Crate blprs 

Fast Berry–Levinsohn–Pakes (BLP) estimation for random coefficients logit models.

This crate provides a demand-side estimator that mirrors the public API of pyBLP while embracing idiomatic Rust. It offers tools to

• manage product-level market data (data module),
• describe simulation draws for heterogeneous consumers (integration module),
• solve the BLP contraction mapping (solving module), and
• assemble a two-step GMM estimator (estimation module).

The implementation focuses on clarity and extensibility. Heavy inline documentation and unit tests illustrate the essential ingredients of BLP: market partitioning, heterogeneous consumer integration, fixed-point solving, and GMM objective evaluation. The goal is to provide a high-quality Rust alternative to pyBLP while maintaining feature parity over time.

Quick start

use blprs::data::{ProductData, ProductDataBuilder};
use blprs::integration::SimulationDraws;
use blprs::{Problem, ProblemOptions};
use nalgebra::{DMatrix, DVector};

// Assume we have N products and K1 linear, K2 nonlinear characteristics.
let market_ids = vec!["m1".to_string(), "m1".to_string(), "m2".to_string()];
let shares = DVector::from_vec(vec![0.3, 0.2, 0.4]);
let x1 = DMatrix::from_row_slice(3, 2, &[1.0, 10.0, 1.0, 15.0, 1.0, 12.0]);
let x2 = DMatrix::from_row_slice(3, 1, &[10.0, 15.0, 12.0]);
let instruments = x1.clone();

let products = ProductDataBuilder::new(market_ids, shares)
    .x1(x1)
    .x2(x2)
    .instruments(instruments)
    .build()
    .expect("validated product data");

let draws = SimulationDraws::standard_normal(200, 1, 1234);

let problem = Problem::builder()
    .products(products)
    .draws(draws)
    .options(ProblemOptions::default())
    .build()
    .expect("well-formed problem");
let sigma = DMatrix::from_row_slice(1, 1, &[2.0]);

let result = problem.solve(&sigma).expect("converged");
println!("Estimated betas: {:?}", result.beta);

The crate is still under heavy development. Supply-side estimation, optimal instruments, and many advanced pyBLP options are tracked in the public roadmap.

Re-exports

pub use estimation::BlpProblem;

pub use estimation::EstimationResult;

pub use estimation::Problem;

pub use estimation::ProblemBuilder;

pub use estimation::ProblemResults;

pub use options::EstimationOptions;

pub use options::GmmOptions;

pub use options::ProblemOptions;

pub use options::WeightingMatrix;

pub use solving::ContractionOptions;

pub use solving::ContractionSummary;

Modules

data

Product-level data containers and validation utilities used by the BLP estimator.

demand

Demand-side primitives: share prediction and the BLP contraction mapping.

error

estimation

High-level demand estimation pipeline that mirrors pyBLP.Problem.

formulation

Lightweight placeholder for pyBLP-style formulas.

integration

Monte Carlo integration helpers for simulating heterogeneous consumer tastes.

options

Configuration structures that mirror pyBLP’s solver and GMM options while remaining idiomatic Rust.

solving

Contraction solver configuration and diagnostics.