怎么把html文件转换成xml（txt版本的标签转换成PASCAL）

1. from xml.dom import minidom,Node
2. import cv2
3. import string
4. res=r'D:\workstation\POD page object detection 2017\Train\Train\Image\\'
5. list = r'D:\workstation\POD page object detection 2017\Train\Train\Image\nameLIST.txt'
6. parsedLabel=r'D:\workstation\POD page object detection 2017\Train\Train\parsedLabel\\'
7. savePath = r'D:\workstation\POD page object detection 2017\Train\Train\VOCVMLLabel\\'
8. f = open(list,'r')
9. count = 0
10. lines = f.readlines()
11. for line in lines[:]:
12. name = line.strip()
13. name = "00" name[-8:-4] '.jpg'
14. im = cv2.imread(res line.strip(),1)
15. w = im.shape[1]
16. h = im.shape[0]
17. d = im.shape[2]
18. #print w,h,d
20. doc = minidom.Document()
22. annotation = doc.createElement('annotation')
23. doc.appendChild(annotation)
25. folder = doc.createElement('folder')
26. folder.appendChild(doc.createTextNode("POD2017"))
27. annotation.appendChild(folder)
29. filename = doc.createElement('filename')
30. filename.appendChild(doc.createTextNode(name))
31. annotation.appendChild(filename)
33. source = doc.createElement('source')
34. database = doc.createElement('database')
35. database.appendChild(doc.createTextNode("The POD2017 Database"))
36. source.appendChild(database)
37. annotation2 = doc.createElement('annotation')
38. annotation2.appendChild(doc.createTextNode("ICDAR POD2017"))
39. source.appendChild(annotation2)
40. image = doc.createElement('image')
41. image.appendChild(doc.createTextNode("image"))
42. source.appendChild(image)
43. flickrid = doc.createElement('flickrid')
44. flickrid.appendChild(doc.createTextNode("NULL"))
45. source.appendChild(flickrid)
46. annotation.appendChild(source)
48. owner = doc.createElement('owner')
49. flickrid = doc.createElement('flickrid')
50. flickrid.appendChild(doc.createTextNode("NULL"))
51. owner.appendChild(flickrid)
52. na = doc.createElement('name')
53. na.appendChild(doc.createTextNode("cxm"))
54. owner.appendChild(na)
55. annotation.appendChild(owner)
57. size = doc.createElement('size')
58. width = doc.createElement('width')
59. width.appendChild(doc.createTextNode("%d" %w))
60. size.appendChild(width)
61. height = doc.createElement('height')
62. height.appendChild(doc.createTextNode("%d" %h))
63. size.appendChild(height)
64. depth = doc.createElement('depth')
65. depth.appendChild(doc.createTextNode("%d" %d))
66. size.appendChild(depth)
67. annotation.appendChild(size)
69. segmented = doc.createElement('segmented')
70. segmented.appendChild(doc.createTextNode("0"))
71. annotation.appendChild(segmented)
73. txtLabel = open(parsedLabel name[:-4] '.txt','r')
74. boxes = txtLabel.readlines()
75. for box in boxes:
76. box = box.strip().split(' ')
77. object = doc.createElement('object')
78. nm = doc.createElement('name')
79. nm.appendChild(doc.createTextNode(box[0]))
80. object.appendChild(nm)
81. pose = doc.createElement('pose')
82. pose.appendChild(doc.createTextNode("undefined"))
83. object.appendChild(pose)
84. truncated = doc.createElement('truncated')
85. truncated.appendChild(doc.createTextNode("0"))
86. object.appendChild(truncated)
87. difficult = doc.createElement('difficult')
88. difficult.appendChild(doc.createTextNode("0"))
89. object.appendChild(difficult)
90. bndbox = doc.createElement('bndbox')
91. xmin = doc.createElement('xmin')
92. xmin.appendChild(doc.createTextNode(box[1]))
93. bndbox.appendChild(xmin)
94. ymin = doc.createElement('ymin')
95. ymin.appendChild(doc.createTextNode(box[2]))
96. bndbox.appendChild(ymin)
97. xmax = doc.createElement('xmax')
98. xmax.appendChild(doc.createTextNode(box[3]))
99. bndbox.appendChild(xmax)
100. ymax = doc.createElement('ymax')
101. ymax.appendChild(doc.createTextNode(box[4]))
102. bndbox.appendChild(ymax)
103. object.appendChild(bndbox)
104. annotation.appendChild(object)
105. savefile = open(savePath name[:-4] '.XML','w')
106. savefile.write(doc.toprettyxml(encoding='utf-8'))
107. savefile.close()
108. count = 1
109. print count

源格式 与 生成的格式 对比如下：

[python] view plain copy