godot-procedural-generation（戈多程序生成）

Procedural Generation

Seeded algorithms, noise functions, and constraint propagation define replayable content generation.

Available Scripts

fast_noise_noise2d_master.gd

Advanced usage of FastNoiseLite with image-based sampling for maximum performance.

cellular_automata_dungeon.gd

The classic 4-5 rule implementation for organic cave and terrain generation.

poisson_disk_sampling_2d.gd

Blue-noise distribution algorithm for non-clumping object and enemy placement.

multi_threaded_chunk_gen.gd

Expert pattern for offloading procedural generation to the WorkerThreadPool.

drunknard_walk_path.gd

Lightweight algorithm for generating winding paths, tunnels, and rivers.

marching_squares_metaballs.gd

Implementing the Marching Squares algorithm for smooth contouring and influential maps.

bsp_tree_rooms.gd

Binary Space Partitioning for generating structured, non-overlapping floor plans.

wave_function_collapse_lite.gd

Foundation for Wave Function Collapse (WFC) using entropy-based adjacency rules.

mesh_gen_infinite_terrain.gd

Runtime 3D terrain generation using ArrayMesh and SurfaceTool with LOD potential.

l_system_tree_gen.gd

L-System string grammar for procedural plant and tree growth in 3D.

wfc_level_generator.gd

Expert Wave Function Collapse implementation with tile adjacency rules.

NEVER Do in Procedural Generation

• NEVER generate chunks on the Main Thread — Proc-gen is CPU intensive and causes frame-rate spikes. Use WorkerThreadPool or a background Thread to keep the UI responsive.
• NEVER query FastNoiseLite every frame — Sampling noise per frame (especially in _process) is a massive waste. Generate your map into an Image or Array once and sample from memory [NoiseSampling].
• NEVER use randi() for reproducible seeds — Always store and reuse a specific seed within your random number generator (RandomNumberGenerator.new()) to ensure consistent world generation.
• NEVER use pure randomness for object placement — Pure random (white noise) causes clumping and overlapping. Use Poisson Disk Sampling or Jittered Grids for natural-looking distributions.
• NEVER forget to bound your loops — Procedural loops (like WFC or Cellular Automata) can easily enter infinite states if constraints are impossible. Always include a max_iterations safety break.
• NEVER instantiate nodes directly from proc-gen threads — You cannot touch the SceneTree from a worker thread. Generate the *data* in the thread, then notify the Main Thread to handle add_child().
• NEVER use complex WFC for simple layouts — Wave Function Collapse is powerful but overkill for simple paths. Use Drunkard's Walk or BSP for lightweight structured layouts.
• NEVER rely on TileMap.set_cell() for large-scale updates — Updating 10,000 cells individually is slow. Prepare a TileMapPattern and use set_pattern() or set_cells_terrain_connect() for batch updates.
• NEVER forget to bake Navigation at the end — Procedurally generated worlds need their navmeshes rebaked at runtime or the AI will walk into walls.
• NEVER ignore data serialization — If you generate a world, you must be able to save the *seed* and any *player modifications*. Don't try to save the entire raw chunk state if avoidable.

func generate_dungeon(width: int, height: int, fill_percent: float = 0.4) -> Array:
    var grid := []
    for y in height:
        var row := []
        for x in width:
            row.append(1)  # 1 = wall
        grid.append(row)

    # Start in center
    var x := width / 2
    var y := height / 2
    var floor_tiles := 0
    var target_floor := int(width * height * fill_percent)

    while floor_tiles < target_floor:
        if grid[y][x] == 1:
            grid[y][x] = 0  # Create floor
            floor_tiles += 1

        # Random walk
        var dir := randi() % 4
        match dir:
            0: x = clampi(x + 1, 0, width - 1)
            1: x = clampi(x - 1, 0, width - 1)
            2: y = clampi(y + 1, 0, height - 1)
            3: y = clampi(y - 1, 0, height - 1)

    return grid

Perlin Noise Terrain

var noise := FastNoiseLite.new()

func generate_terrain(width: int, height: int) -> Array:
    noise.seed = randi()
    noise.frequency = 0.05

    var terrain := []
    for y in height:
        var row := []
        for x in width:
            var value := noise.get_noise_2d(x, y)

            # Map noise to tile types
            var tile: int
            if value < -0.2:
                tile = 0  # Water
            elif value < 0.2:
                tile = 1  # Grass
            else:
                tile = 2  # Mountain

            row.append(tile)
        terrain.append(row)

    return terrain

BSP Rooms

class_name BSPRoom

var x: int
var y: int
var width: int
var height: int
var left: BSPRoom = null
var right: BSPRoom = null

func split(min_size: int = 6) -> bool:
    if left or right:
        return false  # Already split

    # Choose split direction
    var split_horizontal := randf() > 0.5

    if width > height and float(width) / float(height) >= 1.25:
        split_horizontal = false
    elif height > width and float(height) / float(width) >= 1.25:
        split_horizontal = true

    var max := (height if split_horizontal else width) - min_size
    if max <= min_size:
        return false  # Too small

    var split_pos := randi_range(min_size, max)

    if split_horizontal:
        left = BSPRoom.new()
        left.x = x
        left.y = y
        left.width = width
        left.height = split_pos

        right = BSPRoom.new()
        right.x = x
        right.y = y + split_pos
        right.width = width
        right.height = height - split_pos
    else:
        left = BSPRoom.new()
        left.x = x
        left.y = y
        left.width = split_pos
        left.height = height

        right = BSPRoom.new()
        right.x = x + split_pos
        right.y = y
        right.width = width - split_pos
        right.height = height

    return true

func generate_bsp_dungeon(width: int, height: int, iterations: int = 4) -> Array[BSPRoom]:
    var root := BSPRoom.new()
    root.x = 0
    root.y = 0
    root.width = width
    root.height = height

    var rooms: Array[BSPRoom] = [root]

    for i in iterations:
        var new_rooms: Array[BSPRoom] = []
        for room in rooms:
            if room.split():
                new_rooms.append(room.left)
                new_rooms.append(room.right)
            else:
                new_rooms.append(room)
        rooms = new_rooms

    return rooms

Random Loot

func generate_loot(loot_level: int) -> Array[Item]:
    var items: Array[Item] = []
    var roll_count := randi_range(1, 3)

    for i in roll_count:
        var rarity := roll_rarity()
        var item := get_random_item(rarity, loot_level)
        items.append(item)

    return items

func roll_rarity() -> String:
    var roll := randf()
    if roll < 0.6:
        return "common"
    elif roll < 0.85:
        return "uncommon"
    elif roll < 0.95:
        return "rare"
    else:
        return "legendary"

Wave Function Collapse

# Simplified WFC for tile patterns
# Load compatible tile adjacency rules
var tile_rules := {
    "grass": ["grass", "path", "water_edge"],
    "water": ["water", "water_edge"],
    "path": ["grass", "path"]
}

func wfc_generate(width: int, height: int) -> Array:
    var grid := []
    for y in height:
        var row := []
        for x in width:
            row.append(null)  # Uncollapsed
        grid.append(row)

    # Collapse cells until complete
    while has_uncollapsed(grid):
        var pos := find_lowest_entropy(grid)
        collapse_cell(grid, pos)
        propagate_constraints(grid, pos)

    return grid

Best Practices

1. Seeding - Use seeds for reproducibility
2. Validation - Ensure playable levels
3. Performance - Generate async if needed

Reference

• Related: godot-tilemap-mastery, godot-resource-data-patterns

Related

• Master Skill: godot-master