# CI (Canopy Index) Analysis Functions
# ====================================
# 
# Functions for loading and analyzing CI data from Planet imagery
# Creates interactive visualizations for this week, last week, and change detection
#
# Author: Timon
# Date: August 2025

# Load required libraries
if (!require(terra)) install.packages("terra")
if (!require(dplyr)) install.packages("dplyr")
if (!require(sf)) install.packages("sf")
if (!require(ggplot2)) install.packages("ggplot2")
if (!require(tmap)) install.packages("tmap")
if (!require(RColorBrewer)) install.packages("RColorBrewer")
if (!require(viridis)) install.packages("viridis")

library(terra)
library(dplyr)
library(sf)
library(ggplot2)
library(tmap)
library(RColorBrewer)
library(viridis)

# CI Data Loading Functions
# =========================

#' Load CI weekly mosaic for a specific week
#' @param week_num Week number (27-34 available)
#' @param year Year (2025)
#' @param data_dir Directory containing weekly CI mosaics
#' @return SpatRaster object
load_ci_weekly <- function(week_num, year = 2025, 
                          data_dir = "../../../laravel_app/storage/app/aura/weekly_mosaic") {
  
  filename <- file.path(data_dir, paste0("week_", week_num, "_", year, ".tif"))
  
  if (!file.exists(filename)) {
    warning(paste("CI weekly mosaic not found:", filename))
    return(NULL)
  }
  
  # Load raster
  ci_raster <- rast(filename)
  
  # Set appropriate names
  names(ci_raster) <- paste0("CI_week_", week_num, "_", year)
  
  cat("✓ Loaded CI for week", week_num, ":", filename, "\n")
  cat("  Dimensions:", dim(ci_raster), "\n")
  cat("  CRS:", as.character(crs(ci_raster)), "\n")
  
  # Get value range
  ci_range <- global(ci_raster, range, na.rm = TRUE)
  cat("  CI value range:", round(ci_range[1,1], 3), "to", round(ci_range[2,1], 3), "\n")
  
  return(ci_raster)
}

#' Get the most recent available CI week
#' @param data_dir Directory containing weekly CI mosaics
#' @return Week number of most recent data
get_latest_ci_week <- function(data_dir = "../../../laravel_app/storage/app/aura/weekly_mosaic") {
  
  # List all available weeks
  files <- list.files(data_dir, pattern = "week_.*_2025\\.tif$", full.names = FALSE)
  
  if (length(files) == 0) {
    stop("No CI weekly mosaics found in ", data_dir)
  }
  
  # Extract week numbers
  week_nums <- as.numeric(gsub("week_(\\d+)_2025\\.tif", "\\1", files))
  latest_week <- max(week_nums)
  
  cat("Available CI weeks:", paste(sort(week_nums), collapse = ", "), "\n")
  cat("Latest week:", latest_week, "\n")
  
  return(latest_week)
}

#' Load current and previous week CI data
#' @param data_dir Directory containing weekly CI mosaics  
#' @return Named list with current_week, previous_week CI rasters and week numbers
load_ci_comparison_data <- function(data_dir = "../../../laravel_app/storage/app/aura/weekly_mosaic") {
  
  # Get latest week
  latest_week <- get_latest_ci_week(data_dir)
  previous_week <- latest_week - 1
  
  # Load both weeks
  current_ci <- load_ci_weekly(latest_week, data_dir = data_dir)
  previous_ci <- load_ci_weekly(previous_week, data_dir = data_dir)
  
  if (is.null(current_ci)) {
    stop("Could not load current week CI data")
  }
  
  if (is.null(previous_ci)) {
    warning("Could not load previous week CI data, using available week")
    # Try to find any available previous week
    available_weeks <- get_available_ci_weeks(data_dir)
    available_weeks <- available_weeks[available_weeks < latest_week]
    if (length(available_weeks) > 0) {
      previous_week <- max(available_weeks)
      previous_ci <- load_ci_weekly(previous_week, data_dir = data_dir)
    }
  }
  
  return(list(
    current_ci = current_ci,
    previous_ci = previous_ci,
    current_week = latest_week,
    previous_week = previous_week
  ))
}

#' Get all available CI weeks
#' @param data_dir Directory containing weekly CI mosaics
#' @return Vector of available week numbers
get_available_ci_weeks <- function(data_dir = "../../../laravel_app/storage/app/aura/weekly_mosaic") {
  
  files <- list.files(data_dir, pattern = "week_.*_2025\\.tif$", full.names = FALSE)
  week_nums <- as.numeric(gsub("week_(\\d+)_2025\\.tif", "\\1", files))
  return(sort(week_nums))
}

# CI Interactive Visualization Functions  
# ======================================

#' Create interactive CI map
#' @param ci_raster CI SpatRaster
#' @param field_boundaries Field boundary polygons (optional)
#' @param title Map title
#' @param color_palette Color palette for CI values
#' @return tmap object
create_interactive_ci_map <- function(ci_raster, field_boundaries = NULL, 
                                     title = "CI Map", 
                                     color_palette = "viridis") {
  
  # Set tmap to interactive mode
  tmap_mode("view")
  
  # Create base CI map
  ci_map <- tm_shape(ci_raster) +
    tm_raster(
      title = "CI Value",
      palette = color_palette,
      style = "cont",
      alpha = 0.8,
      breaks = seq(0, 1, 0.1)
    ) +
    tm_layout(
      title = title,
      title.size = 1.2
    )
  
  # Add field boundaries if provided
  if (!is.null(field_boundaries)) {
    ci_map <- ci_map +
      tm_shape(field_boundaries) +
      tm_polygons(
        alpha = 0.1,
        col = NA,
        border.col = "white",
        border.lwd = 2,
        popup.vars = c("Field" = "field", "Sub Field" = "sub_field"),
        id = "field"
      )
  }
  
  return(ci_map)
}

#' Create CI change detection map
#' @param current_ci Current week CI raster
#' @param previous_ci Previous week CI raster  
#' @param field_boundaries Field boundary polygons (optional)
#' @param title Map title
#' @return tmap object
create_ci_change_map <- function(current_ci, previous_ci, field_boundaries = NULL,
                                title = "CI Change Detection") {
  
  # Calculate change (current - previous)
  ci_change <- current_ci - previous_ci
  
  # Set names for clarity
  names(ci_change) <- "CI_Change"
  
  # Create change map
  change_map <- tm_shape(ci_change) +
    tm_raster(
      title = "CI Change",
      palette = "RdBu", 
      style = "cont",
      alpha = 0.8,
      midpoint = 0,
      breaks = seq(-0.5, 0.5, 0.1)
    ) +
    tm_layout(
      title = title,
      title.size = 1.2
    )
  
  # Add field boundaries if provided
  if (!is.null(field_boundaries)) {
    change_map <- change_map +
      tm_shape(field_boundaries) +
      tm_polygons(
        alpha = 0.1,
        col = NA,
        border.col = "yellow",
        border.lwd = 2,
        popup.vars = c("Field" = "field", "Sub Field" = "sub_field"),
        id = "field"
      )
  }
  
  return(change_map)
}
  
  # Get the second most recent file
  file_info <- file.info(rds_files)
  sorted_files <- rds_files[order(file_info$mtime, decreasing = TRUE)]
  previous_file <- sorted_files[2]
  
  cat("Loading previous CI from:", basename(previous_file), "\n")
  
  ci_data <- readRDS(previous_file)
  
  if (inherits(ci_data, "SpatRaster")) {
    return(ci_data)
  } else if (is.list(ci_data) && "ci_map" %in% names(ci_data)) {
    return(ci_data$ci_map)
  } else {
    warning("Unexpected data format in previous RDS file")
    return(NULL)
  }
}

#' Calculate CI change between current and previous week
#' @param current_ci Current week CI raster
#' @param previous_ci Previous week CI raster  
#' @return SpatRaster showing CI change
calculate_ci_change <- function(current_ci, previous_ci) {
  
  if (is.null(current_ci) || is.null(previous_ci)) {
    warning("Both current and previous CI data required for change calculation")
    return(NULL)
  }
  
  cat("Calculating CI change...\n")
  
  # Ensure both rasters have the same extent and resolution
  if (!compareGeom(current_ci, previous_ci)) {
    cat("Resampling rasters to match...\n")
    previous_ci <- resample(previous_ci, current_ci)
  }
  
  # Calculate change
  ci_change <- current_ci - previous_ci
  names(ci_change) <- "CI_Change"
  
  return(ci_change)
}

# CI Visualization Functions  
# ==========================

#' Create interactive CI map
#' @param ci_raster CI raster data
#' @param field_boundaries Field boundary polygons
#' @param title Map title
#' @param palette Color palette to use
#' @return tmap object
create_interactive_ci_map <- function(ci_raster, field_boundaries = NULL, 
                                     title = "Canopy Index", 
                                     palette = "RdYlGn") {
  
  if (is.null(ci_raster)) {
    cat("No CI data available for mapping\n")
    return(NULL)
  }
  
  # Set tmap to interactive mode
  tmap_mode("view")
  
  # Create base map
  ci_map <- tm_shape(ci_raster) +
    tm_raster(
      title = "CI",
      palette = palette,
      style = "cont",
      alpha = 0.8,
      n = 10
    ) +
    tm_layout(
      title = title,
      title.size = 1.2
    )
  
  # Add field boundaries if available
  if (!is.null(field_boundaries)) {
    ci_map <- ci_map +
      tm_shape(field_boundaries) +
      tm_polygons(
        alpha = 0.2,
        col = "white",
        border.col = "black",
        border.lwd = 1,
        popup.vars = c("Field" = "field", "Sub Field" = "sub_field"),
        id = "field"
      )
  }
  
  return(ci_map)
}

#' Create CI change map
#' @param change_raster CI change raster
#' @param field_boundaries Field boundary polygons
#' @return tmap object
create_ci_change_map <- function(change_raster, field_boundaries = NULL) {
  
  if (is.null(change_raster)) {
    cat("No CI change data available for mapping\n")
    return(NULL)
  }
  
  tmap_mode("view")
  
  change_map <- tm_shape(change_raster) +
    tm_raster(
      title = "CI Change",
      palette = "RdBu",
      style = "cont", 
      alpha = 0.8,
      midpoint = 0,
      n = 11
    ) +
    tm_layout(
      title = "CI Change: This Week vs Last Week",
      title.size = 1.2
    )
  
  # Add field boundaries
  if (!is.null(field_boundaries)) {
    change_map <- change_map +
      tm_shape(field_boundaries) +
      tm_polygons(
        alpha = 0.2,
        col = "white", 
        border.col = "yellow",
        border.lwd = 2,
        popup.vars = c("Field" = "field", "Sub Field" = "sub_field"),
        id = "field"
      )
  }
  
  return(change_map)
}

#' Load field boundaries
#' @param geojson_path Path to field boundaries geojson
#' @return sf object
load_field_boundaries <- function(geojson_path = "../../../pivot.geojson") {
  
  if (!file.exists(geojson_path)) {
    warning("Field boundaries file not found:", geojson_path)
    return(NULL)
  }
  
  cat("Loading field boundaries from:", basename(geojson_path), "\n")
  boundaries <- st_read(geojson_path, quiet = TRUE)
  
  return(boundaries)
}

#' Calculate CI statistics by field
#' @param ci_raster CI raster data
#' @param field_boundaries Field boundary polygons
#' @return data.frame with field statistics
calculate_field_ci_stats <- function(ci_raster, field_boundaries) {
  
  if (is.null(ci_raster) || is.null(field_boundaries)) {
    return(NULL)
  }
  
  cat("Calculating CI statistics by field...\n")
  
  # Extract CI values for each field
  field_stats <- extract(ci_raster, field_boundaries, fun = c(mean, sd, min, max), na.rm = TRUE)
  
  # Combine with field information
  stats_df <- data.frame(
    field = field_boundaries$field,
    sub_field = field_boundaries$sub_field,
    ci_mean = field_stats[,2],  # mean
    ci_sd = field_stats[,3],    # sd  
    ci_min = field_stats[,4],   # min
    ci_max = field_stats[,5]    # max
  )
  
  return(stats_df)
}

cat("CI analysis functions loaded successfully!\n")