Add proxy cache for PyPI

Investigate using apache proxy caching for PyPI instead of a full
mirror. Make the /pypi path proxy directly to pypi.org upstream instead
of to the rackspace AFS mirror.

We have to use the substitute module (already enabled) to rewrite the
embedded links in the html content returned, because instead of relative
links to something under /pypi which get redirected, the CDN links are
direct.

Change-Id: Id68ff2914ba7fb588092cff113e8d1536f7d5a56

modules/openstack_project/templates/mirror.vhost.erb

@@ -163,8 +163,19 @@ LogFormat "%h %l %u %t \"%r\" %>s %b %{cache-status}e \"%{Referer}i\" \"%{User-a
 
     # pypi
     CacheEnable disk  "/pypi"
-    ProxyPass "/pypi/" "http://mirror.dfw.rax.openstack.org/pypi/" ttl=120 keepalive=On retry=0
-    ProxyPassReverse "/pypi/" "http://mirror.dfw.rax.openstack.org/pypi/"
+    ProxyPass "/pypi/" "https://pypi.org/" ttl=120 keepalive=On retry=0
+    ProxyPassReverse "/pypi/" "https://pypi.org/
+
+    # files.pythonhosted.org
+    CacheEnable disk  "/pypifiles"
+    ProxyPass "/pypifiles/" "https://files.pythonhosted.org/" ttl=120 keepalive=On retry=0
+    ProxyPassReverse "/pypifiles/" "https://files.pythonhosted.org/"
+
+    # Rewrite the locations of the actual files
+    <Location /pypi>
+      SetOutputFilter SUBSTITUTE
+      Substitute "s|https://files.pythonhosted.org/|/pypifiles/|ni"
+    </Location>
 
     # images.linuxcontainers.org
     CacheEnable disk "/images.linuxcontainers"
@@ -225,6 +236,7 @@ LogFormat "%h %l %u %t \"%r\" %>s %b %{cache-status}e \"%{Referer}i\" \"%{User-a
     CacheEnable disk "/alpine"
     ProxyPass "/alpine/" "http://dl-cdn.alpinelinux.org/alpine/" ttl=120 keepalive=On retry=0
     ProxyPassReverse "/alpine/" "http://dl-cdn.alpinelinux.org/alpine/
+
 </VirtualHost>
 
 <VirtualHost <%= @vhost_name %>:8081>