I have a usecase in which my entire RAG system would have to be hosted locally.

I would be running a local weaviate server along with my own huggingface embeddings model. Weaviate’s documentation is a weird labyrinth understandable by only those that have either worked with weaviate for over a couple of years or their developers themselves.

Here is what I want to do:

1. Create an embeddings model with sentence-transformers library

from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim
#since this embeddings model runs some custom code, we need to pass in a sentence transfomrer client with permission to trust the remote code.
embeddings = SentenceTransformer('Alibaba-NLP/gte-large-en-v1.5', trust_remote_code=True)

2. Use the above model to embed the documents and ingest them in the local weavite server running at 8080
3. Use the same embeddings model to encode some query (“This is a test query”)
4. Retrieve from weaviate db the relevant documents that are similar to the query

The most basic most canonical usecase for a vector db. However, much to my surprise, I cannot find weaviate documentation to do this. I am stuck with weaviate for the time being. Can you help me set up this use case