Release v0.2.14

multimput 0.2.14

• In case all imputed values are identical, model_impute() only runs a single
  model on one imputation.
  It reports the mean and standard errors based on the single model as-is.
• model_impute() handles empty data.
• model_impute() can filter the covariates with a user supplied function.
• model_impute() gains a timeout argument.
• Bugfix in generating zero-inflated negative binomial data.