Cox Model Deviance, Gradient, Information and Score Sampling

Fast computation of the deviance of the Cox proportional hazards model (Cox (1972) ), its gradient and the diagonal of its Hessian with respect to the linear predictor, the information and its derivative in a direction as matrix-free linear operators, and draws of the score (multinomial, Bernoulli, Fisher noncentral hypergeometric or wild), with the ties of Breslow (1974) or Efron (1977) , start times (left truncation), strata and sample weights. The 'C++' code is header-only, shared with the 'Python' package 'coxdev', and installed for packages that use it through 'LinkingTo'.