abstreet/headless/examples/abst_helpers.py

83 lines
2.4 KiB
Python
Raw Normal View History

import requests
import statistics
def get(args, cmd, **kwargs):
resp = requests.get(args.api + cmd, **kwargs)
if resp.status_code != requests.codes.ok:
raise Exception(resp.text)
return resp
def post(args, cmd, **kwargs):
resp = requests.post(args.api + cmd, **kwargs)
if resp.status_code != requests.codes.ok:
raise Exception(resp.text)
return resp
# Returns Results
def run_sim(args, modifiers=[], edits=None):
post(args, '/sim/load', json={
'scenario': 'data/system/scenarios/{}/weekday.bin'.format(args.map_name),
'modifiers': modifiers,
'edits': edits,
})
post(args, '/sim/goto-time',
params={'t': '{}:00:00'.format(args.hours)})
raw_trips = get(args, '/data/get-finished-trips').json()
# Map trip ID to the duration (in seconds) of the trip. Filter out aborted
# (failed) trips.
num_aborted = 0
trip_times = {}
capped_trips = set()
for trip in raw_trips:
if trip['mode'] is None:
num_aborted += 1
else:
trip_times[trip['id']] = trip['duration']
if trip['capped']:
capped_trips.add(trip['id'])
return Results(num_aborted, trip_times, capped_trips)
class Results:
def __init__(self, num_aborted, trip_times, capped_trips):
self.num_aborted = num_aborted
# Maps trip ID to seconds
self.trip_times = trip_times
# A set of trip IDs
self.capped_trips = capped_trips
# self is the baseline, results2 is the experiment
def compare(self, results2):
faster = []
slower = []
for trip, after_dt in results2.trip_times.items():
before_dt = self.trip_times.get(trip)
if not before_dt:
# The trip didn't finish in time in the baseline run
continue
if before_dt:
if before_dt > after_dt:
faster.append(before_dt - after_dt)
elif after_dt > before_dt:
slower.append(after_dt - before_dt)
print('{:,} trips faster, average {:.1f}s savings'.format(
len(faster), avg(faster)))
print('{:,} trips slower, average {:.1f}s loss'.format(
len(slower), avg(slower)))
print('{:,} trips aborted before, {:,} after'.format(
self.num_aborted, results2.num_aborted))
def avg(data):
if data:
return statistics.mean(data)
else:
return 0.0