def _climb_and_split(src, dest, chapters):
for child in src.iterchildren():
if child.tag == 'hr' and child.get('class') == MARKER_CLASS:
log('got a marker')
new = copy_element(src, lxml.html.Element)
#build a new tree to this point
root = new
for a in src.iterancestors():
a2 = copy_element(a, root.makeelement)
a2.append(root)
root = a2
chapters.append(root)
#trim the tail of the finished one.
dest.tail = None
for a in dest.iterancestors():
a.tail = None
#now the new tree is the destination
dest = new
else:
new = copy_element(child, dest.makeelement)
new.text = child.text
dest.append(new)
new2 = _climb_and_split(child, new, chapters)
if new2 != new:
dest = new2.getparent()
return dest
def _climb_and_split(src, dest, chapters):
for child in src.iterchildren():
if child.tag == 'hr' and child.get('class') == MARKER_CLASS:
log('got a marker')
new = copy_element(src, lxml.html.Element)
#build a new tree to this point
root = new
for a in src.iterancestors():
a2 = copy_element(a, root.makeelement)
a2.append(root)
root = a2
chapters.append(root)
#trim the tail of the finished one.
dest.tail = None
for a in dest.iterancestors():
a.tail = None
#now the new tree is the destination
dest = new
else:
new = copy_element(child, dest.makeelement)
new.text = child.text
dest.append(new)
new2 = _climb_and_split(child, new, chapters)
if new2 != new:
dest = new2.getparent()
return dest
def jostle_markers(root):
"""If a marker is not separating block level elements, try to
move it out until it is, without completely ruining everything."""
stacks = []
for hr in root.iter(tag='hr'):
if hr.get('class') == MARKER_CLASS:
stack = frozenset(x for x in hr.iterancestors())
stacks.append((hr, stack))
for i, (hr, stack) in enumerate(stacks):
if hr.get('class') == MARKER_CLASS:
while True:
parent = hr.getparent()
log('i is %s hr is %s, parent is %s' % (i, hr, parent))
if parent.tag in ('html', 'body'):
log('hit body')
break
#don't allow two stacks to merge
if ((i > 0 and parent in stacks[i - 1][1]) or
(i + 1 < len(stacks) and parent in stacks[i + 1][1])):
log('hit neighbour')
break
#unless hr is right before the closing tag, don't jump
#out of div, center or blockquote.
if (parent.tag in INESCAPABLE_TAGS
and not (hr.getnext() is None and not hr.tail)):
log('hit %s' % parent.tag)
break
parent.addnext(hr)
continue
def jostle_markers(root):
"""If a marker is not separating block level elements, try to
move it out until it is, without completely ruining everything."""
stacks = []
for hr in root.iter(tag='hr'):
if hr.get('class') == MARKER_CLASS:
stack = frozenset(x for x in hr.iterancestors())
stacks.append((hr, stack))
for i, (hr, stack) in enumerate(stacks):
if hr.get('class') == MARKER_CLASS:
while True:
parent = hr.getparent()
log('i is %s hr is %s, parent is %s' %(i, hr, parent))
if parent.tag in ('html', 'body'):
log('hit body')
break
#don't allow two stacks to merge
if ((i > 0 and parent in stacks[i - 1][1]) or
(i + 1 < len(stacks) and parent in stacks[i + 1][1])):
log('hit neighbour')
break
#unless hr is right before the closing tag, don't jump
#out of div, center or blockquote.
if (parent.tag in INESCAPABLE_TAGS and
not (hr.getnext() is None and not hr.tail)):
log('hit %s' % parent.tag)
break
parent.addnext(hr)
continue
def split_file(fn, splitter):
f = open(fn)
html = f.read()
f.close()
compressed_size = get_compressed_size(html)
splits = max(compressed_size // config.EPUB_COMPRESSED_SIZE_MAX,
len(html) // config.EPUB_FILE_SIZE_MAX)
log("uncompressed: %s, compressed: %s, splits: %s" %
(len(html), compressed_size, splits))
if splits:
target = len(html) // (splits + 1)
s = 0
fragments = []
for i in range(splits):
e = html.find('<', target * (i + 1))
fragments.append(html[s:e])
fragments.append('<hr class="%s" id="split_%s" />' %
(MARKER_CLASS, i))
s = e
fragments.append(html[s:])
log([len(x) for x in fragments])
tree = lxml.html.fromstring(''.join(fragments))
jostle_markers(tree)
html2 = etree.tostring(tree, encoding='UTF-8', method='html')
f = open('/tmp/marked.html', 'w')
f.write(html2)
f.close()
t = time.time()
chapters, name = splitter(tree)
print "%s took %s" % (splitter, time.time() - t)
log(chapters)
for i, c in enumerate(chapters):
f = open('/tmp/%s_%s.html' % (
name,
i + 1,
), 'w')
f.write(etree.tostring(c, encoding='UTF-8', method='html'))
f.close()
def split_file(fn, splitter):
f = open(fn)
html = f.read()
f.close()
compressed_size = get_compressed_size(html)
splits = max(compressed_size // config.EPUB_COMPRESSED_SIZE_MAX,
len(html) // config.EPUB_FILE_SIZE_MAX)
log("uncompressed: %s, compressed: %s, splits: %s" % (len(html), compressed_size, splits))
if splits:
target = len(html) // (splits + 1)
s = 0
fragments = []
for i in range(splits):
e = html.find('<', target * (i + 1))
fragments.append(html[s:e])
fragments.append('<hr class="%s" id="split_%s" />' % (MARKER_CLASS, i))
s = e
fragments.append(html[s:])
log([len(x) for x in fragments])
tree = lxml.html.fromstring(''.join(fragments))
jostle_markers(tree)
html2 = etree.tostring(tree, encoding='UTF-8', method='html')
f = open('/tmp/marked.html', 'w')
f.write(html2)
f.close()
t = time.time()
chapters, name = splitter(tree)
print "%s took %s" % (splitter, time.time() - t)
log(chapters)
for i, c in enumerate(chapters):
f = open('/tmp/%s_%s.html' % (name, i + 1,), 'w')
f.write(etree.tostring(c, encoding='UTF-8', method='html'))
f.close()