Fix errors detected by Ruff (#3918)

### What problem does this PR solve?

Fix errors detected by Ruff

### Type of change

- [x] Refactoring

deepdoc/vision/__init__.py

@@ -18,7 +18,6 @@ from .recognizer import Recognizer
 from .layout_recognizer import LayoutRecognizer
 from .table_structure_recognizer import TableStructureRecognizer
 
-
 def init_in_out(args):
     from PIL import Image
     import os
@@ -47,7 +46,7 @@ def init_in_out(args):
         try:
             images.append(Image.open(fnm))
             outputs.append(os.path.split(fnm)[-1])
-        except Exception as e:
+        except Exception:
             traceback.print_exc()
 
     if os.path.isdir(args.inputs):
@@ -56,6 +55,16 @@ def init_in_out(args):
     else:
         images_and_outputs(args.inputs)
 
-    for i in range(len(outputs)): outputs[i] = os.path.join(args.output_dir, outputs[i])
+    for i in range(len(outputs)):
+        outputs[i] = os.path.join(args.output_dir, outputs[i])
 
-    return images, outputs
+    return images, outputs
+
+
+__all__ = [
+    "OCR",
+    "Recognizer",
+    "LayoutRecognizer",
+    "TableStructureRecognizer",
+    "init_in_out",
+]

deepdoc/vision/layout_recognizer.py

@@ -42,7 +42,7 @@ class LayoutRecognizer(Recognizer):
                     get_project_base_directory(),
                     "rag/res/deepdoc")
             super().__init__(self.labels, domain, model_dir)
-        except Exception as e:
+        except Exception:
             model_dir = snapshot_download(repo_id="InfiniFlow/deepdoc",
                                           local_dir=os.path.join(get_project_base_directory(), "rag/res/deepdoc"),
                                           local_dir_use_symlinks=False)
@@ -77,7 +77,7 @@ class LayoutRecognizer(Recognizer):
                     "page_number": pn,
                     } for b in lts if float(b["score"]) >= 0.8 or b["type"] not in self.garbage_layouts]
             lts = self.sort_Y_firstly(lts, np.mean(
-                [l["bottom"] - l["top"] for l in lts]) / 2)
+                [lt["bottom"] - lt["top"] for lt in lts]) / 2)
             lts = self.layouts_cleanup(bxs, lts)
             page_layout.append(lts)
 

deepdoc/vision/ocr.py

@@ -19,7 +19,9 @@ from huggingface_hub import snapshot_download
 
 from api.utils.file_utils import get_project_base_directory
 from .operators import *
+import math
 import numpy as np
+import cv2
 import onnxruntime as ort
 
 from .postprocess import build_post_process
@@ -484,7 +486,7 @@ class OCR(object):
                         "rag/res/deepdoc")
                 self.text_detector = TextDetector(model_dir)
                 self.text_recognizer = TextRecognizer(model_dir)
-            except Exception as e:
+            except Exception:
                 model_dir = snapshot_download(repo_id="InfiniFlow/deepdoc",
                                               local_dir=os.path.join(get_project_base_directory(), "rag/res/deepdoc"),
                                               local_dir_use_symlinks=False)

deepdoc/vision/operators.py

@@ -232,7 +232,7 @@ class LinearResize(object):
         """
         assert len(self.target_size) == 2
         assert self.target_size[0] > 0 and self.target_size[1] > 0
-        im_channel = im.shape[2]
+        _im_channel = im.shape[2]
         im_scale_y, im_scale_x = self.generate_scale(im)
         im = cv2.resize(
             im,
@@ -255,7 +255,7 @@ class LinearResize(object):
             im_scale_y: the resize ratio of Y
         """
         origin_shape = im.shape[:2]
-        im_c = im.shape[2]
+        _im_c = im.shape[2]
         if self.keep_ratio:
             im_size_min = np.min(origin_shape)
             im_size_max = np.max(origin_shape)
@@ -581,7 +581,7 @@ class SRResize(object):
             return data
 
         images_HR = data["image_hr"]
-        label_strs = data["label"]
+        _label_strs = data["label"]
         transform = ResizeNormalize((imgW, imgH))
         images_HR = transform(images_HR)
         data["img_hr"] = images_HR

deepdoc/vision/postprocess.py

@@ -121,7 +121,7 @@ class DBPostProcess(object):
         outs = cv2.findContours((bitmap * 255).astype(np.uint8), cv2.RETR_LIST,
                                 cv2.CHAIN_APPROX_SIMPLE)
         if len(outs) == 3:
-            img, contours, _ = outs[0], outs[1], outs[2]
+            _img, contours, _ = outs[0], outs[1], outs[2]
         elif len(outs) == 2:
             contours, _ = outs[0], outs[1]
 

deepdoc/vision/recognizer.py

@@ -13,15 +13,18 @@
 
 import logging
 import os
+import math
+import numpy as np
+import cv2
 from copy import deepcopy
 
+
 import onnxruntime as ort
 from huggingface_hub import snapshot_download
 
 from api.utils.file_utils import get_project_base_directory
 from .operators import *
 
-
 class Recognizer(object):
     def __init__(self, label_list, task_name, model_dir=None):
         """
@@ -277,7 +280,8 @@ class Recognizer(object):
             return
         min_dis, min_i = 1000000, None
         for i,b in enumerate(boxes):
-            if box.get("layoutno", "0") != b.get("layoutno", "0"): continue
+            if box.get("layoutno", "0") != b.get("layoutno", "0"):
+                continue
             dis = min(abs(box["x0"] - b["x0"]), abs(box["x1"] - b["x1"]), abs(box["x0"]+box["x1"] - b["x1"] - b["x0"])/2)
             if dis < min_dis:
                 min_i = i
@@ -402,7 +406,8 @@ class Recognizer(object):
         scores = np.max(boxes[:, 4:], axis=1)
         boxes = boxes[scores > thr, :]
         scores = scores[scores > thr]
-        if len(boxes) == 0: return []
+        if len(boxes) == 0:
+            return []
 
         # Get the class with the highest confidence
         class_ids = np.argmax(boxes[:, 4:], axis=1)
@@ -432,7 +437,8 @@ class Recognizer(object):
         for i in range(len(image_list)):
             if not isinstance(image_list[i], np.ndarray):
                 imgs.append(np.array(image_list[i]))
-            else: imgs.append(image_list[i])
+            else:
+                imgs.append(image_list[i])
 
         batch_loop_cnt = math.ceil(float(len(imgs)) / batch_size)
         for i in range(batch_loop_cnt):