Tidy.py

#!/usr/bin/env python

from pyquery import PyQuery as pq
import codecs

def convert_image(i, node):
    print node
    node = pq(node)
    uri = node.attr("src")
    if uri:
        print "converting", uri
        src = "".replace("M", "X")
        node.attr("src", src)

with codecs.open("old-html-tmp/Modelica.Fluid.html", encoding="utf-8") as fh:
    html = fh.read()

doc = pq(html)
doc.find("img").each(convert_image)

print doc.__html__() # __html__ method avoids escaping -- TODO: retain doctype