Switch to Sphinx's InventoryFile when retrieving URLs

The interface to fetch_inventory changed around the 7-8 major version to
take a Sphinx config object.  However, we're not working with a sphinx
site and just want it to load and parse an inventory file for us.
Luckily, the project ships the InventoryFile class which provides an
interface to do exactly this.

Author: ghickman

cbv/management/commands/fetch_docs_urls.py

@@ -1,8 +1,10 @@
+import posixpath
+
 import requests
 from blessings import Terminal
 from django.conf import settings
 from django.core.management.base import BaseCommand
-from sphinx.ext.intersphinx import fetch_inventory
+from sphinx.util.inventory import InventoryFile
 
 from cbv.models import Klass, ProjectVersion
 
@@ -51,7 +53,7 @@ def handle(self, *args, **options):
             # the arg `r.raw` should be a Sphinx instance object..
             r = requests.get(ver_inv_url, stream=True)
             r.raise_for_status()
-            invdata = fetch_inventory(r.raw, ver_url, ver_inv_url)
+            invdata = InventoryFile.load(r.raw, ver_url, posixpath.join)
             # we only want classes..
             for item in invdata["py:class"]:
                 # ..which come from one of our sources