-
Notifications
You must be signed in to change notification settings - Fork 513
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
chore(ai-help): use index.json instead of plain.html for embedding sources #12182
Conversation
…n instead of plainhtml
Bundle StatsHey there, this message comes from a github action that helps you and reviewers to understand how these changes affect the size of this project's bundle. As this PR is updated, I'll keep you updated on how the bundle size is impacted. Total
View detailed bundle breakdownAdded No assets were added Removed No assets were removed Bigger No assets were bigger Smaller No assets were smaller Unchanged
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let's deploy this to stage, so that @mirunacurtean can check AI Help for regressions.
@mirunacurtean this is already on staging |
@caugner Only saw this today, but I did start doing a check. I see where on a deadline with rari beta as it will be released today. is this part of it? And what should I do a quick sanity check of? So far I have:
When it comes to giving an answer I noticed an older issue, where AI Help starts producing the consulted content list and then it gives a "I'm sorry, but I can't answer questions outside web development." response. This also occurs on production, so it's not a regression. Although maybe it might be happening more with questions, for example "How can you retry or fallback to alternative external APIs in case of failures?" still gets a comprehensive response in prod but not in stage. |
Summary
Use index.json instead of plain.html for generating text sources for embedding
(MP-1743)
Problem
rari/yari parity
Solution
Added a new code path to assemble an interim html document directly form the json data. The old behaviour is still available behind a flag:
--use-plain-html
.How did you test this change?
Locally, tried a few documents and diffed the text corpus output beween old and new behaviour.