Fixes #5500 - Add a Github Actions workflow that builds the docs

[martin-g]

Which issue does this PR close?

Closes #5500

Rationale for this change

Deploy the Datafusion website whenever a change is made in the docs/ folder.

What changes are included in this PR?

A new Github Action is introduced that builds the docs and copies them to arrow-site/datafusion

Are these changes tested?

Yes, I have tested them in my fork.
The only part that is not really tested is the git push to arrow-site. But arrow-datafusion-python uses the same code

Are there any user-facing changes?

No

[martin-g]

CC @kou @andygrove

[martin-g]

Converted the PR to a Draft one because of this change.
See #5500 (comment)

[alamb]

@andygrove or @houqp do you remember what this is used for or if it is critical to fix?

[alamb]

(I personally think that as long as this removal doesn't render the docs inoperable, it would be fine to remove this code to get the docs publishing and fix it afterwards)

[kou]

How about using publish.subdir in .asf.yaml like ADBC instead of pushing to apache/arrow-site?

https://cwiki.apache.org/confluence/display/INFRA/Git+-+.asf.yaml+features#Git.asf.yamlfeatures-Specifyingasub-directorytopublishto

If we use publish.subdir, we can publish the DataFusion docs to https://arrow.apache.org/datafusion/ by pushing to the asf-site branch in apache/arrow-datafusion (not apache/arrow-site).

See also:

• ADBC's .asf.yaml: https://github.com/apache/arrow-adbc/blob/main/.asf.yaml#L35-L37
• ADBC's GitHub Actions: https://github.com/apache/arrow-adbc/blob/main/.github/workflows/nightly-website.yml (filename includes "nightly" but it's not nightly...)
• ADBC's asf-site branch: https://github.com/apache/arrow-adbc/tree/asf-site
• ADBC's site: https://arrow.apache.org/adbc/

[kou]

We used publish.subdir for arrow-datafusion-python too.

[martin-g]

It is easy for me to make the change if the team prefers this way!

…

On Sat, Mar 18, 2023, 00:22 Sutou Kouhei ***@***.***> wrote: We used publish.subdir for arrow-datafusion-python too. — Reply to this email directly, view it on GitHub <#5631 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABYUQVQE4W4BYIP67IQVMLW4TP2BANCNFSM6AAAAAAV6TBM2A> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[alamb]

It is easy for me to make the change if the team prefers this way!

I think this would be a great idea. Thank you @martin-g and @kou -- getting the docs building regularly is so exciting!

[martin-g]

@alamb (or anyone else with permissions) Please create a branch for the site, e.g. asf-site in arrow-datafusion repo.
You could remove all files and folders in it. Or leave that to the Github Action workflow. It will remove all old content before copying the newly generated HTMLs, as in arrow-datafusion-python

[kou]

@martin-g @alamb I've created the orphan asf-site branch: https://github.com/apache/arrow-datafusion/tree/asf-site