diff options
Diffstat (limited to 'dashboard_website/router.py')
| -rw-r--r-- | dashboard_website/router.py | 53 |
1 files changed, 30 insertions, 23 deletions
diff --git a/dashboard_website/router.py b/dashboard_website/router.py index 99c83d2..5bafd1e 100644 --- a/dashboard_website/router.py +++ b/dashboard_website/router.py @@ -2,7 +2,7 @@ import numpy as np import requests from sklearn.cluster import KMeans -from dashboard_website import db +from datastructs import * host = "http://acetyl.net:5000" # queries acetyl.net:5000, the OSRM engine @@ -37,21 +37,27 @@ def getZSP(bike, home, clue_cluster): # determines clusters based on current bikes and clues def getClusters(bikes, clues, endpoint): - clusters = [[] for bike in bikes] + + clusters = [[] for bike in bikes ] + route_geos = [[] for bike in bikes ] + active_indices = [i for i in range(len(bikes)) if bikes[i].status == "ACTIVE"] active_bikes = [bike for bike in bikes if bike.status == "ACTIVE"] + if len(active_bikes) == 0: + return clusters, route_geos active_clues = [clue for clue in clues if clue.status == "UNVISITED"] # select only active bikes # select only unvisited clues - clusters_t, route_geo = cluster_and_optimize(active_clues, active_bikes, endpoint) - for cluster in clusters_t: - clusters[i] = cluster + clusters_t, route_geos_t = cluster_and_optimize(active_clues, active_bikes, endpoint) + for i in range(len(active_indices)): + route_geos[active_indices[i]] = route_geos_t[i] + clusters[active_indices[i]] = clusters_t[i] # return list of clue clusters corresponding to bikes - pass + return clusters, route_geos # utility functions (internal) -def cluster_and_optimize(clues: [db.Clue], bikes: [db.Bike], end: db.Point, time_diff=0.25, max_time=24, n=2): +def cluster_and_optimize(clues: [Clue], bikes: [Bike], end: Point, time_diff=0.25, max_time=24, n=2): """ Takes a dataframe of gps coordinates, a list of centroids, and an end point and returns a dataframe with a cluster :param clues: a list of clues @@ -65,7 +71,7 @@ def cluster_and_optimize(clues: [db.Clue], bikes: [db.Bike], end: db.Point, time # Create a new column with normalized gps coordinates and centroids normalized_points, norm_centroids = __normalize_points(clues, bikes) - + print(norm_centroids) # Cluster the coordinates kmeans = KMeans(n_clusters=len(norm_centroids), init=norm_centroids) kmeans.fit(normalized_points) @@ -82,21 +88,22 @@ def cluster_and_optimize(clues: [db.Clue], bikes: [db.Bike], end: db.Point, time routes[i] = __remove_longest_waypoints(routes[i], bikes[i], end, max_time) # Get the json of the routes - route_geo = [] + route_waypoints = [] + geometries = [] for i, route in enumerate(routes): - route_geo.append( - __get_json(__clues_to_string(route), __clues_to_string([bikes[i]]), __clues_to_string([end])[:-1])[ - 'waypoints']) + route_json = __get_json(__clues_to_string(route), __clues_to_string([bikes[i]]), __clues_to_string([end])[:-1]) + geometries.append(route_json['trips'][0]['geometry']['coordinates']) + route_waypoints.append(route_json['waypoints']) # Use the waypoint_index to reorder each route for i, route in enumerate(routes): - route = [route[j] for j in route_geo[i][0]['waypoint_index']] + route = [ route[ j['waypoint_index']-1 ] for j in route_waypoints[i] if route_waypoints[i].index(j) < (len(route_waypoints[i])-1) ] routes[i] = route - return routes + return routes, geometries -def __clues_to_string(points: [db.Clue]): +def __clues_to_string(points: [Clue]): """ Takes a list of points and returns a string of the list of points :param points: a list of points @@ -146,7 +153,7 @@ def __get_trip_time(coordinate_string, num_waypoints, start, end, time_per_waypo return total_time_hours -def __minimize_route_time_diff(routes: [db.Clue], starts: [db.Point], end: db.Point, time_diff, n): +def __minimize_route_time_diff(routes: [Clue], starts: [Point], end: Point, time_diff, n): """ Takes a list of lists of coordinates, a list of start points, an end point, a time difference, and a number of routes :param routes: the list of lists of coordinates @@ -186,7 +193,7 @@ def __minimize_route_time_diff(routes: [db.Clue], starts: [db.Point], end: db.Po return routes -def __remove_longest_waypoints(route_coordinates: [db.Clue], start: db.Bike, end: db.Point, max_time): +def __remove_longest_waypoints(route_coordinates: [Clue], start: Bike, end: Point, max_time): """ Takes a list of coordinates, a start point, an end point, and a maximum time and returns a list of coordinates :param route_coordinates: the list of coordinates @@ -210,7 +217,7 @@ def __remove_longest_waypoints(route_coordinates: [db.Clue], start: db.Bike, end return route_coordinates -def __normalize_points(clues: [db.Clue], bikes: [db.Bike]): +def __normalize_points(clues: [Clue], bikes: [Bike]): """ Takes a list of coordinates and a list of centroids and returns a list of normalized coordinates and a list of normalized centroids @@ -254,7 +261,7 @@ def __min_max_normalize(value, min_value, max_value): return (value - min_value) / (max_value - min_value) -def __find_closest_coordinate(clues: [db.Clue], centroid: db.Point): +def __find_closest_coordinate(clues: [Clue], centroid: Point): """ Takes a list of coordinates and a centroid and returns the clue in the list that is closest to the centroid :param clues: the list of coordinates @@ -273,7 +280,7 @@ def __find_closest_coordinate(clues: [db.Clue], centroid: db.Point): return closest_coordinate -def __find_farthest_coordinate(clues: [db.Clue], centroid: db.Point): +def __find_farthest_coordinate(clues: [Clue], centroid: Point): """ Takes a list of coordinates and a centroid and returns the clue in the list that is farthest from the centroid :param clues: the list of coordinates @@ -291,17 +298,17 @@ def __find_farthest_coordinate(clues: [db.Clue], centroid: db.Point): return farthest_coordinate -def __mean_center(clues: [db.Clue]): +def __mean_center(clues: [Clue]): """ Takes a list of coordinates and returns the mean center of the coordinates :param clues: the list of coordinates :return: the mean center of the coordinates """ - return db.Point(np.mean([coordinate.latitude for coordinate in clues]), + return Point(np.mean([coordinate.latitude for coordinate in clues]), np.mean([coordinate.longitude for coordinate in clues])) -def __distance(coordinate1: db.Clue, coordinate2: db.Point): +def __distance(coordinate1: Clue, coordinate2: Point): """ Takes two coordinates and returns the distance between them :param coordinate1: the first coordinate |