Python inline of XML or ASCII string/template?

Question

I am generating complex XML files through Python scripts that require a bunch of conditional statements (example http://repository.azgs.az.gov/uri_gin/azgs/dlio/536/iso19139.xml). I am working with multiple XML or ASCII metadata standards that often have poor schema or are quite vague.

In PHP, I just wrote out the XML and inserted PHP snippets where needed. Is there an easy way to do that in Python? I am trying to avoid having to escape all that XML. The inline method is also very helpful for tweaking the template without much rewrite.

I have looked a bit into Python templeting solutions but they appeared either too static or were overkill. Moving the whole XML into an XML object is a lot of work at a cost of flexibility when changing the XML or ASCII template.

Thanks for the newbie support!

Answer 1

Try any modern non-XML-based Python templating engine, e.g. Mako or Jinja2. They are fairly easy to integrate into your project, and then you will be able to write things such as:

<sometag>
%if a > b:
   <anothertag />
%endif
</sometag>

You can also use inline python, including assignments.

Answer 2

@Gintautas' suggestion of using a good template engine would also be my first choice (especially Mako, whose "templating language" is basically Python plus some markup;-).

However, alternatives many people prefer to build XML files include writing them from the DOM, e.g. using (to stick with something in the standard library) ElementTree or (a third-party package, but zero trouble to install and very popular) BeautifulSoup (by all means stick with its 3.0.x release unless you're using Python 3!), specifically the BeautifulStoneSoup class (the BeautifulSoup class is for processing HTML, the stone one for XML).

Answer 3

wgrunberg,

I use Python's built-in string.Template class like so:

from string import Template
the_template = Template("<div id='$section_id'>First name: $first</div>")
print the_template.substitute(section_id="anID", first="Sarah")

The output of the above is:

<div id='anID'>First name: Sarah</div>

In the above example I showed an XML template, but any "template" that you can describe as a string would work.

To do conditionals you can do something like:

print the_template.substitute(section_id="theID", first="Sarah" if 0==0 else "John")

If your conditionals are complex, instead of expressing them inline as above, consider breaking them out into closures/functions.

Answer 4

Going for a template is the sensible thing to do. Using the build-in string template over external libraries would be my preferred method because I need to be able to pass on the Python ETL scripts easily to other people. However, I could get away with using templates by putting the XML string inside single quotes and using multi-line strings. Following is an example of this crude method.

    for row in metadata_dictionary:
        iso_xml = ''
        iso_xml += ' *** a bunch of XML *** '
        iso_xml += '\
<gmd:contact> \n\
    <gmd:CI_ResponsibleParty> \n'
        if row['MetadataContactName']:
            iso_xml += '\
      <gmd:individualName> \n\
        <gco:CharacterString>'+row['MetadataContactName'].lower().strip()+'</gco:CharacterString> \n\
      </gmd:individualName> \n'
        if row['MetadataContactOrganisation']:
            iso_xml += '\
      <gmd:organisationName> \n\
        <gco:CharacterString>'+row['MetadataContactOrganisation'].lower().strip()+'</gco:CharacterString> \n\
      </gmd:organisationName> \n'
        iso_xml += ' *** more XML *** '