gamrs 0.11.2

GAM core (Rust) with sklearn-style Python bindings — beta

gamrs

Generalised Additive Models in Rust — clean-room reimplementation built on six composable trait layers (Basis, BasisTransform, Loss/Link/VarianceFn, InnerSolver, ScoreDerivatives, OuterSolver). Designed for parity with R's mgcv.

Status: beta (v0.10). Faster than mgcv_rust 0.23 at every tested fixture and scale (see Performance), and mgcv R-parity on µ across all ten families. Multi-smooth additive (y ~ s(x0) + s(x1)), n-margin tensor products (te(x0, x1, …) / ti(…)) and thin-plate splines (s(x0, x1, bs="tp")) all ship. NegBin and Tweedie fit multi-smooth, with Tweedie offering both profile-p (tw()) and fixed-p (Tweedie(p)).

Install

pip install gamrs            # base wheel
pip install gamrs[quantile]  # + scipy for SHASH-calibrated quantile fits

Quickstart

from gamrs import Gam, CrTerm, TeTerm

# Single 1-D smooth, Gaussian
g = Gam(family="gaussian").fit(X, y)
mu  = g.predict(X)
mu, lo, hi = g.predict_ci(X, level=0.95)

# Multi-smooth additive
g = Gam(terms=[CrTerm("x0", k=10), CrTerm("x1", k=15)]).fit(df, df["y"])

# Tensor product
g = Gam(terms=[TeTerm(cols=("x0", "x1"), k=(5, 5))]).fit(df, df["y"])

# Large-n GLM — switch to the bam()-style fREML optimiser
g = Gam(family="poisson", method="fREML").fit(X_big, y_big)

Full walkthrough: docs/quickstart.md. Optimiser & large-n notes: docs/perf.md.

Families

All ten families land 1-D parity against mgcv:

Family	Link	Inner solver	Outer optimiser	Parity (µ rel-err)
Gaussian	identity	one-Cholesky	1-D Newton	~3e-6
Bernoulli	logit	PIRLS	1-D Newton	~1e-3
Poisson	log	PIRLS	1-D Newton	~8e-5
QuasiPoisson	log	PIRLS	1-D Newton (prof φ)	~2e-4
QuasiBinomial	logit	PIRLS	1-D Newton (prof φ)	~7e-5
Gamma	log	PIRLS	1-D Newton (prof φ)	~2e-2
InverseGaussian	log	PIRLS	1-D Newton (prof φ)	~3e-4
NegBin	log	PIRLS	2-D joint Newton	~9e-7
Tweedie	log	PIRLS	3-D joint Newton	~5e-3
TDist (scat)	identity	PIRLS	3-D joint Newton	~2e-2
Ocat	logit	gam.fit5	joint β + threshold	smoke
Quantile (ELF)	identity	Armijo BT	1-D Newton	smoke

Multi-smooth (s(x0) + s(x1) + …) ships with mgcv R parity tests for Gaussian / Bernoulli / Poisson / QuasiPoisson / QuasiBinomial / Gamma / InvGauss / NegBin / Tweedie. scat / TDist multi-smooth fits run and converge; reference parity tests are pending. Quantile/ELF is single-smooth-only.

Multi-smooth Ocat fits run, produce well-defined predict_proba, and reach 99%+ classification accuracy on synthetic fixtures.

v0.10 ports the full mgcv R outer-Newton stabilisation stack (smart θ-init from category frequencies, diagonal Hessian preconditioning, Gill-Murray-Wright eigen-fix, subset Newton, rank-deficient KKT convergence check). After the ports, single-smooth ocat converges cleanly on every tested seed. Multi-smooth still hits a flat coordinated-shift ridge on the most pathological synthetic fixtures (both θ axes against the bound) where converged_=False lingers despite correct predictions — same regime mgcv R itself bails out of with a "did not converge after 200 iterations" warning. See ~/ObsidianVault/Projects/gamrs/gamrs - mgcv outer-Newton stabilisation techniques (port catalogue) 2026-06-03.md for the full port story.

Smooths

• Single 1-D — s(x0) via CrTerm (cubic regression spline default).
• Additive multi — s(x0) + s(x1) + s(x2).
• Tensor product — te(x0, x1, …) via TeTerm, ti(…) via TiTerm, any n-margin.
• Thin-plate — s(x0, x1, bs="tp") via TpsTerm.
• Random effects — s(g, bs="re") via ReTerm.
• Parametric (linear) — unsmoothed raw column via ParametricTerm or predictor_basis_map={"x": "parametric"} (alias "linear"). Use for 0/1 indicators, counts, or anything you want unpenalised. mgcv R's "pterms" block.

Performance

gamrs vs mgcv_rust 0.23.2, single-smooth, best-of-20 wall time after long warm-up. Numbers >1× mean gamrs is faster. Measured on a single 12th-gen Intel core. See scripts/bench_matters.py.

Single-smooth at gamrs-default k=10, n=2K:

family	speedup vs mgcv_rust
Gaussian	1.6×
Poisson	1.9×
Bernoulli	2.1×
Tweedie	2.6×
NegBin	1.0×
scat	0.77×
ocat	14× (†)

(†) ocat speedup is fixture-dependent; mgcv_rust's ocat path is slow on multi-category outputs by construction.

scat closed substantially in v0.11.1 — from v0.10.0's 0.07× to 0.77× of mgcv_rust. Two structural wins drove it: (a) analytic gradient + Level-2 analytic Hessian + observed-W PIRLS + warm-start (v0.11.0), and (b) broadcast-expression conversion + batched-h_diag matmul across the IFT chain (v0.11.1). The remaining ~25 % gap is per-pair work in the Hessian assembly that's already broadcast-friendly — closing it would need either algorithmic shortcuts or fundamentally different BLAS dispatch at small p.

For GLM families at large n, set method="fREML" (mgcv R's bam() equivalent — Wood & Fasiolo 2017 Fellner-Schall multiplicative updates with single-step IRLS per outer iteration). The defaults are sensible at small/medium n; the perf guide covers when to switch.

Rust API

use gamrs::{TermSpec, MarginKind, DesignStrategy};
use ndarray::Array2;

let x: Array2<f64> = /* (n, n_input_dims) */;
let y = /* Array1<f64> */;

let fit = gamrs::fit(gamrs::family::gaussian_identity(), x.view(), y.view(), None, 10)?;

let fit = gamrs::fit_with_design(
    gamrs::family::gaussian_identity(),
    DesignStrategy::Additive { terms: vec![
        TermSpec::Cr { col: 0, k: 10 },
        TermSpec::Cr { col: 1, k: 15 },
    ]},
    x.view(), y.view(), None,
)?;

let fit = gamrs::fit_with_design(
    gamrs::family::gaussian_identity(),
    DesignStrategy::Additive { terms: vec![
        TermSpec::Tensor { col_a: 0, col_b: 1, k_a: 5, k_b: 5,
                           bs_a: MarginKind::Cr, bs_b: MarginKind::Cr },
    ]},
    x.view(), y.view(), None,
)?;

let mu = fit.predict(x.view())?;

Python API

PyO3 bindings + numpy. sklearn-like surface: fit / predict / predict_ci / predict_diff / vcov_ / coef_ / lambda_ / edf_ / fit_stats_, plus serialize / deserialize and GamPredictor for inference-only deployment.

Architecture

Trait layering (src/traits.rs):

Layer 1   Basis              ←  CrBasis, RandomEffectsBasis, TensorProductBasis<A, B>
Layer 1.5 BasisTransform     ←  SumToZero, StableReparam
Layer 2   Loss/Link/Variance ←  10 families (see table above)
Layer 3   InnerSolver        ←  GaussianClosedFormInner, PirlsInner, GamFit5Inner, ArmijoInner
Layer 4   ScoreDerivatives   ←  EnvelopeScore, ShapeAwareEnvelopeScore
Layer 5   OuterSolver        ←  NewtonWithHalving, FellnerSchall
Layer 6   FittedGam          ←  predict, predict_ci, predict_diff, vcov, serialize

Outer optimisers (Newton, Fellner-Schall) and per-family tolerances are selected through Loss::outer_tuning() and Loss::allows_no_refresh(), so adding a family is a Loss impl, not a fork of the optimiser.

Versioning

Beta (0.10.x). The API is stabilising; minor bumps may carry breaking changes until the remaining shape-aware families (scat/Ocat/ELF) gain multi-smooth support and the 1.0 surface is locked.

License

MIT.