# $Id$
# stuff to manipulate (X)HTML block structures

import cElementTree as ET

BLOCK_TAGS = (
    "p", "h1", "h2", "h3", "h4", "h5", "h6", "ul", "ol", "pre", "dl",
    "div", "noscript", "blockquote", "form", "hr", "table", "fieldset",
    "address"
)

try:
    BLOCK_TAGS = set(BLOCK_TAGS)
except NameError:
    pass

def isblock(elem):
    return elem.tag in BLOCK_TAGS

##
# Fixes tail block texts.  This function looks for block elements with
# trailing text contents, and moves that text (and following non-block
# elements) to a new block element.

def fix_tail(elem, tag="p"):
    for elem in elem.getiterator():
        out = []; se = None
        i = 0
        while i < len(elem):
            e = elem[i]
            out.append(e)
            i += 1
            if isblock(e) and e.tail and e.tail.strip():
                # found block element with tail text; insert new element
                se = ET.Element(tag)
                se.text = e.tail
                e.tail = None
                out.append(se)
                # move subsequent non-blocks to this element
                while i < len(elem) and not isblock(elem[i]):
                    se.append(elem[i])
                    i += 1
                continue
        if se is not None:
            elem[:] = out

# --------------------------------------------------------------------
#

def test(text):
    elem = ET.XML(text)
    fix_tail(elem)
    print ET.tostring(elem)

test("<div>one</div>")
test("<div><p>one</p>two</div>")

# FIXME: what about inline text in block contents, such as the
# "one" in the following example?
test("<div>one<p>two</p>three</div>")

# also see
# http://www.w3.org/TR/REC-CSS2/visuren.html#box-gen