Search function provides too many results

[dev00]

Currently search function seems to perform full text search on the whole text, which usually generates too much output and is not usable at all.

To limit output in a more usable way, we could try to generate short summaries automatically with a ML based approach like LLama or ChatGPT. Another approach might be to switch to a simple tag based search and try generating tags automatically.

Expecting customers to write OpenSearch queries is probably a bit too much.

[SebastianGode]

@dev00 Is this still relevant?
In the mean time we changed a bit of stuff on the search and also added exact matches using "value" format.

Running an AI Model on Ollama can be considered for the future.