GH-45665: [Docs] Add kapa AI bot to the docs

[thisisnic]

Rationale for this change

Add kapa AI bot to the docs to make it easier to find answers to questions

What changes are included in this PR?

Config for enabling bot

Are these changes tested?

Manually, yep

Are there any user-facing changes?

Yep, adds widget to header in docs pages

• GitHub Issue: [Docs] Add kapa AI bot to the docs #45665

[thisisnic]

@github-actions crossbow submit preview-docs

[github-actions]

Revision: 64873ba

Submitted crossbow builds: ursacomputing/crossbow @ actions-2469749b9e

Task	Status
preview-docs

[assignUser]

http://crossbow.voltrondata.com/pr_docs/45667/ sadly I only get There was an error in generating the answer. Please refresh the page and try again. probably because it's not coming from the expected url?

[thisisnic]

Exactly that; have enabled there now. The button I created didn't work, but the other one at the bottom right does so I'll experiment with making that a bit nicer!

[thisisnic]

@github-actions crossbow submit preview-docs

[github-actions]

Revision: b88e50b

Submitted crossbow builds: ursacomputing/crossbow @ actions-a4683479a2

Task	Status
preview-docs

[thisisnic]

http://crossbow.voltrondata.com/pr_docs/45667

[assignUser]

Works for me now, cool!

[assignUser]

I put the recent ML question about encryption into it and received an accurate (no hallucination I could detect) answer that even expanded on the answers given on the ML with page_checksum_verification. 🚀 very nice!

[AlenkaF]

This is great! I like how it lists the sources! I inquired about the struct layout and got a nice answer. Not too long! ;) Asking about the PyCapsule protocol implementation I like that it emphasises the fact that it is currently experimental and can change in the future.

Thank you for working on this @thisisnic!

[AlenkaF]

Aha, and the button is nice!

[thisisnic]

It wasn't working with the R docs, but now appears to be sorted - this is ready for review! Once it's merged and is up on the dev docs, I'll put a note on on the dev mailing list asking folks to test it out and give feedback on the answers to see where we can improve them.

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 784aa6f.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 59 possible false positives for unstable benchmarks that are known to sometimes produce them.