It's a stupid way but I don't found any possible method currently:
u = r"\u9038"
u.encode("ascii").decode("unicode-escape")
done.