fix max_nb_journeys with best_olympics when aggregating

source/jormungandr/jormungandr/scenarios/new_default.py

@@ -1121,7 +1121,7 @@ def get_crowfly_air_pollutants(origin, destination):
 
 def aggregate_journeys(journeys):
     """
-    when building candidates_pool, we should take into count the similarity of journeys, which means, we add a journey
+    when building candidates_pool, we should take into account the similarity of journeys, which means, we add a journey
     into the pool only when there are no other "similar" journey already existing in the pool.
 
     the similarity is defined by a tuple of journeys sections.
@@ -1130,20 +1130,29 @@ def aggregate_journeys(journeys):
     aggregated_journeys = list()
     remaining_journeys = list()
 
+    def to_retain(j):
+        return j.type in JOURNEY_TYPES_TO_RETAIN or set(j.tags) & set(JOURNEY_TAGS_TO_RETAIN)
+
+    journeys_to_retain = (j for j in journeys if to_retain(j))
+    journeys_not_to_retain = (j for j in journeys if not to_retain(j))
+
     # we pick out all journeys that must be kept:
-    for j in (j for j in journeys if j.type in JOURNEY_TYPES_TO_RETAIN):
+    for j in journeys_to_retain:
         section_id = tuple(_get_section_id(s) for s in j.sections if s.type in SECTION_TYPES_TO_RETAIN)
         aggregated_journeys.append(j)
         added_sections_ids.add(section_id)
 
-    for j in (j for j in journeys if j.type not in JOURNEY_TYPES_TO_RETAIN):
+    for j in journeys_not_to_retain:
         section_id = tuple(_get_section_id(s) for s in j.sections if s.type in SECTION_TYPES_TO_RETAIN)
 
         if section_id in added_sections_ids:
             remaining_journeys.append(j)
         else:
             aggregated_journeys.append(j)
             added_sections_ids.add(section_id)
+
+    # aggregated_journeys will be passed to culling algorithm,
+    # remaining_journeys are the redundant ones to be removed
     return aggregated_journeys, remaining_journeys
 
 

source/jormungandr/jormungandr/scenarios/tests/new_default_tests.py

@@ -318,6 +318,26 @@ def aggregate_journeys_test():
     assert len(aggregated_journeys) == 17
     assert len(remaining_journeys) == 2
 
+    journeys_uris = {(tuple(s.uris.line for s in j.sections), j.arrival_date_time) for j in aggregated_journeys}
+    # J19 is dominated by J3, because it arrives later than J3
+    # J3 should be found in final result
+    assert ((u'uri_2', u'uri_3', u'uri_4', u'walking'), 1444905600) in journeys_uris
+    # J3 should NOT be found in final result
+    assert ((u'uri_2', u'uri_3', u'uri_4', u'walking'), 1444905720) not in journeys_uris
+
+    mocked_pb_response = build_mocked_response()
+    mocked_pb_response.journeys[18].tags.append("best_olympics")
+
+    aggregated_journeys, remaining_journeys = new_default.aggregate_journeys(mocked_pb_response.journeys)
+    assert len(aggregated_journeys) == 17
+    assert len(remaining_journeys) == 2
+    journeys_uris = {(tuple(s.uris.line for s in j.sections), j.arrival_date_time) for j in aggregated_journeys}
+    # J19 is dominated by J3, BUT it has joker because it has been tagged "best_olympics"
+    # J3 should be NOT found in final result
+    assert ((u'uri_2', u'uri_3', u'uri_4', u'walking'), 1444905600) not in journeys_uris
+    # J3 should be found in final result
+    assert ((u'uri_2', u'uri_3', u'uri_4', u'walking'), 1444905720) in journeys_uris
+
 
 def merge_responses_on_errors_test():
     """