Use max instead of sum on the times count for the denominator

gsarrco · Jan 17, 2024 · 46dd530 · 46dd530
1 parent 47a5706
commit 46dd530
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 7 deletions.
diff --git a/server/GTFS/source.py b/server/GTFS/source.py
@@ -183,14 +183,14 @@ def upload_stops_clusters_to_db(self, force=False) -> bool:
         ''')
         stops = self.get_all_stops()
         stops_clusters = get_clusters_of_stops(stops)
-        total_times_count = sum([cluster.times_count for cluster in stops_clusters])
+        max_times_count = max([cluster.times_count for cluster in stops_clusters])
 
         new_stations = []
         new_stops = []
 
         for cluster in stops_clusters:
-            times_count = round(cluster.times_count / total_times_count,
-                                int(math.log10(total_times_count)) + 1)
+            times_count = round(cluster.times_count / max_times_count,
+                                int(math.log10(max_times_count)) + 1)
             ids = ','.join([str(stop.id) for stop in cluster.stops])
             station = Station(id=cluster.name, name=cluster.name, lat=cluster.lat, lon=cluster.lon, ids=ids,
                               times_count=times_count, source=self.name)

diff --git a/server/trenitalia/source.py b/server/trenitalia/source.py
@@ -41,21 +41,23 @@ def __init__(self, session, typesense, location='', force_update_stations=False)
     def save_data(self):
         stations = self.session.scalars(select(Station).filter_by(source=self.name, active=True)).all()
 
-        total_times_count = 0
+        max_times_count = 0
         times_count = []
 
         tqdm_stations = tqdm(enumerate(stations), total=len(stations), desc=f'Uploading {self.name} data')
 
         for i, station in tqdm_stations:
             tqdm_stations.set_description(f'Processing station {station.name}')
             stop_times = self.get_stop_times_from_station(station)
-            total_times_count += len(stop_times)
-            times_count.append(len(stop_times))
+            stop_times_count = len(stop_times)
+            if stop_times_count > max_times_count:
+                max_times_count = stop_times_count
+            times_count.append(stop_times_count)
             for stop_time in stop_times:
                 self.upload_trip_stop_time_to_postgres(stop_time)
 
         for i, station in enumerate(stations):
-            station.times_count = round(times_count[i] / total_times_count, int(math.log10(total_times_count)) + 1)
+            station.times_count = round(times_count[i] / max_times_count, int(math.log10(max_times_count)) + 1)
         self.sync_stations_db(stations)
 
     def get_stop_times_from_station(self, station) -> list[TripStopTime]: