generate a full XML file with python

Question

I need to machine generate (not to parse!) a (possibly complex) XML file in Python.

I am (relatively) familiar with xml and lxml modules, but it's unclear to me how to generate a valid XML checked against an xsl spec.

What I need to build is something like:

<?xml version='1.0' encoding='utf-8'?>
<package xmlns="http://www.idpf.org/2007/opf" unique-identifier="uuid_id" version="2.0">
    <metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf">
        <dc:identifier opf:scheme="calibre" id="calibre_id">4117</dc:identifier>
        <dc:identifier opf:scheme="uuid" id="uuid_id">d06a2234-67b4-40db-8f4a-136e52057101</dc:identifier>
        <dc:title>La Fine Di Alice</dc:title>
        <dc:creator opf:file-as="Homes, A. M." opf:role="aut">A. M. Homes</dc:creator>
        <dc:contributor opf:file-as="calibre" opf:role="bkp">calibre (3.10.0) [https://calibre-ebook.com]</dc:contributor>
        <dc:date>2005-11-15T00:00:00+00:00</dc:date>
        <dc:publisher>Minimum Fax</dc:publisher>
        <dc:identifier opf:scheme="ISBN">9788875210649</dc:identifier>
        <dc:language>en</dc:language>
        <meta content="{&quot;A. M. Homes&quot;: &quot;&quot;}" name="calibre:author_link_map"/>
        <meta content="2017-11-07T07:34:41.217796+00:00" name="calibre:timestamp"/>
        <meta content="La Fine Di Alice" name="calibre:title_sort"/>
    </metadata>
    <guide>
        <reference href="cover.jpg" title="Cover" type="cover"/>
    </guide>
</package>

The complete grammar is here.

I tried something along the lines:

from lxml import etree as ET

et = ET.Element('package', attrib={'version': "2.0", 'xmlns': "http://www.idpf.org/2007/opf", 'unique-identifier': "BookId"})
md = ET.SubElement(et, 'metadata', attrib={'xmlns:dc': "http://purl.org/dc/elements/1.1/", 'xmlns:opf': "http://www.idpf.org/2007/opf"})
au = ET.SubElement(md, 'dc:title')
au.text = bk['Title']

s = ET.tostring(et, pretty_print=True)

... but it fails miserably with: "ValueError: Invalid attribute name 'xmlns:dc'"

Any pointer welcome.

Answer 1

Reference: http://lxml.de/tutorial.html#namespaces

You mustn't specify the namespace as a :-qualified string. Instead you may use {http://url.url/url}tag form, or QName form.

Here is your program, using namespaces:

from lxml import etree as ET

NS_DC = "http://purl.org/dc/elements/1.1/"
NS_OPF = "http://www.idpf.org/2007/opf"
nsmap = {
    "dc": NS_DC,
    None: NS_OPF,
}
PACKAGE = ET.QName(NS_OPF, 'package')
METADATA = ET.QName(NS_OPF, 'metadata')
TITLE = ET.QName(NS_DC, 'title')

et = ET.Element(PACKAGE,
                attrib={'version': "2.0",
                        'unique-identifier': "BookId"},
                nsmap=nsmap)

md = ET.SubElement(et, METADATA)
au = ET.SubElement(md, TITLE)
au.text = "A Tale of Two Cities"

s = ET.tostring(et, pretty_print=True)
print(s.decode('utf-8'))

---

---

You might choose to use lxml.builder.ElementMaker. Here is a program that creates a portion of your example.

Notes:

• Note the use of dicts to represent attribute names that are not valid Python names.
• Note the use of QName to represent namespace-qualified names.
• Note the use of validator to validate the resulting tree.

 

from lxml import etree as ET
from lxml.builder import ElementMaker

NS_DC = "http://purl.org/dc/elements/1.1/"
NS_OPF = "http://www.idpf.org/2007/opf"
SCHEME = ET.QName(NS_OPF, 'scheme')
FILE_AS = ET.QName(NS_OPF, "file-as")
ROLE = ET.QName(NS_OPF, "role")
opf = ElementMaker(namespace=NS_OPF, nsmap={"opf": NS_OPF, "dc": NS_DC})
dc = ElementMaker(namespace=NS_DC)
validator = ET.RelaxNG(ET.parse("opf-schema.xml"))

tree = (
    opf.package(
        {"unique-identifier": "uuid_id", "version": "2.0"},
        opf.metadata(
            dc.identifier(
                {SCHEME: "uuid", "id": "uuid_id"},
                "d06a2234-67b4-40db-8f4a-136e52057101"),
            dc.creator({FILE_AS: "Homes, A. M.", ROLE: "aut"}, "A. M. Homes"),
            dc.title("My Book"),
            dc.language("en"),
        ),
        opf.manifest(
            opf.item({"id": "foo", "href": "foo.pdf", "media-type": "foo"})
        ),
        opf.spine(
            {"toc": "uuid_id"},
            opf.itemref({"idref": "uuid_id"}),
        ),
        opf.guide(
            opf.reference(
                {"href": "cover.jpg", "title": "Cover", "type": "cover"})
        ),
    )
)
validator.assertValid(tree)

print(ET.tostring(tree, pretty_print=True).decode('utf-8'))