A proof of concept hyphenator for HTML files using pyphen, based on https://github.com/dmalinovsky/kindle-hyphens/blob/master/main.py

hyphenate-poc.py

import codecs
import sys

import lxml.html
from lxml import etree
import pyphen

lang="en"

def insert_hyphens(node, hyphenator):
    textattrs = ('text', 'tail')
    if isinstance(node, etree._Entity):
        # HTML entities have no .text
        textattrs = ('tail',)
    for attr in textattrs:
        text = getattr(node, attr)
        if not text:
            continue
        new_data = ' '.join([hyphenator.inserted(w, hyphen=u'\u00AD')
            for w in text.split()])
        # Spaces are trimmed, we have to add them manually back
        if text[0].isspace():
            new_data = ' ' + new_data
        if text[-1].isspace():
            new_data += ' '
        setattr(node, attr, new_data)

    for child in node.iterchildren():
        insert_hyphens(child, hyphenator)

dom = lxml.html.parse(sys.argv[1])
hyphenator = pyphen.Pyphen(lang=lang)
for tag in ('p', 'div', 'li', 'span'):
    for node in dom.xpath("//%s" % tag):
        insert_hyphens(node, hyphenator)

with codecs.open('foo.html', 'w+', 'utf-8') as f:
    f.write(lxml.html.tostring(dom, encoding='unicode'))