Datasets:
Languages:
English
Size:
100K<n<1M
Tags:
additive-manufacturing
laser-powder-bed-fusion
smoothed-particle-hydrodynamics
melt-pool
keyhole
physics-simulation
DOI:
License:
| #!/usr/bin/env python3 | |
| """Apply border recoloring to side/front frames; copy top frames as-is. | |
| Input: final_data/sim_NNNNN/frames/{front,side,top}/frame_NNNNN.png | |
| Output: final_data_processed/sim_NNNNN/frames/{front,side,top}/frame_NNNNN.png | |
| """ | |
| from __future__ import annotations | |
| import os | |
| import shutil | |
| from pathlib import Path | |
| import numpy as np | |
| from PIL import Image | |
| # ── constants (from prepare.py / cvae_sph2img) ─────────────────────────────── | |
| PURE_GREEN = np.array([0, 197, 0], dtype=np.uint8) | |
| PURE_BLUE = np.array([0, 0, 189], dtype=np.uint8) | |
| PURE_BLACK = np.array([0, 0, 0], dtype=np.uint8) | |
| PURE_GRAY = np.array([98, 93, 90], dtype=np.uint8) | |
| BORDER_THICKNESS_PX = 10 | |
| SIDE_BOTTOM_HEIGHT_PX = 188 | |
| SIDE_BUFFER = 3 | |
| # ── image processing ────────────────────────────────────────────────────────── | |
| def detect_side_columns(rgb: np.ndarray) -> tuple[int, int]: | |
| h, w, _ = rgb.shape | |
| if h == 0 or w == 0: | |
| return (0, 0) | |
| probe_y = max(0, h - SIDE_BOTTOM_HEIGHT_PX) | |
| row = rgb[probe_y, :, :3] | |
| mask = np.all(row == PURE_BLACK, axis=1) | np.all(row == PURE_GRAY, axis=1) | |
| left = 0 | |
| while left < w and mask[left]: | |
| left += 1 | |
| right = 0 | |
| idx = w - 1 | |
| while idx >= 0 and mask[idx]: | |
| right += 1 | |
| idx -= 1 | |
| return (left, right) | |
| def fixed_border_recolor(rgb: np.ndarray, left_columns: int, right_columns: int) -> np.ndarray: | |
| out = rgb[:, :, :3].copy() | |
| h, w, _ = out.shape | |
| bt = min(BORDER_THICKNESS_PX, h, w) | |
| sbh = min(SIDE_BOTTOM_HEIGHT_PX, h) | |
| out[:bt, :, :] = PURE_BLUE | |
| out[h - bt:, :, :] = PURE_GREEN | |
| if left_columns > 0: | |
| lw = min(left_columns, w) | |
| out[:, :lw, :] = PURE_BLUE | |
| out[h - sbh:, :lw, :] = PURE_GREEN | |
| if right_columns > 0: | |
| rw = min(right_columns, w) | |
| out[:, w - rw:, :] = PURE_BLUE | |
| out[h - sbh:, w - rw:, :] = PURE_GREEN | |
| return out | |
| def process_image(src: Path, dst: Path, is_side: bool) -> None: | |
| rgb = np.array(Image.open(src).convert("RGB")) | |
| h, w, _ = rgb.shape | |
| default_cols = min(BORDER_THICKNESS_PX, h, w) | |
| if is_side: | |
| left, right = detect_side_columns(rgb) | |
| if left > 0: left = min(w, left + SIDE_BUFFER) | |
| if right > 0: right = min(w, right + SIDE_BUFFER) | |
| else: | |
| left, right = default_cols, default_cols | |
| out = fixed_border_recolor(rgb, left, right) | |
| Image.fromarray(out).save(dst) | |
| # ── main ────────────────────────────────────────────────────────────────────── | |
| def main() -> None: | |
| here = Path(__file__).parent | |
| src_root = here / "final_data" | |
| dst_root = here / "final_data_processed" | |
| sim_dirs = sorted(src_root.iterdir()) | |
| print(f"Found {len(sim_dirs)} simulations") | |
| for sim_dir in sim_dirs: | |
| if not sim_dir.is_dir(): | |
| continue | |
| frames_src = sim_dir / "frames" | |
| if not frames_src.is_dir(): | |
| print(f" SKIP {sim_dir.name} — no frames/ directory") | |
| continue | |
| for view in ("front", "side", "top"): | |
| view_src = frames_src / view | |
| view_dst = dst_root / sim_dir.name / "frames" / view | |
| if not view_src.is_dir(): | |
| continue | |
| view_dst.mkdir(parents=True, exist_ok=True) | |
| for png in sorted(view_src.glob("*.png")): | |
| dst_path = view_dst / png.name | |
| if view == "top": | |
| shutil.copy2(png, dst_path) | |
| else: | |
| process_image(png, dst_path, is_side=(view == "side")) | |
| print(f" {sim_dir.name} done") | |
| print("All done.") | |
| if __name__ == "__main__": | |
| main() | |