grouplasso: Do the local adaptive grouped regularization step.
In wrbrooks/lagr: Local adaptive grouped regularization

Do the local adaptive grouped regularization step.

grouplasso(data, index, family, weights = NULL, maxit = 1000,
  thresh = 0.001, min.frac = 0.1, nlam = 20, delta = 2,
  optim.step.size = 0.8, verbose = FALSE, reset = 10, lambda = NULL,
  unpenalized = NULL)

`data`	list containing `x`, a matrix of local covariates, and `y`, the corresponding vector of observed responses
`index`	vector indicating the group membership of each column of the covariate vector
`weights`	vector of observation weights
`maxit`	maximum iterations to run blockwise coordinate descent
`thresh`	iterate blockwise coordinate descent until the norm of the coefficient vector changes by less than this threshold
`min.frac`	ratio between the smallest and largest lambdas (lasso tuning parameters)
`nlam`	number of different lambdas (lasso tuning parameters) at which to fit the coefficients
`delta`	exponent of the unpenalized group coefficient norm in the adaptive penalty weight
`verbose`	print detailed information about model fitting?
`reset`
`lambda`	vector of prespecified lasso tuning parameters - leave NULL to have the lambdas calculated automatically.
`unpenalized`	index of any unpenalized groups
`gamma`