library(here)
library(sf)
library(terra)
library(tidyverse)
library(lubridate)
library(exactextractr)
library(readxl)


# Vang alle command line argumenten op
args <- commandArgs(trailingOnly = TRUE)


# Converteer het tweede argument naar een string waarde
project_dir <- as.character(args[1])

# Controleer of data_dir een geldige waarde is
if (!is.character(project_dir)) {
  project_dir <- "sony"
}


laravel_storage_dir <- here("laravel_app/storage/app", project_dir)
#preparing directories
planet_tif_folder <- here(laravel_storage_dir, "merged_tif")
merged_final <- here(laravel_storage_dir, "merged_final_tif")

new_project_question = FALSE

data_dir <- here(laravel_storage_dir, "Data")
extracted_CI_dir <- here(data_dir, "extracted_ci")
daily_CI_vals_dir <- here(extracted_CI_dir, "daily_vals")
cumulative_CI_vals_dir <- here(extracted_CI_dir, "cumulative_vals")

weekly_CI_mosaic <- here(laravel_storage_dir, "weekly_mosaic")
daily_vrt <- here(data_dir, "vrt")
harvest_dir <-  here(data_dir, "HarvestData")

source("parameters_project.R")
source("ci_extraction_utils.R")


pivot_stats2 <- readRDS(here(cumulative_CI_vals_dir,"combined_CI_data.rds")) %>%
    ungroup() %>%
    group_by(field, sub_field) %>%
    summarise(across(everything(), ~ first(na.omit(.))), .groups = "drop")

   #%>% drop_na(pivot_quadrant)

# gather data into long format for easier calculation and visualisation
pivot_stats_long <- pivot_stats2 %>%
  tidyr::gather("Date", value, -field, -sub_field ) %>%
  mutate(#Date = right(Date, 8),
         Date = lubridate::ymd(Date)
  ) %>%
  drop_na(c("value","Date")) %>%
  mutate(value = as.numeric(value))%>%
  filter_all(all_vars(!is.infinite(.))) %>%
  # rename(field = pivot_quadrant,
  #        sub_field = field) %>%
  unique()


# #2021
# pivot_select_model_Data_2021 <- harvesting_data %>% filter(Year == 2021) %>% pull(field)
#
# pivot_select_model_Data_2022 <- harvesting_data %>% filter(Year == 2022) %>% pull(field)

# pivot_select_model_Data_2023 <- harvesting_data %>% filter(year == 2023) %>% filter(!is.na(season_start))  %>% pull(sub_field)

pivot_select_model_Data_2024 <- harvesting_data %>% filter(year == 2024)%>% filter(!is.na(season_start)) %>% pull(sub_field)
message(pivot_select_model_Data_2024)
# pivots_dates_Data_2022 <- pivots_dates0 %>% filter(!is.na(season_end_2022))
# pivot_select_model_Data_2022 <- unique(pivots_dates_Data_2022$pivot_quadrant )
#
# pivots_dates_Data_2023 <- pivots_dates0 %>% filter(!is.na(season_start_2023))
# pivot_select_model_Data_2023 <- unique(pivots_dates_Data_2023$pivot_quadrant)
#
# pivots_dates_Data_2024 <- pivots_dates0 %>% filter(!is.na(season_start_2024))
# pivot_select_model_Data_2024 <- unique(pivots_dates_Data_2024$pivot_quadrant)

extract_CI_data <- function(field_names, harvesting_data, field_CI_data, season) {
  #  field_names =  "Nandi A1a"
  #  field_names =  "1.13A"
  #  harvesting_data = harvesting_data
  # field_CI_data = pivot_stats_long
  #  season= 2024

  filtered_harvesting_data <- harvesting_data %>%
  #  na.omit() %>%
    filter(year == season, sub_field %in% field_names)

  filtered_field_CI_data <- field_CI_data %>%
    filter(sub_field %in% field_names)
   if (nrow(filtered_field_CI_data) == 0) {
     return(data.frame())  # Return an empty data frame if no data is found
   }

  # CI <- map_df(field_names, ~ {
  ApproxFun <- approxfun(x = filtered_field_CI_data$Date, y = filtered_field_CI_data$value)
  Dates <- seq.Date(ymd(min(filtered_field_CI_data$Date)), ymd(max(filtered_field_CI_data$Date)), by = 1)
  LinearFit <- ApproxFun(Dates)

  CI <-  data.frame(Date = Dates, FitData = LinearFit) %>%
    left_join(., filtered_field_CI_data, by = "Date") %>%
    filter(Date > filtered_harvesting_data$season_start & Date < filtered_harvesting_data$season_end) %>%
    mutate(DOY = seq(1, n(), 1),
           model = paste0("Data", season, " : ", field_names),
           season = season,
           sub_field = field_names)
  # }) #%>%

  return(CI)
}

## Extracting the correct CI values
# Data_2021 <- map(pivot_select_model_Data_2021, ~ extract_CI_data(.x, harvesting_data = harvesting_data, field_CI_data = pivot_stats_long, season = 2021)) %>% list_rbind()
# message('2021')
# Data_2022 <- map(pivot_select_model_Data_2022, ~ extract_CI_data(.x, harvesting_data = harvesting_data, field_CI_data = pivot_stats_long, season = 2022)) %>% list_rbind()
# message('2022')
# Data_2023 <- map(pivot_select_model_Data_2023, ~ extract_CI_data(.x, harvesting_data = harvesting_data, field_CI_data = pivot_stats_long, season = 2023)) %>% list_rbind()
# message('2023')
Data_2024 <- map(pivot_select_model_Data_2024, ~ extract_CI_data(.x, harvesting_data = harvesting_data, field_CI_data = pivot_stats_long, season = 2024)) %>% list_rbind()
message('2024')
head(Data_2024)


#CI_all <- rbind(Data_2023, Data_2024)
CI_all <-  Data_2024
message('CI_all created')
#CI_all <-  Data_2023

CI_all <- CI_all %>% group_by(model) %>% mutate(CI_per_day = FitData - lag(FitData),
                                                cumulative_CI = cumsum(FitData))
message('CI_all cumulative')

head(CI_all)
message('show head')

saveRDS(CI_all, here(cumulative_CI_vals_dir,"All_pivots_Cumulative_CI_quadrant_year_v2.rds"))
message('rds saved')