ALIGNS: Unlocking nomological networks in psychological measurement through a large language model