wim/parfor
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

- add allow_output option

Browse Source 
- update readme
This commit is contained in:
Wim Pomp
2026-01-08 10:42:50 +01:00
parent 6341561380
commit 43a0cf68b5
3 changed files with 47 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
README.md
View File

@@ -4,36 +4,18 @@
Used to parallelize for-loops using parfor in Matlab? This package allows you to do the same in python.
Take any normal serial but parallelizable for-loop and execute it in parallel using easy syntax.
Don't worry about the technical details of using the multiprocessing module, race conditions, queues,
parfor handles all that.
parfor handles all that. Now powered by [ray](https://pypi.org/project/ray/).
Tested on linux, Windows and OSX with python 3.10 and 3.12.
## Why is parfor better than just using multiprocessing?
- Easy to use
- Using dill instead of pickle: a lot more objects can be used when parallelizing
- Progress bars are built-in
- Automatically use multithreading instead of multiprocessing when the GIL is disabled
- Retry the task in the main process upon failure for easy debugging
## How it works
This depends on whether the GIL is currently disabled or not. Disabling the GIL in Python is currently an experimental
feature in Python3.13, and not the standard.
### Python with GIL enabled
The work you want parfor to do is divided over a number of processes. These processes are started by parfor and put
together in a pool. This pool is reused when you want parfor to do more work, or shut down when no new work arrives
within 10 minutes.
A handle to each bit of work is put in a queue from which the workers take work. The objects needed to do the work are
stored in a memory manager in serialized form (using dill) and the manager hands out an object to a worker when the
worker is requesting it. The manager deletes objects automatically when they're not needed anymore.
When the work is done the result is sent back for collection in the main process.
### Python with GIL disabled
The work you want parfor to do is given to a new thread. These threads are started by parfor and put together in a pool.
The threads and pool are not reused and closed automatically when done.
When the work is done a message is sent to the main thread to update the status of the pool.
[Ray](https://pypi.org/project/ray/) does all the heavy lifting. Parfor now is just a wrapper around ray, adding
some ergonomics.
## Installation
`pip install parfor`
@@ -43,13 +25,7 @@ Parfor decorates a functions and returns the result of that function evaluated i
an iterator.
## Requires
tqdm, dill
## Limitations
If you're using Python with the GIL enabaled, then objects passed to the pool need to be dillable (dill needs to
serialize them). Generators and SwigPyObjects are examples of objects that cannot be used. They can be used however, for
the iterator argument when using parfor, but its iterations need to be dillable. You might be able to make objects
dillable anyhow using `dill.register` or with `__reduce__`, `__getstate__`, etc.
numpy, ray, tqdm
## Arguments
To functions `parfor.parfor`, `parfor.pmap` and `parfor.gmap`.
@@ -66,11 +42,11 @@ To functions `parfor.parfor`, `parfor.pmap` and `parfor.gmap`.
bar: bool enable progress bar,
or a callback function taking the number of passed iterations as an argument
serial: execute in series instead of parallel if True, None (default): let pmap decide
length: deprecated alias for total
n_processes: number of processes to use,
the parallel pool will be restarted if the current pool does not have the right number of processes
yield_ordered: return the result in the same order as the iterable
yield_index: return the index of the result too
allow_output: allow output from subprocesses
**bar_kwargs: keyword arguments for tqdm.tqdm
### Return
@@ -185,3 +161,6 @@ Split a long iterator in bite-sized chunks to parallelize
More low-level accessibility to parallel execution. Submit tasks and request the result at any time,
(although to avoid breaking causality, submit first, then request), use different functions and function
arguments for different tasks.
## `SharedArray`
A numpy arrow that can be shared among processes.

53
parfor/__init__.py
View File

@@ -4,6 +4,7 @@ import logging
import os
import warnings
from contextlib import ExitStack, redirect_stdout, redirect_stderr
from io import StringIO
from functools import wraps
from importlib.metadata import version
from multiprocessing.shared_memory import SharedMemory
@@ -57,7 +58,7 @@ class SharedArray(np.ndarray):
shm = SharedMemory(create=True, size=int(np.prod(shape) * np.dtype(dtype).itemsize)) # type: ignore
new = super().__new__(cls, shape, dtype, shm.buf, offset, strides, order)
new.shm = shm
return new
return new # type: ignore
def __reduce__(
self,
@@ -193,17 +194,23 @@ class ExternalBar(Iterable):
@ray.remote
def worker(task):
try:
with (
warnings.catch_warnings(),
redirect_stdout(open(os.devnull, "w")),
redirect_stderr(open(os.devnull, "w")),
):
warnings.simplefilter("ignore", category=FutureWarning)
with ExitStack() as stack: # noqa
if task.allow_output:
out = StringIO()
err = StringIO()
stack.enter_context(redirect_stdout(out))
stack.enter_context(redirect_stderr(err))
else:
stack.enter_context(redirect_stdout(open(os.devnull, "w")))
stack.enter_context(redirect_stderr(open(os.devnull, "w")))
try:
task()
task.status = ("done",)
except Exception: # noqa
task.status = "task_error", format_exc()
if task.allow_output:
task.out = out.getvalue()
task.err = err.getvalue()
except KeyboardInterrupt: # noqa
pass
@@ -217,6 +224,7 @@ class Task:
fun: Callable[[Any, ...], Any],
args: tuple[Any, ...] = (),
kwargs: dict[str, Any] = None,
allow_output: bool = False,
) -> None:
self.handle = handle
self.fun = fun
@@ -225,8 +233,11 @@ class Task:
self.name = fun.__name__ if hasattr(fun, "__name__") else None
self.done = False
self.result = None
self.out = None
self.err = None
self.future = None
self.status = "starting"
self.allow_output = allow_output
@property
def fun(self) -> Callable[[Any, ...], Any]:
@@ -285,11 +296,13 @@ class ParPool:
kwargs: dict[str, Any] = None,
n_processes: int = None,
bar: Bar = None,
allow_output: bool = False,
):
self.handle = 0
self.tasks = {}
self.bar = bar
self.bar_lengths = {}
self.allow_output = allow_output
self.fun = fun
self.args = args
self.kwargs = kwargs
@@ -318,6 +331,7 @@ class ParPool:
kwargs: dict[str, Any] = None,
handle: Hashable = None,
barlength: int = 1,
allow_output: bool = False,
) -> Optional[int]:
if handle is None:
new_handle = self.handle
@@ -331,6 +345,7 @@ class ParPool:
fun or self.fun,
args or self.args,
kwargs or self.kwargs,
allow_output or self.allow_output,
)
task.future = worker.remote(task)
self.tasks[new_handle] = task
@@ -359,6 +374,10 @@ class ParPool:
def finalize_task(self, task: Task) -> Any:
code, *args = task.status
if task.out:
print(task.out, end="")
if task.err:
print(task.err, end="")
getattr(self, code)(task, *args)
self.tasks.pop(task.handle)
return task.result
@@ -366,13 +385,14 @@ class ParPool:
def get_newest(self) -> Optional[Any]:
"""Request the newest handle and result and delete its record. Wait if result not yet available."""
while True:
for handle, task in self.tasks.items():
if handle in self.bar_lengths:
try:
task = ray.get(task.future, timeout=0.01)
return task.handle, self.finalize_task(task)
except ray.exceptions.GetTimeoutError:
pass
if self.tasks:
for handle, task in self.tasks.items():
if handle in self.bar_lengths:
try:
task = ray.get(task.future, timeout=0.01)
return task.handle, self.finalize_task(task)
except ray.exceptions.GetTimeoutError:
pass
def task_error(self, task: Task, error: Exception) -> None:
if task.handle in self:
@@ -424,6 +444,7 @@ def gmap(
n_processes: int = None,
yield_ordered: bool = True,
yield_index: bool = False,
allow_output: bool = False,
**bar_kwargs: Any,
) -> Generator[Any, None, None]:
"""map a function fun to each iteration in iterable
@@ -442,11 +463,11 @@ def gmap(
bar: bool enable progress bar,
or a callback function taking the number of passed iterations as an argument
serial: execute in series instead of parallel if True, None (default): let pmap decide
length: deprecated alias for total
n_processes: number of processes to use,
the parallel pool will be restarted if the current pool does not have the right number of processes
yield_ordered: return the result in the same order as the iterable
yield_index: return the index of the result too
allow_output: allow output from subprocesses
**bar_kwargs: keywords arguments for tqdm.tqdm
output:
@@ -550,7 +571,7 @@ def gmap(
bar = stack.enter_context(ExternalBar(callback=bar)) # noqa
else:
bar = stack.enter_context(tqdm(**bar_kwargs))
with ParPool(chunk_fun, args, kwargs, n_processes, bar) as p: # type: ignore
with ParPool(chunk_fun, args, kwargs, n_processes, bar, allow_output) as p: # type: ignore
for i, (j, l) in enumerate(zip(iterable, iterable.lengths)): # add work to the queue
p(j, handle=i, barlength=l)
if bar.total is None or bar.total < i + 1:

2
pyproject.toml
View File

@@ -1,6 +1,6 @@
[project]
name = "parfor"
version = "2026.1.0"
version = "2026.1.1"
description = "A package to mimic the use of parfor as done in Matlab."
authors = [
{ name = "Wim Pomp-Pervova", email = "wimpomp@gmail.com" }