match_inputs.R

#' matching all inputs:
#' 1. combine in and out of camp data for each, HH and loops 
#' 2. put together questionnaire
#' 3. prepare sampling frames:
#'     3.1 prepare columns in out of camp cluster level sampling frame
#'     3.2 aggregate out-of-camp to stratum level
#'     3.3.make strata id for in-camp sampling frame
#'     3.4.combine the stratum sampling frames
#'     3.5.add strata  ids to the dataset
#'     3.6.add cluster ids to the dataset
#'     3.6. throw error if any don't match




# rbind in and out of camp
idp_in_camp$population_group <- "idp_in_camp"
response <- rbind.fill(response, idp_in_camp)
loop <- rbind.fill(loop, loop_in_camp)


  
# questionnaire

questionnaire <- load_questionnaire(response,questions,choices)



# prepare samplingframes

## cluster sampling frame: make values that can be matched with data
samplingframe$popgroup[samplingframe$popgroup=="out_of_camp"]<-"idp_out_camp"
samplingframe$district <- to_alphanumeric_lowercase(samplingframe$district)
samplingframe$stratum <- paste0(samplingframe$district, samplingframe$popgroup)
#samplingframe$cluster_ID <- cluster_lookup_table$new_ID[match(samplingframe$psu, cluster_lookup_table$psu)]
#samplingframe$cluster_strata_ID <- paste(samplingframe$cluster_ID, samplingframe$popgroup, sep = "_")


## add cluster ids to data:

### any pop group or cluster ids that can't be found individually?
`%find those not in%`<-function(x,y){x[!(x%in%y)] %>% unique}

response$population_group %find those not in% samplingframe$popgroup
#response$cluster_location_id[response$population_group != "idp_in_camp"] %find those not in% samplingframe$cluster_ID

### create id in samplingframe and in data
#samplingframe$cluster_strata_ID <- paste(samplingframe$cluster_ID, samplingframe$popgroup, sep = "_")
response$cluster_id <- paste(response$cluster_location_id, response$population_group, sep = "_")

# any id that can't be found in combination?
#response$cluster_id %find those not in% samplingframe$cluster_strata_ID %>% paste0(collapse="\n") %>% cat


### reverse not matching?
#samplingframe$cluster_strata_ID %find those not in% response$cluster_id
## unique cluster ids for IDP in camp (unique):

in_camp <- response$population_group=="idp_in_camp"
in_camp_ids<-paste0("NO_CLUSTER_",1:length(which(response$population_group=="idp_in_camp")),"_idp_in_camp")
in_camp_cluster_sampling_frame<-data.frame(cluster_strata_ID= in_camp_ids,pop=1)

response$cluster_id[in_camp]<-in_camp_ids














# samplingframe$stratum<-gsub("idp$","idp_out_camp",samplingframe$stratum) # should not be run; sampling frame  


## in camp: make values match data
samplingframe_in_camp$stratum<-paste0(samplingframe_in_camp$camp,"idp_in_camp")

## in camp: make columns match other sampling frame:
samplingframe_in_camp <- samplingframe_in_camp %>% dplyr::select(stratum,"population" = population..july.cccm.)
## combine in camp and out of camp sampling frames
samplingframe_strata <- rbind(samplingframe[,c("stratum", "population")],samplingframe_in_camp)

# add strata names to data

## out of camp:
response <- response %>% 
  mutate(strata = paste0(district,population_group))

## in camp: (replace district with camp name for idp_in_camp population group)
response$strata[response$population_group=="idp_in_camp"]<- (paste0(response$camp,response$population_group))[response$population_group=="idp_in_camp"]



## add rows to cluster samplingframe where no cluster sampling was used
# samplingframe<-bind_rows(samplingframe,in_camp_cluster_sampling_frame)




## Check if all match:

if(any(!(response$strata %in% samplingframe_strata$stratum))){
  warning("some strata not found in samplingframe")
  warning(which(!(response$strata %in% samplingframe_strata$stratum)) %>% length)
}

if(any(is.na(response$strata))){
  warning("strata can not be NA")
}