Performance evaluation add TODOs

libeos/event_analysis.py

@@ -62,6 +62,7 @@ class CalculateWavelength(EventDataAction):
         #lamdaMax = const.lamdaCut+1.e13*dataset.timing.tau*const.hdm/(dataset.geometry.chopperDetectorDistance+124.)
 
         # lambda
+        # TODO: one of the most time consuming actions, could be implemented in numba, instead?
         lamda = (1.e13*const.hdm)*d.events.tof/(dataset.geometry.chopperDetectorDistance+d.events.detXdist)
 
         final_events = append_fields(d.events, [('lamda', np.float64)])

libeos/event_handling.py

@@ -157,9 +157,9 @@ class ApplyMask(EventDataAction):
         self.bitmask_filter = bitmask_filter
 
     def perform_action(self, dataset: EventDatasetProtocol) ->None:
+        # TODO: why is this action time consuming?
         d = dataset.data
-        logging.info(f'      number of events: total = {d.events.shape[0]:7d}, '
-                     f'filtered = {(d.events.mask!=0).sum():7d}')
+        pre_filter = d.events.shape[0]
         if logging.getLogger().level == logging.DEBUG:
             # only run this calculation if debug level is actually active
             filtered_by_mask = {}
@@ -173,3 +173,5 @@ class ApplyMask(EventDataAction):
             # this means that all bits that are set in bitmask_filter will NOT be used to filter events
             fltr = (d.events.mask & (~self.bitmask_filter)) == 0
             d.events = d.events[fltr]
+        post_filter = d.events.shape[0]
+        logging.info(f'      number of events: total = {pre_filter:7d}, filtered = {pre_filter-post_filter:7d}')