Extract stats correctly when extra '.' in path

It appears that a recent cbcollect change adds a redundant '.' to the
path that the Prometheus stats snaphot is added to the zip file with.

E.g we now see this in the path in the zip file

[email protected]_20220610-011859/./stats_snapshot

whereas previously we saw this:

[email protected]_20220610-011859/stats_snapshot

It probably would be preferable to not have this additional, unnecessary
current directory in the path. But it should be the case that Promtimer
should handle it regardless.

Author: dave-finlay

promtimer/cbstats.py

@@ -202,7 +202,8 @@ def snapshot_dir_exists(candidate_cbcollect_dir):
         candidate_cbcollect_dir exists.
         :type candidate_cbcollect_dir: ,lib.Path
         """
-        return CBCollect.make_snapshot_dir_path(candidate_cbcollect_dir).exists()
+        return CBCollect.make_snapshot_dir_path(candidate_cbcollect_dir).exists() or \
+               CBCollect.make_snapshot_dir_path(candidate_cbcollect_dir  / '.').exists()
 
     @staticmethod
     def is_cbcollect_dir(candidate_path):