euromomo: EuroMoMo algorithm

### Start by copying the defaults-example.txt and edit it to suit your needs
if(!file.exists("momomaster.R"))
  if(file.exists("dev"))
    setwd("dev")
getwd()
list.files()
### Example of the workflow in the country
library("euromomo")


### Now using the options
parseDefaultsFile("defaults.txt")
checkOptions()

# Read in the raw data
momoFile <- readmomofile()

# Create the working directory
week.dir<-directories(lastFullWeek=momoFile$dLastFullWeek)
cat("Results are stored in ",week.dir,"\n")

#Create the groups (as stored in the option file)
momo <- makeGroups(momoFile$momo)

### read holidays HERE
holiday.file<-holiday(holiday.filename=getOption("euromomo")$HolidayFile)

### The groups are determined from the options
groups<-names(getOption("euromomo")[["groups"]])

### This is where the results go
results.list<-list()

for (i in groups) {
  #i<-"momodefault5"
  groupOpts <- getOption("euromomo")[["groups"]][[i]]
  cat("Group",groupOpts["label"],"\n")

  groupIndicator <- momo[, paste("group_",i,sep="")]
  back <- as.numeric(groupOpts["back"])

  #Generate reporting triangle
  rTList <- df2ReportingTriangle(momo, groupIndicator, back, dWeeks=momoFile$dWeeks, dLastFullWeek=momoFile$dLastFullWeek) # something about the group

  #Compute diagnostics by illustrating the reporting triangle
  #using larger delays than the requested, say 3*back.
  backDiagnostic <- 3*back
  rTList.diagnostic <- df2ReportingTriangle(momo, groupIndicator, backDiagnostic, dWeeks=momoFile$dWeeks, dLastFullWeek=momoFile$dLastFullWeek)

  #Show delays for 0,...,(group specific) back as function over time
  plotDelayDiagnostics2File(rTList=rTList.diagnostic, w=1, quantile=c(0.25,0.50,0.75,0.9,0.95,0.99),
                            main=groupOpts["label"], week.dir=week.dir)

  #Extract cumulative version from list and show
  rTDF <- rT2DataFrame(rTList$cumRT)
  cat("Group",groupOpts["label"]," reporting triangle (cumulative):\n")
  print(head(rTDF))

  # Delay adjustment
  #drTDF<-delay(rTDF,method="negbin",holiday=holiday.file)
  drTDF<-delay.nb(rTDF,holiday=holiday.file)
  cat("Group",groupOpts["label"]," with delay correction\n")
  print(tail(drTDF,20))

  # Add conditions for the baseline estimation
  data2<-addconditions(drTDF,
                       spring=getOption("euromomo")$spring,
                       autumn=getOption("euromomo")$autumn,
                       delay=back,
                       last=getOption("euromomo")$DayOfAggregation,
                       seasons=getOption("euromomo")$BaselineSeasons)

  #Add meta data (as a column replicating all the information).
  data2WithMD <- addMetaData(df=data2, groupName=i, groupOptions=groupOpts)

  #summary(data2)

  # Estimate the baseline
  data3 <- baseline(data2WithMD,groupOptions=groupOpts,clean=TRUE)
  cat("Group",groupOpts["label"]," with baseline\n")
  print(tail(data3))

  # Calculate Z-scores
  data4 <- zscore(data3)

  # Calculate excess
  data5 <- excess(data4,type=getOption("euromomo")$Delayvariance)
  #tail(data5)

  # Generate output
  try(output(data5))

  # Create diagnostic plots
  try(diagnostic.plots(data5))

  # Store the results
  results.list[[i]]<-data5
}

# on most system, you can check the outputs using
# system(paste("open",week.dir))
final <- do.call("rbind",results.list)
rownames(final) <- NULL

#Write files to hub
writeHUBFiles(final=final, dLastFullWeek=momoFile$dLastFullWeek)