semantic_embedder.py•649 B
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from app.schemas.vector_store_schemas import Embeddings
class SemanticEmbedder:
def __init__(self, model_name: str):
self.model_name = model_name
self._model: HuggingFaceEmbedding | None = None
# Lazy loading of the model
@property
def model(self) -> HuggingFaceEmbedding:
if self._model is None:
self._model = HuggingFaceEmbedding(model_name=self.model_name)
return self._model
def embed_texts(self, texts: list[str]) -> Embeddings:
return Embeddings(vectors=self.model.get_text_embedding_batch(texts))