feat(datasets): YOLO 风格标注数据集与几何/像素增强

- yolo: images/ 与同层 labels/ 的路径约定、`class_id cx cy w h` 归一化行解析
  （兼容分割多边形取外接矩形）、data.yaml 读写；缺失或空 txt 即无缺陷样本
- film_defect: 抛出图像、框、图像级标签；label_mode 决定多热向量还是互斥下标
- transforms: 图框同步的 Resize / Letterbox / CenterCrop / RandomResizedCrop /
  翻转 / ColorJitter / GaussianBlur / PerImageStandardize
- target_encoder: 高斯半径与 heatmap / wh / offset / mask 监督信号编码

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

datasets/__init__.py

@@ -0,0 +1,22 @@
+"""数据集：抛出样本的 Dataset 类与相应的数据变换、标签编码。"""
+
+from datasets.build import build_dataloader, build_dataset
+from datasets.film_defect import FilmDefectDataset, collate_fn
+from datasets.registry import DATASETS
+from datasets.target_encoder import CenterTargetEncoder
+from datasets.transforms import build_transforms
+from datasets.yolo import format_label_lines, load_classes, parse_label_file, write_data_yaml
+
+__all__ = [
+    "build_dataset",
+    "build_dataloader",
+    "FilmDefectDataset",
+    "collate_fn",
+    "DATASETS",
+    "CenterTargetEncoder",
+    "build_transforms",
+    "parse_label_file",
+    "format_label_lines",
+    "load_classes",
+    "write_data_yaml",
+]