Crate simsam 

simsam — sample from custom discrete and continuous distributions.

Build a distribution from a PDF or CDF (closures, histograms, location-scale transforms, or simsym symbolic expressions), then draw samples via inverse transform sampling — similar to SciPy rv_continuous.

Features

• pdf, cdf, ppf, sf, isf, logpdf, logcdf, logsf
• mean, var, std, median, entropy, expect, interval
• Fast sampling via HermitePpfTable (BuildOptions::with_hermite) or TDR (BuildOptions::with_tdr)
• LocScale, Truncated, HistogramPdf

Limitations

• Continuous distributions require finite support [lo, hi] (use Truncated on a wide interval).
• Inverse transform assumes a unimodal CDF on that interval.

Example

use simsam::{from_pdf_fn, Interval};

let support = Interval::new(0.0, 1.0).unwrap();
let sampler = from_pdf_fn(|x| 3.0 * x * x, support).unwrap();
let x = sampler.sample().unwrap();
assert!((0.0..=1.0).contains(&x));
assert!((sampler.mean().unwrap() - 0.75).abs() < 1e-2);

Structs

AffineCdf

Normalized CDF built from a user-provided Cdf, affine-scaled to [0, 1] on support.

BuildOptions

Options for building a continuous sampler from PDF or CDF.

CdfFn

Closure-backed CDF with explicit support.

CdfMcEstimator

Monte-Carlo estimator for normalization and CDF of an N-D PDF on a bounded support.

CdfMcOptions

ConditionalFactorSampler

ContinuousSampler

Continuous distribution sampler.

DiscreteSampler

Finite discrete distribution sampler via inverse transform.

DpdfFn

Closure-backed dPDF with explicit support.

GaussianCopula

Gaussian copula sampler.

GibbsOptions

GibbsSamplerNd

HermitePpfTable

Precomputed inverse-CDF table with monotone cubic Hermite interpolation (fast ppf).

HistogramPdf

Piecewise-constant PDF from a histogram (SciPy rv_histogram).

HmcOptions

HmcSamplerNd

HyperRect

Axis-aligned bounded support in R^n: lo[i] <= x[i] <= hi[i].

IntegratedPdf

CDF obtained by integrating and normalizing a PDF on support.

Interval

Finite support interval [lo, hi] with lo < hi.

InvertOptions

LocScale

Location-scale transform: Y = loc + scale * X for base PDF on base_support.

MetropolisHastingsNd

Random-walk Metropolis–Hastings sampler on a bounded hyper-rectangle.

MhOptions

PdfFn

Closure-backed PDF with explicit support.

PdfNdFn

Closure-backed ND PDF with explicit support.

RejectionOptions

RejectionSamplerNd

Generic N-D rejection sampler on a bounded hyper-rectangle.

TdrBuildConfig

TDR configuration used when SampleMethod::Tdr is selected.

TdrHat

Built TDR hat function; immutable after construction.

TdrOptions

TdrSampler

1D Transformed Density Rejection sampler.

Truncated

Truncate a base PDF to [lo, hi] and renormalize.

Enums

BuildError

Error while constructing a distribution or sampler.

PpfMethod

How to invert the CDF when computing ppf (inverse transform sampling path).

SampleError

Error during sampling (e.g. invalid quantile).

SampleMethod

How to draw samples from a continuous distribution.

TdrTransform

Transformation parameter for TDR.

Traits

Cdf

Cumulative distribution function, monotone on support with cdf(lo) ≈ 0, cdf(hi) ≈ 1.

CdfDiscrete

Discrete cumulative distribution at support index k (inclusive).

CdfSource

Internal CDF source for inverse-transform sampling.

ConditionalFactorization

Conditional factorization x0 ~ f0(), x1 ~ f1(x0), …, x_{n-1} ~ f_{n-1}(x0..x_{n-2}).

ConditionalSampler

Trait for sampling coordinate i conditional on other coordinates.

Dpdf

Derivative of a probability density function dpdf = d/dx pdf(x).

HasSupport

Distribution support.

HasSupportNd

LogPdfNd

Log-density (up to additive constant). Prefer this for MCMC.

Pdf

Probability density function on a finite support.

PdfNd

Joint probability density function on a bounded hyper-rectangle support.

Pmf

Probability mass function on a finite support.

Functions

default_quad_tol

from_cdf_fn

Build a sampler from a CDF closure on [lo, hi].

from_histogram

Build from a histogram (HistogramPdf).

from_pdf_dpdf_fn

Build a sampler from PDF and explicit dPDF closures.

from_pdf_fn

Build a sampler from a PDF closure on [lo, hi] with default options.

from_pdf_fn_with_options

Build a sampler from a PDF closure with custom BuildOptions.

from_pdf_loc_scale

Build from a base PDF with y = loc + scale * x.

tdr_from_fns

Convenience constructor from closures.