Scope of eval_metric in binary classification with sparkxgb

[mzorko]

Hello, so the documentation states:

# eval_metric	
# Evaluation metrics for validation data, a default metric will be assigned according to objective 
# (rmse for regression, and error for classification, mean average precision for ranking). 
# options: rmse, mae, logloss, error, merror, mlogloss, auc, aucpr, ndcg, map, gamma-deviance

# maximize_evaluation_metrics	
# Whether to maximize evaluation metrics. Defaults to FALSE (for minization.)

When I try eval_metric = "auc" or eval_metric = "aucpr" or eval_metric = "error" the code below never completes. Do I have to set up another parameter, or is this option not supported at the moment? (eval_metric = "merror" and eval_metric = "mlogloss" work)

library(sparkxgb)
library(sparklyr)

sc <- spark_connect(master = "local")

my_data <- data.frame(x = 1:10, y = 11:20, target = sample(c("a", "b"), 10, TRUE)) %>% 
  sdf_copy_to(sc, .)

xgb_model <- xgboost_classifier(
  my_data, 
  target ~ .,
  num_class = 2,
  num_round = 5, 
  max_depth = 4,
  eval_metric = "auc",
  maximize_evaluation_metrics = TRUE
)

Looks like this is the problem with binary classification.

Where can I find out more about scope of all of the parameters regarding sparkxgb?
What is the best place to start if I want to help with problems regarding sparklyr extensions?