VLC_HTTP_LAUNCHER/scrapper.py
2023-05-16 23:25:05 +02:00

32 lines
934 B
Python

from bs4 import BeautifulSoup
from urllib.parse import quote, unquote
from utils import get_url
import requests
def get_files(site:dict, path:str) -> []:
url = get_url(site) + quote(path)
if (url in ["/../", "../"]):
return ([])
response = requests.get(url)
if (response.status_code != 200):
print("connection:", response.reason)
sys.exit(1)
soup = BeautifulSoup(response.text, 'html.parser')
files = []
for element in soup.findAll("a"):
files.append(unquote(element["href"]))
return (files)
def get_uri(url: str) -> []:
if (url in ["/../", "../"]):
return ([])
try:
response = requests.get(url)
if (response.status_code != 200):
print("connection:", response.reason)
sys.exit(1)
soup = BeautifulSoup(response.text, 'html.parser')
return(soup.find("h1").text[9:])
except:
return ("")