alex/TUDown
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Added folder support for moodle.

Browse Source
This commit is contained in:
Alexander Weidinger
2015-04-20 12:16:43 +02:00
parent 0fb433d325
commit c5e397e028
1 changed files with 17 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
tudown.py
View File

@@ -39,20 +39,36 @@ def download_files(session, files):
def resolve_direct_links(session, hrefs): def resolve_direct_links(session, hrefs):
links = [] links = []
for href in hrefs: for href in hrefs:
links.append(session.head(href).headers['Location']) tmp = session.head(href).headers
if 'Location' in tmp:
links.append(tmp['Location'])
return links return links
def get_links_from_folder(session, urls):
hrefs = []
for url in urls:
response = session.get(url)
hrefs += findall(compile('https\:\/\/www\.moodle\.tum\.de\/pluginfile\.php\/\d{6}\/mod_folder\/content\/0\/(?:[\w\d\_\-]*\/)*[\w\d\_\-\.]{1,}'), response.text)
return hrefs
def get_file_links(session, url, files): def get_file_links(session, url, files):
links = [] links = []
response = session.get(url) response = session.get(url)
if 'www.moodle.tum.de' in url: if 'www.moodle.tum.de' in url:
# get file links
hrefs = findall(compile('https\:\/\/www\.moodle\.tum\.de\/mod\/resource\/view\.php\?id\=\d{6}'), response.text) hrefs = findall(compile('https\:\/\/www\.moodle\.tum\.de\/mod\/resource\/view\.php\?id\=\d{6}'), response.text)
# resolve all links to direct to the files
hrefs = resolve_direct_links(session, hrefs) hrefs = resolve_direct_links(session, hrefs)
# get folder links
folders = findall(compile('https\:\/\/www\.moodle\.tum\.de\/mod\/folder\/view\.php\?id\=\d{6}'), response.text)
if folders:
hrefs += get_links_from_folder(session, folders)
else: else:
hrefs = html.fromstring(response.text).xpath('//a/@href') hrefs = html.fromstring(response.text).xpath('//a/@href')
# ---------------
for f in files: for f in files:
reg = compile(f[0]) reg = compile(f[0])