data/gushi/getdata.py

# _*_ coding:utf-8 _*_
import glob
import json
datas=glob.glob("chinese-poetry/全唐诗/poet.*json")

for data in datas:
    with open(data,"r",encoding='utf-8') as fp:
        tangshi=json.load(fp)
        for each_shi in tangshi:
            if len(each_shi["paragraphs"])==2 and len(each_shi["paragraphs"][0])==16 and  len(each_shi["paragraphs"][1])==16:
                with open('data/gushi/gushi.txt',"a",encoding='utf-8') as f:
                    f.write("".join(each_shi["paragraphs"]))
                    f.write("\n")