Fix bug in monitoring of the number of candidates

The current implementation of the n_candidates plots the number of candidates for every candidate instead of only once. This MR uses now a new algorithm to plot event variables. The monitoring of the number of candidates is explicitly added to the Topo lines.

Needs Rec!3555 (merged)

Hlt/Moore/python/Moore/lines.py

@@ -152,7 +152,8 @@ class DecisionLine(object):
 
         The producer is defined as the last child in the control flow node,
         i.e. the last item passed as the `algs` argument to the
-        `DecisionLine` constructor.
+        `DecisionLine` constructor. In cases where the last algorithms are monitoring
+        algorithm, the last producer is searched for.
 
         How the output data of the producer is used depends on the application.
 
@@ -160,7 +161,7 @@ class DecisionLine(object):
         """
         for c in reversed(node.children):
             if isinstance(c, Algorithm):
-                if c.typename != "Monitor__ParticleRange":
+                if c.typename != "Monitor__ParticleRange" and c.typename != "Monitor__EventVariable":
                     last = c
                     break
 
@@ -480,7 +481,7 @@ class Hlt2Line(DecisionLine):
                 the location tuple (an N-tuple of str).
             persistreco (bool): If True, this line requests the full HLT2
                 reconstruction be persisted along with its other outputs.
-            tagging_particles (bool): If True, this line requests the proto 
+            tagging_particles (bool): If True, this line requests the proto
                 particles for tagging  be persisted along with its other outputs.
 
         Returns: