Lint with ruff's default rules

deepaneseg/data/generators.py

@@ -22,7 +22,7 @@ def generate_transforms(trans=None, rot=None, center=None, disp=None):
         affine = None
     else:
         t = np.zeros(3)
-        if not trans is None:
+        if trans is not None:
             if np.isscalar(trans):
                 t = trans * np.ones(3)
             elif len(trans) == 3:
@@ -31,7 +31,7 @@ def generate_transforms(trans=None, rot=None, center=None, disp=None):
                 raise TypeError("trans argument should be either a scalar or a 3-vector")
         t = t * (2 * (np.random.random_sample(3) - 0.5))
         r = np.zeros(3)
-        if not rot is None:
+        if rot is not None:
             if np.isscalar(rot):
                 r = rot * np.ones(3)
             elif len(rot) == 3:
@@ -44,7 +44,7 @@ def generate_transforms(trans=None, rot=None, center=None, disp=None):
         affine[:3, :3] = Rotation.from_euler("zyx", r, degrees=True).as_matrix()
         affine[:3, 3] = t if center is None else t + center.ravel() @ (np.eye(3) - affine[:3, :3]).T
 
-    if not disp is None:
+    if disp is not None:
         if np.isscalar(disp):
             amp = disp * np.ones(3)
         elif len(disp) == 3:

deepaneseg/data/io.py

@@ -2,7 +2,8 @@
 import os
 import nibabel as ni
 import pandas as pd
-import json, glob
+import json
+import glob
 
 from deepaneseg.volume.selection import select_points
 from deepaneseg.utils import get_logger
@@ -105,12 +106,12 @@ def read_patient_data(vol_file, ane_file, normalize=None, vessel_file=None):
         - 'Normal': rescaling to 0-mean and unit-variance data
     """
     d, a = read_nii_from_file(vol_file)
-    if not vessel_file is None:
+    if vessel_file is not None:
         seg, _ = read_nii_from_file(vessel_file)
         seg = np.where(seg > 0, 1, 0).astype(np.uint8)  # just in case
     else:
         seg = None
-    if not ane_file is None:
+    if ane_file is not None:
         s = read_points_from_csv(ane_file)
     else:
         s = None
@@ -126,7 +127,7 @@ def read_patient_data_from_config(cfg_name, volume="init volume", normalize=None
     dir = os.path.dirname(cfg_name)
     with open(cfg_name, "r") as f:
         c = json.load(f)
-    if not vessel is None and os.path.isfile(os.path.join(dir, dir.split("/")[-1] + ".nii.gz")):  # c['vessel']))
+    if vessel is not None and os.path.isfile(os.path.join(dir, dir.split("/")[-1] + ".nii.gz")):  # c['vessel']))
         vessel_file = os.path.join(dir, dir.split("/")[-1] + ".nii.gz")  # c['vessel'])
     else:
         vessel_file = None
@@ -214,8 +215,8 @@ def csv2fcsv(csv_pathname):
     head = f"# Markups fiducial file version = 4.10\n# CoordinateSystem = 0\n# columns = {','.join(df)}\n"
     with open(fcsv_file_out, "w") as f:
         f.write(head)
-        for i, l in df.iterrows():
-            f.write(",".join([f"{v}" for v in l]) + "\n")
+        for i, row in df.iterrows():
+            f.write(",".join([f"{v}" for v in row]) + "\n")
 
 
 def fcsv2csv(fcsv_file):
@@ -224,13 +225,13 @@ def fcsv2csv(fcsv_file):
     csv_file_out = filename + ".csv"
 
     with open(fcsv_file, "r") as f:
-        l = f.readlines()
+        lines = f.readlines()
     # coordinate system
-    cs = l[1].strip().split("=")[1].strip()
+    cs = lines[1].strip().split("=")[1].strip()
     # column names
-    cn = list(map(str.strip, l[2].split("=")[1].split(",")))
+    cn = list(map(str.strip, lines[2].split("=")[1].split(",")))
     # data
-    data = [s.strip().split(",") for s in l[3:]]
+    data = [s.strip().split(",") for s in lines[3:]]
     # check coordinate systems and column names
     # RAS: if cs='RAS' or cs='0'
     # LPS: if cs='LPS' or cs='1'

deepaneseg/inference/evaluation.py

@@ -26,7 +26,7 @@ def confusion_matrix(pred, vox2met, spheres, min_size=None, tp_proportion=0.3):
     """
     # label CC in prediction
     label_pred = skme.label(pred.astype(np.uint8))
-    if not min_size is None:  # also perform relabelling to keep consecutive labels
+    if min_size is not None:  # also perform relabelling to keep consecutive labels
         label_pred = skme.label(skmo.remove_small_objects(label_pred, min_size=min_size))
     ncc_pred = np.max(label_pred.ravel())
 
@@ -86,12 +86,12 @@ def adam_evaluation(pred, vox2met, spheres, pretraited=None, min_size=None):
     """
     # label CC in prediction
     label_pred = skme.label(pred.astype(np.uint8))
-    if not min_size is None:  # also perform relabelling to keep consecutive labels
+    if min_size is not None:  # also perform relabelling to keep consecutive labels
         label_pred = skme.label(skmo.remove_small_objects(label_pred, min_size=min_size))
     ncc_pred = np.max(label_pred.ravel())
 
     # Exculude pretraited aneurysms
-    if not pretraited is None:
+    if pretraited is not None:
         print("pretraited")
         for i in range(1, ncc_pred + 1):
             tmp = label_pred.copy()

deepaneseg/models/building_blocks.py

@@ -20,7 +20,7 @@ def create_conv(in_channels, out_channels, kernel_size, order, padding):
         elif char == "l":
             modules.append(("LeakyReLU", nn.LeakyReLU(inplace=True)))
         elif char == "c":
-            bias = not ("b" in order)
+            bias = "b" not in order
             modules.append(("conv", create_conv3d(in_channels, out_channels, kernel_size, padding=padding, bias=bias)))
         elif char == "b":
             is_before_conv = i < order.index("c")

deepaneseg/training/dataset.py

@@ -1,5 +1,6 @@
 import math
-import torch, random
+import torch
+import random
 import numpy as np
 
 from torch.utils.data import DataLoader

deepaneseg/utils.py

@@ -8,6 +8,7 @@
 
 import numpy as np
 import torch
+from torch.optim.lr_scheduler import ReduceLROnPlateau, _LRScheduler
 
 from hydra.utils import instantiate
 
@@ -183,8 +184,6 @@ def init_model(m):
 
 
 ########### WarmUp scheduler ########
-from torch.optim.lr_scheduler import _LRScheduler
-from torch.optim.lr_scheduler import ReduceLROnPlateau
 
 
 class GradualWarmupScheduler(_LRScheduler):
@@ -242,7 +241,7 @@ def step_reduce_lr_on_plateau(self, metrics, epoch=None):
             self.after_scheduler.step(metrics)  # , None)
 
     def step(self, epoch=None, metrics=None):
-        if type(self.after_scheduler) != ReduceLROnPlateau:
+        if not isinstance(self.after_scheduler, ReduceLROnPlateau):
             if self.finished and self.after_scheduler:
                 if epoch is None:
                     self.after_scheduler.step(None)

deepaneseg/volume/patch.py

@@ -16,12 +16,12 @@ def inv_trans_point(p, center, size, affine=None, disp=None, order=3):
     """
     x = p.copy()
     # apply affine transform
-    if not affine is None:
+    if affine is not None:
         D = np.linalg.inv(affine)
         x = D[:3, :] @ np.vstack((x, np.ones(x.shape[1])))
 
     # then apply inverse of deformation
-    if not disp is None:
+    if disp is not None:
         center = np.asarray(center)
         size = np.asarray(size)
         # determine abscissa relative to the cube
@@ -52,10 +52,10 @@ def trans_point(p, center, size, affine=None, disp=None):
     p stored in an 3xN array
     """
     x = p.copy()
-    if not disp is None:
+    if disp is not None:
         ores = sopt.minimize(_euclidean_dist, x.ravel(), method="CG", args=(p.ravel(), center, size, affine, disp))
         x = ores.x.reshape((3, -1))
-    elif not affine is None:
+    elif affine is not None:
         x = affine[:3, :] @ np.vstack((x, np.ones(x.shape[1])))
     return x
 

pyproject.toml

@@ -78,7 +78,8 @@ extend-exclude = ["notebooks"]
 line-ending = "lf"
 
 [tool.ruff.lint]
-select = ["E9", "F"]
+# ruff's default rules: pycodestyle errors and pyflakes
+select = ["E4", "E7", "E9", "F"]
 
 [tool.ruff.lint.per-file-ignores]
 # still references TensorFlow from an earlier port and computes metrics it never reports;

scripts/experimental/evaluate.py

@@ -19,9 +19,8 @@ def main():
     try:
         with open(cfg, "r") as f:
             config = json.load(f)
-    except:
-        print(f"No such config file {cfg}")
-        exit()
+    except OSError:
+        sys.exit(f"No such config file {cfg}")
 
     logger = get_logger("Evaluation")
 

scripts/experimental/evaluate_perf.py

@@ -80,7 +80,7 @@
     pred, vox2met = dio.read_nii_from_file(fname)
     pred = np.where(pred < threshold, 0, 1).astype(np.uint8)
 
-    if not vmin is None:
+    if vmin is not None:
         pred = skimo.remove_small_objects(skime.label(pred), min_size=vmin)
         pred = np.where(pred > 0, 1, 0)