fix warnings

app/services/stops.py

@@ -169,7 +169,7 @@ def _find_stops_around_transformed(self, stopsDataFrame, transformedLine, distan
 
     def _convert_to_dataframe(self, stops):
         return gpd.GeoDataFrame([[stop.name, stop.lon, stop.lat,
-            stop.id, Point(self.projection(stop.lon, stop.lat))] for stop in stops], columns = ['stop_name','x','y','id','geometry'])
+            stop.id, Point(self.projection(stop.lon, stop.lat))] for stop in stops], columns = ['stop_name','x','y','id','geometry'], crs=self.internal_projection)
 
     def _sort_by_distance(self, stops, transformedLine):
         stops['distance']=stops.apply(lambda row: transformedLine.project(row['geometry']), axis=1)

app/services/trips.py

@@ -3,11 +3,11 @@
 from app.services.gtfs_constants import *
 from app.services.routing import RoutingService, RoutingException
 from app.services.stops import is_carpooling_stop
-from app.utils.utils import assert_folder_exists, is_older_than_days, yesterday
+from app.utils.utils import assert_folder_exists, is_older_than_days, yesterday, geodesic_distance_in_m
 from shapely.geometry import Point, LineString, box
 from geojson_pydantic.geometries import LineString as GeoJSONLineString
 from datetime import datetime, timedelta
-
+import numpy as np
 import os
 import json
 import logging
@@ -99,6 +99,10 @@ def put_carpool(self, carpool: Carpool):
             if existing_carpool and existing_carpool.lastUpdated == carpool.lastUpdated:
                 enhanced_carpool = existing_carpool
             else:
+                if len(carpool.stops) < 2 or self.distance_in_m(carpool) < 1000:
+                    logger.warning("Failed to add carpool %s:%s to TripStore, distance too low", carpool.agency, carpool.id)
+                    self.handle_failed_carpool_enhancement(carpool)
+                    return
                 enhanced_carpool = self.transformer.enhance_carpool(carpool)
                 # TODO should only store enhanced_carpool, if it has 2 or more stops
                 assert_folder_exists(f'data/enhanced/{carpool.agency}/')
@@ -109,14 +113,22 @@ def put_carpool(self, carpool: Carpool):
             return self._load_as_trip(enhanced_carpool)
         except RoutingException as err:
             logger.warning("Failed to add carpool %s:%s to TripStore due to RoutingException %s", carpool.agency, carpool.id, getattr(err, 'message', repr(err)))
-            assert_folder_exists(f'data/failed/{carpool.agency}/')
-            with open(f'data/failed/{carpool.agency}/{carpool.id}.json', 'w', encoding='utf-8') as f:
-                f.write(carpool.json()) 
+            self.handle_failed_carpool_enhancement(carpool)
         except Exception as err:
             logger.error("Failed to add carpool %s:%s to TripStore.", carpool.agency, carpool.id, exc_info=True)
-            assert_folder_exists(f'data/failed/{carpool.agency}/')
-            with open(f'data/failed/{carpool.agency}/{carpool.id}.json', 'w', encoding='utf-8') as f:
-                f.write(carpool.json())
+            self.handle_failed_carpool_enhancement(carpool)
+
+    def handle_failed_carpool_enhancement(sellf, carpool: Carpool):
+        assert_folder_exists(f'data/failed/{carpool.agency}/')
+        with open(f'data/failed/{carpool.agency}/{carpool.id}.json', 'w', encoding='utf-8') as f:
+            f.write(carpool.json())
+
+    def distance_in_m(self, carpool):
+        if len(carpool.stops) < 2:
+            return 0
+        s1 = carpool.stops[0]
+        s2 = carpool.stops[-1]
+        return geodesic_distance_in_m((s1.lon, s1.lat),(s2.lon, s2.lat)) 
 
     def recently_added_trips(self):
         return list(self.recent_trips.values())
@@ -229,8 +241,8 @@ def enhance_carpool(self, carpool):
         if len(virtual_stops) > MAX_STOPS_PER_TRIP:
             # in case we found more than MAX_STOPS_PER_TRIP, we retain first and last 
             # half of MAX_STOPS_PER_TRIP
-            virtual_stops = virtual_stops[:int(MAX_STOPS_PER_TRIP/2)]+virtual_stops[int(-MAX_STOPS_PER_TRIP/2):]
-
+            virtual_stops = virtual_stops.iloc[np.r_[0:int(MAX_STOPS_PER_TRIP/2), int(MAX_STOPS_PER_TRIP/2):]]
+            
         trip_id = f"{carpool.agency}:{carpool.id}"
         stop_times = self._stops_and_stop_times(carpool.departureTime, trip_id, virtual_stops)
 

app/utils/utils.py

@@ -1,6 +1,7 @@
 import os
 import re
 from datetime import datetime, date, timedelta
+from pyproj import Geod
 
 def assert_folder_exists(foldername):
     if not os.path.isdir(foldername):
@@ -30,4 +31,10 @@ def yesterday():
     return date_days_ago(1)
 
 def date_days_ago(number_of_days):
-    return date.today() - timedelta(days=number_of_days)
+    return date.today() - timedelta(days=number_of_days)
+
+def geodesic_distance_in_m(coord1, coord2):
+    geod = Geod(ellps="WGS84")
+    lons = [coord1[0], coord2[0]]
+    lats = [coord1[1], coord2[1]]
+    return geod.line_lengths(lons, lats)[0]