src/E_Utils.py

from sentence_transformers import SentenceTransformer
import faiss

def load_faiss_index(file_name="vector_db.index"):
    index = faiss.read_index(file_name)
    return index

def normalize_embeddings(embeddings):
    faiss.normalize_L2(embeddings)
    return embeddings


def train_model(model_name='dumitrescustefan/bert-base-romanian-cased-v1'):
    model = SentenceTransformer(model_name)
    return model

def get_embeddings(model, texts):
    embeddings = model.encode(texts, convert_to_tensor=True)
    return embeddings

def save_embeddings(embeddings, file_name="vector_db.index"):
    import faiss
    dimension = embeddings.shape[1]
    index = faiss.IndexFlatL2(dimension)
    faiss.normalize_L2(embeddings)
    index.add(embeddings.cpu().numpy())
    faiss.write_index(index, file_name)
    return index