##################################################################
### Functions to estimate probabilities and delays from OSIRIS ###
##################################################################
#30/10/2020: nieuw pad voor files

#####################
### Read the data ###
#####################

# files <- list.files("___UITZONDERINGSGROND_6___", full.names = TRUE)
# files <- c(files, list.files("___UITZONDERINGSGROND_6___", full.names = TRUE))
files <- list.files("___UITZONDERINGSGROND_6___", full.names = TRUE)
files <- c(files, list.files("___UITZONDERINGSGROND_6___", full.names = TRUE))
files <- list.files("/rivm/r/COVID-19/Surveillance/Data/OSIRIS/Geschoond/Previous", full.names = TRUE)
files <- c(files, list.files("/rivm/r/COVID-19/Surveillance/Data/OSIRIS/Geschoond", full.names = TRUE))
files <- files[grepl(files, pattern = "rds")]
filedates <- as.Date(
  substr(files, regexpr("[[:digit:]]{8}", files), regexpr("[[:digit:]]{8}", files) + 7),
  "%Y%m%d")
if(exists("analysisdate")) {
  file.name <- files[filedates == analysisdate]
  filedate <- analysisdate
} else {
  file.name <- files[filedates %>% which.max]
  filedate <- max(filedates)
}
dataOsiris <- readRDS(file.name)


##########################
### Symptom to hospital ###
###########################
get_S2H_probs <- function() {
  dataOsiris %>%
    filter(ZIE1eZiekteDt < as.Date("2020-03-20")) %>%
    select(NCOVpatZhs, Leeftijdsgroep) %>%
    separate(Leeftijdsgroep, c("onder", "boven")) %>% 
    filter(onder != "Niet") %>%
    mutate(onder = as.numeric(onder),
           age_class = floor(onder / 10),
           age_class = if_else(age_class > 8, 8, age_class)) %>% 
    group_by(age_class) %>%
    summarise(totaalS2Hp = n(),
              hospS2Hp = sum(NCOVpatZhs == "J", na.rm = T),
              probS2H = hospS2Hp/totaalS2Hp) %>% 
    select(totaalS2Hp, hospS2Hp, probS2H) %>%
    as.list()
}

get_S2H_delays <- function() {
  dataOsiris %>%
    select(ZIE1eZiekteDt, NCOVdat1ezkhopn, Leeftijdsgroep) %>%
    filter(!is.na(ZIE1eZiekteDt) & !is.na(NCOVdat1ezkhopn)) %>%
    mutate(interval = as.numeric(NCOVdat1ezkhopn - ZIE1eZiekteDt)) %>%
    filter(interval >= 0 & interval < 31) %>% 
    separate(Leeftijdsgroep, c("onder", "boven")) %>% 
    filter(onder != "Niet") %>%
    mutate(onder = as.numeric(onder),
           age_class = floor(onder / 10),
           age_class = if_else(age_class > 8, 8, age_class)) %>%
    group_by(age_class) %>%
    summarise(meanS2H = mean(interval),
              varS2H = var(interval),
              pS2H = 1 - meanS2H/varS2H,
              rS2H = meanS2H * (1-pS2H)/pS2H) %>%
    select(meanS2H, pS2H, rS2H) %>%
    as.list()
}

########################
### Symptom to death ###
########################
get_S2D_delays <- function() {
  dataOsiris %>%
    filter(Ziekenhuisopname != "Ja" & ICU != "Ja") %>%
    select(ZIE1eZiekteDt, ZIEDtOverlijden) %>%
    filter(!is.na(ZIE1eZiekteDt) & !is.na(ZIEDtOverlijden)) %>%
    mutate(interval = as.numeric(ZIEDtOverlijden - ZIE1eZiekteDt)) %>%
    filter(interval >= 0) %>%
    summarise(meanS2D = mean(interval),
              varS2D = var(interval),
              pS2D = 1 - meanS2D/varS2D,
              rS2D = meanS2D * (1-pS2D)/pS2D) %>%
    select(meanS2D, pS2D, rS2D) %>%
    as.list()
}

get_S2D_probs <- function() {
  dataOsiris %>%
    filter(Ziekenhuisopname != "Ja" & ICU != "Ja") %>%
    select(Overleden, Leeftijdsgroep) %>%
    separate(Leeftijdsgroep, c("onder", "boven")) %>% 
    filter(onder != "Niet") %>%
    mutate(onder = as.numeric(onder),
           age_class = floor(onder / 10),
           age_class = if_else(age_class > 8, 8, age_class)) %>%
    group_by(age_class) %>%
    summarise(totaalS2Dp = n(),
              gestorvenS2D = sum(Overleden == "Ja"),
              probS2D = gestorvenS2D/totaalS2Dp) %>%
    select(totaalS2Dp, gestorvenS2D, probS2D) %>%
    as.list()
}

########################
### incidence curves ###
########################

inc_osiris_report <- function(ages = FALSE, wROAZ = "all") {
  dataOsiris %>%
    select(ZIE1eZiekteDt) %>%
    filter(ZIE1eZiekteDt <= filedate) %>%
    mutate(sympdag = as.numeric(ZIE1eZiekteDt - as.Date("2020-02-12"))) %>%
    pull(sympdag) %>% tabulate()
}

inc_osiris_hosp <- function(ages = FALSE, wROAZ = "all") {
  if(is.numeric(wROAZ)) {
    wROAZ <- sort(unique(ROAZdata$Regio_ROAZ))[-9][wROAZ]
  }
  if(wROAZ == "all") {
    if(ages) {
      hospdata <- dataOsiris %>%
        select(NCOVdat1ezkhopn, Leeftijdsgroep) %>%
        filter(NCOVdat1ezkhopn <= filedate) %>%
        separate(Leeftijdsgroep, c("onder", "boven")) %>% 
        filter(onder != "Niet") %>%
        mutate(onder = as.numeric(onder),
               age_class = floor(onder / 10),
               age_class = if_else(age_class > 8, 8, age_class)) %>%
        group_by(age_class) %>%
        mutate(hospdag = as.numeric(NCOVdat1ezkhopn - as.Date("2020-02-12"))) 
      toreturn <- lapply(0:8, function(x) tabulate(hospdata %>% 
                                                     filter(age_class == x) %>% 
                                                     pull(hospdag), nbins = as.numeric(filedate - as.Date("2020-02-12"))))
      names(toreturn) <- c("[0,10)", "[10,20)", "[20,30)", "[30,40)",
                           "[40,50)", "[50,60)", "[60,70)", "[70,80)", "[80,Inf]")
      return(toreturn)
    } else {
      dataOsiris %>%
        select(NCOVdat1ezkhopn) %>%
        filter(NCOVdat1ezkhopn <= filedate) %>%
        mutate(hospdag = as.numeric(NCOVdat1ezkhopn - as.Date("2020-02-12"))) %>%
        pull(hospdag) %>% tabulate(nbins = as.numeric(filedate - as.Date("2020-02-12")))
    }
  } else {
    if(ages) {
      hospdata <- dataOsiris %>%
        filter(ROAZ == wROAZ) %>%
        select(NCOVdat1ezkhopn, Leeftijdsgroep) %>%
        filter(NCOVdat1ezkhopn <= filedate) %>%
        separate(Leeftijdsgroep, c("onder", "boven")) %>% 
        filter(onder != "Niet") %>%
        mutate(onder = as.numeric(onder),
               age_class = floor(onder / 10),
               age_class = if_else(age_class > 8, 8, age_class)) %>%
        group_by(age_class) %>%
        mutate(hospdag = as.numeric(NCOVdat1ezkhopn - as.Date("2020-02-12"))) 
      toreturn <- lapply(0:8, function(x) tabulate(hospdata %>% 
                                                     filter(age_class == x) %>% 
                                                     pull(hospdag), nbins = as.numeric(filedate - as.Date("2020-02-12"))))
      names(toreturn) <- c("[0,10)", "[10,20)", "[20,30)", "[30,40)",
                           "[40,50)", "[50,60)", "[60,70)", "[70,80)", "[80,Inf]")
      return(toreturn)
    } else {
      dataOsiris %>%
        filter(ROAZ == wROAZ) %>%
        select(NCOVdat1ezkhopn) %>%
        filter(NCOVdat1ezkhopn <= filedate) %>%
        mutate(hospdag = as.numeric(NCOVdat1ezkhopn - as.Date("2020-02-12"))) %>%
        pull(hospdag) %>% tabulate(nbins = as.numeric(filedate - as.Date("2020-02-12")))
    }
  }
}

inc_osiris_mort <- function(ages = FALSE, wROAZ = "all") {
  if(is.numeric(wROAZ)) {
    wROAZ <- sort(unique(ROAZdata$Regio_ROAZ))[-9][wROAZ]
  }
  if(wROAZ == "all") {
    if(ages) {
      mortdata <- dataOsiris %>%
        select(ZIEDtOverlijden, Leeftijdsgroep) %>%
        filter(ZIEDtOverlijden <= filedate) %>%
        separate(Leeftijdsgroep, c("onder", "boven")) %>% 
        filter(onder != "Niet") %>%
        mutate(onder = as.numeric(onder),
               age_class = floor(onder / 10),
               age_class = if_else(age_class > 8, 8, age_class)) %>%
        group_by(age_class) %>%
        mutate(mortdag = as.numeric(ZIEDtOverlijden - as.Date("2020-02-12"))) 
      toreturn <- lapply(0:8, function(x) tabulate(mortdata %>% 
                                                     filter(age_class == x) %>% 
                                                     pull(mortdag), nbins = as.numeric(filedate - as.Date("2020-02-12"))))
      names(toreturn) <- c("[0,10)", "[10,20)", "[20,30)", "[30,40)",
                           "[40,50)", "[50,60)", "[60,70)", "[70,80)", "[80,Inf]")
      return(toreturn)
    } else {
      dataOsiris %>%
        select(ZIEDtOverlijden) %>%
        filter(ZIEDtOverlijden <= filedate) %>%
        mutate(mortdag = as.numeric(ZIEDtOverlijden - as.Date("2020-02-12"))) %>%
        pull(mortdag) %>% tabulate(nbins = as.numeric(filedate - as.Date("2020-02-12")))
    }
  } else {
    if(ages) {
      mortdata <- dataOsiris %>%
        filter(ROAZ == wROAZ) %>%
        select(ZIEDtOverlijden, Leeftijdsgroep) %>%
        filter(ZIEDtOverlijden <= filedate) %>%
        separate(Leeftijdsgroep, c("onder", "boven")) %>% 
        filter(onder != "Niet") %>%
        mutate(onder = as.numeric(onder),
               age_class = floor(onder / 10),
               age_class = if_else(age_class > 8, 8, age_class)) %>%
        group_by(age_class) %>%
        mutate(mortdag = as.numeric(ZIEDtOverlijden - as.Date("2020-02-12"))) 
      toreturn <- lapply(0:8, function(x) tabulate(mortdata %>% 
                                                     filter(age_class == x) %>% 
                                                     pull(mortdag), nbins = as.numeric(filedate - as.Date("2020-02-12"))))
      names(toreturn) <- c("[0,10)", "[10,20)", "[20,30)", "[30,40)",
                           "[40,50)", "[50,60)", "[60,70)", "[70,80)", "[80,Inf]")
      return(toreturn)
    } else {
      dataOsiris %>%
        filter(ROAZ == wROAZ) %>%
        select(ZIEDtOverlijden) %>%
        filter(ZIEDtOverlijden <= filedate) %>%
        mutate(mortdag = as.numeric(ZIEDtOverlijden - as.Date("2020-02-12"))) %>%
        pull(mortdag) %>% tabulate(nbins = as.numeric(filedate - as.Date("2020-02-12")))
    }
  }
}