Using pinecone vector db from Rails with OpenAI for semantic search

Using pinecone vector db from Rails with OpenAI for semantic search
#rubyonrails #ruby #rails #gpt3 #openai #pinecone

[@dfriis]

Thank you so much for this video! Not only did I learn about pinecone, but as a novice programmer, it's extremely insightful to watch you work through the problem.

[@jslime]

So what were you doing wrong at the end, did you figure out why it kept kicking back the same results?

[@cocoarecords]

Beast

[@stpaquet]

why not using TypeSense for this type of search?

[@cjav_dev]

I’m not sure what type sense is, but the goal is that someone can search for something semantically instead of with exact match. For instance, they could search for “dog” and that would match videos about poodles that never mention the word dog.

[@stpaquet]

@@cjav_dev TypeSens, Meilisearch, Algol are quite the same and are usually good for that use case. TypeSense and Meilisearch are open source (while Algolia is not)

[@cjav_dev]

@@stpaquet Ah okay, I didn't realize those had semantic search. I thought those were all close or exact match search based on regex.

[@stpaquet]

@@cjav_dev I agree it's not obvious but they do offer out of the box semantic search and based on my experience it's very decent for what you are looking forward to implement.
You still need to manage the indexes... but that seems to be way less work that what you are trying to achieve using GPT and pinecone (or Postgres and some modules)