SmartCane/r_app/10_create_master_grid_and_split_tiffs.R

#' Combined: Create master grid and split TIFFs into tiles
#' ====================================================================
#'
#' Purpose:
#' 1. Check all daily TIFFs for matching extents
#' 2. Create master 5×5 grid covering all TIFFs
#' 3. Split each daily TIFF into 25 tiles using the master grid
#' 4. Save tiles in date-specific folders: daily_tiles/[DATE]/[DATE]_[TILE_ID].tif

library(terra)
library(sf)

# ============================================================================
# CONFIGURATION
# ============================================================================

PROJECT <- "angata"
TIFF_FOLDER <- file.path("laravel_app", "storage", "app", PROJECT, "merged_tif_8b")

# GRID SIZE CONFIGURATION - Change this to use different grid sizes
# Options: 5x5 (25 tiles), 10x10 (100 tiles), etc.
# This determines the subfolder: daily_tiles_split/5x5/, daily_tiles_split/10x10/, etc.
GRID_NROWS <- 5
GRID_NCOLS <- 5

# Construct grid-specific subfolder path
GRID_SIZE_LABEL <- paste0(GRID_NCOLS, "x", GRID_NROWS)
OUTPUT_FOLDER <- file.path("laravel_app", "storage", "app", PROJECT, "daily_tiles_split", GRID_SIZE_LABEL)

# Load field boundaries for overlap checking
GEOJSON_PATH <- file.path("laravel_app", "storage", "app", PROJECT, "Data", "pivot.geojson")

cat("Combined: Create Master Grid (", GRID_SIZE_LABEL, ") and Split TIFFs into Tiles\n", sep = "")
cat("Grid subfolder: daily_tiles_split/", GRID_SIZE_LABEL, "/\n", sep = "")

# ============================================================================
# PART 1: CHECK TIFF EXTENTS AND CREATE MASTER GRID
# ============================================================================

cat("\n[PART 1] Creating Master Grid\n")

# Load field boundaries for overlap checking
cat("\n[1] Loading field boundaries from GeoJSON...\n")

if (!file.exists(GEOJSON_PATH)) {
  stop("GeoJSON file not found at: ", GEOJSON_PATH, "\n",
       "Please ensure ", PROJECT, " has a pivot.geojson file.")
}

field_boundaries_sf <- st_read(GEOJSON_PATH, quiet = TRUE)
field_boundaries_vect <- terra::vect(GEOJSON_PATH)

cat("  ✓ Loaded ", nrow(field_boundaries_sf), " field(s)\n", sep = "")

# Try to find a name column (could be 'name', 'field', 'field_name', etc.)
field_names <- NA
if ("name" %in% names(field_boundaries_sf)) {
  field_names <- field_boundaries_sf$name
} else if ("field" %in% names(field_boundaries_sf)) {
  field_names <- field_boundaries_sf$field
} else if ("field_name" %in% names(field_boundaries_sf)) {
  field_names <- field_boundaries_sf$field_name
} else {
  field_names <- 1:nrow(field_boundaries_sf)  # Fall back to indices
}

cat("  Fields: ", paste(field_names, collapse = ", "), "\n", sep = "")

# Helper function: Check if a tile overlaps with any field (simple bbox overlap)
tile_overlaps_fields <- function(tile_extent, field_geoms) {
  tryCatch({
    # Simple bounding box overlap test - no complex geometry operations
    # Two boxes overlap if: NOT (box1.xmax < box2.xmin OR box1.xmin > box2.xmax OR 
    #                             box1.ymax < box2.ymin OR box1.ymin > box2.ymax)
    
    # For each field geometry, check if it overlaps with tile bbox
    for (i in seq_len(length(field_geoms))) {
      # Skip empty geometries
      if (st_is_empty(field_geoms[i])) {
        next
      }
      
      # Get field bbox
      field_bbox <- st_bbox(field_geoms[i])
      
      # Check bbox overlap (simple coordinate comparison)
      x_overlap <- !(tile_extent$xmax < field_bbox$xmin || tile_extent$xmin > field_bbox$xmax)
      y_overlap <- !(tile_extent$ymax < field_bbox$ymin || tile_extent$ymin > field_bbox$ymax)
      
      if (x_overlap && y_overlap) {
        return(TRUE)  # Found overlap!
      }
    }
    
    return(FALSE)  # No overlap found
    
  }, error = function(e) {
    cat("    ⚠️ Error checking overlap: ", e$message, "\n", sep = "")
    return(TRUE)  # Default to including tile if there's an error
  })
}

cat("\n[2] Checking TIFF extents...\n")

tiff_files <- list.files(TIFF_FOLDER, pattern = "\\.tif$", full.names = FALSE)
tiff_files <- sort(tiff_files)

if (length(tiff_files) == 0) {
  stop("No TIFF files found in ", TIFF_FOLDER)
}

cat("  Found ", length(tiff_files), " TIFF file(s)\n", sep = "")
cat("  Checking extents... (this may take a while)\n")

# Load all extents - ONE TIME, upfront
extents <- list()
for (i in seq_along(tiff_files)) {
  tiff_path <- file.path(TIFF_FOLDER, tiff_files[i])
  raster <- terra::rast(tiff_path)
  ext <- terra::ext(raster)
  extents[[i]] <- ext
  
  # Progress indicator every 50 files
  if (i %% 50 == 0) {
    cat("    Checked ", i, "/", length(tiff_files), " files\n", sep = "")
  }
}

cat("  ✓ All extents loaded\n")

# Check if all extents match
cat("\n[3] Comparing extents...\n")

tolerance <- 1e-8
all_match <- TRUE
first_ext <- extents[[1]]

for (i in 2:length(extents)) {
  curr_ext <- extents[[i]]
  match <- (
    abs(curr_ext$xmin - first_ext$xmin) < tolerance &&
    abs(curr_ext$xmax - first_ext$xmax) < tolerance &&
    abs(curr_ext$ymin - first_ext$ymin) < tolerance &&
    abs(curr_ext$ymax - first_ext$ymax) < tolerance
  )
  if (!match) {
    all_match <- FALSE
    cat("  ✗ Extent mismatch: ", tiff_files[1], " vs ", tiff_files[i], "\n", sep = "")
    cat("    File 1: X [", round(first_ext$xmin, 6), ", ", round(first_ext$xmax, 6), "] ", 
        "Y [", round(first_ext$ymin, 6), ", ", round(first_ext$ymax, 6), "]\n", sep = "")
    cat("    File ", i, ": X [", round(curr_ext$xmin, 6), ", ", round(curr_ext$xmax, 6), "] ",
        "Y [", round(curr_ext$ymin, 6), ", ", round(curr_ext$ymax, 6), "]\n", sep = "")
  }
}

if (all_match) {
  cat("  ✓ All TIFF extents MATCH perfectly!\n")
} else {
  cat("  ⚠️ Extents differ - creating master extent covering all\n")
}

# Create master extent
cat("\n[4] Creating master extent...\n")

master_xmin <- min(sapply(extents, function(e) e$xmin))
master_xmax <- max(sapply(extents, function(e) e$xmax))
master_ymin <- min(sapply(extents, function(e) e$ymin))
master_ymax <- max(sapply(extents, function(e) e$ymax))

x_range_m <- (master_xmax - master_xmin) * 111320
y_range_m <- (master_ymax - master_ymin) * 111320

cat("  Master extent: X [", round(master_xmin, 6), ", ", round(master_xmax, 6), "] ", 
    "Y [", round(master_ymin, 6), ", ", round(master_ymax, 6), "]\n", sep = "")
cat("  Coverage: ", round(x_range_m / 1000, 1), "km × ", round(y_range_m / 1000, 1), "km\n", sep = "")

# Auto-determine grid size based on ROI dimensions
if (x_range_m < 10000 && y_range_m < 10000) {
  cat("\n  ⚠️ ROI is small (< 10×10 km). Using single tile (1×1 grid) - no splitting needed!\n")
  GRID_NROWS <- 1
  GRID_NCOLS <- 1
} else {
  cat("\n  ROI size allows tiling. Using 5×5 grid (25 tiles per date).\n")
  GRID_NROWS <- 5
  GRID_NCOLS <- 5
}

N_TILES <- GRID_NROWS * GRID_NCOLS

# Check if master grid already exists
cat("\n[5] Checking if master grid exists...\n")

master_grid_file <- file.path(OUTPUT_FOLDER, paste0("master_grid_", GRID_SIZE_LABEL, ".geojson"))

if (file.exists(master_grid_file)) {
  cat("  ✓ Master grid exists! Loading existing grid...\n")
  master_grid_sf <- st_read(master_grid_file, quiet = TRUE)
  master_grid_vect <- terra::vect(master_grid_file)
  cat("  ✓ Loaded grid with ", nrow(master_grid_sf), " tiles\n", sep = "")
} else {
  cat("  Grid does not exist. Creating new master grid...\n")
  
  # Create 5×5 grid
  cat("\n[6] Creating ", GRID_NCOLS, "×", GRID_NROWS, " master grid...\n", sep = "")
  
  master_bbox <- st_bbox(c(
    xmin = master_xmin,
    xmax = master_xmax,
    ymin = master_ymin,
    ymax = master_ymax
  ), crs = 4326)
  
  bbox_sf <- st_as_sfc(master_bbox)
  
  master_grid <- st_make_grid(
    bbox_sf,
    n = c(GRID_NCOLS, GRID_NROWS),
    what = "polygons"
  )
  
  master_grid_sf <- st_sf(
    tile_id = sprintf("%02d", 1:length(master_grid)),
    geometry = master_grid
  )
  
  cat("  ✓ Created grid with ", length(master_grid), " cells\n", sep = "")
  
  # Convert to SpatVector for use in makeTiles
  master_grid_vect <- terra::vect(master_grid_sf)
  
  # Save master grid
  if (!dir.exists(OUTPUT_FOLDER)) {
    dir.create(OUTPUT_FOLDER, recursive = TRUE, showWarnings = FALSE)
  }
  st_write(master_grid_sf, master_grid_file, delete_dsn = TRUE, quiet = TRUE)
  cat("  ✓ Master grid saved to: master_grid_", GRID_SIZE_LABEL, ".geojson\n", sep = "")
}

# ============================================================================
# PART 2: CREATE FILTERED GRID (ONLY OVERLAPPING TILES)
# ============================================================================

cat("\n[PART 2] Creating Filtered Grid (only overlapping tiles)\n")

cat("\n[7] Filtering master grid to only overlapping tiles...\n")

# Check which tiles overlap with any field
overlapping_tile_indices <- c()
for (tile_idx in 1:nrow(master_grid_sf)) {
  tile_geom <- master_grid_sf[tile_idx, ]
  
  # Check overlap with any field
  if (tile_overlaps_fields(st_bbox(tile_geom$geometry), field_boundaries_sf$geometry)) {
    overlapping_tile_indices <- c(overlapping_tile_indices, tile_idx)
  }
}

cat("  Found ", length(overlapping_tile_indices), " overlapping tiles out of ", N_TILES, "\n", sep = "")
cat("  Reduction: ", N_TILES - length(overlapping_tile_indices), " empty tiles will NOT be created\n", sep = "")

# Create filtered grid with only overlapping tiles
filtered_grid_sf <- master_grid_sf[overlapping_tile_indices, ]
filtered_grid_sf$tile_id <- sprintf("%02d", overlapping_tile_indices)

# Convert to SpatVector for makeTiles
filtered_grid_vect <- terra::vect(filtered_grid_sf)

cat("  ✓ Filtered grid ready: ", nrow(filtered_grid_sf), " tiles to create per date\n", sep = "")

# ============================================================================
# PART 3: SPLIT EACH TIFF INTO TILES (INDEPENDENT, PER-DATE, RESUMABLE)
# ============================================================================

cat("\n[PART 3] Tiling Individual Dates (Per-Date Processing)\n")
cat("\n[8] Processing each date independently...\n")
cat("    (This process is RESUMABLE - you can stop and restart anytime)\n\n")

total_tiles_created <- 0
dates_skipped <- 0
dates_processed <- 0

for (file_idx in seq_along(tiff_files)) {
  tiff_file <- tiff_files[file_idx]
  date_str <- gsub("\\.tif$", "", tiff_file)
  
  # Create date-specific output folder
  date_output_folder <- file.path(OUTPUT_FOLDER, date_str)
  
  # CHECK: Skip if date already processed (RESUME-SAFE)
  if (dir.exists(date_output_folder)) {
    existing_tiles <- list.files(date_output_folder, pattern = "\\.tif$")
    existing_tiles <- existing_tiles[!grepl("master_grid", existing_tiles)]
    
    if (length(existing_tiles) > 0) {
      cat("[", file_idx, "/", length(tiff_files), "] SKIP: ", date_str, 
          " (", length(existing_tiles), " tiles already exist)\n", sep = "")
      dates_skipped <- dates_skipped + 1
      next  # Skip this date
    }
  }
  
  cat("[", file_idx, "/", length(tiff_files), "] Processing: ", date_str, "\n", sep = "")
  dates_processed <- dates_processed + 1
  
  # Load TIFF for this date only
  tiff_path <- file.path(TIFF_FOLDER, tiff_file)
  raster <- terra::rast(tiff_path)
  
  dims <- dim(raster)
  cat("            Dimensions: ", dims[2], "×", dims[1], " pixels\n", sep = "")
  
  # Create date-specific output folder
  if (!dir.exists(date_output_folder)) {
    dir.create(date_output_folder, recursive = TRUE, showWarnings = FALSE)
  }
  
  cat("            Creating ", length(overlapping_tile_indices), " tiles...\n", sep = "")
  
  # Use makeTiles with FILTERED grid (only overlapping tiles)
  tiles_list <- terra::makeTiles(
    x = raster,
    y = filtered_grid_vect,
    filename = file.path(date_output_folder, "tile.tif"),
    overwrite = TRUE
  )
  
  # Rename tiles to [DATE]_[TILE_ID].tif
  for (tile_idx in seq_along(tiles_list)) {
    source_file <- file.path(date_output_folder, paste0("tile", tile_idx, ".tif"))
    tile_id <- filtered_grid_sf$tile_id[tile_idx]
    final_file <- file.path(date_output_folder, paste0(date_str, "_", tile_id, ".tif"))
    
    if (file.exists(source_file)) {
      file.rename(source_file, final_file)
    }
  }
  
  cat("            ✓ Created ", length(tiles_list), " tiles\n", sep = "")
  total_tiles_created <- total_tiles_created + length(tiles_list)
}

# ============================================================================
# VERIFICATION
# ============================================================================

cat("\n[9] Verifying output...\n")

# Count tiles per date folder
date_folders <- list.dirs(OUTPUT_FOLDER, full.names = FALSE, recursive = FALSE)
date_folders <- sort(date_folders[date_folders != "."])

total_tile_files <- 0
for (date_folder in date_folders) {
  tiles_in_folder <- list.files(file.path(OUTPUT_FOLDER, date_folder), 
                                pattern = "\\.tif$")
  tiles_in_folder <- tiles_in_folder[!grepl("master_grid", tiles_in_folder)]
  total_tile_files <- total_tile_files + length(tiles_in_folder)
  cat("  ", date_folder, ": ", length(tiles_in_folder), " tiles\n", sep = "")
}

# ============================================================================
# SUMMARY
# ============================================================================

cat("\n\n========== SUMMARY ==========\n")

cat("\nGrid Configuration:\n")
cat("  - Dimensions: ", GRID_NCOLS, "×", GRID_NROWS, " = ", N_TILES, " total tile positions\n", sep = "")
cat("  - Storage subfolder: daily_tiles_split/", GRID_SIZE_LABEL, "/\n", sep = "")
cat("  - Master grid file: master_grid_", GRID_SIZE_LABEL, ".geojson\n", sep = "")

cat("\nField Filtering:\n")
cat("  - Field boundaries loaded from pivot.geojson\n")
cat("  - Only overlapping tiles created (empty tiles deleted)\n")
cat("  - Significant storage savings for sparse fields!\n")

cat("\nProcessing Summary:\n")
cat("  - Total TIFF files: ", length(tiff_files), "\n", sep = "")
cat("  - Dates skipped (already processed): ", dates_skipped, "\n", sep = "")
cat("  - Dates processed: ", dates_processed, "\n", sep = "")
cat("  - Total tiles created: ", total_tiles_created, "\n", sep = "")
if (dates_processed > 0) {
  avg_tiles_per_date <- total_tiles_created / dates_processed
  cat("  - Average tiles per date: ", round(avg_tiles_per_date, 1), "\n", sep = "")
}

cat("\nDirectory Structure:\n")
cat("  laravel_app/storage/app/", PROJECT, "/daily_tiles_split/\n", sep = "")
cat("  └── ", GRID_SIZE_LABEL, "/\n", sep = "")
cat("      ├── master_grid_", GRID_SIZE_LABEL, ".geojson\n", sep = "")
cat("      ├── 2024-01-15/\n")
cat("      │   ├── 2024-01-15_01.tif  (only overlapping tiles)\n")
cat("      │   ├── 2024-01-15_05.tif\n")
cat("      │   └── ...\n")
cat("      ├── 2024-01-16/\n")
cat("      │   └── ...\n")
cat("      └── ...\n")

cat("\n⭐ Key Benefits:\n")
cat("  ✓ Overlap-filtered: No wasted empty tiles\n")
cat("  ✓ Skip existing dates: Resume-safe, idempotent\n")
cat("  ✓ Grid versioning: Future 10x10 grids stored separately\n")
cat("  ✓ Disk efficient: Storage reduced for sparse ROIs\n")

# ============================================================================
# WRITE TILING CONFIGURATION METADATA
# ============================================================================
# This metadata file is read by parameters_project.R to determine mosaic mode
# It allows script 40 to know what script 10 decided without re-computing

cat("\n[10] Writing tiling configuration metadata...\n")

config_file <- file.path(OUTPUT_FOLDER, "tiling_config.json")
config_json <- paste0(
  '{\n',
  '  "project": "', PROJECT, '",\n',
  '  "has_tiles": ', tolower(N_TILES > 1), ',\n',
  '  "grid_size": "', GRID_SIZE_LABEL, '",\n',
  '  "grid_rows": ', GRID_NROWS, ',\n',
  '  "grid_cols": ', GRID_NCOLS, ',\n',
  '  "roi_width_km": ', round(x_range_m / 1000, 1), ',\n',
  '  "roi_height_km": ', round(y_range_m / 1000, 1), ',\n',
  '  "created_date": "', Sys.Date(), '",\n',
  '  "created_time": "', format(Sys.time(), "%H:%M:%S"), '"\n',
  '}\n'
)

writeLines(config_json, config_file)
cat("  ✓ Metadata saved to: tiling_config.json\n")
cat("    - has_tiles: ", tolower(N_TILES > 1), "\n", sep = "")
cat("    - grid_size: ", GRID_SIZE_LABEL, "\n", sep = "")

cat("\n✓ Script complete!\n")