add dataset path chaeck

paddlex/cv/datasets/dataset.py

@@ -55,6 +55,7 @@ def get_encoding(path):
     f = open(path, 'rb')
     data = f.read()
     file_encoding = chardet.detect(data).get('encoding')
+    f.close()
     return file_encoding
 
 

paddlex/cv/datasets/easydata_cls.py

@@ -14,6 +14,7 @@
 
 from __future__ import absolute_import
 import os.path as osp
+import platform
 import random
 import copy
 import json
@@ -64,10 +65,18 @@ class EasyDataCls(ImageNet):
                 item = line.strip()
                 self.labels.append(item)
         logging.info("Starting to read file list from dataset...")
+        win_sep = "\\"
+        other_sep = "/"
         with open(file_list, encoding=get_encoding(file_list)) as f:
             for line in f:
                 img_file, json_file = [osp.join(data_dir, x) \
                         for x in line.strip().split()[:2]]
+                if platform.system() == "Windows":
+                    img_file = win_sep.join(img_file.split(other_sep))
+                    json_file = win_sep.join(json_file.split(other_sep))
+                else:
+                    img_file = other_sep.join(img_file.split(win_sep))
+                    json_file = other_sep.join(json_file.split(win_sep))
                 if not is_pic(img_file):
                     continue
                 if not osp.isfile(json_file):

paddlex/cv/datasets/easydata_det.py

@@ -14,6 +14,7 @@
 
 from __future__ import absolute_import
 import os.path as osp
+import platform
 import random
 import copy
 import json
@@ -83,10 +84,18 @@ class EasyDataDet(VOCDetection):
         from pycocotools.mask import decode
         ct = 0
         ann_ct = 0
+        win_sep = "\\"
+        other_sep = "/"
         with open(file_list, encoding=get_encoding(file_list)) as f:
             for line in f:
                 img_file, json_file = [osp.join(data_dir, x) \
                         for x in line.strip().split()[:2]]
+                if platform.system() == "Windows":
+                    img_file = win_sep.join(img_file.split(other_sep))
+                    json_file = win_sep.join(json_file.split(other_sep))
+                else:
+                    img_file = other_sep.join(img_file.split(win_sep))
+                    json_file = other_sep.join(json_file.split(win_sep))
                 if not is_pic(img_file):
                     continue
                 if not osp.isfile(json_file):

paddlex/cv/datasets/easydata_seg.py

@@ -14,6 +14,7 @@
 
 from __future__ import absolute_import
 import os.path as osp
+import platform
 import random
 import copy
 import json
@@ -61,6 +62,8 @@ class EasyDataSeg(Dataset):
         from pycocotools.mask import decode
         cname2cid = {}
         label_id = 0
+        win_sep = "\\"
+        other_sep = "/"
         with open(label_list, encoding=get_encoding(label_list)) as fr:
             for line in fr.readlines():
                 cname2cid[line.strip()] = label_id
@@ -71,6 +74,12 @@ class EasyDataSeg(Dataset):
             for line in f:
                 img_file, json_file = [osp.join(data_dir, x) \
                         for x in line.strip().split()[:2]]
+                if platform.system() == "Windows":
+                    img_file = win_sep.join(img_file.split(other_sep))
+                    json_file = win_sep.join(json_file.split(other_sep))
+                else:
+                    img_file = other_sep.join(img_file.split(win_sep))
+                    json_file = other_sep.join(json_file.split(win_sep))
                 if not is_pic(img_file):
                     continue
                 if not osp.isfile(json_file):

paddlex/cv/datasets/imagenet.py

@@ -14,6 +14,7 @@
 
 from __future__ import absolute_import
 import os.path as osp
+import platform
 import random
 import copy
 import paddlex.utils.logging as logging
@@ -63,9 +64,15 @@ class ImageNet(Dataset):
                 item = line.strip()
                 self.labels.append(item)
         logging.info("Starting to read file list from dataset...")
+        win_sep = "\\"
+        other_sep = "/"
         with open(file_list, encoding=get_encoding(file_list)) as f:
             for line in f:
                 items = line.strip().split()
+                if platform.system() == "Windows":
+                    items[0] = win_sep.join(items[0].split(other_sep))
+                else:
+                    items[0] = other_sep.join(items[0].split(win_sep))
                 if not is_pic(items[0]):
                     continue
                 full_path = osp.join(data_dir, items[0])

paddlex/cv/datasets/seg_dataset.py

@@ -61,10 +61,17 @@ class SegDataset(Dataset):
                 for line in f:
                     item = line.strip()
                     self.labels.append(item)
-
+        win_sep = "\\"
+        other_sep = "/"
         with open(file_list, encoding=get_encoding(file_list)) as f:
             for line in f:
                 items = line.strip().split()
+                if platform.system() == "Windows":
+                    items[0] = win_sep.join(items[0].split(other_sep))
+                    items[1] = win_sep.join(items[1].split(other_sep))
+                else:
+                    items[0] = other_sep.join(items[0].split(win_sep))
+                    items[1] = other_sep.join(items[1].split(win_sep))
                 if not is_pic(items[0]):
                     continue
                 full_path_im = osp.join(data_dir, items[0])

paddlex/cv/datasets/voc.py

@@ -16,6 +16,7 @@ from __future__ import absolute_import
 import copy
 import os
 import os.path as osp
+import platform
 import random
 import re
 import numpy as np
@@ -85,6 +86,8 @@ class VOCDetection(Dataset):
             })
         ct = 0
         ann_ct = 0
+        win_sep = "\\"
+        other_sep = "/"
         with open(file_list, 'r', encoding=get_encoding(file_list)) as fr:
             while True:
                 line = fr.readline()
@@ -92,6 +95,12 @@ class VOCDetection(Dataset):
                     break
                 img_file, xml_file = [osp.join(data_dir, x) \
                         for x in line.strip().split()[:2]]
+                if platform.system() == "Windows":
+                    img_file = win_sep.join(img_file.split(other_sep))
+                    xml_file = win_sep.join(xml_file.split(other_sep))
+                else:
+                    img_file = other_sep.join(img_file.split(win_sep))
+                    xml_file = other_sep.join(xml_file.split(win_sep))
                 if not is_pic(img_file):
                     continue
                 if not osp.isfile(xml_file):
@@ -106,8 +115,11 @@ class VOCDetection(Dataset):
                     ct = int(tree.find('id').text)
                     im_id = np.array([int(tree.find('id').text)])
                 pattern = re.compile('<object>', re.IGNORECASE)
-                obj_tag = pattern.findall(
-                    str(ET.tostringlist(tree.getroot())))[0][1:-1]
+                obj_match = pattern.findall(
+                    str(ET.tostringlist(tree.getroot())))
+                if len(obj_match) == 0:
+                    continue
+                obj_tag = obj_match[0][1:-1]
                 objs = tree.findall(obj_tag)
                 pattern = re.compile('<size>', re.IGNORECASE)
                 size_tag = pattern.findall(

paddlex/tools/x2coco.py

@@ -18,6 +18,7 @@ import cv2
 import json
 import os
 import os.path as osp
+import platform
 import shutil
 import numpy as np
 import PIL.ImageDraw
@@ -100,6 +101,12 @@ class LabelMe2COCO(X2COCO):
         image["height"] = json_info["imageHeight"]
         image["width"] = json_info["imageWidth"]
         image["id"] = image_id + 1
+        win_sep = "\\"
+        other_sep = "/"
+        if platform.system() == "Windows":
+            json_info["imagePath"] = win_sep.join(json_info["imagePath"].split(other_sep))
+        else:
+            json_info["imagePath"] = other_sep.join(json_info["imagePath"].split(win_sep))
         image["file_name"] = osp.split(json_info["imagePath"])[-1]
         return image
     
@@ -187,6 +194,12 @@ class EasyData2COCO(X2COCO):
         image["height"] = img.shape[0]
         image["width"] = img.shape[1]
         image["id"] = image_id + 1
+        win_sep = "\\"
+        other_sep = "/"
+        if platform.system() == "Windows":
+            img_path = win_sep.join(img_path.split(other_sep))
+        else:
+            img_path = other_sep.join(img_path.split(win_sep))
         image["file_name"] = osp.split(img_path)[-1]
         return image
     
@@ -268,6 +281,12 @@ class JingLing2COCO(X2COCO):
         image["height"] = json_info["size"]["height"]
         image["width"] = json_info["size"]["width"]
         image["id"] = image_id + 1
+        win_sep = "\\"
+        other_sep = "/"
+        if platform.system() == "Windows":
+            json_info["path"] = win_sep.join(json_info["path"].split(other_sep))
+        else:
+            json_info["path"] = other_sep.join(json_info["path"].split(win_sep))
         image["file_name"] = osp.split(json_info["path"])[-1]
         return image