Use joblib for parallel processing

pyproject.toml

@@ -15,10 +15,10 @@ classifiers = [
 ]
 dependencies = [
     "dask==2024.6.2",
+    "joblib==1.4.2",
     "loguru==0.7.2",
     "numpy==1.26.4",
     "scikit-image==0.24.0",
-    "tqdm",
     "zarr==2.18.2",
 ]
 description = "Convert stacks of images to chunked datasets"

src/stack_to_chunk/_array_helpers.py

@@ -2,9 +2,11 @@
 import numpy as np
 import skimage.measure
 import zarr
+from joblib import delayed
 from loguru import logger
 
 
+@delayed  # type: ignore[misc]
 def _copy_slab(arr_zarr: zarr.Array, slab: da.Array, zstart: int, zend: int) -> None:
     """
     Copy a single slab of data to a zarr array.
@@ -31,6 +33,7 @@ def _copy_slab(arr_zarr: zarr.Array, slab: da.Array, zstart: int, zend: int) ->
     logger.info(f"Finished copying z={zstart} -> {zend-1}")
 
 
+@delayed  # type: ignore[misc]
 def _downsample_block(
     arr_in: zarr.Array, arr_out: zarr.Array, block_idx: tuple[int, int, int]
 ) -> None:

src/stack_to_chunk/main.py

@@ -2,14 +2,13 @@
 Main code for converting stacks to chunks.
 """
 
-from multiprocessing import Pool
 from pathlib import Path
 from typing import Any, Literal
 
 import numpy as np
-import tqdm
 import zarr
 from dask.array.core import Array
+from joblib import Parallel
 from loguru import logger
 from numcodecs import blosc
 from numcodecs.abc import Codec
@@ -225,17 +224,8 @@ def add_full_res_data(
         blosc_use_threads = blosc.use_threads
         blosc.use_threads = 0
 
-        # Use try/finally pattern to allow code coverage to be collected
-        if n_processes == 1:
-            for args in all_args:
-                _copy_slab(*args)
-        else:
-            p = Pool(n_processes)
-            try:
-                p.starmap(_copy_slab, all_args)
-            finally:
-                p.close()
-                p.join()
+        jobs = [_copy_slab(*args) for args in all_args]
+        Parallel(n_jobs=n_processes)(jobs)
 
         blosc.use_threads = blosc_use_threads
         logger.info("Finished full resolution copy to zarr.")
@@ -299,13 +289,8 @@ def add_downsample_level(self, level: int, *, n_processes: int) -> None:
         blosc_use_threads = blosc.use_threads
         blosc.use_threads = 0
 
-        # Use try/finally pattern to allow code coverage to be collected
-        p = Pool(n_processes)
-        try:
-            p.starmap(_downsample_block, tqdm.tqdm(all_args, total=len(all_args)))
-        finally:
-            p.close()
-            p.join()
+        jobs = [_downsample_block(*args) for args in all_args]
+        Parallel(n_jobs=n_processes, verbose=10)(jobs)
 
         self._add_level_metadata(level)
         blosc.use_threads = blosc_use_threads