Beautiful Soup 4 Custom Attribute Order Output - python-3.x

Question

I want to create a custom output formatter in BS4 that will rearrange the order of attributes of tags in an XML in a specific way this is not alphabetical order.
For instance, I want to output the following tag:
<word form="συ" head="2610" id="2357" lemma="συ" postag="p-s----n-" relation="ExD_AP"/>
as:
<word id="2357" head="2610" postag="p-s----n-" form="συ" lemma="συ" relation="ExD_AP"/>
BS4's documentation offers a clue as to where to begin. They give the following example:
from bs4.formatter import HTMLFormatter
class UnsortedAttributes(HTMLFormatter):
def attributes(self, tag):
for k, v in tag.attrs.items():
if k == 'm':
continue
yield k, v
print(attr_soup.p.encode(formatter=UnsortedAttributes()))
This will make a custom HTML output formatter that will leave attributes in the order they were input and also ignore certain tags, but I don't know how to alter this so that it will output in whatever order I would like. Can anyone help me out?

Answer 1

How about this?
from simplified_scrapy.simplified_doc import SimplifiedDoc
html ='''
<word form="συ" head="2610" id="2357" lemma="συ" postag="p-s----n-" relation="ExD_AP"/>
'''
def toString(ele):
order = ['id','head','postag','from','lemma','relation']
result = '<'+ele.tag
for p in order:
result+=' {}="{}"'.format(p,ele[p])
return result+'/>'
doc = SimplifiedDoc(html)
ele = doc.word
print (toString(ele))
Result:
<word id="2357" head="2610" postag="p-s----n-" from="None" lemma="συ" relation="ExD_AP"/>

Answer 2

Strictly speaking, I have an answer to my own question, but it's going to take more work to actually implement it in a way I'd like. Here's how to do it.
Make a sub-class of the XMLFormatter (or HTMLFormatter if you're working with HTML), name it what you want. I chose "SortAttributes." Write the function "attributes" so that it will return a list of tuples: [(attribute1, value1), (attribute2, value2), etc.] in the order you want. Mine may look verbose, but I do it this way because I work with very inconsistent XML.
from bs4 import BeautifulSoup
from bs4.formatter import XMLFormatter
class SortAttributes(XMLFormatter):
def attributes(self, tag):
"""Reorder a tag's attributes however you want."""
attrib_order = ['id', 'head', 'postag', 'relation', 'form', 'lemma']
new_order = []
for element in attrib_order:
if element in tag.attrs:
new_order.append((element, tag[element]))
for pair in tag.attrs.items():
if pair not in new_order:
new_order.append(pair)
return new_order
xml_string = '''
<word form="συ" head="2610" id="2357" lemma="συ" postag="p-s----n-" relation="ExD_AP"/>
'''
soup = BeautifulSoup(xml_string, 'xml')
print(soup.encode(formatter=SortAttributes()))
This will output what I want:
<word id="2357" head="2610" postag="p-s----n-" relation="ExD_AP" form="συ" lemma="συ"/>
Conveniently, I can do this for an entire document with the same encode method. But if I write that to a file as a string, then all the tags are just strung together end to end. A sample would be like such:
<sentence id="783"><word id="2357" head="2610" postag="p-s----n-" relation="ExD_AP" form="συ" lemma="συ"/><word id="2358" head="2610" postag="p-s----n-" relation="ExD_AP" form="συ" lemma="συ"/><word id="2359" head="2610" postag="p-s----n-" relation="ExD_AP" form="συ" lemma="συ"/></sentence>
Instead of something I'd prefer:
<sentence id="783">
<word id="2357" head="2610" postag="p-s----n-" relation="ExD_AP" form="συ" lemma="συ"/>
<word id="2358" head="2610" postag="p-s----n-" relation="ExD_AP" form="συ" lemma="συ"/>
<word id="2359" head="2610" postag="p-s----n-" relation="ExD_AP" form="συ" lemma="συ"/>
</sentence>
To fix that, I can't just .prettify it because prettify rearranges the attributes back to alphabetical order. I'll have to go into more details with the XMLFormatter subclass instead. I hope someone finds this helpful in the future!