#!/usr/bin/env python3
# Source: t4-optimization.html
import numpy as np

def soft(z, t):
    return np.sign(z) * np.maximum(np.abs(z) - t, 0.0)

def lasso_cd(X, y, lam, n_iter=200):
    n, d = X.shape
    w = np.zeros(d)
    r = y - X @ w                     # residual
    col_sq = (X ** 2).sum(axis=0) / n
    for _ in range(n_iter):
        for j in range(d):
            r += X[:, j] * w[j]       # remove feature j
            rho = X[:, j] @ r / n
            w[j] = soft(rho, lam) / col_sq[j]
            r -= X[:, j] * w[j]       # add it back
    return w

rng = np.random.default_rng(0)
X = rng.normal(size=(200, 10))
w_true = np.array([3, -2, 0, 0, 1.5, 0, 0, 0, 0, 0.0])
y = X @ w_true + 0.5 * rng.normal(size=200)
print(np.round(lasso_cd(X, y, lam=0.1), 2) + 0.0)
# [ 2.93 -1.88  0.    0.    1.37 -0.01  0.    0.    0.    0.  ]
