# 80_UTILS_COMMON.R
# ============================================================================
# SHARED UTILITY FUNCTIONS FOR ALL CLIENT TYPES (SCRIPT 80)
# 
# Contains helper and infrastructure functions used by both AURA and ANGATA workflows:
#   - Statistical categorization and calculations
#   - Tile operations and data loading
#   - Field statistics extraction
#   - Week/year calculations for consistent date handling
#   - Excel/CSV/RDS export utilities
#   - Yield prediction using ML models (Random Forest with Feature Selection)
#
# Used by: 80_calculate_kpis.R, all client-specific utils files
#
# NOTE: Libraries required by yield prediction (caret, CAST, here) are loaded
#   in the main script 80_calculate_kpis.R, not here. This keeps dependencies
#   centralized in the orchestrator script.
# ============================================================================

# ============================================================================
# CONSTANTS (from 80_calculate_kpis.R)
# ============================================================================

# Four-week trend thresholds
FOUR_WEEK_TREND_STRONG_GROWTH_MIN <- 0.5
FOUR_WEEK_TREND_GROWTH_MIN <- 0.1
FOUR_WEEK_TREND_GROWTH_MAX <- 0.5
FOUR_WEEK_TREND_NO_GROWTH_RANGE <- 0.1
FOUR_WEEK_TREND_DECLINE_MAX <- -0.1
FOUR_WEEK_TREND_DECLINE_MIN <- -0.5
FOUR_WEEK_TREND_STRONG_DECLINE_MAX <- -0.5

# CV Trend thresholds (8-week slope interpretation)
CV_SLOPE_STRONG_IMPROVEMENT_MIN <- -0.03
CV_SLOPE_IMPROVEMENT_MIN <- -0.02
CV_SLOPE_IMPROVEMENT_MAX <- -0.01
CV_SLOPE_HOMOGENOUS_MIN <- -0.01
CV_SLOPE_HOMOGENOUS_MAX <- 0.01
CV_SLOPE_PATCHINESS_MIN <- 0.01
CV_SLOPE_PATCHINESS_MAX <- 0.02
CV_SLOPE_SEVERE_MIN <- 0.02

# Percentile calculations
CI_PERCENTILE_LOW <- 0.10
CI_PERCENTILE_HIGH <- 0.90

# Phase definitions (used by get_phase_by_age)
PHASE_DEFINITIONS <- data.frame(
  phase = c("Germination", "Tillering", "Grand Growth", "Maturation"),
  age_start = c(0, 4, 17, 39),
  age_end = c(6, 16, 39, 200),
  stringsAsFactors = FALSE
)

# ============================================================================
# WEEK/YEAR CALCULATION HELPERS (Consistent across all scripts)
# ============================================================================

#' Calculate week and year for a given lookback offset
#' This function handles ISO 8601 week numbering with proper year wrapping
#' when crossing year boundaries (e.g., week 01/2026 -> week 52/2025)
#'
#' @param current_week ISO week number (1-53)
#' @param current_year ISO week year (from format(..., "%G"))
#' @param offset_weeks Number of weeks to go back (0 = current week, 1 = previous week, etc.)
#'
#' @return List with: week (ISO week number), year (ISO week year)
#'
#' @details
#' This is the authoritative week/year calculation function.
#' Used by:
#'   - load_historical_field_data() - to find RDS/CSV files for 4-week lookback
#'   - Script 80 main - to calculate previous week with year wrapping
#'   - Any other script needing to walk backwards through weeks
#'
#' Example: Week 01/2026, offset=1 -> returns list(week=52, year=2025)
calculate_target_week_and_year <- function(current_week, current_year, offset_weeks = 0) {
  target_week <- current_week - offset_weeks
  target_year <- current_year
  
  # Handle wrapping: when going back from week 1, wrap to week 52 of previous year
  while (target_week < 1) {
    target_week <- target_week + 52
    target_year <- target_year - 1
  }
  
  return(list(week = target_week, year = target_year))
}

# ============================================================================
# TILE-AWARE HELPER FUNCTIONS
# ============================================================================

#' Get tile IDs that intersect with a field geometry
get_tile_ids_for_field <- function(field_geom, tile_grid, field_id = NULL) {
  if (inherits(field_geom, "sf")) {
    field_bbox <- sf::st_bbox(field_geom)
    field_xmin <- field_bbox["xmin"]
    field_xmax <- field_bbox["xmax"]
    field_ymin <- field_bbox["ymin"]
    field_ymax <- field_bbox["ymax"]
  } else if (inherits(field_geom, "SpatVector")) {
    field_bbox <- terra::ext(field_geom)
    field_xmin <- field_bbox$xmin
    field_xmax <- field_bbox$xmax
    field_ymin <- field_bbox$ymin
    field_ymax <- field_bbox$ymax
  } else {
    stop("field_geom must be sf or terra::vect object")
  }
  
  intersecting_tiles <- tile_grid$id[
    !(tile_grid$xmax < field_xmin | 
      tile_grid$xmin > field_xmax | 
      tile_grid$ymax < field_ymin | 
      tile_grid$ymin > field_ymax)
  ]
  
  return(as.numeric(intersecting_tiles))
}

#' Load and merge tiles for a specific field
load_tiles_for_field <- function(field_geom, tile_ids, week_num, year, mosaic_dir) {
  if (length(tile_ids) == 0) {
    return(NULL)
  }
  
  tiles_list <- list()
  for (tile_id in sort(tile_ids)) {
    tile_filename <- sprintf("week_%02d_%d_%02d.tif", week_num, year, tile_id)
    tile_path <- file.path(mosaic_dir, tile_filename)
    
    if (file.exists(tile_path)) {
      tryCatch({
        tile_rast <- terra::rast(tile_path)
        ci_band <- terra::subset(tile_rast, 5)
        tiles_list[[length(tiles_list) + 1]] <- ci_band
      }, error = function(e) {
        message(paste("  Warning: Could not load tile", tile_id, ":", e$message))
      })
    }
  }
  
  if (length(tiles_list) == 0) {
    return(NULL)
  }
  
  if (length(tiles_list) == 1) {
    return(tiles_list[[1]])
  } else {
    tryCatch({
      rsrc <- terra::sprc(tiles_list)
      merged <- terra::mosaic(rsrc, fun = "max")
      return(merged)
    }, error = function(e) {
      message(paste("  Warning: Could not merge tiles:", e$message))
      return(tiles_list[[1]])
    })
  }
}

#' Build tile grid from available tile files
build_tile_grid <- function(mosaic_dir, week_num, year) {
  # Handle grid-size subdirectories (e.g., weekly_tile_max/5x5/)
  detected_grid_size <- NA
  if (dir.exists(mosaic_dir)) {
    subfolders <- list.dirs(mosaic_dir, full.names = FALSE, recursive = FALSE)
    grid_patterns <- grep("^\\d+x\\d+$", subfolders, value = TRUE)
    
    if (length(grid_patterns) > 0) {
      detected_grid_size <- grid_patterns[1]
      mosaic_dir <- file.path(mosaic_dir, detected_grid_size)
      message(paste("  Using grid-size subdirectory:", detected_grid_size))
    }
  }
  
  tile_pattern <- sprintf("week_%02d_%d_([0-9]{2})\\.tif", week_num, year)
  tile_files <- list.files(mosaic_dir, pattern = tile_pattern, full.names = TRUE)
  
  if (length(tile_files) == 0) {
    stop(paste("No tile files found for week", week_num, year, "in", mosaic_dir))
  }
  
  tile_grid <- data.frame(
    id = integer(),
    xmin = numeric(),
    xmax = numeric(),
    ymin = numeric(),
    ymax = numeric(),
    stringsAsFactors = FALSE
  )
  
  for (tile_file in tile_files) {
    tryCatch({
      matches <- regmatches(basename(tile_file), regexpr("_([0-9]{2})\\.tif$", basename(tile_file)))
      if (length(matches) > 0) {
        tile_id <- as.integer(sub("_|\\.tif", "", matches[1]))
        tile_rast <- terra::rast(tile_file)
        tile_ext <- terra::ext(tile_rast)
        tile_grid <- rbind(tile_grid, data.frame(
          id = tile_id,
          xmin = tile_ext$xmin,
          xmax = tile_ext$xmax,
          ymin = tile_ext$ymin,
          ymax = tile_ext$ymax,
          stringsAsFactors = FALSE
        ))
      }
    }, error = function(e) {
      message(paste("  Warning: Could not process tile", basename(tile_file), ":", e$message))
    })
  }
  
  if (nrow(tile_grid) == 0) {
    stop("Could not extract extents from any tile files")
  }
  
  return(list(
    tile_grid = tile_grid,
    mosaic_dir = mosaic_dir,
    grid_size = detected_grid_size
  ))
}

# ============================================================================
# STATISTICAL CATEGORIZATION FUNCTIONS
# ============================================================================

#' Categorize four-week CI trend
categorize_four_week_trend <- function(ci_values_list) {
  if (is.null(ci_values_list) || length(ci_values_list) < 2) {
    return(NA_character_)
  }
  
  ci_values_list <- ci_values_list[!is.na(ci_values_list)]
  if (length(ci_values_list) < 2) {
    return(NA_character_)
  }
  
  weekly_changes <- diff(ci_values_list)
  avg_weekly_change <- mean(weekly_changes, na.rm = TRUE)
  
  if (avg_weekly_change >= FOUR_WEEK_TREND_STRONG_GROWTH_MIN) {
    return("strong growth")
  } else if (avg_weekly_change >= FOUR_WEEK_TREND_GROWTH_MIN && 
             avg_weekly_change < FOUR_WEEK_TREND_GROWTH_MAX) {
    return("growth")
  } else if (abs(avg_weekly_change) <= FOUR_WEEK_TREND_NO_GROWTH_RANGE) {
    return("no growth")
  } else if (avg_weekly_change <= FOUR_WEEK_TREND_DECLINE_MIN && 
             avg_weekly_change > FOUR_WEEK_TREND_STRONG_DECLINE_MAX) {
    return("decline")
  } else if (avg_weekly_change < FOUR_WEEK_TREND_STRONG_DECLINE_MAX) {
    return("strong decline")
  } else {
    return("no growth")
  }
}

#' Round cloud coverage to interval categories
round_cloud_to_intervals <- function(cloud_pct_clear) {
  if (is.na(cloud_pct_clear)) {
    return(NA_character_)
  }
  
  if (cloud_pct_clear < 10) return("0-10%")
  if (cloud_pct_clear < 20) return("10-20%")
  if (cloud_pct_clear < 30) return("20-30%")
  if (cloud_pct_clear < 40) return("30-40%")
  if (cloud_pct_clear < 50) return("40-50%")
  if (cloud_pct_clear < 60) return("50-60%")
  if (cloud_pct_clear < 70) return("60-70%")
  if (cloud_pct_clear < 80) return("70-80%")
  if (cloud_pct_clear < 90) return("80-90%")
  if (cloud_pct_clear < 95) return("90-95%")
  return("95-100%")
}

#' Get CI percentile range (10th to 90th)
get_ci_percentiles <- function(ci_values) {
  if (is.null(ci_values) || length(ci_values) == 0) {
    return(NA_character_)
  }
  
  ci_values <- ci_values[!is.na(ci_values)]
  if (length(ci_values) == 0) {
    return(NA_character_)
  }
  
  p10 <- quantile(ci_values, CI_PERCENTILE_LOW, na.rm = TRUE)
  p90 <- quantile(ci_values, CI_PERCENTILE_HIGH, na.rm = TRUE)
  
  return(sprintf("%.1f-%.1f", p10, p90))
}

#' Calculate short-term CV trend (current week vs previous week)
calculate_cv_trend <- function(cv_current, cv_previous) {
  if (is.na(cv_current) || is.na(cv_previous)) {
    return(NA_real_)
  }
  return(round(cv_current - cv_previous, 4))
}

#' Calculate four-week CI trend
calculate_four_week_trend <- function(mean_ci_values) {
  if (is.null(mean_ci_values) || length(mean_ci_values) == 0) {
    return(NA_real_)
  }
  
  ci_clean <- mean_ci_values[!is.na(mean_ci_values)]
  
  if (length(ci_clean) < 2) {
    return(NA_real_)
  }
  
  trend <- ci_clean[length(ci_clean)] - ci_clean[1]
  return(round(trend, 2))
}

#' Categorize CV slope (8-week regression) into field uniformity interpretation
categorize_cv_slope <- function(slope) {
  if (is.na(slope)) {
    return(NA_character_)
  }
  
  if (slope <= CV_SLOPE_IMPROVEMENT_MIN) {
    return("Excellent uniformity")
  } else if (slope < CV_SLOPE_HOMOGENOUS_MIN) {
    return("Homogenous growth")
  } else if (slope <= CV_SLOPE_HOMOGENOUS_MAX) {
    return("Homogenous growth")
  } else if (slope <= CV_SLOPE_PATCHINESS_MAX) {
    return("Minor patchiness")
  } else {
    return("Severe fragmentation")
  }
}

#' Calculate 8-week CV trend via linear regression slope
calculate_cv_trend_long_term <- function(cv_values) {
  if (is.null(cv_values) || length(cv_values) == 0) {
    return(NA_real_)
  }
  
  cv_clean <- cv_values[!is.na(cv_values)]
  
  if (length(cv_clean) < 2) {
    return(NA_real_)
  }
  
  weeks <- seq_along(cv_clean)
  
  tryCatch({
    lm_fit <- lm(cv_clean ~ weeks)
    slope <- coef(lm_fit)["weeks"]
    return(round(as.numeric(slope), 4))
  }, error = function(e) {
    return(NA_real_)
  })
}

#' Calculate Gap Filling Score KPI (2σ method)
#' @param ci_raster Current week CI raster (single band)
#' @param field_boundaries Field boundaries (sf or SpatVector)
#' @return List with summary data frame and field-level results data frame
calculate_gap_filling_kpi <- function(ci_raster, field_boundaries) {
  
  # Handle both sf and SpatVector inputs
  if (!inherits(field_boundaries, "SpatVector")) {
    field_boundaries_vect <- terra::vect(field_boundaries)
  } else {
    field_boundaries_vect <- field_boundaries
  }

  field_results <- data.frame()

  for (i in seq_len(nrow(field_boundaries))) {
    field_name <- if ("field" %in% names(field_boundaries)) field_boundaries$field[i] else NA_character_
    sub_field_name <- if ("sub_field" %in% names(field_boundaries)) field_boundaries$sub_field[i] else NA_character_
    field_vect <- field_boundaries_vect[i]

    # Extract CI values using helper function
    ci_values <- extract_ci_values(ci_raster, field_vect)
    valid_values <- ci_values[!is.na(ci_values) & is.finite(ci_values)]

    if (length(valid_values) > 1) {
      # Gap score using 2σ below median to detect outliers
      median_ci <- median(valid_values)
      sd_ci <- sd(valid_values)
      outlier_threshold <- median_ci - (2 * sd_ci)
      low_ci_pixels <- sum(valid_values < outlier_threshold)
      total_pixels <- length(valid_values)
      gap_score <- round((low_ci_pixels / total_pixels) * 100, 2)

      # Classify gap severity
      gap_level <- dplyr::case_when(
        gap_score < 10 ~ "Minimal",
        gap_score < 25 ~ "Moderate", 
        TRUE ~ "Significant"
      )

      field_results <- rbind(field_results, data.frame(
        field = field_name,
        sub_field = sub_field_name,
        gap_level = gap_level,
        gap_score = gap_score,
        mean_ci = mean(valid_values),
        outlier_threshold = outlier_threshold
      ))
    } else {
      field_results <- rbind(field_results, data.frame(
        field = field_name,
        sub_field = sub_field_name,
        gap_level = NA_character_,
        gap_score = NA_real_,
        mean_ci = NA_real_,
        outlier_threshold = NA_real_
      ))
    }
  }
  
  # Summarize results
  gap_summary <- field_results %>%
    dplyr::group_by(gap_level) %>%
    dplyr::summarise(field_count = n(), .groups = 'drop') %>%
    dplyr::mutate(percent = round((field_count / sum(field_count)) * 100, 1))

  return(list(summary = gap_summary, field_results = field_results))
}

#' Calculate gap filling scores for all per-field mosaics (wrapper)
#' 
#' This wrapper handles per-field mosaic structure by iterating over
#' individual field files and calling the basic KPI function
#' 
#' @param per_field_files Character vector of paths to per-field mosaic TIFFs
#' @param field_boundaries_sf sf object with field geometries
#' @return data.frame with Field_id and gap_score columns
calculate_gap_scores <- function(per_field_files, field_boundaries_sf) {
  message("\nCalculating gap filling scores (2σ method)...")
  message(paste("  Using per-field mosaics for", length(per_field_files), "fields"))
  
  field_boundaries_by_id <- split(field_boundaries_sf, field_boundaries_sf$field)
  
  process_gap_for_field <- function(field_file) {
    field_id <- basename(dirname(field_file))
    field_bounds <- field_boundaries_by_id[[field_id]]
    
    if (is.null(field_bounds) || nrow(field_bounds) == 0) {
      return(data.frame(Field_id = field_id, gap_score = NA_real_))
    }
    
    tryCatch({
      field_raster <- terra::rast(field_file)
      ci_band_name <- "CI"
      if (!(ci_band_name %in% names(field_raster))) {
        return(data.frame(Field_id = field_id, gap_score = NA_real_))
      }
      field_ci_band <- field_raster[[ci_band_name]]
      names(field_ci_band) <- "CI"
      
      gap_result <- calculate_gap_filling_kpi(field_ci_band, field_bounds)
      
      if (is.null(gap_result) || is.null(gap_result$field_results) || nrow(gap_result$field_results) == 0) {
        return(data.frame(Field_id = field_id, gap_score = NA_real_))
      }
      
      gap_scores <- gap_result$field_results
      gap_scores$Field_id <- gap_scores$field
      gap_scores <- gap_scores[, c("Field_id", "gap_score")]
      
      stats::aggregate(gap_score ~ Field_id, data = gap_scores, FUN = function(x) mean(x, na.rm = TRUE))
    }, error = function(e) {
      message(paste("  WARNING: Gap score failed for field", field_id, ":", e$message))
      data.frame(Field_id = field_id, gap_score = NA_real_)
    })
  }
  
  # Process fields sequentially with progress bar
  message("  Processing gap scores for ", length(per_field_files), " fields...")
  pb <- utils::txtProgressBar(min = 0, max = length(per_field_files), style = 3, width = 50)
  
  results_list <- lapply(seq_along(per_field_files), function(idx) {
    result <- process_gap_for_field(per_field_files[[idx]])
    utils::setTxtProgressBar(pb, idx)
    result
  })
  close(pb)
  
  gap_scores_df <- dplyr::bind_rows(results_list)
  
  if (!is.null(gap_scores_df) && nrow(gap_scores_df) > 0) {
    gap_scores_df <- gap_scores_df %>%
      dplyr::group_by(Field_id) %>%
      dplyr::summarise(gap_score = mean(gap_score, na.rm = TRUE), .groups = "drop")
    
    message(paste("  ✓ Calculated gap scores for", nrow(gap_scores_df), "fields"))
    message(paste("    Gap score range:", round(min(gap_scores_df$gap_score, na.rm=TRUE), 2), "-", 
                  round(max(gap_scores_df$gap_score, na.rm=TRUE), 2), "%"))
  } else {
    message("  WARNING: No gap scores calculated from per-field mosaics")
    gap_scores_df <- NULL
  }
  
  return(gap_scores_df)
}

# ============================================================================
# HELPER FUNCTIONS
# ============================================================================

#' Get crop phase by age in weeks
get_phase_by_age <- function(age_weeks) {
  if (is.na(age_weeks)) return(NA_character_)
  for (i in seq_len(nrow(PHASE_DEFINITIONS))) {
    if (age_weeks >= PHASE_DEFINITIONS$age_start[i] && 
        age_weeks <= PHASE_DEFINITIONS$age_end[i]) {
      return(PHASE_DEFINITIONS$phase[i])
    }
  }
  return("Unknown")
}

#' Get status trigger based on CI values and field age
get_status_trigger <- function(ci_values, ci_change, age_weeks) {
  if (is.na(age_weeks) || length(ci_values) == 0) return(NA_character_)
  
  ci_values <- ci_values[!is.na(ci_values)]
  if (length(ci_values) == 0) return(NA_character_)
  
  pct_above_2 <- sum(ci_values > 2) / length(ci_values) * 100
  pct_at_or_above_2 <- sum(ci_values >= 2) / length(ci_values) * 100
  ci_cv <- if (mean(ci_values, na.rm = TRUE) > 0) sd(ci_values) / mean(ci_values, na.rm = TRUE) else 0
  mean_ci <- mean(ci_values, na.rm = TRUE)
  
  if (age_weeks >= 0 && age_weeks <= 6) {
    if (pct_at_or_above_2 >= 70) {
      return("germination_complete")
    } else if (pct_above_2 > 10) {
      return("germination_started")
    }
  }
  
  if (age_weeks >= 45) {
    return("harvest_ready")
  }
  
  if (age_weeks > 6 && !is.na(ci_change) && ci_change < -1.5 && ci_cv < 0.25) {
    return("stress_detected_whole_field")
  }
  
  if (age_weeks > 6 && !is.na(ci_change) && ci_change > 1.5) {
    return("strong_recovery")
  }
  
  if (age_weeks >= 4 && age_weeks < 39 && !is.na(ci_change) && ci_change > 0.2) {
    return("growth_on_track")
  }
  
  if (age_weeks >= 39 && age_weeks < 45 && mean_ci > 3.5) {
    return("maturation_progressing")
  }
  
  return(NA_character_)
}

#' Extract planting dates from harvesting data
extract_planting_dates <- function(harvesting_data, field_boundaries_sf = NULL) {
  if (is.null(harvesting_data) || nrow(harvesting_data) == 0) {
    message("Warning: No harvesting data available - planting dates will be NA.")
    if (!is.null(field_boundaries_sf)) {
      return(data.frame(
        field_id = field_boundaries_sf$field,
        planting_date = rep(as.Date(NA), nrow(field_boundaries_sf)),
        stringsAsFactors = FALSE
      ))
    }
    return(NULL)
  }
  
  tryCatch({
    planting_dates <- harvesting_data %>%
      arrange(field, desc(season_start)) %>%
      distinct(field, .keep_all = TRUE) %>%
      select(field, season_start) %>%
      rename(field_id = field, planting_date = season_start) %>%
      filter(!is.na(planting_date)) %>%
      as.data.frame()
    
    message(paste("Extracted planting dates for", nrow(planting_dates), "fields from harvest.xlsx"))
    return(planting_dates)
  }, error = function(e) {
    message(paste("Error extracting planting dates:", e$message))
    return(NULL)
  })
}

# ============================================================================
# DATA LOADING HELPERS
# ============================================================================

#' Load and validate harvest data from harvest.xlsx
#'
#' Encapsulates harvest data loading with validation, type coercion, and error handling.
#' Returns a data frame with required columns (field, year, tonnage_ha) or an empty 
#' data frame with proper structure if loading fails.
#'
#' @param data_dir Path to data directory containing harvest.xlsx
#'
#' @return data.frame with columns: field (character), year (numeric), tonnage_ha (numeric)
#'   - On success: data frame with N rows of harvest records
#'   - On failure: empty data frame with correct structure (0 rows, 3 columns)
#'
#' @details
#' **File Location**: Expected at `file.path(data_dir, "harvest.xlsx")`
#'
#' **Validation**:
#'   - Checks file existence before reading
#'   - Validates required columns: field, year, tonnage_ha
#'   - Coerces year and tonnage_ha to numeric
#'   - Logs status messages for debugging
#'
#' **Error Handling**:
#'   - Missing file: Returns empty DF (logs NOTE)
#'   - Missing columns: Returns empty DF (logs WARNING)
#'   - Read errors: Returns empty DF (logs WARNING)
#'   - Always returns valid data frame structure (won't return NULL)
#'
#' **Usage**:
#'   ```r
#'   harvesting_data <- load_harvest_data(setup$data_dir)
#'   # harvesting_data is guaranteed to be a data.frame with 3 columns
#'   # even if harvest.xlsx is unavailable or invalid
#'   ```
load_harvest_data <- function(data_dir) {
  harvesting_data <- NULL
  harvest_file <- file.path(data_dir, "harvest.xlsx")
  
  if (file.exists(harvest_file)) {
    tryCatch({
      harvesting_data <- readxl::read_excel(harvest_file)
      
      # Ensure required columns are present
      required_cols <- c("field", "year", "tonnage_ha")
      if (all(required_cols %in% names(harvesting_data))) {
        # Convert to data frame and ensure column types
        harvesting_data <- as.data.frame(harvesting_data)
        harvesting_data$year <- as.numeric(harvesting_data$year)
        harvesting_data$tonnage_ha <- as.numeric(harvesting_data$tonnage_ha)
        
        message(paste("  ✓ Loaded harvest data:", nrow(harvesting_data), "records from harvest.xlsx"))
        return(harvesting_data)
      } else {
        message(paste("  WARNING: harvest.xlsx missing required columns. Expected: field, year, tonnage_ha"))
        harvesting_data <- NULL
      }
    }, error = function(e) {
      message(paste("  WARNING: Could not read harvest.xlsx:", e$message))
    })
  } else {
    message(paste("  NOTE: harvest.xlsx not found at", harvest_file))
  }
  
  # Fallback: create empty data frame if loading failed
  if (is.null(harvesting_data)) {
    message("  WARNING: No harvest data available. TCH yield prediction will use graceful fallback (NA values)")
    harvesting_data <- data.frame(field = character(), year = numeric(), tonnage_ha = numeric())
  }
  
  return(harvesting_data)
}

# ============================================================================
# FIELD STATISTICS EXTRACTION
# ============================================================================

#' Extract CI statistics for all fields from a single CI raster band
extract_field_statistics_from_ci <- function(ci_band, field_boundaries_sf) {
  extract_result <- terra::extract(ci_band, field_boundaries_sf)
  
  stats_list <- list()
  
  for (field_idx in seq_len(nrow(field_boundaries_sf))) {
    field_pixels <- extract_result[extract_result$ID == field_idx, 2]
    pixels <- as.numeric(field_pixels[!is.na(field_pixels)])
    
    if (length(pixels) == 0) {
      stats_list[[field_idx]] <- data.frame(
        field_idx = field_idx,
        mean_ci = NA_real_,
        cv = NA_real_,
        p10 = NA_real_,
        p90 = NA_real_,
        min_ci = NA_real_,
        max_ci = NA_real_,
        pixel_count_valid = 0,
        pixel_count_total = 0,
        stringsAsFactors = FALSE
      )
      next
    }
    
    mean_val <- mean(pixels, na.rm = TRUE)
    cv_val <- if (mean_val > 0) sd(pixels, na.rm = TRUE) / mean_val else NA_real_
    p10_val <- quantile(pixels, probs = CI_PERCENTILE_LOW, na.rm = TRUE)[[1]]
    p90_val <- quantile(pixels, probs = CI_PERCENTILE_HIGH, na.rm = TRUE)[[1]]
    min_val <- min(pixels, na.rm = TRUE)
    max_val <- max(pixels, na.rm = TRUE)
    
    stats_list[[field_idx]] <- data.frame(
      field_idx = field_idx,
      mean_ci = mean_val,
      cv = cv_val,
      p10 = p10_val,
      p90 = p90_val,
      min_ci = min_val,
      max_ci = max_val,
      pixel_count_valid = length(pixels),
      pixel_count_total = nrow(extract_result[extract_result$ID == field_idx, ]),
      stringsAsFactors = FALSE
    )
  }
  
  return(dplyr::bind_rows(stats_list))
}

# ============================================================================
# EXPORT FUNCTIONS (USED BY ALL CLIENTS)
# ============================================================================

#' Generate summary statistics from field analysis data
generate_field_analysis_summary <- function(field_df) {
  message("Generating summary statistics...")
  
  total_acreage <- sum(field_df$Acreage, na.rm = TRUE)
  
  germination_acreage <- sum(field_df$Acreage[field_df$Phase == "Germination"], na.rm = TRUE)
  tillering_acreage <- sum(field_df$Acreage[field_df$Phase == "Tillering"], na.rm = TRUE)
  grand_growth_acreage <- sum(field_df$Acreage[field_df$Phase == "Grand Growth"], na.rm = TRUE)
  maturation_acreage <- sum(field_df$Acreage[field_df$Phase == "Maturation"], na.rm = TRUE)
  unknown_phase_acreage <- sum(field_df$Acreage[field_df$Phase == "Unknown"], na.rm = TRUE)
  
  harvest_ready_acreage <- sum(field_df$Acreage[field_df$Status_trigger == "harvest_ready"], na.rm = TRUE)
  stress_acreage <- sum(field_df$Acreage[field_df$Status_trigger == "stress_detected_whole_field"], na.rm = TRUE)
  recovery_acreage <- sum(field_df$Acreage[field_df$Status_trigger == "strong_recovery"], na.rm = TRUE)
  growth_on_track_acreage <- sum(field_df$Acreage[field_df$Status_trigger == "growth_on_track"], na.rm = TRUE)
  germination_complete_acreage <- sum(field_df$Acreage[field_df$Status_trigger == "germination_complete"], na.rm = TRUE)
  germination_started_acreage <- sum(field_df$Acreage[field_df$Status_trigger == "germination_started"], na.rm = TRUE)
  no_trigger_acreage <- sum(field_df$Acreage[is.na(field_df$Status_trigger)], na.rm = TRUE)
  
  clear_fields <- sum(field_df$Cloud_category == "Clear view", na.rm = TRUE)
  partial_fields <- sum(field_df$Cloud_category == "Partial coverage", na.rm = TRUE)
  no_image_fields <- sum(field_df$Cloud_category == "No image available", na.rm = TRUE)
  total_fields <- nrow(field_df)
  
  clear_acreage <- sum(field_df$Acreage[field_df$Cloud_category == "Clear view"], na.rm = TRUE)
  partial_acreage <- sum(field_df$Acreage[field_df$Cloud_category == "Partial coverage"], na.rm = TRUE)
  no_image_acreage <- sum(field_df$Acreage[field_df$Cloud_category == "No image available"], na.rm = TRUE)
  
  summary_df <- data.frame(
    Category = c(
      "--- PHASE DISTRIBUTION ---",
      "Germination",
      "Tillering",
      "Grand Growth",
      "Maturation",
      "Unknown phase",
      "--- STATUS TRIGGERS ---",
      "Harvest ready",
      "Stress detected",
      "Strong recovery",
      "Growth on track",
      "Germination complete",
      "Germination started",
      "No trigger",
      "--- CLOUD COVERAGE (FIELDS) ---",
      "Clear view",
      "Partial coverage",
      "No image available",
      "--- CLOUD COVERAGE (ACREAGE) ---",
      "Clear view",
      "Partial coverage",
      "No image available",
      "--- TOTAL ---",
      "Total Acreage"
    ),
    Acreage = c(
      NA,
      round(germination_acreage, 2),
      round(tillering_acreage, 2),
      round(grand_growth_acreage, 2),
      round(maturation_acreage, 2),
      round(unknown_phase_acreage, 2),
      NA,
      round(harvest_ready_acreage, 2),
      round(stress_acreage, 2),
      round(recovery_acreage, 2),
      round(growth_on_track_acreage, 2),
      round(germination_complete_acreage, 2),
      round(germination_started_acreage, 2),
      round(no_trigger_acreage, 2),
      NA,
      paste0(clear_fields, " fields"),
      paste0(partial_fields, " fields"),
      paste0(no_image_fields, " fields"),
      NA,
      round(clear_acreage, 2),
      round(partial_acreage, 2),
      round(no_image_acreage, 2),
      NA,
      round(total_acreage, 2)
    ),
    stringsAsFactors = FALSE
  )
  
  return(summary_df)
}

#' Export field analysis to Excel, CSV, and RDS formats
export_field_analysis_excel <- function(field_df, summary_df, project_dir, current_week, year, reports_dir) {
  message("Exporting per-field analysis to Excel, CSV, and RDS...")
  
  field_df_rounded <- field_df %>%
    mutate(across(where(is.numeric), ~ round(., 2)))
  
  # Handle NULL summary_df
  summary_df_rounded <- if (!is.null(summary_df)) {
    summary_df %>%
      mutate(across(where(is.numeric), ~ round(., 2)))
  } else {
    NULL
  }
  
  output_dir <- file.path(reports_dir)
  if (!dir.exists(output_dir)) {
    dir.create(output_dir, recursive = TRUE)
  }
  
  excel_filename <- paste0(project_dir, "_field_analysis_week", sprintf("%02d_%d", current_week, year), ".xlsx")
  excel_path <- file.path(output_dir, excel_filename)
  excel_path <- normalizePath(excel_path, winslash = "\\", mustWork = FALSE)
  
  # Build sheets list dynamically
  sheets <- list(
    "Field Data" = field_df_rounded
  )
  if (!is.null(summary_df_rounded)) {
    sheets[["Summary"]] <- summary_df_rounded
  }
  
  write_xlsx(sheets, excel_path)
  message(paste("✓ Field analysis Excel exported to:", excel_path))
  
  kpi_data <- list(
    field_analysis = field_df_rounded,
    field_analysis_summary = summary_df_rounded,
    metadata = list(
      current_week = current_week,
      year = year,
      project = project_dir,
      created_at = Sys.time()
    )
  )
  
  rds_filename <- paste0(project_dir, "_field_analysis_week", sprintf("%02d_%d", current_week, year), ".rds")
  rds_path <- file.path(output_dir, rds_filename)
  
  saveRDS(kpi_data, rds_path)
  message(paste("✓ Field analysis RDS exported to:", rds_path))
  
  csv_filename <- paste0(project_dir, "_field_analysis_week", sprintf("%02d_%d", current_week, year), ".csv")
  csv_path <- file.path(output_dir, csv_filename)
  write_csv(field_df_rounded, csv_path)
  message(paste("✓ Field analysis CSV exported to:", csv_path))
  
  return(list(excel = excel_path, rds = rds_path, csv = csv_path))
}

# ============================================================================
# ADDITIONAL CRITICAL FUNCTIONS FROM 80_weekly_stats_utils.R (REQUIRED BY 80_calculate_kpis.R)
# ============================================================================

#' Calculate statistics for all fields from weekly mosaics
calculate_field_statistics <- function(field_boundaries_sf, week_num, year, 
                                        mosaic_dir, report_date = Sys.Date()) {
  
  message(paste("Calculating statistics for all fields - Week", week_num, year))
  
  # Per-field mode: look in per-field subdirectories
  single_file_pattern <- sprintf("week_%02d_%d\\.tif", week_num, year)
  
  # Find all field subdirectories with mosaics for this week
  field_dirs <- list.dirs(mosaic_dir, full.names = FALSE, recursive = FALSE)
  field_dirs <- field_dirs[field_dirs != ""]
  
  per_field_files <- list()
  for (field in field_dirs) {
    field_mosaic_dir <- file.path(mosaic_dir, field)
    files <- list.files(field_mosaic_dir, pattern = single_file_pattern, full.names = TRUE)
    if (length(files) > 0) {
      per_field_files[[field]] <- files[1]  # Take first match for this field
    }
  }
  
  if (length(per_field_files) == 0) {
    stop(paste("No per-field mosaic files found for week", week_num, year, "in", mosaic_dir))
  }
  
  message(paste("  Found", length(per_field_files), "per-field mosaic file(s) for week", week_num))
  results_list <- list()
  
  # Initialize progress bar
  pb <- progress::progress_bar$new(
    format = "  [:bar] :percent | Field :current/:total",
    total = length(per_field_files),
    width = 60
  )
  
  # Process each field's mosaic
  for (field_idx in seq_along(per_field_files)) {
    pb$tick()  # Update progress bar
    field_name <- names(per_field_files)[field_idx]
    field_file <- per_field_files[[field_name]]
    
    tryCatch({
      current_rast <- terra::rast(field_file)
      ci_band <- current_rast[["CI"]]
      
      if (is.null(ci_band) || !inherits(ci_band, "SpatRaster")) {
        message(paste("    [SKIP] Field", field_name, "- CI band not found"))
        next
      }
      
      # Extract CI values for this field
      field_boundary <- field_boundaries_sf[field_boundaries_sf$field == field_name, ]
      
      if (nrow(field_boundary) == 0) {
        message(paste("    [SKIP] Field", field_name, "- not in field boundaries"))
        next
      }
      
      extracted <- terra::extract(ci_band, field_boundary, na.rm = FALSE)
      
      if (nrow(extracted) == 0 || all(is.na(extracted$CI))) {
        message(paste("    [SKIP] Field", field_name, "- no CI values found"))
        next
      }
      
      ci_vals <- extracted$CI[!is.na(extracted$CI)]
      
      if (length(ci_vals) == 0) {
        next
      }
      
      # Calculate statistics
      mean_ci <- mean(ci_vals, na.rm = TRUE)
      ci_std <- sd(ci_vals, na.rm = TRUE)
      cv <- if (mean_ci > 0) ci_std / mean_ci else NA_real_
      range_min <- min(ci_vals, na.rm = TRUE)
      range_max <- max(ci_vals, na.rm = TRUE)
      range_str <- sprintf("%.1f-%.1f", range_min, range_max)
      ci_percentiles_str <- get_ci_percentiles(ci_vals)
      
      num_pixels_total <- length(ci_vals)
      num_pixels_gte_2 <- sum(ci_vals >= 2)
      pct_pixels_gte_2 <- if (num_pixels_total > 0) round((num_pixels_gte_2 / num_pixels_total) * 100, 1) else 0
      
      num_total <- nrow(extracted)
      num_data <- sum(!is.na(extracted$CI))
      pct_clear <- if (num_total > 0) round((num_data / num_total) * 100, 1) else 0
      cloud_cat <- if (num_data == 0) "No image available"
                   else if (pct_clear >= 95) "Clear view"
                   else "Partial coverage"
      
      # Add to results
      results_list[[length(results_list) + 1]] <- data.frame(
        Field_id = field_name,
        Mean_CI = round(mean_ci, 2),
        CV = round(cv * 100, 2),
        CI_range = range_str,
        CI_Percentiles = ci_percentiles_str,
        Pct_pixels_CI_gte_2 = pct_pixels_gte_2,
        Cloud_pct_clear = pct_clear,
        Cloud_category = cloud_cat,
        stringsAsFactors = FALSE
      )
      
    }, error = function(e) {
      message(paste("    [ERROR] Field", field_name, ":", e$message))
    })
  }
  
  if (length(results_list) == 0) {
    stop(paste("No fields processed successfully for week", week_num))
  }
  
  stats_df <- dplyr::bind_rows(results_list)
  message(paste("  ✓ Successfully calculated statistics for", nrow(stats_df), "fields"))
  
  return(stats_df)
}

#' Load or calculate weekly statistics (with RDS caching)
load_or_calculate_weekly_stats <- function(week_num, year, project_dir, field_boundaries_sf, 
                                           mosaic_dir, reports_dir, report_date = Sys.Date()) {
  
  rds_filename <- sprintf("%s_field_stats_week%02d_%d.rds", project_dir, week_num, year)
  rds_path <- file.path(reports_dir, "field_stats", rds_filename)
  
  if (file.exists(rds_path)) {
    message(paste("Loading cached statistics from:", basename(rds_path)))
    return(readRDS(rds_path))
  }
  
  message(paste("Cached RDS not found, calculating statistics from tiles for week", week_num))
  stats_df <- calculate_field_statistics(field_boundaries_sf, week_num, year, mosaic_dir, report_date)
  
  output_dir <- file.path(reports_dir, "field_stats")
  if (!dir.exists(output_dir)) {
    dir.create(output_dir, recursive = TRUE, showWarnings = FALSE)
  }
  
  saveRDS(stats_df, rds_path)
  message(paste("Saved weekly statistics RDS:", basename(rds_path)))
  
  csv_filename <- sprintf("%s_field_stats_week%02d_%d.csv", project_dir, week_num, year)
  csv_path <- file.path(output_dir, csv_filename)
  readr::write_csv(stats_df, csv_path)
  message(paste("Saved weekly statistics CSV:", basename(csv_path)))
  
  return(stats_df)
}

#' Load historical field data from CSV (4-week lookback)
load_historical_field_data <- function(project_dir, current_week, current_year, reports_dir, num_weeks = 4, auto_generate = TRUE, field_boundaries_sf = NULL, daily_vals_dir = NULL) {
  
  historical_data <- list()
  loaded_weeks <- c()
  missing_weeks <- c()
  
  for (lookback in 0:(num_weeks - 1)) {
    target <- calculate_target_week_and_year(current_week, current_year, offset_weeks = lookback)
    target_week <- target$week
    target_year <- target$year
    
    csv_filename <- paste0(project_dir, "_field_analysis_week", sprintf("%02d_%d", target_week, target_year), ".csv")
    csv_path <- file.path(reports_dir, "field_analysis", csv_filename)
    
    if (file.exists(csv_path)) {
      tryCatch({
        data <- readr::read_csv(csv_path, show_col_types = FALSE, 
                                col_types = readr::cols(.default = readr::col_character()))
        historical_data[[lookback + 1]] <- list(
          week = target_week,
          year = target_year,
          data = data
        )
        loaded_weeks <- c(loaded_weeks, paste0("week", sprintf("%02d_%d", target_week, target_year)))
      }, error = function(e) {
        message(paste("  Warning: Could not load week", target_week, "/", target_year, ":", e$message))
        missing_weeks <<- c(missing_weeks, paste0("week", sprintf("%02d_%d", target_week, target_year)))
      })
    } else {
      missing_weeks <- c(missing_weeks, paste0("week", sprintf("%02d_%d", target_week, target_year)))
    }
  }
  
  if (length(historical_data) == 0) {
    message(paste("Error: No historical field data found"))
    return(NULL)
  }
  
  message(paste("✓ Loaded", length(historical_data), "weeks of historical data:", 
                paste(loaded_weeks, collapse = ", ")))
  
  return(historical_data)
}

#' Calculate KPI trends (CI change, CV trends, 4-week and 8-week trends)
calculate_kpi_trends <- function(current_stats, prev_stats = NULL, 
                                  project_dir = NULL, reports_dir = NULL, 
                                  current_week = NULL, year = NULL) {
  
  message("Calculating KPI trends from current and previous week data")
  
  current_stats$Weekly_ci_change <- NA_real_
  current_stats$CV_Trend_Short_Term <- NA_real_
  current_stats$Four_week_trend <- NA_real_
  current_stats$CV_Trend_Long_Term <- NA_real_
  current_stats$nmr_of_weeks_analysed <- 1L
  
  if (is.null(prev_stats) || nrow(prev_stats) == 0) {
    message("  No previous week data available - using defaults")
    return(current_stats)
  }
  
  message(paste("  prev_stats has", nrow(prev_stats), "rows and", ncol(prev_stats), "columns"))
  
  prev_lookup <- setNames(seq_len(nrow(prev_stats)), prev_stats$Field_id)
  prev_field_analysis <- NULL
  
  tryCatch({
    analysis_dir <- file.path(reports_dir, "field_analysis")
    if (dir.exists(analysis_dir)) {
      analysis_files <- list.files(analysis_dir, pattern = "_field_analysis_week.*\\.csv$", full.names = TRUE)
      if (length(analysis_files) > 0) {
        recent_file <- analysis_files[which.max(file.info(analysis_files)$mtime)]
        prev_field_analysis <- readr::read_csv(recent_file, show_col_types = FALSE,
                                               col_types = readr::cols(.default = readr::col_character()),
                                               col_select = c(Field_id, nmr_of_weeks_analysed, Phase))
      }
    }
  }, error = function(e) {
    message(paste("  Note: Could not load previous field_analysis for nmr_weeks tracking:", e$message))
  })
  
  if (!is.null(prev_field_analysis) && nrow(prev_field_analysis) > 0) {
    message(paste("  Using previous field_analysis to track nmr_of_weeks_analysed"))
  }
  
  historical_4weeks <- list()
  historical_8weeks <- list()
  
  if (!is.null(project_dir) && !is.null(reports_dir) && !is.null(current_week)) {
    message("  Loading historical field_stats for 4-week and 8-week trends...")
    
    for (lookback in 1:4) {
      target_week <- current_week - lookback
      target_year <- year
      if (target_week < 1) {
        target_week <- target_week + 52
        target_year <- target_year - 1
      }
      
      rds_filename <- sprintf("%s_field_stats_week%02d_%d.rds", project_dir, target_week, target_year)
      rds_path <- file.path(reports_dir, "field_stats", rds_filename)
      
      if (file.exists(rds_path)) {
        tryCatch({
          stats_data <- readRDS(rds_path)
          historical_4weeks[[length(historical_4weeks) + 1]] <- list(week = target_week, stats = stats_data)
        }, error = function(e) {
          message(paste("    Warning: Could not load week", target_week, ":", e$message))
        })
      }
    }
    
    for (lookback in 1:8) {
      target_week <- current_week - lookback
      target_year <- year
      if (target_week < 1) {
        target_week <- target_week + 52
        target_year <- target_year - 1
      }
      
      rds_filename <- sprintf("%s_field_stats_week%02d_%d.rds", project_dir, target_week, target_year)
      rds_path <- file.path(reports_dir, "field_stats", rds_filename)
      
      if (file.exists(rds_path)) {
        tryCatch({
          stats_data <- readRDS(rds_path)
          historical_8weeks[[length(historical_8weeks) + 1]] <- list(week = target_week, stats = stats_data)
        }, error = function(e) {
          # Silently skip
        })
      }
    }
    
    if (length(historical_4weeks) > 0) {
      message(paste("    Loaded", length(historical_4weeks), "weeks for 4-week trend"))
    }
    if (length(historical_8weeks) > 0) {
      message(paste("    Loaded", length(historical_8weeks), "weeks for 8-week CV trend"))
    }
  }
  
  cv_trends_calculated <- 0
  four_week_trends_calculated <- 0
  cv_long_term_calculated <- 0
  
  for (i in seq_len(nrow(current_stats))) {
    field_id <- current_stats$Field_id[i]
    prev_idx <- prev_lookup[field_id]
    
    if (!is.na(prev_idx) && prev_idx > 0 && prev_idx <= nrow(prev_stats)) {
      prev_row <- prev_stats[prev_idx, , drop = FALSE]
      
      prev_ci <- prev_row$Mean_CI[1]
      if (!is.na(prev_ci) && !is.na(current_stats$Mean_CI[i])) {
        current_stats$Weekly_ci_change[i] <- round(current_stats$Mean_CI[i] - prev_ci, 2)
      }
      
      prev_cv <- prev_row$CV[1]
      if (!is.na(prev_cv) && !is.na(current_stats$CV[i])) {
        current_stats$CV_Trend_Short_Term[i] <- calculate_cv_trend(current_stats$CV[i], prev_cv)
        cv_trends_calculated <- cv_trends_calculated + 1
      }
      
      if (length(historical_4weeks) > 0) {
        ci_values_4week <- numeric()
        for (hist_idx in rev(seq_along(historical_4weeks))) {
          hist_data <- historical_4weeks[[hist_idx]]$stats
          hist_field <- which(hist_data$Field_id == field_id)
          if (length(hist_field) > 0 && !is.na(hist_data$Mean_CI[hist_field[1]])) {
            ci_values_4week <- c(ci_values_4week, hist_data$Mean_CI[hist_field[1]])
          }
        }
        ci_values_4week <- c(ci_values_4week, current_stats$Mean_CI[i])
        
        if (length(ci_values_4week) >= 2) {
          current_stats$Four_week_trend[i] <- calculate_four_week_trend(ci_values_4week)
          four_week_trends_calculated <- four_week_trends_calculated + 1
        }
      }
      
      if (length(historical_8weeks) > 0) {
        cv_values_8week <- numeric()
        for (hist_idx in rev(seq_along(historical_8weeks))) {
          hist_data <- historical_8weeks[[hist_idx]]$stats
          hist_field <- which(hist_data$Field_id == field_id)
          if (length(hist_field) > 0 && !is.na(hist_data$CV[hist_field[1]])) {
            cv_values_8week <- c(cv_values_8week, hist_data$CV[hist_field[1]])
          }
        }
        cv_values_8week <- c(cv_values_8week, current_stats$CV[i])
        
        if (length(cv_values_8week) >= 2) {
          slope <- calculate_cv_trend_long_term(cv_values_8week)
          current_stats$CV_Trend_Long_Term[i] <- slope
          cv_long_term_calculated <- cv_long_term_calculated + 1
        }
      }
      
      if (!is.null(prev_field_analysis) && nrow(prev_field_analysis) > 0) {
        prev_analysis_row <- prev_field_analysis %>% dplyr::filter(Field_id == field_id)
        
        if (nrow(prev_analysis_row) > 0) {
          prev_nmr_weeks_analysis <- prev_analysis_row$nmr_of_weeks_analysed[1]
          if (!is.na(prev_nmr_weeks_analysis)) {
            current_stats$nmr_of_weeks_analysed[i] <- prev_nmr_weeks_analysis + 1L
          } else {
            current_stats$nmr_of_weeks_analysed[i] <- 1L
          }
        }
      }
    }
  }
  
  message(paste("  ✓ Calculated CV_Trend_Short_Term:", cv_trends_calculated, "fields"))
  message(paste("  ✓ Calculated Four_week_trend:", four_week_trends_calculated, "fields"))
  message(paste("  ✓ Calculated CV_Trend_Long_Term:", cv_long_term_calculated, "fields"))
  return(current_stats)
}

# ============================================================================
# INTERNAL HELPER FUNCTIONS (from 80_kpi_utils.R) - DO NOT DELETE
# ============================================================================

# Spatial autocorrelation thresholds for field pattern analysis
MORAN_THRESHOLD_HIGH <- 0.95    # Very strong clustering (problematic patterns)
MORAN_THRESHOLD_MODERATE <- 0.85 # Moderate clustering
MORAN_THRESHOLD_LOW <- 0.7      # Normal field continuity

#' Calculate coefficient of variation for uniformity assessment
calculate_cv <- function(values) {
  values <- values[!is.na(values) & is.finite(values)]
  if (length(values) < 2) return(NA)
  cv <- sd(values) / mean(values)
  return(cv)
}

#' Calculate percentage of field with positive vs negative change
calculate_change_percentages <- function(current_values, previous_values) {
  if (length(current_values) != length(previous_values)) {
    return(list(positive_pct = NA, negative_pct = NA, stable_pct = NA))
  }
  
  change_values <- current_values - previous_values
  valid_changes <- change_values[!is.na(change_values) & is.finite(change_values)]
  
  if (length(valid_changes) < 2) {
    return(list(positive_pct = NA, negative_pct = NA, stable_pct = NA))
  }
  
  positive_pct <- sum(valid_changes > 0) / length(valid_changes) * 100
  negative_pct <- sum(valid_changes < 0) / length(valid_changes) * 100
  stable_pct <- sum(valid_changes == 0) / length(valid_changes) * 100
  
  return(list(
    positive_pct = positive_pct,
    negative_pct = negative_pct,
    stable_pct = stable_pct
  ))
}

#' Calculate spatial autocorrelation (Moran's I) for a field
calculate_spatial_autocorrelation <- function(ci_raster, field_boundary) {
  tryCatch({
    field_raster <- terra::crop(ci_raster, field_boundary)
    field_raster <- terra::mask(field_raster, field_boundary)
    raster_points <- terra::as.points(field_raster, na.rm = TRUE)
    
    if (length(raster_points) < 10) {
      return(list(morans_i = NA, p_value = NA, interpretation = "insufficient_data"))
    }
    
    points_sf <- sf::st_as_sf(raster_points)
    coords <- sf::st_coordinates(points_sf)
    k_neighbors <- min(8, max(4, floor(nrow(coords) / 10)))
    
    knn_nb <- spdep::knearneigh(coords, k = k_neighbors)
    knn_listw <- spdep::nb2listw(spdep::knn2nb(knn_nb), style = "W", zero.policy = TRUE)
    
    ci_values <- points_sf[[1]]
    moran_result <- spdep::moran.test(ci_values, knn_listw, zero.policy = TRUE)
    
    morans_i <- moran_result$estimate[1]
    p_value <- moran_result$p.value
    
    interpretation <- if (is.na(morans_i)) {
      "insufficient_data"
    } else if (p_value > 0.05) {
      "random"
    } else if (morans_i > MORAN_THRESHOLD_HIGH) {
      "very_strong_clustering"
    } else if (morans_i > MORAN_THRESHOLD_MODERATE) {
      "strong_clustering"
    } else if (morans_i > MORAN_THRESHOLD_LOW) {
      "normal_continuity"
    } else if (morans_i > 0.3) {
      "weak_clustering"
    } else if (morans_i < -0.3) {
      "dispersed"
    } else {
      "low_autocorrelation"
    }
    
    return(list(morans_i = morans_i, p_value = p_value, interpretation = interpretation))
  }, error = function(e) {
    warning(paste("Error calculating spatial autocorrelation:", e$message))
    return(list(morans_i = NA, p_value = NA, interpretation = "error"))
  })
}

#' Extract CI band from multi-band raster
extract_ci_values <- function(ci_raster, field_vect) {
  extracted <- terra::extract(ci_raster, field_vect, fun = NULL)
  
  if ("CI" %in% names(extracted)) {
    return(extracted[, "CI"])
  } else if (ncol(extracted) > 1) {
    return(extracted[, ncol(extracted)])
  } else {
    return(extracted[, 1])
  }
}

#' Calculate current and previous week numbers using ISO 8601
calculate_week_numbers <- function(report_date = Sys.Date()) {
  current_week <- as.numeric(format(report_date, "%V"))
  current_year <- as.numeric(format(report_date, "%G"))
  
  previous_week <- current_week - 1
  previous_year <- current_year
  
  if (previous_week < 1) {
    previous_week <- 52
    previous_year <- current_year - 1
  }
  
  return(list(
    current_week = current_week,
    previous_week = previous_week,
    current_year = current_year,
    previous_year = previous_year
  ))
}

#' Load field CI raster (handles single-file and per-field architectures)
load_field_ci_raster <- function(ci_raster_or_obj, field_name, field_vect = NULL) {
  is_per_field <- !is.null(attr(ci_raster_or_obj, "is_per_field")) && attr(ci_raster_or_obj, "is_per_field")
  
  if (is_per_field) {
    per_field_dir <- attr(ci_raster_or_obj, "per_field_dir")
    week_file <- attr(ci_raster_or_obj, "week_file")
    field_mosaic_path <- file.path(per_field_dir, field_name, week_file)
    
    if (file.exists(field_mosaic_path)) {
      tryCatch({
        field_mosaic <- terra::rast(field_mosaic_path)
        if (terra::nlyr(field_mosaic) >= 5) {
          return(field_mosaic[[5]])
        } else {
          return(field_mosaic[[1]])
        }
      }, error = function(e) {
        return(NULL)
      })
    } else {
      return(NULL)
    }
  } else {
    if (!is.null(field_vect)) {
      return(terra::crop(ci_raster_or_obj, field_vect, mask = TRUE))
    } else {
      return(ci_raster_or_obj)
    }
  }
}

#' Load weekly CI mosaic (single-file or per-field)
load_weekly_ci_mosaic <- function(week_num, year, mosaic_dir) {
  week_file <- sprintf("week_%02d_%d.tif", week_num, year)
  week_path <- file.path(mosaic_dir, week_file)
  
  if (file.exists(week_path)) {
    tryCatch({
      mosaic_raster <- terra::rast(week_path)
      ci_raster <- mosaic_raster[[5]]
      names(ci_raster) <- "CI"
      return(ci_raster)
    }, error = function(e) {
      return(NULL)
    })
  }
  
  if (dir.exists(mosaic_dir)) {
    field_dirs <- list.dirs(mosaic_dir, full.names = FALSE, recursive = FALSE)
    field_dirs <- field_dirs[field_dirs != ""]
    
    found_any <- FALSE
    for (field in field_dirs) {
      field_mosaic_path <- file.path(mosaic_dir, field, week_file)
      if (file.exists(field_mosaic_path)) {
        found_any <- TRUE
        break
      }
    }
    
    if (found_any) {
      dummy_raster <- terra::rast(nrow=1, ncol=1, vals=NA)
      attr(dummy_raster, "per_field_dir") <- mosaic_dir
      attr(dummy_raster, "week_file") <- week_file
      attr(dummy_raster, "is_per_field") <- TRUE
      return(dummy_raster)
    }
  }
  
  return(NULL)
}

#' Prepare predictions with consistent naming
prepare_predictions <- function(predictions, newdata) {
  return(predictions %>%
    as.data.frame() %>%
    dplyr::rename(predicted_Tcha = ".") %>%
    dplyr::mutate(
      sub_field = newdata$sub_field,
      field = newdata$field,
      Age_days = newdata$DOY,
      total_CI = round(newdata$cumulative_CI, 0),
      predicted_Tcha = round(predicted_Tcha, 0),
      season = newdata$season
    ) %>%
    dplyr::select(field, sub_field, Age_days, predicted_Tcha, season) %>%
    dplyr::left_join(., newdata, by = c("field", "sub_field", "season"))
  )
}

# ============================================================================
# YIELD PREDICTION KPI (SHARED ML-BASED MODEL FOR ALL CLIENT TYPES)
# ============================================================================

#' Helper function for graceful fallback when training data unavailable
#'
#' @param field_boundaries Field boundaries (sf or SpatVector)
#' @return List with summary and field_results (both with NA values)
create_fallback_result <- function(field_boundaries) {
  # Convert to SpatVector if needed (for terra::project)
  if (!inherits(field_boundaries, "SpatVector")) {
    field_boundaries <- terra::vect(field_boundaries)
  }
  field_boundaries_projected <- terra::project(field_boundaries, "EPSG:6933") # Equal Earth projection
  field_areas <- terra::expanse(field_boundaries_projected) / 10000  # Convert m² to hectares
  total_area <- sum(field_areas)

  summary_result <- data.frame(
    field_groups = c("Top 25%", "Average", "Lowest 25%", "Total area forecasted"),
    count = c(0, 0, 0, nrow(field_boundaries)),
    value = c(NA_real_, NA_real_, NA_real_, round(total_area, 1)),
    stringsAsFactors = FALSE
  )
  
  field_results <- data.frame(
    field = character(0),
    sub_field = character(0),
    Age_days = numeric(0),
    yield_forecast_t_ha = numeric(0),
    season = numeric(0),
    stringsAsFactors = FALSE
  )

  return(list(summary = summary_result, field_results = field_results))
}

#' Calculate yield prediction KPI using Random Forest with Feature Selection
#'
#' Trains a Random Forest model on historical harvest data with cumulative CI,
#' days of year (DOY), and CI-per-day as predictors. Uses CAST::ffs() for
#' Forward Feature Selection. Predicts yields for mature fields (DOY >= 240).
#' 
#' @param field_boundaries Field boundaries (sf or SpatVector)
#' @param harvesting_data Data frame with harvest data including tonnage_ha column
#' @param cumulative_CI_vals_dir Directory containing "All_pivots_Cumulative_CI_quadrant_year_v2.rds"
#'
#' @return List with:
#'   - summary: Data frame with field_groups, count, value (quartiles and total area)
#'   - field_results: Data frame with field-level yield forecasts (yield_forecast_t_ha)
#'
#' @details
#' **Training Data Requirements:**
#'   - cumulative_CI_vals_dir must contain "All_pivots_Cumulative_CI_quadrant_year_v2.rds"
#'   - harvesting_data must have tonnage_ha column with numeric yield values
#'   - Training stops gracefully if either is missing (returns NA values, not fake numbers)
#'
#' **Model Specifications:**
#'   - Algorithm: Random Forest (caret + CAST)
#'   - Feature Selection: Forward Feature Selection (CAST::ffs)
#'   - Cross-validation: 5-fold CV
#'   - Predictors: cumulative_CI, DOY, CI_per_day
#'   - Mature field threshold: DOY >= 240 (8 months)
#'   - Output: Field-level yield forecasts grouped by quartile
#'
#' **Error Handling:**
#'   - Missing tonnage_ha: Returns graceful fallback with NA (not zero) values
#'   - No training data: Logs WARNING, returns empty field_results
#'   - RDS file missing: Returns graceful fallback
#'   - Prediction errors: Wrapped in tryCatch, returns fallback on failure
calculate_yield_prediction_kpi <- function(field_boundaries, harvesting_data, cumulative_CI_vals_dir) {
  safe_log("Calculating yield prediction KPI using Random Forest with Feature Selection")

  tryCatch({
    # Check if tonnage_ha column is present and has valid data
    if (is.null(harvesting_data) || 
        !("tonnage_ha" %in% names(harvesting_data)) || 
        all(is.na(harvesting_data$tonnage_ha))) {
      safe_log("No harvest data available: lacking tonnage_ha column or all values are NA", "WARNING")
      return(create_fallback_result(field_boundaries))
    }

    # Check if CI quadrant RDS file exists
    ci_quadrant_path <- file.path(cumulative_CI_vals_dir, "All_pivots_Cumulative_CI_quadrant_year_v2.rds")
    if (!file.exists(ci_quadrant_path)) {
      safe_log(paste("CI quadrant file not found at:", ci_quadrant_path), "WARNING")
      return(create_fallback_result(field_boundaries))
    }

    # Load CI quadrant data and fill missing field/sub_field values
    CI_quadrant <- readRDS(ci_quadrant_path) %>%
      dplyr::group_by(model) %>%
      tidyr::fill(field, sub_field, .direction = "downup") %>%
      dplyr::ungroup()

    # Rename year column to season for consistency
    harvesting_data_renamed <- harvesting_data %>% dplyr::rename(season = year)

    # Join CI and yield data
    CI_and_yield <- dplyr::left_join(CI_quadrant, harvesting_data_renamed, 
                                      by = c("field", "sub_field", "season")) %>%
      dplyr::group_by(sub_field, season) %>%
      dplyr::slice(which.max(DOY)) %>%
      dplyr::select(field, sub_field, tonnage_ha, cumulative_CI, DOY, season, sub_area) %>%
      dplyr::mutate(CI_per_day = cumulative_CI / DOY)

    # Define predictors and response variables
    predictors <- c("cumulative_CI", "DOY", "CI_per_day")
    response <- "tonnage_ha"

    # Prepare training dataset (fields with harvest data)
    CI_and_yield_test <- CI_and_yield %>%
      as.data.frame() %>%
      dplyr::filter(!is.na(tonnage_ha))

    # Check if we have minimum training data
    if (nrow(CI_and_yield_test) == 0) {
      safe_log("No training data available: no fields with tonnage_ha observations", "WARNING")
      return(create_fallback_result(field_boundaries))
    }

    # Pre-clean training data: remove rows with any NAs in predictors or response
    # This is required because CAST::ffs doesn't support na.rm parameter
    CI_and_yield_test <- CI_and_yield_test %>%
      dplyr::filter(!dplyr::if_any(dplyr::all_of(c(predictors, response)), is.na))

    if (nrow(CI_and_yield_test) == 0) {
      safe_log("No complete training data after removing NAs in predictors/response", "WARNING")
      return(create_fallback_result(field_boundaries))
    }

    # Prepare prediction dataset (fields without harvest data, mature fields only)
    prediction_yields <- CI_and_yield %>%
      as.data.frame() %>%
      dplyr::filter(is.na(tonnage_ha) & DOY >= 240)  # Mature fields only

    # Configure model training parameters
    ctrl <- caret::trainControl(
      method = "cv",
      savePredictions = TRUE,
      allowParallel = TRUE,
      number = 5,
      verboseIter = FALSE
    )

    # Train the model with forward feature selection
    set.seed(202) # For reproducibility
    safe_log("Training Random Forest model with Forward Feature Selection...")
    model_ffs_rf <- CAST::ffs(
      CI_and_yield_test[, predictors],
      CI_and_yield_test[, response],
      method = "rf",
      trControl = ctrl,
      importance = TRUE,
      withinSE = TRUE,
      tuneLength = 5
    )

    # Predict yields on validation data (same as training data for RMSE calculation)
    pred_ffs_rf <- prepare_predictions(
      stats::predict(model_ffs_rf, newdata = CI_and_yield_test), 
      CI_and_yield_test
    )

    # Extract cross-validated RMSE from the model object (more honest than in-sample RMSE)
    # The CAST::ffs model stores CV results in $results data frame
    # We extract the RMSE from the best model (lowest RMSE across folds)
    if (!is.null(model_ffs_rf$results) && "RMSE" %in% names(model_ffs_rf$results)) {
      # Get minimum RMSE from cross-validation results (best model from feature selection)
      rmse_value <- min(model_ffs_rf$results$RMSE, na.rm = TRUE)
      safe_log(paste("Yield prediction RMSE (cross-validated):", round(rmse_value, 2), "t/ha"))
    } else {
      # Fallback: compute in-sample RMSE if CV results unavailable, but label it clearly
      rmse_value <- sqrt(mean((pred_ffs_rf$predicted_Tcha - CI_and_yield_test$tonnage_ha)^2, na.rm = TRUE))
      safe_log(paste("Yield prediction RMSE (in-sample/training):", round(rmse_value, 2), "t/ha"))
    }

    # Predict yields for current season (mature fields >= 240 days)
    if (nrow(prediction_yields) > 0) {
      pred_rf_current_season <- prepare_predictions(
        stats::predict(model_ffs_rf, newdata = prediction_yields), 
        prediction_yields
      ) %>%
        dplyr::filter(Age_days >= 240) %>%
        dplyr::select(c("field", "Age_days", "predicted_Tcha", "season"))
    } else {
      pred_rf_current_season <- data.frame()
    }

    # Calculate summary statistics for KPI
    if (nrow(pred_rf_current_season) > 0) {
      safe_log(paste("Number of fields with yield predictions:", nrow(pred_rf_current_season)))
      safe_log(paste("Predicted yield range:", 
                     round(min(pred_rf_current_season$predicted_Tcha, na.rm = TRUE), 1),
                     "-", 
                     round(max(pred_rf_current_season$predicted_Tcha, na.rm = TRUE), 1),
                     "t/ha"))

      # Calculate quartiles for grouping
      yield_quartiles <- quantile(pred_rf_current_season$predicted_Tcha, 
                                  probs = c(0.25, 0.5, 0.75), na.rm = TRUE)

      # Count fields in each quartile
      top_25_count <- sum(pred_rf_current_season$predicted_Tcha >= yield_quartiles[3], na.rm = TRUE)
      average_count <- sum(pred_rf_current_season$predicted_Tcha >= yield_quartiles[1] & 
                           pred_rf_current_season$predicted_Tcha < yield_quartiles[3], na.rm = TRUE)
      lowest_25_count <- sum(pred_rf_current_season$predicted_Tcha < yield_quartiles[1], na.rm = TRUE)

      # Calculate total area
      if (!inherits(field_boundaries, "SpatVector")) {
        field_boundaries_vect <- terra::vect(field_boundaries)
      } else {
        field_boundaries_vect <- field_boundaries
      }

      # Handle both sf and SpatVector inputs for area calculation
      if (inherits(field_boundaries, "sf")) {
        field_boundaries_projected <- sf::st_transform(field_boundaries, "EPSG:6933")
        field_areas <- sf::st_area(field_boundaries_projected) / 10000  # m² to hectares
      } else {
        field_boundaries_projected <- terra::project(field_boundaries_vect, "EPSG:6933")
        field_areas <- terra::expanse(field_boundaries_projected) / 10000
      }
      total_area <- sum(as.numeric(field_areas))

      safe_log(paste("Total area:", round(total_area, 1), "hectares"))

      # Build summary result
      result <- data.frame(
        field_groups = c("Top 25%", "Average", "Lowest 25%", "Total area forecasted"),
        count = c(top_25_count, average_count, lowest_25_count, nrow(field_boundaries)),
        value = c(round(yield_quartiles[3], 1), round(yield_quartiles[2], 1), 
                 round(yield_quartiles[1], 1), round(total_area, 1)),
        stringsAsFactors = FALSE
      )

      # Prepare field-level results
      field_level_results <- pred_rf_current_season %>%
        dplyr::select(field, Age_days, predicted_Tcha, season) %>%
        dplyr::rename(yield_forecast_t_ha = predicted_Tcha)

      safe_log("✓ Yield prediction complete")
      return(list(summary = result, field_results = field_level_results))
    } else {
      safe_log("No fields meet maturity threshold (DOY >= 240) for prediction", "WARNING")
      return(list(summary = create_fallback_result(field_boundaries)$summary, 
                 field_results = data.frame()))
    }

  }, error = function(e) {
    safe_log(paste("Error in yield prediction:", e$message), "ERROR")
    return(create_fallback_result(field_boundaries))
  })
}