API Reference

adelie.bcd

objective(beta, *, quad, linear, l1, l2)	Computes the BCD objective.
root(*, quad, linear, l1[, tol, max_iters, ...])	Solves the non-negative root of the BCD root function.
root_function(h, *, quad, linear, l1)	Computes the BCD root function.
root_lower_bound(*, quad, linear, l1)	Computes a lower bound on the root of BCD root function.
root_upper_bound(*, quad, linear, l1[, zero_tol])	Computes an upper bound on the root of BCD root function.
solve(*, quad, linear, l1, l2[, tol, ...])	Solves the BCD update.

adelie.configs

set_configs(name[, value])

Sets global configurations.

adelie.constraint

box(lower, upper, *[, method, configs, dtype])	Creates a box constraint.
linear(A, lower, upper, *[, vars, copy, ...])	Creates a linear constraint.
lower(b, **kwargs)	Creates a lower bound constraint.
one_sided(D, b, *[, method, configs, dtype])	Creates a one-sided bound constraint.
upper(b, **kwargs)	Creates an upper bound constraint.

adelie.cv

CVGrpnetResult(lmdas, losses, avg_losses, ...)	Result of K-fold CV group elastic net.
cv_grpnet(X, glm, *[, n_threads, ...])	Solves cross-validated group elastic net via naive method.

adelie.data

dense(n, p, G, *[, K, glm, equal_groups, ...])	Creates a dense dataset.
snp_phased_ancestry(n, s, A, *[, K, glm, ...])	Creates a SNP phased, ancestry dataset.
snp_unphased(n, p, *[, K, glm, sparsity, ...])	Creates a SNP unphased dataset.

adelie.diagnostic

DiagnosticCov(state)	Diagnostic class for covariance states.
DiagnosticNaive(state)	Diagnostic class for naive states.
coefficient(lmda, betas, intercepts, lmdas)	Computes the coefficient at \(\lambda\) using linear interpolation of solutions.
diagnostic(state)	Creates a diagnostic class object appropriate for the state.
gradients(X, resids, *[, n_threads])	Computes the gradients.
gradient_norms(grads, betas, duals, lmdas, *)	Computes the group-wise gradient norms.
gradient_scores(grad_norms, lmdas, *[, ...])	Computes the gradient scores.
objective(X, glm, betas, intercepts, lmdas, *)	Computes the group elastic net objective.
plot_benchmark(total_time, benchmark_screen, ...)	Plots benchmark times.
plot_coefficients(betas, lmdas, groups, ...)	Plots the coefficient profile.
plot_devs(lmdas, devs)	Plots the deviance profile.
plot_kkt(lmdas, scores, *[, idx, relative])	Plots KKT failures.
plot_set_sizes(groups, screen_sizes, ...[, ...])	Plots the set sizes.
predict(X, betas, intercepts, *[, offsets, ...])	Computes the linear predictions.
residuals(glm, etas)	Computes the residuals.

adelie.glm

binomial(y, *[, weights, link, dtype])	Creates a Binomial GLM family object.
cox(start, stop, status, *[, strata, ...])	Creates a Cox GLM family object.
gaussian(y, *[, weights, dtype, opt])	Creates a Gaussian GLM family object.
multigaussian(y, *[, weights, dtype, opt])	Creates a MultiGaussian GLM family object.
multinomial(y, *[, weights, dtype])	Creates a Multinomial GLM family object.
poisson(y, *[, weights, dtype])	Creates a Poisson GLM family object.

adelie.io

snp_phased_ancestry(filename[, read_mode])	IO handler for SNP phased, ancestry matrix.
snp_unphased(filename[, read_mode])	IO handler for SNP unphased matrix.

adelie.matrix

block_diag(mats, *[, method, n_threads])	Creates a block-diagonal matrix given by the list of matrices.
concatenate(mats, *[, axis, n_threads])	Creates a concatenation of the matrices.
convex_relu(mat, mask, *[, gated, copy, ...])	Creates a feature matrix for the convex relu problem.
dense(mat, *[, method, copy, n_threads])	Creates a viewer of a dense matrix.
eager_cov(mat[, n_threads])	Creates an eager covariance matrix.
interaction(mat, intr_map[, levels, copy, ...])	Creates a matrix with pairwise interactions.
kronecker_eye(mat, K, *[, copy, n_threads])	Creates a Kronecker product with identity matrix.
lazy_cov(mat, *[, copy, n_threads])	Creates a lazy covariance matrix.
one_hot(mat[, levels, copy, n_threads])	Creates a one-hot encoded matrix.
snp_phased_ancestry(io, *[, n_threads, dtype])	Creates a SNP phased, ancestry matrix.
snp_unphased(io, *[, n_threads, dtype])	Creates a SNP unphased matrix.
sparse(mat, *[, method, copy, n_threads])	Creates a viewer of a sparse matrix.
standardize(mat[, centers, scales, ddof, ...])	Creates a standardized matrix.
subset(mat, indices, *[, axis, n_threads])	Creates a subset of the matrix along an axis.

adelie.sklearn

CSSModelSelection(alpha[, n_inits, n_sims, ...])	Column Subset Selection estimator for model selection with scikit-learn compatible API.
GroupElasticNet([solver, family])	Group Elastic Net estimator with scikit-learn compatible API.

adelie.solver

bvls(X, y, lower, upper, *[, weights, ...])	Solves bounded variable least squares.
css_cov(S[, subset_size, subset, method, ...])	Solves column subset selection via covariance method.
gaussian_cov(A, v, *[, constraints, groups, ...])	Solves Gaussian group elastic net via covariance method.
grpnet(X, glm, *[, constraints, groups, ...])	Solves group elastic net via naive method.
pinball(A, S, v, penalty_neg, penalty_pos, *)	Solves pinball least squares.

adelie.state

bvls(X, y_var, X_vars, lower, upper, ...)	Creates a BVLS state object.
css_cov(S, subset_size, subset, method, ...)	Creates a CSS covariance state object.
gaussian_cov(*, A, v, constraints, groups, ...)	Creates a Gaussian, covariance method state object.
gaussian_naive(*, X, y, X_means, y_mean, ...)	Creates a Gaussian, naive method state object.
gaussian_pin_cov(*, A, constraints, groups, ...)	Creates a Gaussian, pin, covariance method state object.
gaussian_pin_naive(*, X, y_mean, y_var, ...)	Creates a Gaussian, pin, naive method state object.
glm_naive(*, X, glm, constraints, groups, ...)	Creates a GLM, naive method state object.
multigaussian_naive(*, X, y, X_means, y_var, ...)	Creates a MultiGaussian, naive method state object.
multiglm_naive(*, X, glm, constraints, ...)	Creates a multi-response GLM, naive method state object.
pinball(A, y_var, S, penalty_neg, ...)	Creates a pinball state object.