Refine module vs. document detection heuristic.

Module listbox links to documents do not have a trailing slash, which skews
results for normal usage patterns.

apachedex/__init__.py

@@ -283,7 +283,9 @@ class ERP5SiteStats(GenericSiteStats):
       module = split[0]
       if module.endswith('_module'):
         super(ERP5SiteStats, self).accumulate(match, url_match, date)
-        self.module[module][len(split) > 2][date].accumulate(match)
+        self.module[module][
+          len(split) > 1 and (split[1] != 'view' and '_view' not in split[1])
+        ][date].accumulate(match)
       else:
         self.no_module[date].accumulate(match)