fasttrips.Assignment¶

class fasttrips.Assignment[source]¶

Bases: object

Assignment class. Documentation forthcoming.

__init__()[source]¶: This does nothing. Assignment methods are static methods for now.

Methods

`__init__`()	This does nothing.
`assign_paths`(output_dir, FT)	Finds the paths for the passengers.
`choose_paths_without_simulation`(FT, …)	Given a pathset for each passernger, choose a path (if relevant).
`compare_choices`(pathset_paths_df, prior_choice)	Reads in a pathset_path_df and the pathset_path_df of a prior iteration and returns the number of new choices made in the current iteration along with a simplified chosen df to use in the next iteration.
`filter_trip_list_to_not_arrived`(…)	Filter the given trip list to only those that have not arrived according to pathset_paths_df.
`find_passenger_vehicle_times`(…)	Given a dataframe of passenger links and a dataframe of vehicle trip links, adds two new columns to the passenger links for board and alight time.
`find_trip_based_pathset`(iteration, …)	Perform trip-based path set search.
`flag_missed_transfers`(pathset_paths_df, …)	Given passenger pathset links with the vehicle board_time and alight_time attached to trip links, this method will add columns to determine if there are missed transfers.
`generate_pathsets`(FT, pathset_paths_df, …)	Figures out which person trips for whom to generate_pathsets, stored in `Passenger.pathfind_trip_list_df`
`initialize_fasttrips_extension`(…)	Initialize the C++ fasttrips extension by passing it the network supply.
`load_passengers_on_vehicles_with_cap`(…)	Check if we have boards on over-capacity vehicles.
`merge_pathsets`(pathfind_trip_list_df, …)	Merge the given new pathset paths and links into the existing
`merge_prior_choices`(pathset_paths_df, …)	Join success or bump flag count information onto a pathset_paths_df and pathset_links_df :param pathset_paths_df: Current iteration of pathset_paths_df.
`number_of_pathsets`(pathset_paths_df)	Counts the number of passenger trips with pathsets and returns it.
`process_weight_qualifiers`(weights)	Qualifiers are used to change the default behavior of weight_names.
`put_passengers_on_vehicles`(pathset_links_df, …)	Puts the chosen passenger trips specified in pathset_links_df onto the transit vehicle trips specified by veh_trip_df.
`read_configuration`(config_fullpath)	Read the configuration parameters from `Assignment.CONFIGURATION_FILE`
`read_functions`(func_file)	Read the functions from :py:attr:`Assignment.CONFIGURATION_FUNCTIONS_FILE
`read_weights`([weights_file])	Read the weights from :py:attr:`Assignment.INPUT_WEIGHTS
`save_choices`(pathset_paths_df, success_df, …)	Returns an updated count of successfully chosen paths and paths that were bumped for each user.
`set_fasttrips_bump_wait`(bump_wait_df)	Sends the bump wait information to the fasttrips extension
`simulate`(FT, output_dir, iteration, …)	Given a pathset for each passenger, choose a path (if relevant) and then actually assign the passengers trips to the vehicles.
`write_configuration`(output_dir)	Write the configuration parameters to function as a record with the output.
`write_vehicle_trips`(output_dir, iteration, …)

Attributes

`BOARD_STATE_CATEGORICAL`	path invalidated before ever chosen due to capacity problems
`BUMP_BUFFER`	Extra time so passengers don’t get bumped (?).
`BUMP_ONE_AT_A_TIME`	Simulation: bump one stop at a time (slower, more accurate)
`CAPACITY_CONSTRAINT`	Route choice configuration: Use vehicle capacity constraints.
`CHOSEN_CATEGORIES`	These will be ordered, so to select chosen, choose those > CHOSEN_NOT_CHOSEN_YET
`CHOSEN_NOT_CHOSEN_YET`	categories for SIM_COL_PAX_CHOSEN
`CHOSEN_REJECTED`
`CONFIGURATION_FILE`	Configuration file for fasttrips
`CONFIGURATION_FUNCTIONS_FILE`	Configuration functions
`CONFIGURATION_OUTPUT_FILE`	Output copy of the configuration file in case anything got overridden (Hmm naming conventions are a bit awkward here)
`CONVERGENCE_GAP`
`CREATE_SKIMS`	Configuration: Create skims flag.
`DEBUG_NUM_TRIPS`	Debug mode: only run this number of trips, -1 to run all.
`DEBUG_OUTPUT_COLUMNS`	Debug: include debug columns in output
`DEBUG_TRACE_ONLY`	Debug mode: only run trace passengers
`FARE_ZONE_SYMMETRY`
`INPUT_DEMAND_DIR`	Configuration: Input demand directory
`INPUT_NETWORK_ARCHIVE`	Configuration: Input network directory
`INPUT_WEIGHTS`	Configuration: Pathweight parameters
`MAX_ITERATIONS`	Configuration: Maximum number of iterations to remove capacity violations.
`MAX_NUM_PATHS`	Route choice configuration: Max number of paths in a pathset.
`MAX_PF_ITERATIONS`
`MAX_SIMULATION_ITERS`	How many Simulation Iterations should we do before going back to path-finding?
`MIN_PATH_PROBABILITY`	Route choice configuration: Minimum path probability for the path to be used.
`MSA_RESULTS`	MSA the results that affect the next iteration to avoid oscillation: boards, alights, overcap onboard at stops
`NETWORK_BUILD_DATE`
`NETWORK_BUILD_DATE_START_TIME`
`NUMBER_OF_PROCESSES`	Number of processes to use for path finding (via `multiprocessing`) Set to 1 to run everything in this process Set to less than 1 to use the result of `multiprocessing.cpu_count()` Set to positive integer greater than 1 to set a fixed number of processes
`OUTPUT_DIR`	Configuration: Run Configuration
`OUTPUT_PASSENGER_TRAJECTORIES`	Configuration: Passenger trajectory output flag.
`OUTPUT_PATHSET_PER_SIM_ITER`	Configuration: If true, outputs pathset every simulation iteration.
`PATHFINDING_EVERYONE`	Are we finding paths for everyone right now? Or just un-arrived folks?
`PATHFINDING_TYPE`	Configuration: Pathfinding Type.
`PATHFINDING_TYPE_DETERMINISTIC`	Find paths deterministically, using shortest path search based on travel time.
`PATHFINDING_TYPE_READ_FILE`	Don’t find paths; read `Passenger.PF_PATHS_CSV` and `Passenger.PF_LINKS_CSV`.
`PATHFINDING_TYPE_STOCHASTIC`	Find paths stochastically using trip-based hyperpath
`PREPEND_ROUTE_ID_TO_TRIP_ID`	Prepend the route id to the trip id? This is for readability in debugging, since route IDs are typically more readable and trip ids are inscrutable
`SIMULATION`	Configuration: Do simulation? It should be True for iterative assignment.
`SIM_COL_MISSED_XFER`	Is this link/path a missed transfer? Set in both pathset links and pathset paths, this is a 1 or 0
`SIM_COL_PAX_ALIGHT_DELAY_MIN`	Delay in alight_time from original pathfinding understanding of alight time
`SIM_COL_PAX_ALIGHT_TIME`	Alight time from the transit vehicle
`SIM_COL_PAX_A_TIME`	Time of arrival at A
`SIM_COL_PAX_BOARD_STATE`	NaN if not relevent, 1 if lucky enough to board at an at- or over-capacity stop, 0 if bumped.
`SIM_COL_PAX_BOARD_TIME`	Board time on the transit vehicle
`SIM_COL_PAX_BUMP_ITER`
`SIM_COL_PAX_B_TIME`	Time of arrival at B
`SIM_COL_PAX_CHOSEN`	Chosen status for path
`SIM_COL_PAX_COST`	Link cost.
`SIM_COL_PAX_DISTANCE`	Link distance
`SIM_COL_PAX_FARE`	Link fare in currency
`SIM_COL_PAX_FARE_PERIOD`	Fare period id
`SIM_COL_PAX_FREE_TRANSFER`	Free transfer? NaN, 0.0 or 1.0, only free based on fare_attributes_ft.txt
`SIM_COL_PAX_LINK_TIME`	Link time (SIM_COL_PAX_B_TIME - SIM_COL_PAX_A_TIME)
`SIM_COL_PAX_LNPS`	log(PathSize)
`SIM_COL_PAX_LOGSUM`	Logsum of all paths
`SIM_COL_PAX_MISSED_XFER`	Is this link a missed transfer
`SIM_COL_PAX_OVERCAP`
`SIM_COL_PAX_OVERCAP_FRAC`	If board at an overcap stop, fraction of boards that are overcap
`SIM_COL_PAX_PROBABILITY`	Probability of this path
`SIM_COL_PAX_WAIT_TIME`	Wait time
`SKIM_END_TIME`	Configuration: End of the time period for which the skim is required (specify as ‘HH:MM’).
`SKIM_START_TIME`	Configuration: Beginning of the time period for which the skim is required.
`SKIP_PERSON_IDS`	Skip these passengers
`STOCH_DISPERSION`	Route choice configuration: Dispersion parameter in the logit function.
`STOCH_MAX_STOP_PROCESS_COUNT`	Route choice configuration: How many times max times should we process a stop during labeling? Use -1 to specify no max.
`STOCH_PATHSET_SIZE`	Route choice configuration: How many stochastic paths will we generate (not necessarily unique) to define a path choice set? Int.
`TIME_WINDOW`	Configuration: Path time-window.
`TRACE_IDS`	Trace these persons/person trips (a list of tuples)
`TRANSFER_FARE_IGNORE_PATHENUM`	In path-enumeration, suppress trying to adjust fares using transfer fare rules.
`TRANSFER_FARE_IGNORE_PATHFINDING`	In path-finding, suppress trying to adjust fares using transfer fare rules.
`UTILS_CONVERSION`	Stop labeling configuration: Multiplies the utilities by this factor so that there are not negative costs labels which can result in lengthy and ineffective path-finding.
`bump_wait`	This is the only simulation state that exists across iterations It’s a dictionary of (trip_id, stop_id) -> earliest time a bumped passenger started waiting
`bump_wait_df`

static assign_paths(output_dir, FT)[source]¶: Finds the paths for the passengers.

static choose_paths_without_simulation(FT, output_dir, iteration, pathfinding_iteration, pathset_paths_df, pathset_links_df, veh_trips_df)[source]¶

Given a pathset for each passernger, choose a path (if relevant). That’s it.

Returns (valid_linked_trips, pathset_paths_df, pathset_links_df)

static compare_choices(pathset_paths_df, prior_choice)[source]¶

Reads in a pathset_path_df and the pathset_path_df of a prior iteration and returns the number of new choices made in the current iteration along with a simplified chosen df to use in the next iteration.

Parameters:

pathset_paths_df – Current iteration of Pathset_Paths_DF
prior_choice – Pathset_Paths_DF of a prior iteration.

Returns:

new_choices_count: Total number of new choices in this iteration: compared to the input prior choice df.
chosen: Simplified DF of ‘person_id’, ‘person_trip_id_num’, and ‘description’: that can be used in subsequent invocations of this method.

static filter_trip_list_to_not_arrived(trip_list_df, pathset_paths_df)[source]¶: Filter the given trip list to only those that have not arrived according to pathset_paths_df.

static find_passenger_vehicle_times(pathset_links_df, veh_trips_df)[source]¶

Given a dataframe of passenger links and a dataframe of vehicle trip links, adds two new columns to the passenger links for board and alight time.

Takes the trip links of pathset_links_df (columns: person_id, trip_list_id_num, pathnum, linkmode, trip_id_num, A_id_num, B_id_num, A_seq, B_seq, pf_A_time, pf_B_time, pf_linktime, A_id, B_id, trip_id)
Joins with vehicle trips on trip id num, A_id, A_seq to add: * board time (Assignment.SIM_COL_PAX_BOARD_TIME) * overcap (Assignment.SIM_COL_PAX_OVERCAP) * overcap_frac (Assignment.SIM_COL_PAX_OVERCAP_FRAC)
Joins with vehicle trips on trip id num, B_id, B_seq to add: * alight time (Assignment.SIM_COL_PAX_ALIGHT_TIME)

Returns the same dataframe but with four additional columns (replacing them if they’re already there).

static find_trip_based_pathset(iteration, pathfinding_iteration, pathset, hyperpath, trace)[source]¶

Perform trip-based path set search.

Will do so either backwards (destination to origin) if PathSet.direction is PathSet.DIR_OUTBOUND or forwards (origin to destination) if PathSet.direction is PathSet.DIR_INBOUND.

Returns (pathdict,: performance_dict)

Where pathdict maps {pathnum:{PATH_KEY_COST:cost, PATH_KEY_PROBABILITY:probability, PATH_KEY_STATES:[state list]}}

Where performance_dict includes:: pathfinding return status, number of label iterations, max number of times a stop was processed, seconds spent in labeling, seconds spend in enumeration

Parameters:	pathset (a `PathSet` instance) – the path to fill in hyperpath (bool) – pass True to use a stochastic hyperpath-finding algorithm, otherwise a deterministic shortest path search algorithm will be use. trace (bool) – pass True if this path should be traced to the debug log

static flag_missed_transfers(pathset_paths_df, pathset_links_df)[source]¶

Given passenger pathset links with the vehicle board_time and alight_time attached to trip links, this method will add columns to determine if there are missed transfers.

This works on all paths in the pathset rather than just the chosen paths because then we can choose a path without missed transfers.

The following columns are used in pathset_links_df: * Passenger.TRIP_LIST_COLUMN_TRIP_LIST_ID_NUM, * Passenger.PF_COL_PATH_NUM, * Passenger.PF_COL_LINK_NUM, * Assignment.SIM_COL_PAX_BOARD_TIME * Assignment.SIM_COL_PAX_ALIGHT_TIME * Passenger.PF_COL_PAX_B_TIME

In particular, the following columns are added (or replaced if they’re already there) to pathset_links_df:

Column Name	Column Description
`Assignment.SIM_COL_PAX_ALIGHT_DELAY_MIN`	Delay in alight_time from original pathfinding understanding of alight time
`Assignment.SIM_COL_PAX_A_TIME`	New A time given the trip board/alight times for the trip links
`Assignment.SIM_COL_PAX_B_TIME`	New B time given the trip board/alight times for the trip links
`Assignment.SIM_COL_PAX_LINK_TIME`	New link time from B time - A time
`Assignment.SIM_COL_PAX_WAIT_TIME`	New waittime given the trip board/alight times for the trip links
`Assignment.SIM_COL_PAX_MISSED_XFER`	Is this link a missed xfer

The column, AssignmentSIM_COL_PAX_MISSED_XFER, is also added to pathset_paths_df.

static generate_pathsets(FT, pathset_paths_df, veh_trips_df, output_dir, iteration, pathfinding_iteration)[source]¶

Figures out which person trips for whom to generate_pathsets, stored in Passenger.pathfind_trip_list_df

Generates paths sets for those person trips using deterministic trip-based shortest path (TBSP) or stochastic trip-based hyperpath (TBHP).

Returns the number of pathsets found.

static initialize_fasttrips_extension(process_number, output_dir, stop_times_df)[source]¶: Initialize the C++ fasttrips extension by passing it the network supply.

static load_passengers_on_vehicles_with_cap(iteration, pathfinding_iteration, simulation_iteration, trips, pathset_paths_df, pathset_links_df, veh_loaded_df)[source]¶

Check if we have boards on over-capacity vehicles. Mark them and mark the boards.

If Assignment.CAPACITY_CONSTRAINT, then bump off overcapacity passengers.

The process is:

Look at which vehicle links are over capacity, adding columns named Trip.SIM_COL_VEH_OVERCAP and py:attr:Trip.SIM_COL_VEH_OVERCAP_FRAC to veh_loaded_df
Look at the stops where the first people board after we’re at capacity (impossible boards) if any
If Assignment.BUMP_ONE_AT_A_TIME, select the first such stop by arrival time Otherwise, select the first such stop for each vehicle trip
Join these stops to pathset_links_df, so pathset_links_df now has column Assignment.SIM_COL_PAX_OVERCAP_FRAC
If not Assignment.CAPACITY_CONSTRAINT, return (and drop the column named Trip.SIM_COL_VEH_OVERCAP from veh_loaded_df)
Figure out which passenger trips are actually getting bumped. Some people can get on at these stops, but not all, so let the first ones that arrive at the stop get on and filter to the ones we’ll actually bump. Update the column named Assignmment.SIM_COL_PAX_BUMP_ITER. If non-null, this represents the iteration the passenger got bumped.

Return (chosen_paths_bumped, pathset_paths_df, pathset_links_df, veh_loaded_df)

static merge_pathsets(pathfind_trip_list_df, pathset_paths_df, pathset_links_df, new_pathset_paths_df, new_pathset_links_df)[source]¶: Merge the given new pathset paths and links into the existing

static merge_prior_choices(pathset_paths_df, pathset_links_df, flag_df)[source]¶: Join success or bump flag count information onto a pathset_paths_df and pathset_links_df :param pathset_paths_df: Current iteration of pathset_paths_df. :param pathset_links_df: Current iteration of pathset_links_df :param flag_df: success or bump df :return: pathset_paths_df and pathset_links_df with additional flag column.

static number_of_pathsets(pathset_paths_df)[source]¶: Counts the number of passenger trips with pathsets and returns it.

static process_weight_qualifiers(weights)[source]¶

Qualifiers are used to change the default behavior of weight_names. Qualifiers are added by adding a period (.) after the weight_name and specifying the qualifier name. Qualifier attributes are specified after a second dot.

For example: depart_early_cost_min.logistic.growth_rate depart_early_cost_min is being qualified as a logistic penalty instead of the default behavior. growth_rate is an attribute of the logistic qualifier. :param weights: vertically oriented qualifiers :return: pivoted weights table with the qualifiers normalized horizontally.

static put_passengers_on_vehicles(pathset_links_df, veh_trips_df)[source]¶

Puts the chosen passenger trips specified in pathset_links_df onto the transit vehicle trips specified by veh_trip_df.

Returns veh_trips_df but with updated columns

static read_configuration(config_fullpath)[source]¶: Read the configuration parameters from Assignment.CONFIGURATION_FILE

static read_functions(func_file)[source]¶: Read the functions from :py:attr:`Assignment.CONFIGURATION_FUNCTIONS_FILE

static read_weights(weights_file=None)[source]¶: Read the weights from :py:attr:`Assignment.INPUT_WEIGHTS

static save_choices(pathset_paths_df, success_df, bump_df)[source]¶: Returns an updated count of successfully chosen paths and paths that were bumped for each user. :param pathset_paths_df: Current iteration of pathset_paths_df after simulation. :param success_df: Count of chosen paths for each user across prior iterations :param bump_df: Count of bumped paths for each user across prior iterations :return:

success_df: Updated count of chosen paths for each user across each iteration bump_df: Updated count of bumped paths for each user across each iteration

static set_fasttrips_bump_wait(bump_wait_df)[source]¶: Sends the bump wait information to the fasttrips extension

static simulate(FT, output_dir, iteration, pathfinding_iteration, pathset_paths_df, pathset_links_df, veh_trips_df)[source]¶

Given a pathset for each passenger, choose a path (if relevant) and then actually assign the passengers trips to the vehicles.

Returns (valid_linked_trips, pathset_paths_df, pathset_links_df, veh_loaded_df)

static write_configuration(output_dir)[source]¶: Write the configuration parameters to function as a record with the output.

static write_vehicle_trips(output_dir, iteration, pathfinding_iteration, simulation_iteration, veh_trips_df)[source]¶

BOARD_STATE_CATEGORICAL = ['board_easy', 'boarded', 'bumped', 'bumped_othertrip', 'bumped_unchosen']¶: path invalidated before ever chosen due to capacity problems

BUMP_BUFFER = None¶: Extra time so passengers don’t get bumped (?). A datetime.timedelta instance.

BUMP_ONE_AT_A_TIME = None¶

Simulation: bump one stop at a time (slower, more accurate)

When addressing capacity constraints in simulation, we look at all the (trip, stop)-pairs where the boards are not allowed since vehicle is over capacity. The faster way to address this is to bump all of those passengers, which means we call the assigned path bad and try to reassign.

However, this could over-bump passengers, because if a passenger rides multiple crowded vehicles, then bumping her frees up space on other vehicles and so some other bumping may not be necessary. Thus, the more accurate (but slower) method is to bump passengers from each (trip,stop) at a time, in order of the full vehicle arrival time, and then recalculate loads, and iterate until we have no capacity issues. Boolean.

CAPACITY_CONSTRAINT = None¶: Route choice configuration: Use vehicle capacity constraints. Boolean.

CHOSEN_CATEGORIES = ['rejected', 'unchosen']¶: These will be ordered, so to select chosen, choose those > CHOSEN_NOT_CHOSEN_YET

CHOSEN_NOT_CHOSEN_YET = 'unchosen'¶: categories for SIM_COL_PAX_CHOSEN

CHOSEN_REJECTED = 'rejected'¶

CONFIGURATION_FILE = None¶: Configuration file for fasttrips

CONFIGURATION_FUNCTIONS_FILE = None¶: Configuration functions

CONFIGURATION_OUTPUT_FILE = 'ft_output_config.txt'¶: Output copy of the configuration file in case anything got overridden (Hmm naming conventions are a bit awkward here)

CONVERGENCE_GAP = None¶

CREATE_SKIMS = None¶: Configuration: Create skims flag. This is specific to the travel demand models (not working in this version). Boolean.

DEBUG_NUM_TRIPS = -1¶: Debug mode: only run this number of trips, -1 to run all. Int.

DEBUG_OUTPUT_COLUMNS = False¶: Debug: include debug columns in output

DEBUG_TRACE_ONLY = False¶: Debug mode: only run trace passengers

FARE_ZONE_SYMMETRY = False¶

INPUT_DEMAND_DIR = None¶: Configuration: Input demand directory

INPUT_NETWORK_ARCHIVE = None¶: Configuration: Input network directory

INPUT_WEIGHTS = None¶: Configuration: Pathweight parameters

MAX_ITERATIONS = None¶: Configuration: Maximum number of iterations to remove capacity violations. When the transit system is not crowded or when capacity constraint is relaxed the model will terminate after the first iteration

MAX_NUM_PATHS = None¶: Route choice configuration: Max number of paths in a pathset. Used in conjuntion with Assignment.MIN_PATH_PROBABILITY

MAX_PF_ITERATIONS = 10¶

MAX_SIMULATION_ITERS = 10¶: How many Simulation Iterations should we do before going back to path-finding?

MIN_PATH_PROBABILITY = None¶: Route choice configuration: Minimum path probability for the path to be used. Used in conjucntion with Assignment.MAX_NUM_PATHS, so it only kicks in if that is specified AND we hit it, then we start dropping using this threshhold.

MSA_RESULTS = False¶: MSA the results that affect the next iteration to avoid oscillation: boards, alights, overcap onboard at stops

NETWORK_BUILD_DATE = datetime.datetime(2019, 2, 27, 21, 11, 42, 808269)¶

NETWORK_BUILD_DATE_START_TIME = datetime.datetime(2019, 2, 27, 0, 0)¶

NUMBER_OF_PROCESSES = None¶: Number of processes to use for path finding (via multiprocessing) Set to 1 to run everything in this process Set to less than 1 to use the result of multiprocessing.cpu_count() Set to positive integer greater than 1 to set a fixed number of processes

OUTPUT_DIR = None¶: Configuration: Run Configuration

OUTPUT_PASSENGER_TRAJECTORIES = None¶: Configuration: Passenger trajectory output flag. Passengers’ path and time will be reported if this flag is on. Note that the simulation flag should be on for passengers’ time. Boolean.

OUTPUT_PATHSET_PER_SIM_ITER = None¶: Configuration: If true, outputs pathset every simulation iteration. If false, outputs pathset every path-finding iteration.

PATHFINDING_EVERYONE = True¶: Are we finding paths for everyone right now? Or just un-arrived folks?

PATHFINDING_TYPE = None¶: Configuration: Pathfinding Type. Should be one of Deterministic, Stochastic or File

PATHFINDING_TYPE_DETERMINISTIC = 'deterministic'¶: Find paths deterministically, using shortest path search based on travel time.

PATHFINDING_TYPE_READ_FILE = 'file'¶: Don’t find paths; read Passenger.PF_PATHS_CSV and Passenger.PF_LINKS_CSV.

PATHFINDING_TYPE_STOCHASTIC = 'stochastic'¶: Find paths stochastically using trip-based hyperpath

PREPEND_ROUTE_ID_TO_TRIP_ID = False¶: Prepend the route id to the trip id? This is for readability in debugging, since route IDs are typically more readable and trip ids are inscrutable

SIMULATION = None¶: Configuration: Do simulation? It should be True for iterative assignment. In a one shot assignment with simulation flag off, the passengers are assigned to paths but are not loaded to the network. Boolean.

SIM_COL_MISSED_XFER = 'missed_xfer'¶: Is this link/path a missed transfer? Set in both pathset links and pathset paths, this is a 1 or 0

SIM_COL_PAX_ALIGHT_DELAY_MIN = 'alight_delay_min'¶: Delay in alight_time from original pathfinding understanding of alight time

SIM_COL_PAX_ALIGHT_TIME = 'alight_time'¶: Alight time from the transit vehicle

SIM_COL_PAX_A_TIME = 'new_A_time'¶: Time of arrival at A

SIM_COL_PAX_BOARD_STATE = 'board_state'¶: NaN if not relevent, 1 if lucky enough to board at an at- or over-capacity stop, 0 if bumped. Set by Assignment.flag_bump_overcap_passengers()

SIM_COL_PAX_BOARD_TIME = 'board_time'¶: Board time on the transit vehicle

SIM_COL_PAX_BUMP_ITER = 'bump_iter'¶

SIM_COL_PAX_B_TIME = 'new_B_time'¶: Time of arrival at B

SIM_COL_PAX_CHOSEN = 'chosen'¶: Chosen status for path

SIM_COL_PAX_COST = 'sim_cost'¶: Link cost. (Cannot be cost because it collides with TAZ.DRIVE_ACCESS_COLUMN_COST)

SIM_COL_PAX_DISTANCE = 'distance'¶: Link distance

SIM_COL_PAX_FARE = 'fare'¶: Link fare in currency

SIM_COL_PAX_FARE_PERIOD = 'fare_period'¶: Fare period id

SIM_COL_PAX_FREE_TRANSFER = 'free_transfer'¶: Free transfer? NaN, 0.0 or 1.0, only free based on fare_attributes_ft.txt

SIM_COL_PAX_LINK_TIME = 'new_linktime'¶: Link time (SIM_COL_PAX_B_TIME - SIM_COL_PAX_A_TIME)

SIM_COL_PAX_LNPS = 'ln_PS'¶: log(PathSize)

SIM_COL_PAX_LOGSUM = 'logsum'¶: Logsum of all paths

SIM_COL_PAX_MISSED_XFER = 'missed_xfer'¶: Is this link a missed transfer

SIM_COL_PAX_OVERCAP = 'overcap'¶

SIM_COL_PAX_OVERCAP_FRAC = 'overcap_frac'¶: If board at an overcap stop, fraction of boards that are overcap

SIM_COL_PAX_PROBABILITY = 'probability'¶: Probability of this path

SIM_COL_PAX_WAIT_TIME = 'new_waittime'¶: Wait time

SKIM_END_TIME = None¶: Configuration: End of the time period for which the skim is required (specify as ‘HH:MM’). A datetime.datetime instance.

SKIM_START_TIME = None¶: Configuration: Beginning of the time period for which the skim is required. (specify as ‘HH:MM’). A datetime.datetime instance.

SKIP_PERSON_IDS = None¶: Skip these passengers

STOCH_DISPERSION = None¶: Route choice configuration: Dispersion parameter in the logit function. Higher values result in less stochasticity. Must be nonnegative. If unknown use a value between 0.5 and 1. Float.

STOCH_MAX_STOP_PROCESS_COUNT = None¶: Route choice configuration: How many times max times should we process a stop during labeling? Use -1 to specify no max. Int. Setting this to a positive value may increase runtime but may decrease pathset quality. (Todo: test/quantify this.)

STOCH_PATHSET_SIZE = None¶: Route choice configuration: How many stochastic paths will we generate (not necessarily unique) to define a path choice set? Int.

TIME_WINDOW = None¶: Configuration: Path time-window. This is the time in which the paths are generated. E.g. with a typical 30 min window, any path within 30 min of the departure time will be checked. A datetime.timedelta instance.

TRACE_IDS = []¶: Trace these persons/person trips (a list of tuples)

TRANSFER_FARE_IGNORE_PATHENUM = None¶: In path-enumeration, suppress trying to adjust fares using transfer fare rules. This is for performance testing.

TRANSFER_FARE_IGNORE_PATHFINDING = None¶: In path-finding, suppress trying to adjust fares using transfer fare rules. This is for performance testing.

UTILS_CONVERSION = None¶: Stop labeling configuration: Multiplies the utilities by this factor so that there are not negative costs labels which can result in lengthy and ineffective path-finding. Must be positive; should be greater than 1.0. Double.

bump_wait = {}¶: This is the only simulation state that exists across iterations It’s a dictionary of (trip_id, stop_id) -> earliest time a bumped passenger started waiting

bump_wait_df = None¶