Hi Vaijanath, That would be great, thank you. And this will implement this formula ?
RRFscore(d ∈ D) = sum r∈R 1 /(k + r(d) ) where r(d) = rank of d i.e. 1,2,....n Thanks again and kind regards, Dan On Sat, 1 Apr 2023 at 17:41, Vaijanath Rao <vaiju1...@gmail.com> wrote: > Hi Dan, > > This was a long time ago, but I remember we use the recip function in solr > to built it out. I will try to reach out to team to see if they can give me > access to it. If not I will try to re-implement it and send it to you. This > was thought for Solr-5/6. > > --Thanks and Regards > Vaijanath > > On Sat, Apr 1, 2023 at 1:33 AM Dan Rosher <rosh...@gmail.com> wrote: > > > Hi Vaijanath, > > > > Was it a custom rank function, or custom rrf function for the sorting? > Are > > you able to provide an example? > > > > Many thanks, > > Dan > > > > On Fri, 31 Mar 2023 at 17:42, Vaijanath Rao <vaiju1...@gmail.com> wrote: > > > > > We implemented it as a sort function instead of Q-parser. It was easier > > and > > > speed wise more efficient. > > > > > > --Thanks and Regards > > > Vaijanath > > > > > > On Fri, Mar 31, 2023 at 8:23 AM Dan Rosher <rosh...@gmail.com> wrote: > > > > > > > Hi, > > > > > > > > Would anyone have an idea of how one might implement Reciprocal Rank > > > Fusion > > > > (RRF) with existing qparsers/function queries (e.g see > > > > https://plg.uwaterloo.ca/~gvcormac/cormacksigir09-rrf.pdf) ? I was > > > > thinking > > > > it would be interesting to test the outcome with a lexical + dense > > vector > > > > search, or would this need to be a custom QParser ? > > > > > > > > Cheers, > > > > Dan > > > > > > > > > > > > > -- > > > I am feeling fine, healthier and Happier, what about you > > > > > > > > -- > I am feeling fine, healthier and Happier, what about you >
