Update search functionality to add pre-processing to query

[joannacknight]

For EmbeddingGemma, we want to be able to pre-fix a prompt to the query.

We already have the clustering prompt as a function in tt_models.py. We would like to be able to add this to the query before embedding so that it is in the same vector space as the document embeddings.

Not sure where we should do this - perhaps @eddableheath can put some hints in this ticket and then it should be something that any of us could pick up.