Parse website sitemap.xml using Ansible XML

There is this website with a /sitemap.xml file such as follows:

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9" xmlns:xhtml="http://www.w3.org/1999/xhtml">
  <url>
    <loc>https://example.com/es/</loc>
    <lastmod>2023-02-15</lastmod>
    <changefreq>monthly</changefreq>
    <priority>0.5</priority>
  </url>
  <url>
    <loc>https://example/en/</loc>
    <lastmod>2023-02-15</lastmod>
    <changefreq>monthly</changefreq>
    <priority>0.5</priority>
  </url>
  <url>
    <loc>https://example.com/en/destinations/</loc>
    <lastmod>2021-09-16</lastmod>
    <changefreq>monthly</changefreq>
    <priority>0.5</priority>
  </url>
[..]
</urlset>

Using Ansible (latest version) I am trying to:

1. Download the file using ansible.builtin.uri and register the result into a variable.
2. Either loop the url nodes inside urlset and create a list with the loc and the priority, or
3. Convert it to JSON and do the same.

I am stuck at points 2-3. This is my current code:

- name: Get the website 'sitemap.xml' file
  ansible.builtin.uri:
    url: "https://example.com/sitemap.xml"
    method: GET
    return_content: true
    headers:
      Accept: "application/xml"
    status_code: 200
    timeout: 5
  register: sitemap
  delegate_to: localhost

- name: Parse the retrieved XML file
  community.general.xml:
    xmlstring: "{{ sitemap.content }}"
    xpath: /s:urlset
    content: text
    namespaces:
      s: http://www.sitemaps.org/schemas/sitemap/0.9
  register: parsedxml
  delegate_to: localhost

Now parsedxml.xmlstring contains the XML of the file sitemap.xml, which is something I already had at the sitemap.content variable. So, basically, I haven't been able to either:

1. Use community.general.xml to somehow build a list of dicts (with loc and priority) by looping the list of url nodes,
2. Or convert the XML file to JSON using the ansible.netcommon.parse_xml filter, but I have not been able to produce a specifications file to be passed as parametre to the filter. And the documentation of such filter seems to be missing.

Any hints on how to loop through all the url nodes and build such list of dictionaries?

So, yeah, ansible isn't great at dealing with xml, and that parsexml module you found is really designed for use by ansible net module authors which explains why it is so terrible to use

This is my approach:

  tasks:
    - vars:
        sitemap:
          content: |
            <?xml version="1.0" encoding="UTF-8"?>
            <urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9" xmlns:xhtml="http://www.w3.org/1999/xhtml">
              <url>
                <loc>https://example.com/es/</loc>
                <lastmod>2023-02-15</lastmod>
                <changefreq>monthly</changefreq>
                <priority>0.5</priority>
              </url>
              <url>
                <loc>https://example/en/</loc>
                <lastmod>2023-02-15</lastmod>
                <changefreq>monthly</changefreq>
                <priority>0.5</priority>
              </url>
            </urlset>            
      with_items: [ loc, priority ]
      register: parsedxml
      community.general.xml:
        xmlstring: '{{ sitemap.content }}'
        xpath: /s:urlset/s:url/s:{{ item }}
        content: text
        namespaces:
          s: "http://www.sitemaps.org/schemas/sitemap/0.9"
    - set_fact:
        things: '{{ tmp | from_yaml }}'
      vars:
        zipped_shape: |
          [
            [ {"loc": "http"}, {"pri": "1.0"} ],
            [ {"loc": "http"}, {"pri": "1.0"} ],
          ]          
        tmp: |
           {% set zipped = parsedxml.results[0].matches 
                     | zip(parsedxml.results[1].matches) %}
           {% for tup in zipped %}
           - url: {{ tup[0].values()|first }}
             pri: {{ tup[1].values()|first }}
           {% endfor %}           

which produces:

{
    "ansible_facts": {
        "things": [
            {
                "pri": 0.5,
                "url": "https://example.com/es/"
            },
            {
                "pri": 0.5,
                "url": "https://example/en/"
            }
        ]
} }

as best I can tell, that .xml: module really is designed for more "surgical" changes than a generic XPath query into a document, and definitely bad for "give me multiple keys". So, we just cheat and compose the XPath more than once, for each child key we wish, and then | zip the two result lists back together.