def run_setup(input_network_dir,
input_demand_dir,
input_weights,
run_config,
iters,
output_dir,
pathfinding_type='stochastic',
input_functions=None,
output_folder=None,
trace_only=False,
skim_config_file=None,
**kwargs):
"""
Reads run configuration files from network and demand input directories.
If additional parameters are input here, they will override the run configuration files.
Named Keyword arguments:
input_network_dir -- the input directory with GTFS-PLUS networks (required)
input_demand_dir -- the input directory with the dyno-demand files (required)
input_weights -- the file with the pathweight parameters (required)
run_config -- the file with further run configs (required)
iters -- number of global iterations, integer (required)
output_dir -- where to put the output folder (required). Will be created if it doesn't already exist.
pathfinding_type -- one of ['stochastic','deterministic','file'] (default: stochastic)
input_functions -- the file with user functions
output_folder -- where to put the outputs. Will be created if it doesn't already exist.
trace_only -- will only run demand for the trace. Will not overwrite other output (default: False)
Unnamed keyword arguments:
num_trips -- if specified, will process this number of trips and ignore the rest
pf_iters -- Integer. If specified, will set the maximum number of pathfinding iterations(default: 10)
dispersion -- theta parameter; essentially the nesting parameter. Good value is between 0.5-1. (default: 1.0)
max_stop_process_count = maximum number of times you will re-processe a node (default: 20)
capacity -- Boolean to activate capacity constraints (default: False)
overlap_variable -- One of ['None','count','distance','time']. Variable to use for overlap penalty calculation (default: 'count')
overlap_split_transit -- Boolean.Split transit for path overlap penalty calculation (default: False)
time_window = Integer. The time window in minutes where a passenger searches for a eminating transit route at each node.
transfer_fare_ignore_pathfinding = Boolean. In path-finding, suppress trying to adjust fares using transfer rules. For performance.
transfer_fare_ignore_pathenum = Boolean. In path-enumeration, suppress trying to adjust fares using transfer rules. For performance.
number_of_processes = Integer. Number of processes to run at once (default: 1)
output_pathset_per_sim_iter = Boolean. Output pathsets per simulation iteration? (default: false)
debug_output_columnns -- boolean to activate extra columns for debugging (default: False)
"""
print(kwargs)
if not input_network_dir:
msg = "Must specify where to find input networks"
FastTripsLogger.fatal(msg)
raise fasttrips.ConfigurationError("external input", msg)
if not input_demand_dir:
msg = "Must specify where to find input demand"
FastTripsLogger.fatal(msg)
raise fasttrips.ConfigurationError("external input", msg)
if not input_weights:
msg = "Must specify where pathweight parameters are"
FastTripsLogger.fatal(msg)
raise fasttrips.ConfigurationError("external input", msg)
if not run_config:
msg = "Must specify file with run configurations"
FastTripsLogger.fatal(msg)
raise fasttrips.ConfigurationError("external input", msg)
if pathfinding_type not in ['deterministic', 'stochastic', 'file']:
msg = "pathfinding.type [%s] not defined. Expected values: %s" % (
pathfinding_type, ['deterministic', 'stochastic', 'file'])
FastTripsLogger.fatal(msg)
raise fasttrips.ConfigurationError("external override", msg)
# Setup Output Directory
if not output_folder:
output_folder = "output_%s_iter%d_%s" % (
pathfinding_type, iters, "cap" if
("capacity" in kwargs and kwargs["capacity"] == True) else "nocap")
# don't override full run results
if trace_only:
output_folder = "%s_trace" % output_folder
if not output_dir:
output_dir = os.path.basename(input_demand_dir)
# create folder if it doesn't already exist
full_output_dir = os.path.join(output_dir, output_folder)
if not os.path.exists(output_dir):
os.mkdir(output_dir)
if not os.path.exists(full_output_dir):
print("Creating full output dir [%s]" % full_output_dir)
os.mkdir(full_output_dir)
# Create fast-trips instance
ft = fasttrips.FastTrips(input_network_dir,
input_demand_dir,
input_weights,
run_config,
full_output_dir,
input_functions=input_functions,
skim_config_file=skim_config_file)
# Read the configuration file and overwrite with any options called with the function call
ft.read_configuration()
if iters > 0:
fasttrips.Assignment.MAX_ITERATIONS = int(iters)
if "pf_iters" in kwargs:
fasttrips.Assignment.MAX_PF_ITERATIONS = kwargs["pf_iters"]
if "number_of_processes" in kwargs:
fasttrips.Assignment.NUMBER_OF_PROCESSES = kwargs[
"number_of_processes"]
if "trace_ids" in list(kwargs.keys()):
fasttrips.Assignment.TRACE_IDS = kwargs["trace_ids"]
if trace_only:
if len(fasttrips.Assignment.TRACE_IDS) == 0:
print(
"Trace only requested but no trace IDs are specified in configuration."
)
sys.exit(2)
fasttrips.Assignment.DEBUG_TRACE_ONLY = True
fasttrips.Assignment.NUMBER_OF_PROCESSES = 1
if "pathfinding_type" in list(kwargs.keys()):
fasttrips.Assignment.PATHFINDING_TYPE = kwargs["pathfinding_type"]
if "learning_convergence" in list(kwargs.keys()):
fasttrips.PathSet.LEARN_ROUTES = kwargs["learning_convergence"]
if "max_stop_process_count" in list(kwargs.keys()):
fasttrips.Assignment.STOCH_MAX_STOP_PROCESS_COUNT = kwargs[
"max_stop_process_count"]
if "debug_output_columns" in list(kwargs.keys()):
fasttrips.Assignment.DEBUG_OUTPUT_COLUMNS = kwargs[
"debug_output_columns"]
if "overlap_variable" in list(kwargs.keys()):
if kwargs["overlap_variable"] not in [
'None', 'count', 'distance', 'time'
]:
msg = "pathfinding.overlap_variable [%s] not defined. Expected values: %s" % (
kwargs["overlap_variable"],
str(fasttrips.PathSet.OVERLAP_VARIABLE_OPTIONS))
fasttrips.FastTripsLogger.fatal(msg)
raise fasttrips.ConfigurationError("external override", msg)
fasttrips.PathSet.OVERLAP_VARIABLE = kwargs["overlap_variable"]
if "overlap_split_transit" in list(kwargs.keys()):
fasttrips.PathSet.OVERLAP_SPLIT_TRANSIT = kwargs[
"overlap_split_transit"]
if "transfer_fare_ignore_pathfinding" in list(kwargs.keys()):
fasttrips.Assignment.TRANSFER_FARE_IGNORE_PATHFINDING = kwargs[
"transfer_fare_ignore_pathfinding"]
if "transfer_fare_ignore_pathenum" in list(kwargs.keys()):
fasttrips.Assignment.TRANSFER_FARE_IGNORE_PATHENUM = kwargs[
"transfer_fare_ignore_pathenum"]
if "time_window" in list(kwargs.keys()):
fasttrips.Assignment.TIME_WINDOW = datetime.timedelta(
minutes=float(kwargs["time_window"]))
if "utils_conversion_factor" in list(kwargs.keys()):
fasttrips.Assignment.UTILS_CONVERSION = kwargs[
"utils_conversion_factor"]
if "dispersion" in list(kwargs.keys()):
fasttrips.Assignment.STOCH_DISPERSION = kwargs["dispersion"]
if "num_trips" in list(kwargs.keys()):
fasttrips.Assignment.DEBUG_NUM_TRIPS = kwargs["num_trips"]
if "capacity" in list(kwargs.keys()):
fasttrips.Assignment.CAPACITY_CONSTRAINT = kwargs["capacity"]
if "output_pathset_per_sim_iter" in list(kwargs.keys()):
fasttrips.Assignment.OUTPUT_PATHSET_PER_SIM_ITER = kwargs[
"output_pathset_per_sim_iter"]
if "user_class_function" in list(kwargs.keys()):
fasttrips.PathSet.USER_CLASS_FUNCTION = kwargs["user_class_function"]
return ft
person_trips_df = person_trips_df.sort_values(['Unique_ID','route_id','agency_id','stop_dist'])
person_trips_df = person_trips_df.groupby(['Unique_ID','route_id']).head(1)
# rename the new columns
person_trips_df.rename(columns={"stop_id" :"%s_stop_id" % location_prefix,
"stop_name" :"%s_stop_name" % location_prefix,
"stop_lat" :"%s_stop_lat" % location_prefix,
"stop_lon" :"%s_stop_lon" % location_prefix,
"stop_dist" :"%s_stop_dist" % location_prefix,
"route_id" :"%s_route_id" % location_prefix,
"agency_id" :"%s_agency_id" % location_prefix}, inplace=True)
return person_trips_df
if __name__ == "__main__":
ft = fasttrips.FastTrips(input_network_dir= NETWORK_DIR,
input_demand_dir = None,
output_dir = ".")
ft.read_input_files()
# Get the full (joined) transit vehicle trip table and add lat/lon for the stops
full_trips_df = ft.trips.get_full_trips()
full_trips_df = ft.stops.add_stop_lat_lon(full_trips_df, id_colname="stop_id", new_lat_colname="stop_lat", new_lon_colname="stop_lon", new_stop_name_colname="stop_name")
# Add operator_type as the mutual field between CHTS and GTFS vehicle trips
full_trips_df["operator_type"] = full_trips_df["agency_id"].replace(NETWORK_AGENCY_TO_OPERATOR_TYPE)
# Read CHTS observed person trips file
df = pd.read_csv('CHTS_ft_output.csv')
# Add operator_type as the mutual field between CHTS and GTFS vehicle trips
df["operator_type"] = df["transit_mode_no"].replace(CHTS_MODE_TO_OPERATOR_TYPE)
# Add Unique_ID to be used for merging later on
df['Unique_ID'] = df.index
'--description',
dest='use_description',
action='store_true',
help=
"Specify this to use path description as ID. Otherwise will use standard pathset fields (person_id, person_trip_id, iteration, pathfinding_iteration, simulation_iteration, pathnum)"
)
args = parser.parse_args()
LOG_DIR = "create_tableau_path_map"
if not os.path.exists(LOG_DIR):
os.mkdir(LOG_DIR)
# Use fasttrips to read the network
ft = fasttrips.FastTrips(input_network_dir=args.input_network_dir[0],
input_demand_dir=None,
output_dir=LOG_DIR)
ft.read_input_files()
# and the pathset files
(pathset_paths_df,
pathset_links_df) = ft.passengers.read_passenger_pathsets(
args.input_path_dir[0], ft.stops, ft.routes.modes_df)
fasttrips.Assignment.TRACE_PERSON_IDS = pathset_paths_df["person_id"].head(
5).tolist()
fasttrips.FastTripsLogger.debug("Read passenger pathsets. head=\n%s" %
str(pathset_links_df.head(100)))
pathset_links_df = add_taz_coords(args.input_network_dir[0],
pathset_links_df)
pathset_links_df = add_stop_coords(ft, pathset_links_df)
else:
test_dir = "%s%s_iter%d_%s" % (
"" if args.input_network_dir == args.input_demand_dir else "%s_" %
os.path.basename(args.input_demand_dir), args.pathfinding_type,
args.iters, "cap" if args.capacity else "nocap")
# don't override full run results
if args.trace_only:
test_dir = "%s_trace" % test_dir
full_output_dir = os.path.join(args.output_loc, test_dir)
if not os.path.exists(full_output_dir):
print "Creating full output dir [%s]" % full_output_dir
os.mkdir(full_output_dir)
ft = fasttrips.FastTrips(args.input_network_dir, args.input_demand_dir,
full_output_dir)
# Read the configuration here so we can overwrite options below
ft.read_configuration()
fasttrips.Assignment.PATHFINDING_TYPE = args.pathfinding_type
fasttrips.Assignment.ITERATION_FLAG = int(args.iters)
if args.max_stop_process_count:
fasttrips.Assignment.STOCH_MAX_STOP_PROCESS_COUNT = args.max_stop_process_count
if args.overlap_variable:
fasttrips.PathSet.OVERLAP_VARIABLE = args.overlap_variable
if args.overlap_split_transit:
fasttrips.PathSet.OVERLAP_SPLIT_TRANSIT = args.overlap_split_transit
rec['agency_id'] = ''
done_queue.put(rec)
except:
FastTripsLogger.exception("Exception")
# call it a day
done_queue.put((worker_num, "EXCEPTION", str(sys.exc_info())))
break
if __name__ == "__main__":
start_time = datetime.datetime.now()
## Read FT network input files and get transit trips information
ft = fasttrips.FastTrips(input_network_dir=NETWORK_DIR,
input_demand_dir=None,
input_weights=None,
run_config=None,
output_dir=".")
fasttrips.FastTripsLogger.info("Reading network from [%s]" % NETWORK_DIR)
ft.read_input_files()
# Get the full (joined) transit vehicle trip table and add lat/lon for the stops
full_trips_df = ft.trips.get_full_trips()
full_trips_df = ft.stops.add_stop_lat_lon(
full_trips_df,
id_colname="stop_id",
new_lat_colname="stop_lat",
new_lon_colname="stop_lon",
new_stop_name_colname="stop_name")
# full_trips_df.to_csv('full_trips.csv', index=False)
# sys.exit()
pd.set_option('display.width', 1000)
pd.set_option('display.max_rows', 1000)
pd.set_option('display.max_columns', 100)
parser = argparse.ArgumentParser()
parser.add_argument("network_dir",
type=str,
help="Location of GTFS-Plus network")
args = parser.parse_args(sys.argv[1:])
# Initialize fasttrips; we'll use it to read the network
fasttrips.FastTripsLogger.info("Reading network from [%s]" %
args.network_dir)
ft = fasttrips.FastTrips(input_network_dir=args.network_dir,
input_demand_dir=None,
output_dir=".")
# Read the observed person trips file
df = pd.read_csv('survey_wSFtaz.csv',
dtype={
"onoff_enter_station": object,
"onoff_exit_station": object,
"persons": object
},
na_values=["None"])
# Keep only the columns we're interested in
df = df[[
'Unique_ID',
'operator',
'transfer_from', # operator