chore: minor chores

Author: okunator

cellseg_models_pytorch/datasets/_base_dataset.py

@@ -47,7 +47,8 @@ def __init__(
             inst_transforms : List[str]
                 A list containg all the transformations that are applied to only the
                 instance labelled masks. Allowed ones: "cellpose", "contour", "dist",
-                "edgeweight", "hovernet", "omnipose", "smooth_dist", "binarize"
+                "edgeweight", "hovernet", "omnipose", "smooth_dist", "binarize",
+                "strong_augment
             normalization : str, optional
                 Apply img normalization after all the transformations. One of "minmax",
                 "norm", "percentile", None.

cellseg_models_pytorch/utils/mask_utils.py

@@ -568,13 +568,11 @@ def remove_debris_semantic(sem_map: np.ndarray, min_size: int = 10000):
         classes = classes[1:]
 
     for i in classes:
-
         area = np.array(res == i, np.int32)
         inst_map = ndi.label(area)[0]
         labels, counts = np.unique(inst_map, return_counts=True)
 
         for label, npixls in zip(labels, counts):
-
             if npixls < min_size:
                 res[inst_map == label] = 0
 
@@ -922,10 +920,7 @@ def majority_vote_parallel(type_map: np.ndarray, inst_map: np.ndarray) -> np.nda
     return tmap
 
 
-def majority_vote_sequential(
-    type_map: np.ndarray,
-    inst_map: np.ndarray,
-) -> np.ndarray:
+def majority_vote_sequential(type_map: np.ndarray, inst_map: np.ndarray) -> np.ndarray:
     """Do a majority voting on the type_map pixels by broadcasting.
 
     Adapted from:

pyproject.toml

@@ -12,14 +12,14 @@ keywords = ["cell segmentation", "nuclei segmentation", "pytorch", "pytorch-ligh
 
 [tool.poetry.dependencies]
 python = "^3.8,<3.10"
-torch = "^1.8.1"
 numpy = "^1.22,<1.24"
-scipy = "^1.7"
-scikit-image = "^0.19"
 pathos = ">=0.2.8,<0.4.0"
-opencv-python = "^4.2.0.32"
 timm = ">=0.5.4,<0.7.0"
 numba = ">=0.55.2,<0.58.0"
+torch = "^1.8.1"
+scipy = "^1.7"
+scikit-image = "^0.19"
+opencv-python = "^4.2.0.32"
 tqdm = "^4.64.0"
 scikit-learn = "^1.0.2"
 pytorch-lightning = {version = "^1.6.0", optional = true}