Datasets:
File size: 3,870 Bytes
8a12fdb | 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 | """materialize.py —— 把 parquet 展开成目录树（按字节原样写出，不重编码）。
目录形状（与原 GenImage 一致，方便直接喂 ImageFolder）：
<out>/<generator>/<split>/ai/<class_name>/*.PNG # 假图
<out>/nature/<split>/<class_name>/*.JPEG # 真图池，全局只有一份
用法：
python scripts/materialize.py --root . --out ./data
python scripts/materialize.py --root . --out ./data --config tiny
python scripts/materialize.py --root . --out ./data --splits test --generators ADM nature
python scripts/materialize.py --root . --out ./data --dry-run
"""
from __future__ import annotations
import argparse
import sys
from pathlib import Path
import pyarrow.parquet as pq
def image_bytes(cell):
"""兼容两种写法：struct{bytes,path} 或裸 binary。"""
if isinstance(cell, dict):
return cell.get("bytes")
if isinstance(cell, (bytes, bytearray, memoryview)):
return bytes(cell)
return None
def guess_ext(orig_path: str, fmt: str) -> str:
suf = Path(orig_path).suffix
if suf:
return suf.lower()
return {"PNG": ".png", "JPEG": ".jpg"}.get(fmt, ".bin")
def main() -> None:
ap = argparse.ArgumentParser()
ap.add_argument("--root", required=True, help="仓库根目录（含 data/ 与 meta/）")
ap.add_argument("--out", required=True, help="输出目录树")
ap.add_argument("--config", default="default", choices=["default", "tiny"])
ap.add_argument("--splits", nargs="*", default=None)
ap.add_argument("--generators", nargs="*", default=None)
ap.add_argument("--limit", type=int, default=None, help="每个 (生成器, split) 最多导出多少张")
ap.add_argument("--dry-run", action="store_true")
args = ap.parse_args()
root = Path(args.root)
sub = "data" if args.config == "default" else "tiny"
files = sorted((root / sub).glob("*.parquet"))
if not files:
sys.exit(f"[!] {root / sub} 下没有 parquet，检查 --root / --config")
out = Path(args.out)
total = written = 0
counts: dict[tuple, int] = {}
for f in files:
pf = pq.ParquetFile(f)
for batch in pf.iter_batches(batch_size=128):
d = batch.to_pydict()
for i in range(batch.num_rows):
split = d["split"][i]
gen = d["generator"][i]
if args.splits and split not in args.splits:
continue
if args.generators and gen not in args.generators:
continue
key = (gen, split)
if args.limit and counts.get(key, 0) >= args.limit:
continue
label = int(d["label"][i])
cls = str(d["class_name"][i] or f"class_{d['class_id'][i]}").replace("/", "_")
orig = d["image_path"][i]
stem = Path(orig).stem
ext = guess_ext(orig, d["format"][i])
branch = "ai" if label == 1 else "nature"
dst = out / gen / split / branch / cls / f"{stem}{ext}"
total += 1
if args.dry_run:
continue
dst.parent.mkdir(parents=True, exist_ok=True)
b = image_bytes(d["image"][i])
if b is None:
continue
dst.write_bytes(b)
written += 1
counts[key] = counts.get(key, 0) + 1
print(f"\r 处理 {f.name} … 已导出 {written:,}", end="", flush=True)
print()
if args.dry_run:
print(f"[dry-run] 将导出 {total:,} 张到 {out}")
return
print(f"完成：{written:,} 张 -> {out.resolve()}")
for (gen, split), n in sorted(counts.items()):
print(f" {gen:10s} {split:6s} {n:>7,}")
if __name__ == "__main__":
main()
|