loo_se_ridgeaug: Use leave out one estimates (placebo gaps) to estimate...
In ebenmichael/ents: Maximum Entropy Synthetic Controls

Use leave out one estimates (placebo gaps) to estimate unit-level variance Do this for ridge-augmented synth

loo_se_ridgeaug(outcomes, metadata, covs = NULL, trt_unit = 1,
  lambda = NULL, scm = T, ridge = T, use_weights = T, hc = -1,
  trt_effect = T, cols = list(unit = "unit", time = "time", outcome =
  "outcome", treated = "treated"))

`outcomes`	Tidy dataframe with the outcomes and meta data
`metadata`	Dataframe of metadata
`covs`	Dataframe of covariates if not null then uses covariates in outcome model; default NULL
`trt_unit`	Treated unit
`lambda`	Ridge hyper-parameter, if NULL use CV
`scm`	Include SCM or not
`ridge`	Include ridge or not
`use_weights`	Whether to use weights in se estimate, default: FALSE
`hc`	Type of HC variance estimate (-1 for homoskedastic)
`trt_effect`	Whether to compute standard errors for treatment effect or counterfactual mean
`cols`	Column names corresponding to the units, time variable, outcome, and treated indicator