alamb commented on issue #6518: URL: https://github.com/apache/datafusion/issues/6518#issuecomment-2585694146
> ```rust > let sql = r#" > CREATE FUNCTION an_llm_function(STRING) > RETURNS STRING > LANGUAGE MODEL > AS 'microsoft/phi-4' > "#; > > ctx.sql(sql).await?.show().await?; > ``` > > greatest way to disrespect all those tears and sweat put into getting datafusion to be as performant as it is 😀 LOL, though I think the idea of combining the fast local excution to prep the data to send could still be compelling. Something crazy like this to have an LLM summarize bad comments 🤔 ```sql SELECT llm_summarize(array_agg(comments)), company_id GROUP BY company_id WHERE company_id IN (1,2,3) and comments ILIKE '%not good%' ``` > > joke aside good job @goldmedal & @alamb! do you plan getting it merged in datafusion or df contrib ? I hadn't thought this far -- I think it would definitely make sense to be in datafusion-contrib. Maybe depending on how much interest there is more broadly we could also potentially put it into datafusion. However, the fact that you can implement async functions via the existing extension mechanisms is pretty neat -- the benefit of putting it into the core seems like it would make it easier to integrate -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org
