如何使用 python 修改 xml 文件中嵌套元素的文本？

Question

Currently I'm working on a corpus/dataset. It's in xml format as you can see the picture below. I'm facing a problem. I want to access all 'ne' elements one by one as shown in below picture. Then I want to access the text of the 'W' elements which are inside the 'ne' elements. Then I want to concatenate thy symbols 'SDi' and 'EDi' with the text of these 'W' elements. 'i' can take any positive whole number starting from 1. In the case of 'SDi' I need only the text of first 'W' element that is inside the 'ne' element. In the case of 'EDi' I need only the text of last 'W' element that is inside the 'ne' element. Currently I don't get anything as output after running the code. I think this is because of the fact that the element 'W' is never accessed. Moreover, i think that element 'W' is not accessed because it is a grandchild of element 'ne' therefore it can't be accessed directly rather it may be possible with the help its father node.

Note1: The number and names of sub elements inside 'ne' elements are not same.

Note2: Only those things are explained here which needed. You may find some other details in the coding/picture but ignore them.

I'm using Spyder (python 3.6) Any help would be appreciated.

A picture from the XML file I'm working on is given below:

Text version of XML file: Click here

Sample/Expected output image (below):

Coding I've done so far:

for i in range(len(List_of_root_nodes)):
true_false = True
current = List_of_root_nodes[i]
start_ID = current.PDante_ID
#print('start:', start_ID)  # For Testing
end_ID = None
number = str(i+1)  # This number will serve as i used with SD and ED that is (SDi and EDi)

discourse_starting_symbol = "SD" + number
discourse_ending_symbol = "ED" + number

while true_false:    
    if current.right_child is None:        
        end_ID = current.PDante_ID
        #print('end:', end_ID)  # For Testing
        true_false = False        
    else:        
        current = current.right_child

# Finding 'ne' element with id='start_ID'
ne_text = None
ne_id = None

for ne in myroot.iter('ne'):    
    ne_id = ne.get('id')

    # If ne_id matches with start_ID means the place where SDi is to be placed is found    
    if ne_id == start_ID:        
        for w in ne.iter('W'):            
            ne_text = str(w.text)            
            boundary_and_text = " " + str(discourse_starting_symbol) + " " + ne_text
            w.text = boundary_and_text
            break

    # If ne_id matches with end_ID means the place where EDi is to be placed is found

    # Some changes Required here: Here the 'EDi' will need to be placed after the last 'W' element.
    # So last 'W' element needs to be accessed
    if ne_id == end_ID:        
        for w in ne.iter('W'):            
            ne_text = str(w.text)            
            boundary_and_text = ne_text + " " + str(discourse_ending_symbol) + " "
            w.text = boundary_and_text
            break

Answer 1

Whenever you need to modify XML with various nuanced needs, consider XSLT , the special-purpose language designed to transform XML files. You can run XSLT 1.0 scripts with Python's third-party module, lxml (not built-in etree ).

Specifically, call the identity transform to copy XML as is and then run the two templates to add SDI to text of very first <W> and very last EDI to text of last <W> . Solution will work if there are 10 or 10,000 <W> nodes, deeply nested or not.

To demonstrate with example data of StackOverflow's top Python and XSLT users, see online demo where SDI and EDI are added to first and last <user> node:

XSLT (save as.xsl file, a special.xml file to be loaded in Python)

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:output indent="yes"/>
  <xsl:strip-space elements="*"/>

  <!-- IDENTITY TRANSFORM -->    
  <xsl:template match="@*|node()">
    <xsl:copy>
      <xsl:apply-templates select="@*|node()"/>
    </xsl:copy>
  </xsl:template>

  <!-- EDIT FIRST W NODE -->    
  <xsl:template match="W[count(preceding::W)=0]">
    <xsl:copy>
      <xsl:copy-of select="@*"/>
      <xsl:value-of select="concat('SDI ', text())"/>
    </xsl:copy>
  </xsl:template>

  <!-- EDIT LAST W NODE -->    
  <xsl:template match="W[count(preceding::W)+1 = count(//W)]">
    <xsl:copy>
      <xsl:copy-of select="@*"/>
      <xsl:value-of select="concat('EDI ', text())"/>
    </xsl:copy>
  </xsl:template>

</xsl:stylesheet>

Python (no loops or if/else logic)

import lxml.etree as et

doc = et.parse('/path/to/Input.xml')
xsl = et.parse('/path/to/Script.xsl')

# CONFIGURE TRANSFORMER
transform = et.XSLT(xsl)    

# TRANSFORM SOURCE DOC
result = transform(doc)

# OUTPUT TO CONSOLE
print(result)

# SAVE TO FILE
with open('Output.xml', 'wb') as f:
    f.write(result)

Answer 2

Something like this (a.xml is the XML you have uploaded):

Note the code is not using any external library.

import xml.etree.ElementTree as ET

SD = 'SD'
ED = 'ED'

root = ET.parse('a.xml')

counter = 1

for ne in root.findall('.//ne'):
    w_lst = ne.findall('.//W')
    if w_lst:
        w_lst[0].text = '{}{} {}'.format(SD, counter, w_lst[0].text)
        if len(w_lst) > 1:
            w_lst[-1].text = '{} {}{}'.format(w_lst[-1].text, ED, counter)
        counter += 1
ET.dump(root)