remove save from default runme (due to rare use and poooor performance)

runme.py

@@ -115,11 +115,6 @@ def run_scenario(input_file, timesteps, scenario, result_dir, plot_periods={}):
         os.path.join(result_dir, '{}.xlsx').format(sce),
         prob.com_demand, prob.sit)
 
-    # store optimisation problem for later re-analysis
-    urbs.save(
-        prob,
-        os.path.join(result_dir, '{}.pgz').format(sce))
-
     urbs.result_figures(
         prob, 
         os.path.join(result_dir, '{}'.format(sce)),

urbs.py

@@ -1525,10 +1525,6 @@ def get_entity(instance, name):
 
         # convert to Series
         results = results[name]
-
-        # if Series has MultiIndex, sort it to avoid PerformanceWarnings
-        if isinstance(results.index, pd.core.index.MultiIndex):
-            results = results.sortlevel()
     return results
 
 
@@ -1781,8 +1777,10 @@ def get_timeseries(instance, com, sit, timesteps=None):
     dsmup = get_entity(instance, 'dsm_up')
     dsmdo = get_entity(instance, 'dsm_down')
 
-    dsmup = dsmup.xs(sit, level = 'sit').xs(com, level = 'com')
-    dsmdo = dsmdo.xs(sit, level = 'sit').xs(com, level = 'com')
+    dsmup = dsmup.xs(sit, level = 'sit')
+    dsmup = dsmup.xs(com, level = 'com')
+    dsmdo = dsmdo.xs(sit, level = 'sit')
+    dsmdo = dsmdo.xs(com, level = 'com')
 
     #  series by summing the first time step set
     dsmdo = dsmdo.unstack().sum(axis=0)