Add model description to the vector database

[kuraisle]

Is this the right issue type?

• Yes, I'm planning work for this project team.

Summary

pgvector doesn't care which model makes any embeddings. Thus, by itself, responsibility for making sure the correct model is used to generate query embeddings is unclear.

An expedient solution is to have a one-row table in the database to describe the model used to generate the embeddings that code interacting with the database can read.

Acceptance Criteria

• The embeddings build container adds a table to the database describing the model

Tasks

• Add environment variable to load embeddings for model description
• Add function to load_embeddings.py that creates a model description table
• Add function to load_embeddings.py that inserts model information read from the environment variable

Confirm creation

• This issue is ready