using CSV, DataFrames, MLJ, Random

data_csv=DataFrame(CSV.File("../CCGRID2022.csv"))
data=groupby(data_csv,[:simkey,:wireless,:wakeupfor,:seed])
data=combine(data,:nDataRcv => sum => :coverage, :energy => sum => :energy)

# Building train/test sets
train_idx=randsubseq(1:size(data,1),0.80)
test_idx=setdiff(1:size(data,1),train_idx)
train_set=data[train_idx,:]
test_set=data[test_idx,:]

# KNN
knn=@load KNNClassifier
y,X=unpack(data,==(:simkey))
knn_machine=machine(knn(),select(X,Not(:wireless),Not(:seed)),select(data,:seed))