Пример #1
0
from crawler.service import Service

class Controller:
    def __init__(self):
        pass

if __name__ == '__main__':
    api = Controller()
    service = Service()
    # url 파싱
    soup = service.naver_cartoon('https://comic.naver.com/webtoon/weekday.nhn')
    
    myfolder = 'E:/Dropbox/Dropbox/Programming/Git/Machine Learning/SBA_project/crawler/webtoon_data/'
    
    # 요일별 폴더 생성
    weekday_dict = service.create_folder_weekend(myfolder)

    # 웹툰 리스트 추출 및 저장
    mylist = service.create_webtoon_list(soup, myfolder, weekday_dict)

    service.save_csv_file(mylist, 'naver_webtoon_list')

    print(type(mylist))
    # for item in mylist:
    #     print(item)





Пример #2
0
import sys
sys.path.insert(0, r'C:\Users\yenaHan\SbaProjects')

from crawler.entity import Entity

from crawler.service import Service


class Controller:
    def __init__(self):
        self.entity = Entity()
        self.service = Service()


if __name__ == '__main__':
    mycolumns = ['타이틀 번호', '요일', '제목', '링크']
    filename = 'cartoon.csv'
    url = 'https://comic.naver.com/webtoon/weekday.nhn'
    new_folder_name = 'newfile'
    tag = 'div'
    attrs = 'thumb'
    replace_str = '/webtoon/list.nhn?'
    api = Controller()
    service = Service()
    service.get_url(url)
    service.create_folder_weekend(new_folder_name)
    service.setting_target(tag, attrs)
    service.loop_fun(replace_str, mycolumns, filename)