Spaces:
Runtime error
Runtime error
import pandas as pd | |
from sentence_transformers import SentenceTransformer, util | |
# Cargar modelo SBERT en español | |
try: | |
model = SentenceTransformer("hiiamsid/sentence_similarity_spanish_es") | |
except Exception as e: | |
print(f"⚠️ Error cargando modelo BETO: {e}") | |
model = None | |
# Cargar CSV de marcas | |
try: | |
df = pd.read_csv("IGS - Consolidado.csv") | |
marca_textos = df["NombreProducto"].astype(str).str.lower().tolist() | |
if model: | |
marca_embeddings = model.encode(marca_textos, convert_to_tensor=True) | |
else: | |
marca_embeddings = None | |
except Exception as e: | |
print(f"⚠️ Error cargando CSV en BETO: {e}") | |
marca_textos = [] | |
marca_embeddings = None | |
def buscar_marcas_similares(input_marca, top_n=5): | |
input_marca = input_marca.lower() | |
input_embedding = model.encode(input_marca, convert_to_tensor=True) | |
similitudes = util.pytorch_cos_sim(input_embedding, marca_embeddings)[0] | |
top_resultados = sorted( | |
zip(marca_textos, similitudes), | |
key=lambda x: x[1], | |
reverse=True | |
)[:top_n] | |
return [(marca, score.item() * 100) for marca, score in top_resultados] # <<< Cambiar aquí | |