Add Sphinx redirects for latest pages

[oraNod]

Fixes #2147

This PR adds the Sphinx reredirects extension so that we can replace the latest redirects with something that is easier to maintain and provides a better user experience than stub pages.

For more detail, also see the conversation here: ansible/docsite#261 (comment)

Here are some comments about the files in the _data subdirectory:

• _data/pages.yaml these are redirect rules for latest|devel that I took from https://github.com/ansible/docsite/blob/main/.htaccess

I'm not sure if we really want to add those redirects to core docs but those are pages that were in core docs at some point and were moved. So it seems reasonable.

• _data/plugins.yaml these are redirect rules for plugins/* pages from https://github.com/ansible/docsite/blob/main/ansible/11/.htaccess

• _data/modules.yaml these are redirect rules for modules/* pages that come from the same .htaccess file as the plugins. Still need to go through and determine the best collection to use as the target redirect for most of these but it's not an impossible task. Rather than pointing to specific modules, I suggest we point to the main collection. This is in keeping with the effort to reduce maintenance burden for all these redirects.

• data/deprecated_modules.yaml these are redirects for modules/* pages that have the deprecation notice. If / when these collections are removed, we can point to the docs.ansible.com/collections.html page.

[oraNod]

I formatted the entire conf.py file, which might not be something we want to do for this PR because it makes it a little hard to see exactly what changed. I can revert that. Although I think formatting at some point would be a good idea.

Also not sure if this is too much cowbell or what. I mean initializing redirects and checking that things exist. Maybe this could be a lot simpler with fewer lines? I'd appreciate feedback on that please.

[oraNod]

Hint: You can see the redirects in action with the RTD preview: https://ansible--2266.org.readthedocs.build/projects/ansible-core/2266/

Append some of the source paths from the _data/pages.yaml to the preview url and you'll see the page redirect to the target. For example: https://github.com/ansible/ansible-documentation/pull/2266/files#diff-06991d000b4e85ba03248d9ad6186e37102d7f0e9e5fe440b0f66e0d06ea069cR1

https://ansible--2266.org.readthedocs.build/projects/ansible-core/2266/ansible.html redirects to https://ansible--2266.org.readthedocs.build/projects/ansible-core/2266/cli/ansible.html

[oraNod]

Ran the package docs build workflow against this PR. Here is the run: https://github.com/ansible/ansible-documentation/actions/runs/12200497630

Generating the redirects doesn't appear to add to the build time (this is actually the second run).

Here are some pages that redirect:

• https://ansible-community.github.io/package-doc-builds/modules/shell_module.html
• https://ansible-community.github.io/package-doc-builds/modules/command_module.html
• https://ansible-community.github.io/package-doc-builds/modules/aci_fabric_node_module.html
• https://ansible-community.github.io/package-doc-builds/modules/gcp_storage_bucket_access_control_module.html
• https://ansible-community.github.io/package-doc-builds/playbook_pathing.html
• https://ansible-community.github.io/package-doc-builds/windows_dsc.html
• https://ansible-community.github.io/package-doc-builds/user_guide/intro_adhoc.html
• https://ansible-community.github.io/package-doc-builds/porting_guide_2.3.html
• https://ansible-community.github.io/package-doc-builds/YAMLSyntax.html
• https://ansible-community.github.io/package-doc-builds/plugins/callback/full_skip.html
• https://ansible-community.github.io/package-doc-builds/plugins/become/doas.html

I've checked the http status codes of all the pages (just the source pages, not the targets). I can see they're returning 200 instead of 301 or 302 but that's expected. As I mentioned elsewhere a 301 would be ideal for SEO but 200 is still much better than 404 and this is arguably a better user experience than redirecting to a single collections.html page.

I suppose a solution here could be to add all the generated source pages to a norobots.txt file so they don't get indexed. This would ensure that they don't cause the actual pages with content to be counted as duplicates.

Anyway here are the reports:

• url_report_sphinx_redirects_20241206_154622.csv
• url_report_sphinx_redirects_20241206_154622.txt