fix: update _summarise_core() in MGTFS to return correct results; cho…

…re: update tests (#233)
datasciencecampus · Jan 17, 2024 · aa98494 · aa98494
1 parent 42e6dd1
commit aa98494
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/src/transport_performance/gtfs/multi_validation.py b/src/transport_performance/gtfs/multi_validation.py
@@ -399,7 +399,9 @@ def _summarise_core(
         if not return_summary:
             return daily_schedule
         if not to_days:
-            dated = daily_schedule[["date", "route_type", group_col]]
+            dated = daily_schedule[
+                ["date", "route_type", group_col]
+            ].drop_duplicates()
             dated = dated.groupby(["date", "route_type"]).count().reset_index()
             dated = dated.rename(columns={group_col: count_col})
             dated = dated.sort_values("date")

diff --git a/tests/gtfs/test_multi_validation.py b/tests/gtfs/test_multi_validation.py
@@ -326,7 +326,7 @@ def test__summarise_core(self, multi_gtfs_fixture):
             {
                 "date": ["2023-06-06", "2023-06-06"],
                 "route_type": [3, 200],
-                "route_count": [151, 25],
+                "route_count": [12, 4],
             },
             index=[2, 3],
         )
@@ -368,8 +368,7 @@ def test__summarise_core(self, multi_gtfs_fixture):
         )
         assert len(dated_sum) == 590, "Dated route counts not as expected"
         assert (
-            dated_sum[dated_sum.date == "2024-04-06"].route_count.iloc[0]
-            == 159
+            dated_sum[dated_sum.date == "2024-04-06"].route_count.iloc[0] == 9
         ), "Unexpecteed number of routes on 2024-04-06"
 
     def test_summarise_trips(self, multi_gtfs_fixture):