Fused MCP Agents

@fused.udf def udf(lat_start = 34.0154145, lng_start = -118.2253804, lat_end = 33.9422, lng_end = -118.4036, cost = 'auto'): import time import pandas as pd import requests from utils import decode, compute_distance #costing options: auto, pedestrian, bicycle, truck, bus, motor_scooter start = time.time() # Get the route directly from Valhalla server url = 'https://valhalla1.openstreetmap.de/route' params = { "locations": [{"lon": lng_start, "lat": lat_start},{"lon": lng_end, "lat": lat_end}] ,"costing": cost ,"units":"miles" } response = requests.post(url, json=params) result = response.json() # Extract driving instructions instructions = [el['instruction'] for el in result['trip']['legs'][0]['maneuvers']] print('Driving Instructions:', instructions) # Get the encoded shape and decode it encoded_shape = result['trip']['legs'][0]['shape'] decoded_shape = decode(encoded_shape) # Create temporary GeoDataFrame to compute distance import geopandas as gpd from shapely.geometry import LineString gdf_route = gpd.GeoDataFrame(columns=['geometry'], geometry=[LineString(decoded_shape)], crs=4326) gdf_route = compute_distance(gdf_route, col_name='separation') # Extract route length route_length_km = gdf_route['separation'].values[0] / 1000 print('time to create route', time.time() - start, 'sec') print('route length:', route_length_km, 'km') # Create a simple DataFrame with the relevant information route_info = { 'start_lat': lat_start, 'start_lng': lng_start, 'end_lat': lat_end, 'end_lng': lng_end, 'transportation_mode': cost, 'route_length_km': route_length_km, 'computation_time_sec': time.time() - start, 'instructions': instructions } return pd.DataFrame([route_info])