-
Notifications
You must be signed in to change notification settings - Fork 106
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
Update Rockset details under rockset.json #151
Conversation
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.
Hi @ankit1khare - thank you for your PR!
Please see my comments next to your changes.
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.
made the requested changes
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.
please provide a reference to docs/code/notebook example in source_url for each field where you are marking support as "full" or "partial"
We combine ANN (vector search) with metadata filtering which is hybrid search in essence. Updated the links. |
"support": "", | ||
"source_url": "", | ||
"comment": "" | ||
"support": "full", |
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.
Please change this to none as Hybrid Search has a specific meaning in this tab le, which is different from what you explained.
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.
Hi @ankit1khare, I'm guessing you mean to point to this section https://github.com/sofia099/search_autocomplete#step-3-create-3-query-lambdas in the new repo you've linked. It seems to combine a set of exact matches with approximate lexical matches.
Hybrid Search in this table refers to "ability to merge between ranked lists of keyword search and vector search results within the DB".
Is it possible to do something like Reciprocal Rank Fusion or take a linear combination of scores from vector search and keyword search in a similar way?
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.
Hey @dhruv-anand-aintech, after all it's SQL we are talking about where we have built our own engine from scratch such that all standard SQL operations can be done on different data modalities. One can build their own weighting mechanism using SQL and get all of the analytical capabilities built into the query language (like linear combination, reciprocal rank fusion, etc.).
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.
Thanks for the explanation. It'll still be good to have an example notebook that does these types of fusion in Rockset, so that users are easily able to migrate to it from other DBs that do it under the hood.
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.
done
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.
Please make the requested changes. Thanks
"support": "full", | ||
"source_url": "https://docs.rockset.com/documentation/reference/user-defined-functions", | ||
"comment": "" | ||
}, | ||
"embeddings_image": { | ||
"support": "", | ||
"source_url": "", | ||
"support": "full", | ||
"source_url": "https://docs.rockset.com/documentation/reference/user-defined-functions", |
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.
Feel free to reach out to me if something is unclear :)