Merge pull request #6 from I-Al-Istannen/master

Hack in support for TI exams
This commit is contained in:
Garmelon 2020-03-01 23:09:32 +00:00 committed by GitHub
commit 9b658776ca
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 90 additions and 0 deletions

View File

@ -6,6 +6,7 @@ from .norbert import *
from .tgi import *
from .tgi_jimbo import *
from .ti import *
from .os_exams import *
from .utils import *
__all__ = ["STYLE", "FORMAT", "DATE_FORMAT", "FORMATTER", "enable_logging"]
@ -17,6 +18,7 @@ __all__ += tgi.__all__
__all__ += tgi_jimbo.__all__
__all__ += ti.__all__
__all__ += utils.__all__
__all__ += os_exams.__all__
STYLE = "{"
FORMAT = "[{levelname:<7}] {message}"

85
PFERD/os_exams.py Normal file
View File

@ -0,0 +1,85 @@
# Operating systems Exams
import getpass
import logging
import pathlib
import re
import bs4
import requests
from .organizer import Organizer
from .utils import stream_to_path, PrettyLogger
__all__ = ["OsExams"]
logger = logging.getLogger(__name__)
pretty = PrettyLogger(logger)
class OsExams:
BASE_URL = "https://os.itec.kit.edu/deutsch/1556.php"
LINK_RE = re.compile(
r"^http://os.itec.kit.edu/downloads_own/sysarch-exam-assandsols"
r".*/(.*\.pdf)$"
)
_credentials = None
def __init__(self, base_path):
self.base_path = base_path
self._session = requests.Session()
def synchronize(self, to_dir, transform=lambda x: x):
pretty.starting_synchronizer(to_dir, "OsExams")
sync_path = pathlib.Path(self.base_path, to_dir)
orga = Organizer(self.base_path, sync_path)
orga.clean_temp_dir()
self._crawl(orga, transform)
orga.clean_sync_dir()
orga.clean_temp_dir()
def _crawl(self, orga, transform):
url = self.BASE_URL
r = self._session.get(url)
soup = bs4.BeautifulSoup(r.text, "html.parser")
for found in soup.find_all("a", href=self.LINK_RE):
url = found["href"]
filename = re.match(self.LINK_RE, url).group(1).replace("/", ".")
logger.debug(f"Found file {filename} at {url}")
old_path = pathlib.PurePath(filename)
new_path = transform(old_path)
if new_path is None:
continue
logger.debug(f"Transformed from {old_path} to {new_path}")
temp_path = orga.temp_file()
self._download(url, temp_path)
orga.add_file(temp_path, new_path)
def _download(self, url, to_path):
while True:
username, password = self._get_credentials()
with self._session.get(url, stream=True, auth=(username, password)) as r:
if r.ok:
stream_to_path(r, to_path)
return
else:
print("Incorrect credentials.")
self._reset_credentials()
def _get_credentials(self):
if self._credentials is None:
print("Please enter OS credentials.")
username = getpass.getpass(prompt="Username: ")
password = getpass.getpass(prompt="Password: ")
self._credentials = (username, password)
return self._credentials
def _reset_credentials(self):
self._credentials = None

View File

@ -67,6 +67,9 @@ class Ti:
if soup.find(href="./Tutorien/Tutorien.php"):
logger.info("Found Tutorien/")
available["Tutorien"] = "/Tutorien/"
if soup.find(href="./AlteKlausuren/AlteKlausuren.php"):
logger.info("Found AlteKlausuren/")
available["AltKlausuren"] = "/AlteKlausuren/"
return available