Switch to ruff with linting

.github/workflows/format-ubuntu.yml

@@ -38,8 +38,8 @@ jobs:
           fi
           set +x -euo pipefail
           sudo apt-get update && sudo apt-get install -y clang-format-14
-          python -m pip install black==24.10.0
+          python -m pip install ruff==0.6.7
           ./scripts/format/clang_format.sh
-          ./scripts/format/black.sh
+          ./scripts/format/python.sh
           git diff --name-only
           git diff --exit-code || (echo "Code formatting failed" && exit 1)

limap/__init__.py

@@ -3,22 +3,21 @@
 sys.path.append("build/limap/_limap")
 from _limap import *
 
-from . import base
-from . import point2d
-from . import line2d
-from . import vplib
-from . import pointsfm
-from . import undistortion
-
-from . import triangulation
-from . import merging
-from . import evaluation
-from . import fitting
-from . import util
-from . import visualize
-from . import structures
-
-from . import features
-from . import optimize
-
-from . import runners
+from . import (
+    base,
+    evaluation,
+    features,
+    fitting,
+    line2d,
+    merging,
+    optimize,
+    point2d,
+    pointsfm,
+    runners,
+    structures,
+    triangulation,
+    undistortion,
+    util,
+    visualize,
+    vplib,
+)

limap/base/__init__.py

@@ -1,6 +1,7 @@
 from _limap._base import *
-from .functions import *
+
 from .align import *
-from .unit_test import *
 from .depth_reader_base import *
+from .functions import *
 from .p3d_reader_base import *
+from .unit_test import *

limap/base/align.py

@@ -14,7 +14,7 @@ def umeyama_alignment(x, y, with_scale=True):
     :return: r, t, c - rotation matrix, translation vector and scale factor
     """
     if x.shape != y.shape:
-        assert False, "x.shape not equal to y.shape"
+        raise AssertionError("x.shape not equal to y.shape")
 
     # m = dimension, n = nr. of data points
     m, n = x.shape
@@ -54,10 +54,7 @@ def umeyama_alignment(x, y, with_scale=True):
 def align_imagecols_umeyama(imagecols_src, imagecols_dst):
     # assertion check
     assert imagecols_src.NumImages() == imagecols_dst.NumImages()
-    assert (
-        np.all(imagecols_src.get_img_ids() == imagecols_dst.get_img_ids())
-        == True
-    )
+    assert np.all(imagecols_src.get_img_ids() == imagecols_dst.get_img_ids())
 
     # fit transformation
     xyz_src = np.array(imagecols_src.get_locations()).transpose()
@@ -74,18 +71,17 @@ def align_imagecols_colmap(
     max_error=0.01,
     tmp_folder="tmp/model_convertion",
 ):
-    import os, shutil
+    import os
+    import subprocess
+
     import numpy as np
-    from limap.pointsfm import convert_imagecols_to_colmap
+
     import limap.util.io as limapio
-    import subprocess
+    from limap.pointsfm import convert_imagecols_to_colmap
 
     # assertion check
     assert imagecols_src.NumImages() == imagecols_dst.NumImages()
-    assert (
-        np.all(imagecols_src.get_img_ids() == imagecols_dst.get_img_ids())
-        == True
-    )
+    assert np.all(imagecols_src.get_img_ids() == imagecols_dst.get_img_ids())
 
     limapio.check_makedirs(tmp_folder)
     src_folder = os.path.join(tmp_folder, "source")
@@ -102,7 +98,7 @@ def align_imagecols_colmap(
         for img_id in imagecols_src.get_img_ids():
             imname = imagecols_src.image_name(img_id)
             pos = imagecols_dst.camview(img_id).pose.center()
-            f.write("{0} {1} {2} {3}\n".format(imname, pos[0], pos[1], pos[2]))
+            f.write(f"{imname} {pos[0]} {pos[1]} {pos[2]}\n")
 
     # call comlap model aligner
     transform_path = os.path.join(tmp_folder, "transform.txt")
@@ -130,7 +126,7 @@ def align_imagecols_colmap(
 
     # read in transformation
     def read_trans(fname):
-        with open(fname, "r") as f:
+        with open(fname) as f:
             lines = f.readlines()
         mat = []
         for idx in range(4):

limap/estimators/__init__.py

@@ -1,2 +1,3 @@
 from _limap._estimators import *
+
 from .absolute_pose import *

limap/estimators/absolute_pose/_pl_estimate_absolute_pose.py

@@ -1,8 +1,9 @@
+import numpy as np
 from _limap import _ceresbase
-import limap.optimize as _optimize
-import limap.estimators as _estimators
+
 import limap.base as _base
-import numpy as np
+import limap.estimators as _estimators
+import limap.optimize as _optimize
 
 
 def _pl_estimate_absolute_pose(

limap/features/__init__.py

@@ -1,3 +1,4 @@
 from _limap._features import *
-from .extractors import *
+
 from .extract_line_patches import *
+from .extractors import *

limap/features/extract_line_patches.py

@@ -1,5 +1,5 @@
-from _limap import _features
 import numpy as np
+from _limap import _features
 
 
 def write_patch(fname, patch, dtype="float16"):
@@ -17,7 +17,7 @@ def write_patch(fname, patch, dtype="float16"):
 
 def load_patch(fname, dtype="float16"):
     # return a PatchInfo_f object
-    patch_info_name = "PatchInfo_f{0}".format(dtype[-2:])
+    patch_info_name = f"PatchInfo_f{dtype[-2:]}"
     with open(fname, "rb") as f:
         data = np.load(f, allow_pickle=True)
         patch = getattr(_features, patch_info_name)(
@@ -28,7 +28,7 @@ def load_patch(fname, dtype="float16"):
 
 def get_extractor(cfg, channels):
     lpe_options = _features.LinePatchExtractorOptions(cfg)
-    patch_extractor_name = "LinePatchExtractor_f64_c{0}".format(channels)
+    patch_extractor_name = f"LinePatchExtractor_f64_c{channels}"
     extractor = getattr(_features, patch_extractor_name)(lpe_options)
     return extractor
 
@@ -39,9 +39,7 @@ def extract_line_patch_oneimage(cfg, track, img_id, camview, feature):
     _features.PatchInfo_fx
     """
     lpe_options = _features.LinePatchExtractorOptions(cfg)
-    patch_extractor_name = "LinePatchExtractor_f64_c{0}".format(
-        feature.shape[2]
-    )
+    patch_extractor_name = f"LinePatchExtractor_f64_c{feature.shape[2]}"
     extractor = getattr(_features, patch_extractor_name)(lpe_options)
     patch = extractor.ExtractOneImage(track, img_id, camview, feature)
     return patch
@@ -53,9 +51,7 @@ def extract_line_patches(cfg, track, p_camviews, p_features):
     list of _features.PatchInfo_fx
     """
     lpe_options = _features.LinePatchExtractorOptions(cfg)
-    patch_extractor_name = "LinePatchExtractor_f64_c{0}".format(
-        p_features[0].shape[2]
-    )
+    patch_extractor_name = f"LinePatchExtractor_f64_c{p_features[0].shape[2]}"
     extractor = getattr(_features, patch_extractor_name)(lpe_options)
     patches = extractor.Extract(track, p_camviews, p_features)
     return patches

limap/features/extractors.py

@@ -1,17 +1,17 @@
 # [NOTE] modified from the pixel-perfect-sfm project
 
-import os, sys
-import torch
+import sys
+import time
+
 import numpy as np
 import PIL
-
-from .models.s2dnet import *
-from .models.vggnet import VGGNet
+import torch
 import torchvision.transforms.functional as tvf
+from _limap import _features
 from torchvision import transforms
-import time
 
-from _limap import _features
+from .models.s2dnet import S2DNet
+from .models.vggnet import VGGNet
 
 RGB_mean = [0.485, 0.456, 0.406]
 RGB_std = [0.229, 0.224, 0.225]
@@ -44,7 +44,7 @@ class Extractor(torch.nn.Module):
     """
 
     def __init__(self, device: str):
-        super(Extractor, self).__init__()
+        super().__init__()
         self.device = device
         self.num_levels = 0
         self.model = None
@@ -87,7 +87,7 @@ def __init__(self, device: str, *args, **kwargs):
         self.num_levels = 1
         self.channels = [128]  # ,128] #[128,128]
         self.l2_normalize = True
-        if "output_channels" in kwargs.keys():
+        if "output_channels" in kwargs:
             self.channels = [min(kwargs["output_channels"], self.channels[0])]
 
     def extract_featuremaps(self, image_batch: torch.Tensor) -> list:
@@ -148,7 +148,7 @@ def __init__(self, device: str, *args, **kwargs):
         self.num_levels = 1  # 2
         self.channels = [64]  # [128,128]
         self.l2_normalize = True
-        if "output_channels" in kwargs.keys():
+        if "output_channels" in kwargs:
             self.channels = [min(kwargs["output_channels"], self.channels[0])]
 
     def extract_featuremaps(self, image_batch: torch.Tensor) -> list:

limap/features/models/base_model.py

@@ -4,9 +4,10 @@
 """
 
 from abc import ABCMeta, abstractmethod
+from copy import copy
+
 from omegaconf import OmegaConf
 from torch import nn
-from copy import copy
 
 
 class BaseModel(nn.Module, metaclass=ABCMeta):
@@ -77,7 +78,7 @@ def freeze_bn(module):
     def forward(self, data):
         """Check the data and call the _forward method of the child model."""
         for key in self.required_data_keys:
-            assert key in data, "Missing key {} in data".format(key)
+            assert key in data, f"Missing key {key} in data"
         return self._forward(data)
 
     @abstractmethod

limap/features/models/s2dnet.py

@@ -1,22 +1,14 @@
+import logging
+import os
+from pathlib import Path
 from typing import List
-import numpy
+
+import numpy as np
 import torch
 import torch.nn as nn
 from torchvision import models
-from pathlib import Path
-import logging
 
-from PIL import Image
 from .base_model import BaseModel
-import os, sys
-from torchvision import transforms
-import numpy as np
-import torch.nn.functional as F
-
-import argparse
-import h5py
-
-from time import time
 
 type_dict = {
     "uint8_t": torch.cuda.ByteTensor,
@@ -90,12 +82,12 @@ def __init__(self, hypercolumn_layers: List[str], output_dim: int = 128):
                 nn.BatchNorm2d(output_dim),
             )
             self.layers.append(layer)
-            self.add_module("adap_layer_{}".format(i), layer)
+            self.add_module(f"adap_layer_{i}", layer)
 
     def forward(self, features: List[torch.tensor]):
         """Apply adaptation layers."""
         for i, _ in enumerate(features):
-            features[i] = getattr(self, "adap_layer_{}".format(i))(features[i])
+            features[i] = getattr(self, f"adap_layer_{i}")(features[i])
         return features
 
 

limap/features/models/vggnet.py

@@ -1,3 +1,7 @@
+import torch
+import torch.nn as nn
+import torchvision.models as models
+
 from limap.features.models.s2dnet import *
 
 

limap/fitting/__init__.py

@@ -1,2 +1,3 @@
 from _limap._fitting import *
+
 from .fitting import *

limap/fitting/fitting.py

@@ -1,6 +1,5 @@
-from _limap import _base, _estimators, _fitting
-import os
 import numpy as np
+from _limap import _estimators, _fitting
 from bresenham import bresenham
 from hloc.localize_inloc import interpolate_scan
 
@@ -64,7 +63,7 @@ def estimate_seg3d_from_points3d(
     var2d=5.0,
 ):
     h, w = camview.h(), camview.w()
-    K, R, T = camview.K(), camview.R(), camview.T()
+    R, T = camview.R(), camview.T()
 
     # get points and depths
     seg1_pts = np.linspace(

limap/line2d/DeepLSD/deeplsd.py

@@ -1,13 +1,15 @@
 import os
+
 import numpy as np
 import torch
 from deeplsd.models.deeplsd_inference import DeepLSD
+
 from ..base_detector import BaseDetector, BaseDetectorOptions
 
 
 class DeepLSDDetector(BaseDetector):
     def __init__(self, options=BaseDetectorOptions()):
-        super(DeepLSDDetector, self).__init__(options)
+        super().__init__(options)
 
         conf = {
             "detect_lines": True,

limap/line2d/GlueStick/extractor.py

@@ -1,18 +1,19 @@
 import os
+
 import numpy as np
-from sklearn.cluster import DBSCAN
 import torch
-from omegaconf import OmegaConf
 from gluestick.models.wireframe import lines_to_wireframe
+from omegaconf import OmegaConf
 
 import limap.util.io as limapio
 from limap.point2d.superpoint.superpoint import SuperPoint, sample_descriptors
+
 from ..base_detector import BaseDetector, BaseDetectorOptions
 
 
 class WireframeExtractor(BaseDetector):
     def __init__(self, options=BaseDetectorOptions(), device=None):
-        super(WireframeExtractor, self).__init__(options)
+        super().__init__(options)
         self.device = "cuda" if device is None else device
         self.sp = (
             SuperPoint({"weight_path": self.weight_path}).eval().to(self.device)
@@ -28,7 +29,7 @@ def get_module_name(self):
         return "wireframe"
 
     def get_descinfo_fname(self, descinfo_folder, img_id):
-        fname = os.path.join(descinfo_folder, "descinfo_{0}.npz".format(img_id))
+        fname = os.path.join(descinfo_folder, f"descinfo_{img_id}.npz")
         return fname
 
     def save_descinfo(self, descinfo_folder, img_id, descinfo):