ingestion/embedder.py

from langchain_community.embeddings import FastEmbedEmbeddings
from app.config import config
from app.utils.logger import logger


class Embedder:
    def __init__(self):
        self.model = None
        
    def get_embeddings(self):
        if self.model is None:
            model_name = config["models"]["embedding"]["model_name"]
            max_length = config["models"]["embedding"]["max_length"]
            
            self.model = FastEmbedEmbeddings(
                model_name=model_name,
                max_length=max_length
            )
            logger.info(f"FastEmbed model loaded: {model_name}")
        
        return self.model
    
    def embed_documents(self, texts: list[str]) -> list[list[float]]:
        embeddings = self.get_embeddings()
        return embeddings.embed_documents(texts)
    
    def embed_query(self, text: str) -> list[float]:
        embeddings = self.get_embeddings()
        return embeddings.embed_query(text)


embedder = Embedder()