function_assessment_part3 <- "global_run_scalingup_assessment_part3_05162016.R"

l_dates <- c("19920101","19920102","19920103","19920701","19920702","19920703") #dates to plot and analze

### Get data information

data_date <- unlist(lapply(list_data_s_fname, function(x){unlist(strsplit(basename(x),"_"))[5]}))

df_data_s_fname <- data.frame(data_date,region_name,dataset="data_s",file=list_data_s_fname)

year_str <- as.character(unlist(lapply(list_data_v_fname, function(x){unlist(strsplit(basename(x),"_"))[5]})))

df_data_v_fname <- data.frame(year_str,region_name,dataset="data_v",file=list_data_s_fname)

df_data_v_fname$year <- df_data_v_fname$year_str

df_data_v_fname$year <- as.character(df_data_v_fname$year_str)

list_dates_val <- as.Date(strptime(l_dates,"%Y%m%d"))

month_str <- format(list_dates_val, "%b") ## Month, char, abbreviated

year_str <- format(list_dates_val, "%Y") ## Year with century

day_str <- as.numeric(format(list_dates_val, "%d")) ## numeric month

for(i in 1:length(unique(year_str))){

  df_data_v_fname$data_date==year_str[i]

  filename_tmp <- as.character(df_data_v_fname$file[9])

  df_stations_tmp <- read.table(filename_tmp,stringsAsFactors=F,sep=",")

}

df_points <- subset(df_stations,df_stations_tmp$date==l_dates[1])

table(df_points$tile_id)

plot(df_points,add=T)

coordinates(df_points) <- cbind(df_points$x,df_points$y)

proj4string(df_points) <- CRS_locs_WGS84

#####################

#select one station based on id or coordinates and find that in the list of data.frame??

id_selected <- "82111099999"

dim(df_points)

### loop through all files and get the time series

extract_from_df <- function(x,col_selected,val_selected){

  df_tmp <- read.table(x,stringsAsFactors=F,sep=",")

  #data_subset <- subset(data_stations,col_selected==val_selected)

  data_subset <- subset(df_tmp,df_tmp$id%in%val_selected)

  return(data_subset)

}

lf_data_s_subset <- mclapply(list_data_s_fname,

                           FUN=extract_from_df,

                           col_selected="id",

                           val_selected=id_selected,

                           mc.preschedule=FALSE,

                           mc.cores = num_cores)                         

lf_data_v_subset <- mclapply(list_data_v_fname,

                           FUN=extract_from_df,

                           col_selected="id",

                           val_selected=id_selected,

                           mc.preschedule=FALSE,

                           mc.cores = num_cores)   

data_v_subset <- do.call(rbind,lf_data_v_subset)

data_s_subset <- do.call(rbind,lf_data_s_subset)

data_s_subset$training <- 1

data_v_subset$training <- 0

data_stations <- rbind(data_s_subset,data_v_subset)

dim(data_s_subset)

dim(data_v_subset)

#rbind.fill(mtcars[c("mpg", "wt")], mtcars[c("wt", "cyl")])

data_stations <- rbind.fill(data_v_subset, data_s_subset)

coordinates(data_stations) <- cbind(data_stations$x,data_stations$y)

proj4string(data_stations) <- CRS_locs_WGS84

dim(data_stations)

##Now grab year year 1992 or matching year...maybe put this in a data.frame??

dim(data_stations)

df_points_extracted <- extract(r_mosaic,data_stations,df=T)

df_points_extracted_fname <- paste0("df_points_extracted_fname",".txt")

write.table(df_points_extracted,file=df_points_extracted_fname)