Python XML 转换为 JSON数据

一时失言乱红尘 2023-02-13 14:40 72阅读 0赞

## Json标注格式 ##

// bool：必须小写的true和false
    { "bool":true}
    // null:必须小写的null
    { "object":null}

*  JSON 规定字符集是UTF-8，字符串必须使用`""`，Object 的键也必须使用`""`
 *  数组或对象的最后一个成员，不能加 **逗号**

数组方式 \[ \]

对象方法 \{ \}

![json数据格式][json]

## XML解析转JSON ##

[如何将JSON数据写入文件][JSON]

*  ensure\_ascii：默认值True，如果dict内含有non-ASCII的字符，则会类似\\uXXXX的显示数据，设置成False后，就能正常显示
 *  indent：应该是一个非负的整型，如果是0，或者为空，则一行显示数据，否则会换行且按照indent的数量显示前面的空白，这样打印出来的json数据也叫pretty-printed json
 *  encoding：默认是UTF-8，设置json数据的编码方式

> track的父节点是annotations

![格式][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzM4Njg5Mzk1_size_16_color_FFFFFF_t_70]  
主要思路：把XML解析成Python对象，之后将其追加到列表或字典中，再用json.dump输出

import untangle
    import os
    import json
    
    root_path = 'C:/Users/Desktop/Xml'
    
    files = os.listdir(root_path)
    for file in files:
      output_dict = { }
      num = 100
      for i in range(num):
        output_dict[i] = { 
          'frame':i,
          'labels':[]
        }
    
      xml_path = os.path.join(root_path, file)
      obj = untangle.parse(xml_path)
      for track in obj.annotations.track:
        for box in track.box:
          box_json = { }
          frame_number = int(box['frame'])
          box_json['category'] = track['label']
          box_json['box2d'] = { 
            'x1':float(box['xtl']),
            'y1':float(box['ytl']),
            'x2':float(box['xbr']),
            'y2':float(box['ybr'])
          }
          output_dict[frame_number]['labels'].append(box_json)
    
      output_list = []
      for i in range(num):
        output_list.append(output_dict[i])
        
      output_path = os.path.join(root_path, file.split('.')[0] + '.json')
      with open(output_path, 'w', encoding='utf-8') as outfile:
        json.dump(output_list, outfile, ensure_ascii=False,indent=2)

![frame][]

[json]: https://img-blog.csdnimg.cn/20200529153225676.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzM4Njg5Mzk1,size_16,color_FFFFFF,t_70
[JSON]: https://codenong.com/12309269/
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzM4Njg5Mzk1_size_16_color_FFFFFF_t_70]: https://img-blog.csdnimg.cn/20200509102248961.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzM4Njg5Mzk1,size_16,color_FFFFFF,t_70
[frame]: https://img-blog.csdnimg.cn/20200529160820582.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzM4Njg5Mzk1,size_16,color_FFFFFF,t_70