Models & types

All request and response types are Pydantic models exported from the package top level:

from humanbaselines import (
    GeofenceSelections, RouteSelections, DepotSelections, DepotPin,
    ComputeResult, PerCellResult, BatchComputeResult, BatchItemResult,
    RouteComputeResult, PerSegmentResult,
    DepotComputeResult, FiltersResponse, FilterDef, FilterOption,
    RegionsResponse, RegionInfo,
)

Generated from the API

These types are generated from the server's OpenAPI schema. The server's Pydantic models are the single source of truth, so the client cannot drift from the API. filters() is the authoritative runtime source for valid values and defaults.

Request models (selections)

One per mode. Pass an instance via selections=, or use keyword args. All use extra="forbid", so an unknown field raises locally before any request.

Model	Used by	Notable fields
GeofenceSelections	compute	outcome, severity, ego_vehicle, road_type, weather, light_condition, crash_year, desk_reports, denominator_vmt, operator_weighting, operator_weight, …
RouteSelections	compute_route	outcome, ego_vehicle (defaults ["combination"]), weather, driver_impairment, ci_method, …
DepotSelections	compute_depot_route	same shape as RouteSelections
DepotPin	compute_depot_route	lat, lon

from humanbaselines import GeofenceSelections, Outcome
sel = GeofenceSelections(outcome=Outcome.fatal, road_type=["interstate"])
hb.compute(selections=sel)

Response models

Model	Returned by	Key fields
ComputeResult	compute	N, D_miles, D_billions, rate, rate_low/high, rate_non_dyn, rate_dyn, multiplier, cells (empty when summary_only)
PerCellResult	ComputeResult.cells	s2_cell, count, vmt, mult_contrib
BatchComputeResult	compute_batch	results: list[BatchItemResult] (same order as requested regions)
BatchItemResult	BatchComputeResult.results	region, result (ComputeResult | None), error (str | None)
RouteComputeResult	compute_route	N, trip_miles, rate, rate_low/high, variance, segments
PerSegmentResult	RouteComputeResult.segments	route, milepost, count, vmt, length_mi
DepotComputeResult	compute_depot_route	access_a, interstate, access_b, total, depot_a/b, vertices (permissive)
FiltersResponse	filters	modes: dict[str, list[FilterDef]]
RegionsResponse	regions	regions: list[RegionInfo], default_region

Enums

String enums for filter values. They accept the enum member or its string value interchangeably:

Outcome, VehicleClass, RoadGroup, SpeedBand, Tier3Mode, InTransport, OperatorWeighting, MultiplierVmt, UnderReporting, WeatherFilter, LightFilter, DriverImpairment, CiMethod.

from humanbaselines import Outcome
hb.compute(outcome=Outcome.fatal)   # same as outcome="fatal"

The newest filters (desk_reports, day_type, time_of_day) have no exported enum yet, so pass their values as strings.

posted_speed_max is not a wire field. The client turns the cap into the posted_speed band list with speed_bands_up_to(), which is also exported:

from humanbaselines import speed_bands_up_to
speed_bands_up_to(45)   # ['le15', 's20', 's25', 's30', 's35', 's40', 's45']

The live option lists and per-mode defaults are on the Discovery page.