# $Id$
# get inner text from HTML element

##
# Extracts inner text from an HTML element.  This version attempts to
# handle newlines according to the standard HTML rule, where newlines
# just after a start tag, and just before an end tag, are ignored.
#
# @param elem HTML element

def html_inner_text(elem):
    # brute-force solution
    # 1) synthesize parser event stream
    START, END = range(2)
    def add_events(elem, events):
        events.append(START)
        if elem.text:
            events.append(elem.text)
        for e in elem:
            add_events(e, events)
            if e.tail:
                events.append(e.tail)
        events.append(END)
        return events
    events = []
    add_events(elem, events)
    # 2) clean up text fragments
    for i, e in enumerate(events):
        if not isinstance(e, basestring):
            continue
        if i and events[i-1] == START and e.startswith("\n"):
            events[i] = events[i][1:]
        if i < len(events)-1 and events[i+1] == END and e.endswith("\n"):
            events[i] = events[i][:-1]
    # return inner text
    return "".join([word for word in events if isinstance(word, basestring)])

# --------------------------------------------------------------------
# demo

import cElementTree as ET

HTML = """\
<p>
some
<b>
bold
</b>
text
</p>
"""

tree = ET.XML(HTML)

def test1():
    return html_inner_text(tree)

print repr(test1())