#MODIFIED ON: 11/15/2016            

  df_files <- data.frame(lf =lf,

generate_raster_number_of_prediction_by_day <- function(i,list_param){

  ##This function generates raster of missing pixels and number of predictions for days with missing tiles for a given region.

  #INPUTS

  #1) list_tiles_predicted_masked 

  #2) df_missing_tiles_day     

  #3) r_overlap_m 

  #4) num_cores

  #5) region_name 

  #6) NA_flag_val 

  #7) out_suffix 

  #8) out_dir 

  ###### Start script #####

  #### read in paramters    

  list_tiles_predicted_masked <- list_param$list_tiles_predicted_masked

  df_missing_tiles_day <- list_param$df_missing_tiles_day    

  r_overlap_m <- list_param$r_overlap_m

  num_cores <- list_param$num_cores # 6 #PARAM 14

  region_name <- list_param$region_name #<- "world" #PARAM 15

  NA_flag_val <-list_param$NA_flag_val

  out_suffix  <- list_param$out_suffix

  out_dir  <- list_param$out_dir

  ### To add or explore later...could have differences between predictions and rmse

  layers_option <- c("var_pred") #arg 17 ,param 17, options are:#res_training, res_testing,ac_training, ac_testing, var_pred

  NA_value <- NA_flag_val 

  #metric_name <- "rmse" #to be added to the code later...

  ##### Select relevant day and create stack of missing tiles

  missing_tiles <- df_missing_tiles_day[i,]

  date_str <- missing_tiles$date

  selected_col <- names(list_tiles_predicted_masked)

  missing_tiles_subset <- subset(missing_tiles,select=selected_col)

  selected_missing <- missing_tiles_subset==1

  list_missing_tiles_raster <- list_tiles_predicted_masked[selected_missing]

  r_tiles_s <- stack(list_missing_tiles_raster)

  ##### Sum missing tiles in the stack and generate number of predictions by pixels

  datasum <- stackApply(r_tiles_s, 1:nlayers(r_tiles_s), fun = sum)

  ### then substract missing tiles...

  r_day_predicted <- r_overlap_m - datasum

  raster_name_number_prediction <- file.path(out_dir_str,paste("r_day_number_of_prediction_sum_day_mosaiced","_",region_name,"_masked_",date_str,file_format,sep=""))

  writeRaster(r_day_predicted, NAflag=NA_flag_val,filename=raster_name_number_prediction,overwrite=TRUE)  

  r_table <- ratify(r_day_predicted) # build the Raster Attibute table

  rat <- levels(r_table)[[1]]#get the values of the unique cell frot the attribute table

  #rat$legend <- paste0("tile_",1:26)

  tb_freq <- as.data.frame(freq(r_table))

  rat$legend <- tb_freq$value

  levels(r_table) <- rat

  res_pix <- 800

  #res_pix <- 480

  col_mfrow <- 1

  row_mfrow <- 1

  png_filename <-  file.path(out_dir,paste("Figure_number_of_predictionds_by_pixel_",date_str,"_",region_name,"_",out_suffix,".png",sep =""))

  title_str <-  paste("Number of predicted pixels for ",variable_name," on ",date_str, sep = "")

  png(filename=png_filename,width = col_mfrow * res_pix,height = row_mfrow * res_pix)

  #my_col=c('blue','red','green')

  my_col <- rainbow(length(tb_freq$value))

  plot(r_table,col=my_col,legend=F,box=F,axes=F,main=title_str)

  legend(x='topright', legend =rat$legend,fill = my_col,cex=0.8)

  dev.off()

  ### Day missing reclass above

  ## do this in gdalcalc or overlay function to go faster?

  r_missing_day <- r_day_predicted == 0

  res_pix <- 800

  #res_pix <- 480

  col_mfrow <- 1

  row_mfrow <- 1

  png_filename <-  file.path(out_dir,paste("Figure_missing_predictionds_by_pixel_",date_str,"_",region_name,"_",out_suffix,".png",sep =""))

  title_str <-  paste("Number of predicted pixels for ",variable_name," on ",date_str, sep = "")

  png(filename=png_filename,width = col_mfrow * res_pix,height = row_mfrow * res_pix)

  #my_col=c('blue','red','green')

  my_col <- c("black","red")

  plot(r_missing_day,col=my_col,legend=F,box=F,axes=F,main=title_str)

  legend(x='topright', legend =c("prediced","missing"),fill = my_col,cex=0.8)

  dev.off()

  ### writeout data

  #extension_str <- extension(lf_files)

  #raster_name_tmp <- gsub(extension_str,"",basename(lf_files))

  #out_suffix_str <- paste0(region_name,"_",out_suffix)

  raster_name_missing <- file.path(out_dir_str,paste("r_missing_day_mosaiced","_",region_name,"_masked_",date_str,file_format,sep=""))

  writeRaster(r_missing_day, NAflag=NA_flag_val,filename=raster_name_missing,overwrite=TRUE)  

  ### generate return object

  obj_number_day_predicted <- list(raster_name_number_prediction,raster_name_missing)

  names(obj_number_day_predicted) <- c("raster_name_number_prediction","raster_name_missing")

  return(r_day_predicted)

}