rr

library(C50);
library(RoughSets);
library(keras);
source("knn.R")


#* Echo back the input
#* @param STG The message to echo
#* @param SCG The second number to add
#* @param STR The second number to add
#* @param LPR The second number to add
#* @param PEG The second number to add
#* @get /api/results
function(STG, SCG, STR, LPR, PEG) {

  STG <- as.double(STG)
  SCG <- as.double(SCG)
  STR <- as.double(STR)
  LPR <- as.double(LPR)
  PEG <- as.double(PEG)

  new_case <- matrix(c(STG, SCG, STR, LPR, PEG), ncol=5, byrow=TRUE)
  colnames(new_case) <- c("STG","SCG","STR","LPR","PEG")
  rownames(new_case) <- c("1")

  model_tree <- get(load("models/c50.RData"))
  model_lem2 <- get(load("models/lem2.RData"))
  model_tf <- load_model_tf("models/deep_learning_model")

  sink("nul")
  model_knn <- func_knn(STG, SCG, STR, LPR, PEG)
  sink()

  predict_tree <- predict(model_tree, new_case)
  predict_lem2 <- predict(model_lem2, SF.asDecisionTable(new_case))
  predict_tf <- predict_classes(model_tf, new_case) + 1
  predict_knn <- model_knn$Prediction

  predict_lem2 <- as.integer(as.vector(predict_lem2$predictions))
  predict_tree <- as.integer(predict_tree)
  predict_tf <- as.integer(predict_tf)
  predict_knn <- as.integer(predict_knn)

  result_list <- list("c50" = predict_tree,
                      "lem2" = predict_lem2,
                      "tf" = predict_tf,
                      "knn" = predict_knn,
                      "knn_stats"=model_knn )
  return(result_list)
}


#library(plumber)
# r <- plumb("prediction.R")  # Where 'plumber.R' is the location of the file shown above
# r$run(port=8000)