update benchmark results

[StefanBogdan]

What's being changed:

The benchmark results: new benchmark results to be published with latest Weaviate version and new infrastructure.

Type of change:

• Documentation updates (non-breaking change to fix/update documentation)

How Has This Been Tested?

• GitHub action – automated build completed without errors
• Local build - the site works as expected when running yarn start

note, you can run yarn verify-links to test site links locally

[StefanBogdan]

For this PR I would need some help with the results we want to show on the documentation page. Currently it has ALL the results that were collected which is more than we would need to show. The results are stored in the new ann-{dataset}.mdx files. The new script collate-results.py from this PR can be used to generate/analyse the results.

Another thing that needs to be checked is the Recommended configuration for each dataset. I chose one that looked pretty balanced to me but let me know if there is a better one I can use.

And the last thing is the QPS vs Recall plots. Currently it is just PNG photo. I can create PNGs that have a transparent background with either black or white axes and text, the only thing I could not figure out is how can I switch between the black or white image depending on the light/dark mode used. Let me know if we want to further investigate how to achieve different plots depending on the mode used.

[weaviate-git-bot]

To avoid any confusion in the future about your contribution to Weaviate, we work with a Contributor License Agreement. If you agree, you can simply add a comment to this PR that you agree with the CLA so that we can merge.

beep boop - the Weaviate bot 👋🤖

PS:
Are you already a member of the Weaviate Slack channel?

[StefanBogdan]

Here is the link to this docs version: https://stefan-update-benchmark--tangerine-buttercream-20c32f.netlify.app/developers/weaviate/benchmarks/ann

[databyjp]

Amazing. I will take a look later this week (was away last week), thanks so much @StefanBogdan !

[StefanBogdan]

Here is the live lint to the website.