File size: 12,988 Bytes
5c8cc9d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 | """
Quadtree Image Engine - Python implementation of the C quadtree image manipulation.
Mirrors the logic from compress.c, decompress.c, filters.c, rotate.c, union.c
"""
import numpy as np
from dataclasses import dataclass, field
from typing import Optional
from PIL import Image
import io
@dataclass
class QtNode:
red: int = 0
green: int = 0
blue: int = 0
area: int = 0
topLeft: Optional['QtNode'] = field(default=None, repr=False)
topRight: Optional['QtNode'] = field(default=None, repr=False)
bottomLeft: Optional['QtNode'] = field(default=None, repr=False)
bottomRight: Optional['QtNode'] = field(default=None, repr=False)
def is_leaf(self):
return (self.topLeft is None and self.topRight is None and
self.bottomLeft is None and self.bottomRight is None)
def get_mean(matrix: np.ndarray, x: int, y: int, size: int):
"""Compute average color and variance score for a region. Mirrors getMean() in suppl.c"""
region = matrix[y:y+size, x:x+size]
red = int(np.mean(region[:, :, 0]))
green = int(np.mean(region[:, :, 1]))
blue = int(np.mean(region[:, :, 2]))
variance = (
np.mean((region[:, :, 0].astype(int) - red)**2) +
np.mean((region[:, :, 1].astype(int) - green)**2) +
np.mean((region[:, :, 2].astype(int) - blue)**2)
) / 3
return red, green, blue, int(variance)
def compress_image(matrix: np.ndarray, x: int, y: int, size: int, threshold: int) -> QtNode:
"""Build quadtree. Mirrors compressImage() in compress.c"""
red, green, blue, score = get_mean(matrix, x, y, size)
node = QtNode(red=red, green=green, blue=blue, area=size * size)
if size > 1 and score > threshold:
half = size // 2
node.topLeft = compress_image(matrix, x, y, half, threshold)
node.topRight = compress_image(matrix, x + half, y, half, threshold)
node.bottomRight = compress_image(matrix, x + half, y + half, half, threshold)
node.bottomLeft = compress_image(matrix, x, y + half, half, threshold)
return node
def decompress_image(node: QtNode, matrix: np.ndarray, x: int, y: int, size: int):
"""Reconstruct pixel matrix from quadtree. Mirrors decompressImage() in decompress.c"""
if node.is_leaf():
matrix[y:y+size, x:x+size, 0] = node.red
matrix[y:y+size, x:x+size, 1] = node.green
matrix[y:y+size, x:x+size, 2] = node.blue
else:
half = size // 2
decompress_image(node.topLeft, matrix, x, y, half)
decompress_image(node.topRight, matrix, x + half, y, half)
decompress_image(node.bottomRight, matrix, x + half, y + half, half)
decompress_image(node.bottomLeft, matrix, x, y + half, half)
# ββ Filters (mirrors filters.c) ββββββββββββββββββββββββββββββββββββββββββββββ
def apply_grayscale(node: QtNode) -> QtNode:
"""True luminance-weighted grayscale.
Compute single luma value and assign to all 3 channels so the
output is actually grey (not green-tinted).
luma = 0.299R + 0.587G + 0.114B (BT.601 standard)
"""
res = QtNode(area=node.area)
if not node.is_leaf():
res.topLeft = apply_grayscale(node.topLeft)
res.topRight = apply_grayscale(node.topRight)
res.bottomLeft = apply_grayscale(node.bottomLeft)
res.bottomRight = apply_grayscale(node.bottomRight)
luma = int(0.299 * node.red + 0.587 * node.green + 0.114 * node.blue)
res.red = luma
res.green = luma
res.blue = luma
return res
def apply_negative(node: QtNode) -> QtNode:
"""Invert channels: 255 - value. Mirrors negativeImage() in filters.c"""
res = QtNode(area=node.area)
if not node.is_leaf():
res.topLeft = apply_negative(node.topLeft)
res.topRight = apply_negative(node.topRight)
res.bottomLeft = apply_negative(node.bottomLeft)
res.bottomRight = apply_negative(node.bottomRight)
res.red = 255 - node.red
res.green = 255 - node.green
res.blue = 255 - node.blue
return res
def apply_sepia(node: QtNode) -> QtNode:
"""Cinematic warm sepia tone. Mirrors sepia() in filters.c"""
res = QtNode(area=node.area)
if not node.is_leaf():
res.topLeft = apply_sepia(node.topLeft)
res.topRight = apply_sepia(node.topRight)
res.bottomLeft = apply_sepia(node.bottomLeft)
res.bottomRight = apply_sepia(node.bottomRight)
r, g, b = node.red, node.green, node.blue
res.red = min(255, int(0.393 * r + 0.769 * g + 0.189 * b))
res.green = min(255, int(0.272 * r + 0.534 * g + 0.131 * b))
res.blue = min(255, int(0.349 * r + 0.686 * g + 0.168 * b))
return res
def apply_brighten(node: QtNode, factor: float = 1.3) -> QtNode:
"""Brighten by scaling channels. Mirrors brighten() in filters.c"""
res = QtNode(area=node.area)
if not node.is_leaf():
res.topLeft = apply_brighten(node.topLeft, factor)
res.topRight = apply_brighten(node.topRight, factor)
res.bottomLeft = apply_brighten(node.bottomLeft, factor)
res.bottomRight = apply_brighten(node.bottomRight, factor)
res.red = min(255, int(node.red * factor))
res.green = min(255, int(node.green * factor))
res.blue = min(255, int(node.blue * factor))
return res
# ββ Spatial Transforms (mirrors rotate.c) ββββββββββββββββββββββββββββββββββββ
def get_water_image(node: QtNode):
"""Vertical flip (top-bottom). Mirrors getWaterImage() in rotate.c"""
if not node.is_leaf():
get_water_image(node.topLeft)
get_water_image(node.topRight)
get_water_image(node.bottomLeft)
get_water_image(node.bottomRight)
node.topLeft, node.bottomLeft = node.bottomLeft, node.topLeft
node.topRight, node.bottomRight = node.bottomRight, node.topRight
def get_mirror_image(node: QtNode):
"""Horizontal mirror (left-right). Mirrors getMirrorImage() in rotate.c"""
if not node.is_leaf():
get_mirror_image(node.topLeft)
get_mirror_image(node.topRight)
get_mirror_image(node.bottomLeft)
get_mirror_image(node.bottomRight)
node.topLeft, node.topRight = node.topRight, node.topLeft
node.bottomLeft, node.bottomRight = node.bottomRight, node.bottomLeft
def rotate_left(node: QtNode):
"""90Β° counter-clockwise. Mirrors rotateLeft() in rotate.c"""
if not node.is_leaf():
rotate_left(node.topLeft)
rotate_left(node.topRight)
rotate_left(node.bottomLeft)
rotate_left(node.bottomRight)
tl, tr, bl, br = node.topLeft, node.topRight, node.bottomLeft, node.bottomRight
node.topLeft = tr
node.topRight = br
node.bottomLeft = tl
node.bottomRight = bl
def rotate_right(node: QtNode):
"""90Β° clockwise. Mirrors rotateRight() in rotate.c"""
if not node.is_leaf():
rotate_right(node.topLeft)
rotate_right(node.topRight)
rotate_right(node.bottomLeft)
rotate_right(node.bottomRight)
tl, tr, bl, br = node.topLeft, node.topRight, node.bottomLeft, node.bottomRight
node.topLeft = bl
node.topRight = tl
node.bottomLeft = br
node.bottomRight = tr
# ββ Union / Blend (mirrors union.c) ββββββββββββββββββββββββββββββββββββββββββ
def union_of_images(t1: QtNode, t2: QtNode) -> QtNode:
"""Average-blend two quadtrees. Mirrors unionOfImages() in union.c"""
res = QtNode(area=min(t1.area, t2.area))
res.red = (t1.red + t2.red) // 2
res.green = (t1.green + t2.green) // 2
res.blue = (t1.blue + t2.blue) // 2
if not t1.is_leaf() and not t2.is_leaf():
res.topLeft = union_of_images(t1.topLeft, t2.topLeft)
res.topRight = union_of_images(t1.topRight, t2.topRight)
res.bottomLeft = union_of_images(t1.bottomLeft, t2.bottomLeft)
res.bottomRight = union_of_images(t1.bottomRight, t2.bottomRight)
elif t1.is_leaf() and not t2.is_leaf():
res.topLeft = union_of_images(t1, t2.topLeft)
res.topRight = union_of_images(t1, t2.topRight)
res.bottomLeft = union_of_images(t1, t2.bottomLeft)
res.bottomRight = union_of_images(t1, t2.bottomRight)
elif not t1.is_leaf() and t2.is_leaf():
res.topLeft = union_of_images(t1.topLeft, t2)
res.topRight = union_of_images(t1.topRight, t2)
res.bottomLeft = union_of_images(t1.bottomLeft, t2)
res.bottomRight = union_of_images(t1.bottomRight, t2)
return res
# ββ I/O Helpers βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
def read_ppm_bytes(data: bytes):
"""
Parse P6 binary PPM data robustly.
Returns numpy array (H, W, 3) uint8.
"""
# First try Pillow (handles PNG/JPG/PPM automatically)
try:
img = Image.open(io.BytesIO(data)).convert("RGB")
return np.array(img, dtype=np.uint8)
except Exception:
pass
# Manual P6 binary PPM parser β reads header byte-by-byte
pos = 0
def read_token():
nonlocal pos
# Skip whitespace and comments
while pos < len(data):
c = data[pos:pos+1]
if c == b'#':
while pos < len(data) and data[pos:pos+1] != b'\n':
pos += 1
elif c in (b' ', b'\t', b'\n', b'\r'):
pos += 1
else:
break
start = pos
while pos < len(data) and data[pos:pos+1] not in (b' ', b'\t', b'\n', b'\r'):
pos += 1
return data[start:pos].decode('ascii')
magic = read_token()
if magic != 'P6':
raise ValueError(f"Not a P6 PPM file (got: {magic!r})")
w = int(read_token())
h = int(read_token())
_maxval = int(read_token())
# skip exactly one whitespace byte after maxval
pos += 1
raw = data[pos:pos + h * w * 3]
arr = np.frombuffer(raw, dtype=np.uint8).reshape(h, w, 3)
return arr.copy() # make writable
def arr_to_pil(arr: np.ndarray) -> Image.Image:
return Image.fromarray(arr.astype(np.uint8), 'RGB')
def next_power_of_two(n: int) -> int:
p = 1
while p < n:
p <<= 1
return p
def pad_to_square_pow2(arr: np.ndarray):
"""Pad image to square power-of-2 as required by the quadtree."""
h, w = arr.shape[:2]
size = next_power_of_two(max(h, w))
padded = np.zeros((size, size, 3), dtype=np.uint8)
padded[:h, :w] = arr
return padded, h, w
def process_image(matrix: np.ndarray, operation: str, threshold: int,
matrix2: Optional[np.ndarray] = None, return_tree: bool = False) -> np.ndarray:
"""
Full pipeline: pad β compress β transform β decompress β crop.
Returns the result as a numpy RGB array, and optionally the tree.
"""
padded, oh, ow = pad_to_square_pow2(matrix)
size = padded.shape[0]
tree = compress_image(padded, 0, 0, size, threshold)
if operation == "grayscale":
tree = apply_grayscale(tree)
elif operation == "negative":
tree = apply_negative(tree)
elif operation == "sepia":
tree = apply_sepia(tree)
elif operation == "brighten":
tree = apply_brighten(tree)
elif operation == "mirror":
get_mirror_image(tree)
elif operation == "water":
get_water_image(tree)
elif operation == "rotate_left":
rotate_left(tree)
elif operation == "rotate_right":
rotate_right(tree)
elif operation == "union" and matrix2 is not None:
padded2, oh2, ow2 = pad_to_square_pow2(matrix2)
s2 = padded2.shape[0]
s_common = max(size, s2)
p1 = np.zeros((s_common, s_common, 3), dtype=np.uint8)
p1[:padded.shape[0], :padded.shape[1]] = padded
p2 = np.zeros((s_common, s_common, 3), dtype=np.uint8)
p2[:padded2.shape[0], :padded2.shape[1]] = padded2
t1 = compress_image(p1, 0, 0, s_common, threshold)
t2 = compress_image(p2, 0, 0, s_common, threshold)
tree = union_of_images(t1, t2)
size = s_common
# For union, we want the bounding box that encompasses both original images
oh = max(oh, oh2)
ow = max(ow, ow2)
out = np.zeros((size, size, 3), dtype=np.uint8)
decompress_image(tree, out, 0, 0, size)
if operation == "mirror":
res = out[:oh, size - ow : size]
elif operation == "water":
res = out[size - oh : size, :ow]
elif operation == "rotate_left":
res = out[size - ow : size, :oh]
elif operation == "rotate_right":
res = out[:ow, size - oh : size]
else:
res = out[:oh, :ow]
if return_tree:
return res, tree
return res
|