On Mon, Mar 29, 2010 at 2:31 AM, Tyler Erickson <[email protected]> wrote:
> Thanks for the ideas, which led me to
> http://effbot.org/zone/element-namespaces.htm . This page notes that the
> standard serializer passes through element names that use the prefix:tag
> notation. So I think a good approach will be to write elements with the
> prefix:tag notation, then iterate through the tree to find out what prefixes
> I have used, and then set the appropriate namespaces on the root element of
> the tree.
>
I had the same issue, except it was a requirement because we were
using ~8 namespaces and it got heller-confusing to look at a document
without sane prefixes... I did it a hack way. I doubt this would work
for really big docs, and performance will suck :)
HTH,
Rob :)
... root is the root element of my generated tree ...
namespaces = set()
for e in root.iter():
for ns in e.nsmap.values():
namespaces.add(ns)
if len(namespaces):
# build up useful prefixes...
nsmap = build_ns_map(namespaces)
# build a new root element with a different nsmap
new_root = etree.Element(root.tag, attrib=root.attrib, nsmap=nsmap)
new_root.text=root.text
new_root.tail=root.tail
for c in root:
new_root.append(c)
root = new_root
tree = etree.ElementTree(root)
tree.write(output_file, encoding='utf8', pretty_print=pretty)
def build_ns_map(namespaces):
nsmap = {}
for ns in namespaces:
# build_ns_prefix() gets or creates a sensible prefix for
a given namespace URL
prefix = build_ns_prefix(ns)
# having a non-populated prefix map is fine, etree will fill it in
# with ns# etc
if prefix and (prefix not in nsmap):
nsmap[prefix] = ns
return nsmap
To unsubscribe from this group, send email to
python-gis-sig+unsubscribegooglegroups.com or reply to this email with the
words "REMOVE ME" as the subject.
