Stash benchmark work

.gitignore

@@ -170,3 +170,6 @@ src/stack_to_chunk/_version.py
 
 docs/auto_examples
 docs/sg_execution_times.rst
+
+memray*.bin
+benchmarks/data

benchmarks/behnchmark.py

@@ -0,0 +1,39 @@
+"""
+Benchmark
+"""
+
+import pathlib
+import shutil
+import time
+
+import dask_image.imread
+import numpy as np
+import tifffile
+
+import stack_to_chunk
+
+if __name__ == "__main__":
+    plane = np.random.randint(low=0, high=2**16, size=(2000, 2000), dtype=np.uint16)
+    image_dir = pathlib.Path(__file__).parent / "data"
+    if not image_dir.exists():
+        image_dir.mkdir()
+        for i in range(64):
+            tifffile.imwrite(image_dir / f"{str(i).zfill(3)}.tif", plane)
+
+    images = dask_image.imread.imread(str(image_dir / "*.tif")).T
+    print(f"Volume size: {images.nbytes / 1e6} MB")
+
+    for n_processes in [1, 2, 3, 4]:
+        shutil.rmtree(image_dir / "chunked.zarr")
+        group = stack_to_chunk.MultiScaleGroup(
+            image_dir / "chunked.zarr",
+            name="my_zarr_group",
+            spatial_unit="centimeter",
+            voxel_size=(3, 4, 5),
+        )
+        t_start = time.time()
+        group.add_full_res_data(
+            images, chunk_size=32, compressor="default", n_processes=n_processes
+        )
+        t_end = time.time()
+        print(f"{n_processes=}, t={t_end - t_start} seconds")