fasttrips.PathSet

class fasttrips.PathSet(trip_list_dict)

Bases: object

Represents a path set for a passenger from an origin TAZ to a destination TAZ through a set of stops.

__init__(trip_list_dict)

Constructor from dictionary mapping attribute to value.

Methods

__init__(trip_list_dict)	Constructor from dictionary mapping attribute to value.
calculate_cost(STOCH_DISPERSION, …[, …])	This is equivalent to the C++ Path::calculateCost() method.
calculate_overlap(pathset_links_to_use)	Given a set of pathset links, returns a the results of overlap calculations.
goes_somewhere()	Does this path go somewhere? Does the destination differ from the origin?
num_paths()	Number of paths in the PathSet
path_found()	Was a a transit path found from the origin to the destination with the constraints?
reset()	Delete my states, something went wrong and it won’t work out.
set_user_class(trip_list_df, new_colname)	Adds a column called user_class by applying the configured user class function.
split_transit_links(pathset_links_df, …)	Splits the transit links to their component links and returns.
verify_weight_config(modes_df, output_dir, …)	Verify that we have complete weight configurations for the user classes and modes in the given DataFrame.
verify_weights(weights)
write_path_times(passengers_df, output_dir)	Write the assigned path times to the given output file.
write_paths(passengers_df, output_dir)	Write the assigned paths to the given output file.

Attributes

ARRIVE_LATE_ALLOWED_MIN	Allow departures and arrivals before / after preferred time
BUMP_EXPERIENCED_COST
BUMP_FLAG_COLUMN
CONFIGURED_FUNCTIONS	Configured functions, indexed by name
CONSTANT_GROWTH_MODEL
DEPART_EARLY_ALLOWED_MIN
DIR_INBOUND	Trips inbound to home have preferred departure times
DIR_OUTBOUND	Trips outbound from home have preferred arrival times
EXP_GROWTH_MODEL
HUGE_COST
LEARN_ROUTES
LEARN_ROUTES_RATE
LOGARITHMIC_GROWTH_MODEL
LOGISTIC_GROWTH_MODEL
MIN_TRANSFER_PENALTY	Configuration: Minimum transfer penalty.
OUTPUT_WEIGHTS_FILE	File with weights for c++
OVERLAP_CHUNK_SIZE	Overlap chunk size.
OVERLAP_COUNT	Overlap variable option: count.
OVERLAP_DISTANCE	Overlap variable option: distance.
OVERLAP_NONE	Overlap variable option: None.
OVERLAP_SCALE_PARAMETER	Configuration: Overlap scale parameter.
OVERLAP_SPLIT_TRANSIT	Overlap option: Split transit leg into component parts? e.g.
OVERLAP_TIME	Overlap variable option: time.
OVERLAP_VARIABLE	Configuration: Overlap variable.
OVERLAP_VARIABLE_OPTIONS	Valid values for OVERLAP_VARAIBLE
PATHS_OUTPUT_FILE	Paths output file
PATH_KEY_COST	path cost according to pathfinder
PATH_KEY_FARE	path fare according to pathfinder
PATH_KEY_INIT_COST	initial cost (in pathfinding, before path was finalized)
PATH_KEY_INIT_FARE	initial fare (in pathfinding, before path was finalized)
PATH_KEY_PROBABILITY	path probability according to pathfinder
PATH_KEY_STATES
PATH_TIMES_OUTPUT_FILE	Path times output file
PENALTY_GROWTH_MODELS
STATE_IDX_ARRDEP	datetime.datetime instance.
STATE_IDX_COST	cost float, for hyperpath/stochastic assignment
STATE_IDX_DEPARR	datetime.datetime instance.
STATE_IDX_DEPARRMODE	mode id
STATE_IDX_LABEL	datetime.timedelta instance
STATE_IDX_LINKCOST	link generalized cost, float for hyperpath/stochastic,
STATE_IDX_LINKDIST	link distance, float
STATE_IDX_LINKFARE	fare cost, float
STATE_IDX_LINKTIME	datetime.timedelta instance
STATE_IDX_SEQ	sequence (for trip)
STATE_IDX_SEQ_SUCCPRED	sequence for successor/predecessor
STATE_IDX_SUCCPRED	stop identifier or TAZ identifier
STATE_IDX_TRIP	trip id
STATE_MODE_ACCESS
STATE_MODE_EGRESS
STATE_MODE_TRANSFER
STATE_MODE_TRIP
SUCCESS_FLAG_COLUMN
USER_CLASS_FUNCTION	Path configuration: Name of the function that defines user class
WEIGHTS_COLUMN_DEMAND_MODE	Weights column: Demand Mode Type
WEIGHTS_COLUMN_DEMAND_MODE_TYPE	Weights column: Demand Mode Type
WEIGHTS_COLUMN_PURPOSE	Weights column: Purpose
WEIGHTS_COLUMN_SUPPLY_MODE	Weights column: Supply Mode
WEIGHTS_COLUMN_SUPPLY_MODE_NUM	Weights column: Supply Mode number
WEIGHTS_COLUMN_USER_CLASS	Weights column: User Class
WEIGHTS_COLUMN_WEIGHT_NAME	Weights column: Weight Name
WEIGHTS_COLUMN_WEIGHT_VALUE	Weights column: Weight Value
WEIGHTS_DF	Path weights
WEIGHTS_FILE	File with weights file.
WEIGHTS_FIXED_WIDTH	Read weights file as fixed-width format.
WEIGHTS_GROWTH_LOGISTIC_MAX	Weights column: Max value for logistic growth function
WEIGHTS_GROWTH_LOGISTIC_MID	Weights column: Midpoint value for logistic growth function
WEIGHTS_GROWTH_LOG_BASE	Weights column: Log Base for logarithmic growth function
WEIGHTS_GROWTH_TYPE	Weights column: Growth Type
WEIGHT_NAME_ARRIVE_EARLY_MIN
WEIGHT_NAME_ARRIVE_LATE_MIN
WEIGHT_NAME_DEPART_EARLY_MIN
WEIGHT_NAME_DEPART_LATE_MIN
WEIGHT_NAME_VALID_NAMES

static calculate_cost(STOCH_DISPERSION, pathset_paths_df, pathset_links_df, veh_trips_df, trip_list_df, routes, tazs, transfers, stops=None, reset_bump_iter=False)

This is equivalent to the C++ Path::calculateCost() method. Would it be faster to do it in C++? It would require us to package up the networks and paths and send back and forth. :p

I think if we can do it using vectorized pandas operations, it should be fast, but we can compare/test.

It’s also messier to have this in two places. Maybe we should delete it from the C++; the overlap calcs are only in here right now.

Returns pathset_paths_df with additional columns, Assignment.SIM_COL_PAX_FARE, Assignment.SIM_COL_PAX_COST, Assignment.SIM_COL_PAX_PROBABILITY, Assignment.SIM_COL_PAX_LOGSUM And pathset_links_df with additional columns, Assignment.SIM_COL_PAX_FARE, Assignment.SIM_COL_PAX_FARE_PERIOD, Assignment.SIM_COL_PAX_COST and Assignment.SIM_COL_PAX_DISTANCE

static calculate_overlap(pathset_links_to_use)

Given a set of pathset links, returns a the results of overlap calculations.

This return dataframe will have colums person_id, person_trip_id, pathnum, and ln_PS

goes_somewhere()

Does this path go somewhere? Does the destination differ from the origin?

num_paths()

Number of paths in the PathSet

path_found()

Was a a transit path found from the origin to the destination with the constraints?

reset()

Delete my states, something went wrong and it won’t work out.

static set_user_class(trip_list_df, new_colname)

Adds a column called user_class by applying the configured user class function.

static split_transit_links(pathset_links_df, veh_trips_df, stops)

Splits the transit links to their component links and returns.

So if a transit trip goes from stop A to D but passes stop B and C in between, the row A->D will now be replaced by rows A->B, B->C, and C->D.

Adds “split_first” bool - True on the first veh link only

Note that this does not renumber the linknum field.

static verify_weight_config(modes_df, output_dir, routes, capacity_constraint, trip_list_df)

Verify that we have complete weight configurations for the user classes and modes in the given DataFrame.

Trips with invalid weight configurations will be dropped from the trip list and warned about.

The parameter mode_df is a dataframe with the user_class, demand_mode_type and demand_mode combinations found in the demand file.

If capacity_constraint is true, make sure there’s an at_capacity weight on the transit supply mode links to enforce it.

Returns updated trip_list_df.

static verify_weights(weights)

static write_path_times(passengers_df, output_dir)

Write the assigned path times to the given output file.

Parameters:

• passengers_df (pandas.DataFrame instance) – Passenger path links
• output_dir (string) – Output directory

static write_paths(passengers_df, output_dir)

Write the assigned paths to the given output file.

Parameters:

• passengers_df (pandas.DataFrame instance) – Passenger paths assignment results
• output_dir (string) – Output directory

ARRIVE_LATE_ALLOWED_MIN = datetime.timedelta(0)

Allow departures and arrivals before / after preferred time

BUMP_EXPERIENCED_COST = 999999

BUMP_FLAG_COLUMN = 'bump_flag'

CONFIGURED_FUNCTIONS = {'generic_user_class': <function generic_user_class>}

Configured functions, indexed by name

CONSTANT_GROWTH_MODEL = 'constant'

DEPART_EARLY_ALLOWED_MIN = datetime.timedelta(0)

DIR_INBOUND = 2

Trips inbound to home have preferred departure times

DIR_OUTBOUND = 1

Trips outbound from home have preferred arrival times

EXP_GROWTH_MODEL = 'exponential'

HUGE_COST = 9999

LEARN_ROUTES = False

LEARN_ROUTES_RATE = 0.05

LOGARITHMIC_GROWTH_MODEL = 'logarithmic'

LOGISTIC_GROWTH_MODEL = 'logistic'

MIN_TRANSFER_PENALTY = None

Configuration: Minimum transfer penalty. Safeguard against having no transfer penalty which can result in terrible paths with excessive transfers.

OUTPUT_WEIGHTS_FILE = 'ft_intermediate_weights.txt'

File with weights for c++

OVERLAP_CHUNK_SIZE = None

Overlap chunk size. How many person’s trips to process at a time in overlap calculations in python simulation

OVERLAP_COUNT = 'count'

Overlap variable option: count. Use leg count overlap pathsize correction.

OVERLAP_DISTANCE = 'distance'

Overlap variable option: distance. Use leg distance overlap pathsize correction.

OVERLAP_NONE = 'None'

Overlap variable option: None. Don’t use overlap pathsize correction.

OVERLAP_SCALE_PARAMETER = None

Configuration: Overlap scale parameter.

OVERLAP_SPLIT_TRANSIT = None

Overlap option: Split transit leg into component parts? e.g. split A-E into A-B-C-D-E for overlap calculations?

OVERLAP_TIME = 'time'

Overlap variable option: time. Use leg time overlap pathsize correction.

OVERLAP_VARIABLE = None

Configuration: Overlap variable. Can be “None”, “count”, “distance”, “time”.

OVERLAP_VARIABLE_OPTIONS = ['None', 'count', 'distance', 'time']

Valid values for OVERLAP_VARAIBLE

PATHS_OUTPUT_FILE = 'ft_output_passengerPaths.txt'

Paths output file

PATH_KEY_COST = 'pf_cost'

path cost according to pathfinder

PATH_KEY_FARE = 'pf_fare'

path fare according to pathfinder

PATH_KEY_INIT_COST = 'pf_initcost'

initial cost (in pathfinding, before path was finalized)

PATH_KEY_INIT_FARE = 'pf_initfare'

initial fare (in pathfinding, before path was finalized)

PATH_KEY_PROBABILITY = 'pf_probability'

path probability according to pathfinder

PATH_KEY_STATES = 'states'

PATH_TIMES_OUTPUT_FILE = 'ft_output_passengerTimes.txt'

Path times output file

PENALTY_GROWTH_MODELS = ['constant', 'exponential', 'logarithmic', 'logistic']

STATE_IDX_ARRDEP = 12

datetime.datetime instance. Arrival if outbound/backwards, departure if inbound/forwards.

STATE_IDX_COST = 11

cost float, for hyperpath/stochastic assignment

STATE_IDX_DEPARR = 1

datetime.datetime instance. Departure if outbound/backwards, arrival if inbound/forwards.

STATE_IDX_DEPARRMODE = 2

mode id

STATE_IDX_LABEL = 0

datetime.timedelta instance

STATE_IDX_LINKCOST = 9

link generalized cost, float for hyperpath/stochastic,

STATE_IDX_LINKDIST = 10

link distance, float

STATE_IDX_LINKFARE = 8

fare cost, float

STATE_IDX_LINKTIME = 7

datetime.timedelta instance

STATE_IDX_SEQ = 5

sequence (for trip)

STATE_IDX_SEQ_SUCCPRED = 6

sequence for successor/predecessor

STATE_IDX_SUCCPRED = 4

stop identifier or TAZ identifier

STATE_IDX_TRIP = 3

trip id

STATE_MODE_ACCESS = 'access'

STATE_MODE_EGRESS = 'egress'

STATE_MODE_TRANSFER = 'transfer'

STATE_MODE_TRIP = 'transit'

SUCCESS_FLAG_COLUMN = 'success_flag'

USER_CLASS_FUNCTION = None

Path configuration: Name of the function that defines user class

WEIGHTS_COLUMN_DEMAND_MODE = 'demand_mode'

Weights column: Demand Mode Type

WEIGHTS_COLUMN_DEMAND_MODE_TYPE = 'demand_mode_type'

Weights column: Demand Mode Type

WEIGHTS_COLUMN_PURPOSE = 'purpose'

Weights column: Purpose

WEIGHTS_COLUMN_SUPPLY_MODE = 'supply_mode'

Weights column: Supply Mode

WEIGHTS_COLUMN_SUPPLY_MODE_NUM = 'supply_mode_num'

Weights column: Supply Mode number

WEIGHTS_COLUMN_USER_CLASS = 'user_class'

Weights column: User Class

WEIGHTS_COLUMN_WEIGHT_NAME = 'weight_name'

Weights column: Weight Name

WEIGHTS_COLUMN_WEIGHT_VALUE = 'weight_value'

Weights column: Weight Value

WEIGHTS_DF = None

Path weights

WEIGHTS_FILE = 'pathweight_ft.txt'

File with weights file. Space delimited table.

WEIGHTS_FIXED_WIDTH = False

Read weights file as fixed-width format. If false, standard CSV format is read.

WEIGHTS_GROWTH_LOGISTIC_MAX = 'logistic_max'

Weights column: Max value for logistic growth function

WEIGHTS_GROWTH_LOGISTIC_MID = 'logistic_mid'

Weights column: Midpoint value for logistic growth function

WEIGHTS_GROWTH_LOG_BASE = 'log_base'

Weights column: Log Base for logarithmic growth function

WEIGHTS_GROWTH_TYPE = 'growth_type'

Weights column: Growth Type

WEIGHT_NAME_ARRIVE_EARLY_MIN = 'arrive_early_min'

WEIGHT_NAME_ARRIVE_LATE_MIN = 'arrive_late_min'

WEIGHT_NAME_DEPART_EARLY_MIN = 'depart_early_min'

WEIGHT_NAME_DEPART_LATE_MIN = 'depart_late_min'

WEIGHT_NAME_VALID_NAMES = ['depart_early_min', 'depart_late_min', 'arrive_early_min', 'arrive_late_min']

pathdict = None

Dict of path-num -> { cost:, probability:, states: [List of (stop_id, stop_state)]}