PGVector + SQLModel

create the embedding and write to the database manually

Adapt the SQLModel model (see word2vec) with a column for the embedding

class Player(SQLModel, table=True):
	id: Optional[int] = Field(default=None, primary_key=True)
	name: str
	foot: str
	club_id: Optional[int] = Field(default=None, foreign_key="team.id")
	club: Optional[Club] = Relationship(back_populates="players")
	quote: Optional[str] = None
	embedding: Optional[List[float]] = Field(default=None, sa_column=Column(Vector(384))) # change embedding size according to model used

Then, use a sentence transformer model to embed the document

ronaldo_embedding = embeddings_model.encode("Siuuuuu")
ronaldo = Player(
	name="Cristiano Ronaldo",
	foot="right",
	club="Real Madrid"
	quote="Siuuuuu"
	embedding=ronaldo_embedding
)

and then write to the database. Might have to run session.exec(text("CREATE EXTENSION IF NOT EXISTS vector)) first.

writing to the database using LangChain

Or use LangChain to write (with less flexible column and table names).

PGVector.from_documents(
	documents=documents, # need to be List[Documents]
	embedding=embeddings_model,
	ids=[documents[i].metadata["player_id"] for i in range(len(documents))], # create `custom_id` column
	connection_string=db_url,
)

We can then create a session without having to rewrite to the database using from_existing index

def get_vector_db():
	db = PGVector.from_existing_index(
		embedding=embeddings_model,
		connection_string=db_url,
	)
	yield db

Session management can be used in a FastAPI function with

def vector_db_func(
	vector_db: VectorStore = Depends(get_vector_db)
):
	pass