fastmob-core
Low-level Rust compute kernels for mobility and trajectory analysis.
fastmob-core powers the Python fastmob
package and the higher-level fastmob-rs DataFrame API. It is suitable for
Rust callers that own their columnar/slice-oriented data and want direct
access to trajectory preparation, measures, models, networks, privacy, and
social-analysis kernels.
Installation
[]
= "0.2"
Example
use haversine_km;
let distance_km = haversine_km;
assert!;
Features
simd(default): enables the acceleratednumkongdependency and thestvd-emdkernels.numkong: enables only the optional acceleration dependency.stvd-emd: enables the STVD EMD and sparse Sinkhorn kernels.
Disable defaults when a portable, dependency-minimal build is preferred:
= { = "0.2", = false }
API and compatibility
This crate exposes low-level kernels over primitive slices and collections.
Functions suffixed _impl are deliberate public entry points for bindings and
advanced Rust callers; they are not private implementation details. Before
1.0, compatibility follows Cargo's pre-1.0 semantic-versioning conventions,
so a minor release may include breaking changes.
License and provenance
fastmob-core is BSD-3-Clause licensed. Some algorithms were ported or
adapted from other projects; see
THIRD_PARTY_NOTICES.md for attribution, license
status, and release requirements.