from __future__ import annotations from sentence_transformers import SentenceTransformer import numpy as np _model = None def get_embedder(model_name: str) -> SentenceTransformer: global _model if _model is None: _model = SentenceTransformer(model_name) return _model def embed_texts(texts: list[str], model_name: str) -> list[list[float]]: model = get_embedder(model_name) vecs = model.encode(texts, normalize_embeddings=True, show_progress_bar=False) if isinstance(vecs, np.ndarray): vecs = vecs.tolist() return vecs def embed_query(query: str, model_name: str) -> list[float]: return embed_texts([query], model_name=model_name)[0] def embedding_dim(model_name: str) -> int: model = get_embedder(model_name) return int(model.get_sentence_embedding_dimension())