Add migration scripts to bioimage-cpp

environment.yaml

@@ -2,7 +2,7 @@ name: patho-sam
 channels:
     - conda-forge
 dependencies:
-    - micro_sam
+    - micro_sam >=1.8.1
     # Note: installing the pytorch package from conda-forge will generally
     # give you the most optmized version for your system, if you have a modern
     # enough OS and CUDA version (CUDA >= 12). For older versions, you can

experiments/benchmarking/cellvit/eval_util.py

@@ -7,7 +7,7 @@
 import numpy as np
 import pandas as pd
 import imageio.v3 as imageio
-from skimage.measure import label
+from bioimage_cpp.segmentation import label
 
 from elf.evaluation import mean_segmentation_accuracy
 

experiments/benchmarking/hovernet/eval_util.py

@@ -6,7 +6,7 @@
 import numpy as np
 import pandas as pd
 import imageio.v3 as imageio
-from skimage.measure import label
+from bioimage_cpp.segmentation import label
 
 from elf.evaluation import mean_segmentation_accuracy
 

experiments/benchmarking/hovernext/evaluate_ais_hover.py

@@ -6,7 +6,7 @@
 import numpy as np
 import pandas as pd
 import imageio.v3 as imageio
-from skimage.measure import label
+from bioimage_cpp.segmentation import label
 
 from elf.evaluation import mean_segmentation_accuracy
 

experiments/benchmarking/instanseg/instanseg_inference.py

@@ -7,7 +7,7 @@
 import numpy as np
 import pandas as pd
 import imageio.v3 as imageio
-from skimage.measure import label
+from bioimage_cpp.segmentation import label
 
 from tukra.io import read_image
 from tukra.inference import segment_using_instanseg

experiments/benchmarking/outdated/cellvitplusplus/eval.py

@@ -6,7 +6,7 @@
 import numpy as np
 import pandas as pd
 import imageio.v3 as imageio
-from skimage.measure import label
+from bioimage_cpp.segmentation import label
 
 from elf.evaluation import mean_segmentation_accuracy
 

experiments/benchmarking/stardist/stardist_inference.py

@@ -7,7 +7,7 @@
 import numpy as np
 import pandas as pd
 import imageio.v3 as imageio
-from skimage.measure import label
+from bioimage_cpp.segmentation import label
 
 from tukra.io import read_image
 from tukra.inference import segment_using_stardist

experiments/patho-sam/per_image_eval.py

@@ -7,7 +7,7 @@
 import numpy as np
 import pandas as pd
 import imageio.v3 as imageio
-from skimage.measure import label
+from bioimage_cpp.segmentation import label
 
 from elf.evaluation import mean_segmentation_accuracy
 

patho_sam/semantic_segmentation.py

@@ -5,7 +5,7 @@
 import numpy as np
 from numpy.typing import ArrayLike
 
-from nifty.tools import blocking
+from bioimage_cpp.utils import Blocking
 
 import torch
 
@@ -238,7 +238,7 @@ def initialize(
             self._predictor, image, image_embeddings, tile_shape, halo,
             verbose=verbose, batch_size=batch_size, mask=mask, i=i,
         )
-        tiling = blocking([0, 0], original_size, tile_shape)
+        tiling = Blocking([0, 0], original_size, tile_shape)
 
         if semantic_segmentation is None:
             semantic_segmentation = np.zeros(original_size, dtype="uint8")
@@ -250,7 +250,7 @@ def initialize(
 
         msg = "Initialize tiled semantic segmentation with decoder"
         if tiles_in_mask is None:
-            n_tiles = tiling.numberOfBlocks
+            n_tiles = tiling.number_of_blocks
             all_tile_ids = list(range(n_tiles))
         else:
             n_tiles = len(tiles_in_mask)
@@ -282,11 +282,11 @@ def initialize(
                 output = np.argmax(output, axis=0)
 
                 # Set the predictions in the output for this tile.
-                block = tiling.getBlockWithHalo(tile_id, halo=list(halo))
+                block = tiling.get_block_with_halo(tile_id, halo=list(halo))
                 local_bb = tuple(
-                    slice(beg, end) for beg, end in zip(block.innerBlockLocal.begin, block.innerBlockLocal.end)
+                    slice(beg, end) for beg, end in zip(block.inner_block_local.begin, block.inner_block_local.end)
                 )
-                inner_bb = tuple(slice(beg, end) for beg, end in zip(block.innerBlock.begin, block.innerBlock.end))
+                inner_bb = tuple(slice(beg, end) for beg, end in zip(block.inner_block.begin, block.inner_block.end))
                 semantic_segmentation[inner_bb] = output[local_bb]
                 pbar_update(1)
 

scripts/test_pannuke.py

@@ -6,7 +6,7 @@
 import h5py
 import numpy as np
 import imageio.v3 as imageio
-from skimage.segmentation import relabel_sequential
+from bioimage_cpp.segmentation import relabel_sequential
 
 from torch_em.data.datasets.histopathology import pannuke, monuseg
 

setup.py

@@ -16,6 +16,7 @@
     url='https://github.com/computational-cell-analytics/patho-sam',
     packages=find_packages(include=['patho_sam', 'patho_sam.*']),
     license="MIT",
+    install_requires=["micro_sam>=1.8.1"],
     entry_points={
         "console_scripts": [
             "patho_sam.example_data=patho_sam.util:get_example_wsi_data",