timeout=None, **options) -> tuple
pmap(fn, iterable, *, workers=None, rate=None, timeout=None,
     context=None, initializer=None, initargs=(), **options) -> tuple

Helper	Queue	Notes
amap	AsyncQueue	Awaitable; call from async code
tmap	ThreadQueue	Blocks the calling thread
pmap	ProcessQueue	Blocks; standard pickling/spawn rules apply

Parameter	Description
fn	The callable to run once per element
iterable	Elements interpreted like map(): tuple → positional args, mapping → keyword args, else a single arg
workers	Forwarded to the queue (None = auto-scale)
rate	Max task attempts per second (None = unlimited)
timeout	Overall run limit in seconds
context / initializer / initargs	pmap only — forwarded to ProcessQueue
**options	Any TaskOptions fields (retries=3, priority=0, ...)

Raises: ExecutionError if any task failed or was cancelled.

---

Examples

import osiiso

# Async — fan out HTTP fetches
pages = await osiiso.amap(fetch, urls, workers=8, retries=2)

# Threads — blocking I/O
sizes = osiiso.tmap(stat_file, paths, workers=8)

# Processes — CPU-bound work
scores = osiiso.pmap(rank, datasets, workers=4, timeout=60)

When you need handles, groups, priorities across mixed workloads, or a reusable queue, use the queue classes directly — the shortcuts are for the "just run this over these inputs" case.