aboutsummaryrefslogtreecommitdiffhomepage
path: root/sec_certs/parallel_processing.py
diff options
context:
space:
mode:
authorJ08nY2021-12-14 11:33:47 +0100
committerJ08nY2021-12-14 11:33:47 +0100
commit8f48d490683ce3bb9d9b96581021168d5d128051 (patch)
tree14432e13cefafe345c734b1d60ed88483a656d59 /sec_certs/parallel_processing.py
parenta3e1b59e984531382ff4b18ad907909e0d85ec3b (diff)
downloadsec-certs-8f48d490683ce3bb9d9b96581021168d5d128051.tar.gz
sec-certs-8f48d490683ce3bb9d9b96581021168d5d128051.tar.zst
sec-certs-8f48d490683ce3bb9d9b96581021168d5d128051.zip
Use thread/process pools from concurrent.futures.
The process pool from concurrent.futures should not start its workers as daemonic processes and so processing inside of a Celery worker should work as per: https://stackoverflow.com/questions/6974695/python-process-pool-non-daemonic
Diffstat (limited to 'sec_certs/parallel_processing.py')
-rw-r--r--sec_certs/parallel_processing.py18
1 files changed, 9 insertions, 9 deletions
diff --git a/sec_certs/parallel_processing.py b/sec_certs/parallel_processing.py
index 40025074..be7b0d4b 100644
--- a/sec_certs/parallel_processing.py
+++ b/sec_certs/parallel_processing.py
@@ -1,26 +1,26 @@
from tqdm import tqdm
-from multiprocessing.pool import Pool, ThreadPool
+from concurrent.futures import ProcessPoolExecutor as ProcessPool, ThreadPoolExecutor as ThreadPool
+# from multiprocessing.pool import Pool, ThreadPool
from typing import Callable, Iterable, Optional
import time
-def process_parallel(func: Callable, items: Iterable, max_workers: int, callback: Optional[Callable] = None,
+def process_parallel(func: Callable, items: Iterable, max_workers: int,
use_threading: bool = True, progress_bar: bool = True, unpack: bool = False,
progress_bar_desc: Optional[str] = None):
- pool = ThreadPool(max_workers) if use_threading else Pool(max_workers)
- results = [pool.apply_async(func, (*i,), callback=callback) for i in items] if unpack else [pool.apply_async(func, (i, ), callback=callback) for i in items]
+ pool = ThreadPool(max_workers) if use_threading else ProcessPool(max_workers)
+ results = [pool.submit(func, *i) for i in items] if unpack else [pool.submit(func, i) for i in items]
if progress_bar is True and items:
bar = tqdm(total=len(results), desc=progress_bar_desc)
- while not all([x.ready() for x in results]):
- done_count = len([x.ready() for x in results if x.ready()])
+ while not all(all_done := [x.done() for x in results]):
+ done_count = len(list(filter(lambda x: x, all_done)))
bar.update(done_count - bar.n)
time.sleep(1)
bar.update(len(results) - bar.n)
bar.close()
- pool.close()
- pool.join()
+ pool.shutdown()
- return [r.get() for r in results]
+ return [r.result() for r in results]