Erik W.
a62da8926d
- bump project/runtime version markers to v2.3.0 (pyproject + TetriNode UI version) - complete behavior-preserving refactor from monoliths into modular Python and JS architecture (tetrinode/ + js/live/) - add parity + standard regression harnesses under qa/ - fix queue panel layout so Queue bottom stays aligned to matrix bottom regardless of Controls visibility - fix music playback fallback when selected track is none, including modal Play behavior and runtime playback path
88 lines
3.0 KiB
Python
88 lines
3.0 KiB
Python
import numpy as np
|
|
import torch
|
|
from PIL import Image, ImageDraw
|
|
|
|
from ..constants import COLORS, PREVIEW_GRID, SHAPES
|
|
from ..game.rng import _new_bag
|
|
|
|
def _render_next_piece(shape, block_size, colors=None):
|
|
palette = colors or COLORS
|
|
grid = PREVIEW_GRID
|
|
size = grid * block_size
|
|
img = Image.new("RGB", (size, size), palette["X"])
|
|
draw = ImageDraw.Draw(img)
|
|
|
|
cells = SHAPES[shape][0]
|
|
min_x = min(x for x, _ in cells)
|
|
min_y = min(y for _, y in cells)
|
|
max_x = max(x for x, _ in cells)
|
|
max_y = max(y for _, y in cells)
|
|
shape_w = max_x - min_x + 1
|
|
shape_h = max_y - min_y + 1
|
|
offset_x = (grid - shape_w) // 2 - min_x
|
|
offset_y = (grid - shape_h) // 2 - min_y
|
|
|
|
for x, y in cells:
|
|
gx = x + offset_x
|
|
gy = y + offset_y
|
|
if 0 <= gx < grid and 0 <= gy < grid:
|
|
x0 = gx * block_size
|
|
y0 = gy * block_size
|
|
draw.rectangle([x0, y0, x0 + block_size - 2, y0 + block_size - 2], fill=palette[shape])
|
|
|
|
arr = np.array(img).astype(np.float32) / 255.0
|
|
return torch.from_numpy(arr)[None, ...]
|
|
|
|
|
|
def _get_upcoming_shapes(state, count):
|
|
if count <= 0:
|
|
return []
|
|
upcoming = [state.get("next_piece_shape")] + list(state.get("bag", []))
|
|
bag_count = state.get("bag_count", 0)
|
|
seed = state.get("seed", 0)
|
|
while len(upcoming) < count:
|
|
bag = _new_bag(seed, bag_count)
|
|
bag_count += 1
|
|
upcoming.extend(bag)
|
|
return upcoming[:count]
|
|
|
|
|
|
def _render_queue(shapes, block_size, colors=None):
|
|
palette = colors or COLORS
|
|
if not shapes:
|
|
size = PREVIEW_GRID * block_size
|
|
img = Image.new("RGB", (size, size), palette["X"])
|
|
arr = np.array(img).astype(np.float32) / 255.0
|
|
return torch.from_numpy(arr)[None, ...]
|
|
gap = block_size
|
|
width = PREVIEW_GRID * block_size
|
|
height = len(shapes) * PREVIEW_GRID * block_size + max(0, len(shapes) - 1) * gap
|
|
img = Image.new("RGB", (width, height), palette["X"])
|
|
draw = ImageDraw.Draw(img)
|
|
for idx, shape in enumerate(shapes):
|
|
if shape not in SHAPES:
|
|
continue
|
|
offset_y = idx * (PREVIEW_GRID * block_size + gap)
|
|
cells = SHAPES[shape][0]
|
|
min_x = min(x for x, _ in cells)
|
|
min_y = min(y for _, y in cells)
|
|
max_x = max(x for x, _ in cells)
|
|
max_y = max(y for _, y in cells)
|
|
shape_w = max_x - min_x + 1
|
|
shape_h = max_y - min_y + 1
|
|
offset_x = (PREVIEW_GRID - shape_w) // 2 - min_x
|
|
offset_cell_y = (PREVIEW_GRID - shape_h) // 2 - min_y
|
|
for x, y in cells:
|
|
gx = x + offset_x
|
|
gy = y + offset_cell_y
|
|
if 0 <= gx < PREVIEW_GRID and 0 <= gy < PREVIEW_GRID:
|
|
x0 = gx * block_size
|
|
y0 = offset_y + gy * block_size
|
|
draw.rectangle(
|
|
[x0, y0, x0 + block_size - 2, y0 + block_size - 2],
|
|
fill=palette[shape],
|
|
)
|
|
arr = np.array(img).astype(np.float32) / 255.0
|
|
return torch.from_numpy(arr)[None, ...]
|
|
|