需要的包:
html 解析包 beautifulsoup
pip install bs4
获取后缀名为 mp4 / wav / avi 的文件:
from bs4 import BeautifulSoup
from os import listdir
root = "E:/software/nginx-1.20.1/html/" #nginx html 路径
resourcefiles = root + "RES/" #需要被代理的资源文件路径
indexfile = root + "index.html" #原始html文件
newindexfile = root + "index.html.new" #新html文件,运行完后用这个文件覆盖原始文件
resourcelist = listdir(resourcefiles) #获取所有文件
appendixlist = [".mp4",".wav",".avi"] #过滤条件
# Press the green button in the gutter to run the script.
if __name__ == '__main__':
#get html obj
bs_obj_indexfile = BeautifulSoup(open(indexfile))
for file in resourcelist:
if(file.endswith(appendixlist[0]) or file.endswith(appendixlist[1]) or file.endswith(appendixlist[2])):
newtag = bs_obj_indexfile.new_tag('a',href = "RES/" + file)
bs_obj_indexfile.body.append(newtag)
newtag.string = file
br = bs_obj_indexfile.new_tag('br')
bs_obj_indexfile.body.append(br)
print(bs_obj_indexfile.prettify())
with open(newindexfile,'w') as newindexfileobj:
newindexfileobj.write(str(bs_obj_indexfile.prettify()))
newindexfileobj.close()