FunASR
FunASR copied to clipboard
如何屏蔽itn中带有货币形式的转换? TP99-->T九十九博茨瓦纳普拉
❓ Questions and Help
Code
from fun_text_processing.text_normalization.normalize import Normalizer
normalizer = Normalizer(input_case='cased', whitelist=None, lang='zh')
text = "取第99%的那个值作为TP99的值。"
text = normalizer.normalize(
text, verbose=False,
punct_pre_process=True,
punct_post_process=False,
)
#取第百分之九十九的那个值作为T九十九博茨瓦纳普拉的值。
tagged_text = ' tokens { name: "取" } tokens { name: "第" } tokens { measure { decimal { integer_part: "九十九" } } } tokens { name: "的" } tokens { name: "那" } tokens { name: "个" } tokens { name: "值" } tokens { name: "作" } tokens { name: "为" } tokens { name: "T" } tokens { money { integer_part: "九十九" fractional_part: "博茨瓦纳普拉" } } tokens { name : "的" } tokens { name: "值" } tokens { name: "。" } '
想要去掉 money的选项, tokens { money { integer_part: "九十九" fractional_part: "博茨瓦纳普拉" }