Do a joined query to improve performance

When running db.list_timeseries_instances with a database containing a
large number of series instances performance can be slow due to lazy
loading running an individual query for each record when building the
data frame. Doing a join means all the records are read in a single
query, which is much more efficient when we are going to use all the
returned values anyway.

Signed-off-by: Andrew MacDonald <andrew@maccas.net>

phildb/database.py

@@ -471,7 +471,9 @@ def list_timeseries_instances(self, **kwargs):
         query_args = self.__parse_attribute_kwargs(**kwargs)
         query_args.update(initial_args)
 
-        records = session.query(TimeseriesInstance).filter_by(**query_args)
+        records = session.query(TimeseriesInstance).options(
+                joinedload(TimeseriesInstance.timeseries)
+            ).filter_by(**query_args)
         instance_list = []
         for record in records:
             instance = {