rjalex
1
I am using my own computed vectors for the item I am searching and am happy with the examples of nearvector, nearoobject and hybrid searches.
what I’m not quite getting is where do I define my generative module, be it an OpenAI API-key based or my own generative on prem modulue.
Can anyone suggest a full example for this?
Hi @rjalex !!
First, you need to enable that module in Weviate:
If you are using WCS, it is enabled by default.
So for example, let’s say you want this module:
Now, with that module installed, you need to create your collection and specify that module on it’s configuration.
Here we have a python v3 version recipe (py v4 should come soon):
Now, for the OpenAI APIKEY, you can set those both on your environment variable, or during query time.
Let me know if this helps 
rjalex
3
Looks like the original page has serious formatting issues 
What formatting? I don’t see it.
rjalex
5
on https://github.com/weaviate/recipes/blob/main/generative-search/generative_search_openai.ipynb
the notebook cells seem jumbled: