Nextcloud crawler via WebDav

trying_to_fix_nextcloud.py

@@ -0,0 +1,56 @@
+import requests
+import xml.etree.ElementTree as ET
+import sys
+
+def propfind(path,auth):
+  headers = {"Depth":"1"}
+  req = r = requests.Request("PROPFIND", path, headers=headers, auth=auth)
+  prepped = req.prepare()
+    
+  s = requests.Session()
+  resp = s.send(prepped)
+  return resp.text
+
+def search_folder(requestreturn):
+  innerfolders = []
+  linkswrongtime = []
+  firstfolder = True
+  temp = "this is sa temporary storage"
+  tree = ET.ElementTree(ET.fromstring(requestreturn))
+  for resp in tree.findall('.//{DAV:}response'):
+    for p in resp:
+      if not(p.text is None):
+        if (p.text[-1] == '/'):
+          if not(firstfolder):
+            innerfolders.append(p.text)
+          else: 
+            firstfolder = False
+          break
+        else:
+          temp = p.text
+      else:
+        for time in p.findall('.//{DAV:}getlastmodified'):
+          year = time.text.split(' ')
+          if int(year[3]) < 1990:
+            linkswrongtime.append(temp)
+  return [innerfolders, linkswrongtime]
+
+user = sys.arcv[1]
+passw = sys.arcv[2]
+auth = requests.auth.HTTPBasicAuth(user,passw)
+prefix_path = "https://kingsx.cs.uni-saarland.de"
+mainpath = "/remote.php/dav/files/"+ user + "/Testrequests/"
+folders = [mainpath]
+wrongtime = []
+
+while folders:
+  path_suffix = folders.pop(0)
+  path = prefix_path + str(path_suffix)
+  r = propfind(path,auth)
+  res = search_folder(r)
+  folders = folders + res[0]
+  wrongtime = wrongtime + res[1]
+
+print(wrongtime)
+
+version_suffix = "/remote.php/dav/versions/USER/versions/FILEID"