library("tidyverse")
library("class")
library("rpart")
library("rpart.plot")
library("viridis")
library("MLmetrics")

## Simulation Parameters:
## simkey         {baseline,extended,hint,hintandextended}
## wireless       {lora,nbiot}
## wakeupfor      {60s,180s}
## seed           [1,200]
## node           on[0,12]
## isSender       {0,1}
## dataSize       {1MB}

## Metrics:
## energy         [0,+inf)
## nDataRcv       [0,+inf)

nseed=200
nwakeupfor=2
nwireless=2
nsimkey=4
nsimulations=nseed*nwakeupfor*nwireless*nsimkey # Must be 3200

## Load data
data=read_csv("../CCGRID2022.csv")%>%distinct() # Note that in the data experiment wireless=="lora",seed==1,wakeupfor==60,simkey=="baseline" is present 2 times in the CSV file
tmp_data_coverage=data%>%group_by(simkey,wireless,wakeupfor,seed)%>%mutate(coverage=sum(nDataRcv))%>%ungroup()%>%filter(isSender==1)%>%select(simkey,wireless,wakeupfor,seed,coverage)
data_seed_isSender=data%>%group_by(simkey,wireless,wakeupfor,seed,isSender)%>%summarize(energy_mean=mean(energy))%>%
    left_join(tmp_data_coverage,by=c("simkey","wireless","wakeupfor","seed"))%>%
    mutate(efficiency=energy_mean/coverage)%>%
    ungroup()
data_seed=data%>%group_by(simkey,wireless,wakeupfor,seed)%>%summarize(energy=sum(energy),coverage=sum(nDataRcv))%>%
    mutate(efficiency=energy/coverage)%>%
    ungroup()

generate_accuracy_for=function(ignore_hint=FALSE,seed_max=200,attempts_max=2,wrl="lora",wuf=180) {
    attempts=seq(1,attempts_max)
    results=sapply(attempts,function(attempt){
    ## Prepare data for traning
    set.seed(1+attempt) # Reproducibility
    wireless_map=c("lora"=1,"nbiot"=2)
    cur_data_seed=data_seed%>%filter(wakeupfor==wuf,wireless==wrl)
    data_ml=cur_data_seed%>%select(-efficiency)%>%mutate(wireless=wireless_map[cur_data_seed$wireless])
    if(ignore_hint){
        data_ml=data_ml%>%filter(simkey!="hint")
    }
    train_set=data_ml%>%filter(seed<=seed_max)%>%select(-seed)     # train data on seed_max*3 days
    test_set=data_ml%>%anti_join(train_set)%>%select(-seed)        # build test_sed excluding training set

    ## KNN training
    knn_predictions=knn(train=train_set%>%select(-simkey),test=test_set%>%select(-simkey),cl=train_set$simkey,k=min(10,NROW(train_set)))
    ## KNN analysis
    knn_cont_table=table(knn_predictions,test_set$simkey)
    knn_accuracy=round((sum(diag(knn_cont_table)/sum(rowSums(knn_cont_table))))*100)
    knn_prop_table=round(prop.table(knn_cont_table),digits=2)
    knn_f1_score=F1_Score(test_set$simkey,knn_predictions)
    knn_recall=Recall(test_set$simkey,knn_predictions)
    knn_precision=Precision(test_set$simkey,knn_predictions)
        
    ## Decision tree
    tree=rpart(
        simkey ~ wireless + wakeupfor + energy + coverage,
        data=train_set,
        method="class",
        minsplit=60,
        minbucket=1)
    tree_predictions=predict(tree,newdata=test_set%>%select(-simkey),type="class")
    tree_cont_table=table(tree_predictions,test_set$simkey)
    tree_accuracy=round((sum(diag(tree_cont_table)/sum(rowSums(tree_cont_table))))*100)
    tree_prop_table=round(prop.table(tree_cont_table),digits=2)
    tree_f1_score=F1_Score(test_set$simkey,tree_predictions)
    tree_recall=Recall(test_set$simkey,tree_predictions)
    tree_precision=Precision(test_set$simkey,tree_predictions)
     list(tibble(seed_max=seed_max,model=c("knn","tree"),accuracy=c(knn_accuracy,tree_accuracy),f1_score=c(knn_f1_score,tree_f1_score),recall=c(knn_recall,tree_recall),precision=c(knn_precision,tree_precision)))
    })
    ## Prints
    results=do.call("rbind",results)
    results%>%mutate(seed_max=seed_max,attempts_max=attempts_max,wireless=wrl,wakeupfor=wuf)

}

generate_accuracy = function(wireless,wakeupfor,steps=1, accuracy=20,ignore_hint=TRUE){
    npolicies=4
    if(ignore_hint){npolicies=npolicies-1}
    ## Generate inputs
    result=tibble()
    for(i in seq(1,160,by=steps)){ # We stop at 80% of the data (this way test set is at least 20%)
        acc=generate_accuracy_for(ignore_hint=ignore_hint,seed=i,attempts_max=accuracy,wrl=wireless,wuf=wakeupfor)
        result=rbind(result,acc)
    }
    result%>%mutate(days=seed_max*npolicies,months=days/30) # Since 3 policies (since ignore_hint=TRUE)
}

# Generate accuracy for each wireless and uptime
accuracy=rbind(generate_accuracy("lora",60),
               generate_accuracy("lora",180),
               generate_accuracy("nbiot",60),
               generate_accuracy("nbiot",180))

## Summarize
result_summary=accuracy%>%group_by(wireless,wakeupfor,months,model)%>%
    summarize(
        mean_accuracy=mean(accuracy),sd_accuracy=sd(accuracy),min_accuracy=min(accuracy),max_accuracy=max(accuracy),
        mean_f1_score=mean(f1_score),sd_f1_score=sd(f1_score),min_f1_score=min(f1_score),max_f1_score=max(f1_score),
        mean_recall=mean(recall),sd_recall=sd(recall),min_recall=min(recall),max_recall=max(recall),
        mean_precision=mean(precision),sd_precision=sd(precision),min_precision=min(precision),max_precision=max(precision))

## Result max
metrics_peak=result_summary%>%group_by(wireless,wakeupfor,model)%>%
    summarize(max_accuracy=max(mean_accuracy),
              max_f1_score=max(mean_f1_score),
              max_recall=max(mean_recall),
              max_precision=max(mean_precision))

## Plot
sapply(c("knn","tree"),function(grp){
    ggplot(result_summary%>%filter(model==grp),aes(months,mean_accuracy))+
        geom_ribbon(aes(ymin=mean_accuracy-sd_accuracy,ymax=mean_accuracy+sd_accuracy),alpha=0.2,color=NA)+
        geom_line(size=1.1)+geom_point(size=3)+xlab("Number of training months")+ylab(paste("Mean",grp,"accuracy"))+ggtitle(paste(grp,"accuracy"))+
#        ylim(c(0,100))+
        geom_hline(data=metrics_peak%>%filter(model==grp),aes(yintercept=max_accuracy),color="red",size=1)+
        geom_text(data=metrics_peak%>%filter(model==grp),x=0,aes(y=max_accuracy,label = round(max_accuracy,digits=1),vjust=-1),color="red")+
        facet_wrap(~wireless+wakeupfor)
        scale_x_continuous(breaks = seq(0, max(result_summary$months), by = 1))
    ggsave(paste0("figures/months_",grp,".pdf"))
})