How to tokenize a sentence in the XML and create new child nodes?

Question

I have XML which looks like this:

<para id="0">
    <se lang="hi">काकेशिया में तब लड़ाई</se>
    <se lang="ru">потом боевые действия на Кавказе</se>
</para>
<para id="1">
...
</para>
<para id="2">
...
</para>

and I want to tokenize the devanagari text by using iNLTK library and get a file which looks like this:

<para id="0">
    <se lang="hi">
        <w>काकेशिया</w> 
        <w>में</w> 
        <w>तब</w>
        <w>लड़ाई</w> 
    </se>
    <se lang="ru">потом боевые действия на Кавказе</se>
</para>
<para id="1">
...
</para>
<para id="2">
...
</para>

I understand how to tokenize the sentence:

paras = body.getElementsByTagName('para')
for para in paras:
    devanagari = para.getElementsByTagName('se')[1].childNodes[0].nodeValue
    print(tokenize(devanagari, 'hi'))

but what I don't know is how to make childnodes xml <w>...</w> for each word and write it into the XML

How can I do that by using xml.etree.ElementTree?