abstreet/headless/examples/generate_traffic.py

66 lines
2.2 KiB
Python
Raw Normal View History

#!/usr/bin/python3
# This example loads an exported JSON map, finds different buildings, and
# generates a simple travel demand model.
#
# 1) cargo run --bin dump_map data/system/us/seattle/maps/montlake.bin > montlake.json
# 2) ./headless/examples/generate_traffic.py --map=montlake.json --out=traffic.json
# 3) cargo run --bin import_traffic -- --map=data/system/us/seattle/maps/montlake.bin --input=traffic.json
# 4) Use data/system/us/seattle/scenarios/montlake/monday.bin in the game or from the API.
#
# Keep this script formatted with autopep8 -i
import argparse
import json
import random
def main():
parser = argparse.ArgumentParser()
parser.add_argument('--map', type=str, required=True)
parser.add_argument('--out', type=str, required=True)
args = parser.parse_args()
# Load the map and find all buildings
residential_building_ids = []
commercial_building_ids = []
with open(args.map, encoding='utf8') as f:
map = json.load(f)
for b in map['buildings']:
# These categories are inferred from OpenStreetMap tags
if 'Residential' in b['bldg_type'] or 'ResidentialCommercial' in b['bldg_type']:
residential_building_ids.append(b['id'])
if 'Commercial' in b['bldg_type'] or 'ResidentialCommercial' in b['bldg_type']:
commercial_building_ids.append(b['id'])
# Randomly generate a few people who take just one trip
scenario = {
'scenario_name': 'monday',
'people': []
}
for _ in range(100):
src = random.choice(residential_building_ids)
dst = random.choice(commercial_building_ids)
scenario['people'].append({
'origin': {
'TripEndpoint': {
'Bldg': src,
}
},
'trips': [{
'departure': 1.0,
'destination': {
'TripEndpoint': {
'Bldg': dst,
}
},
'mode': 'Bike'
}]
})
with open(args.out, 'w') as f:
f.write(json.dumps(scenario, indent=2))
if __name__ == '__main__':
main()