JSON数据处理

import pandas as pd
import os
import json

file_data=pd.read_csv('文件.txt')
file_data.head()

print(file_data.isnull().sum())

file_data.dropna(subset=['title', 'reply'], how='any', inplace=True) 

shuffled=file_data.sample(frac=1)

cut=int(file_data.shape[0]*.98)

test_data=shuffled[cut:]

with open('train.jsonl', mode='w',encoding='utf-8') as file:
    for index, row in train_data.iterrows():
        line={"input":row['title'],"target":row['reply']}
        str=json.dumps(line,ensure_ascii=False)
        file.write(str+'\n')