refactor(preprocessing): rename ImageDataset as `ExperimentVideoDat…

…aset`

boiling_learning/preprocessing/cases.py

@@ -1,11 +1,11 @@
 from typing import Optional
 
 from boiling_learning.preprocessing.experiment_video import ExperimentVideo
-from boiling_learning.preprocessing.experiment_video_dataset import ImageDataset
+from boiling_learning.preprocessing.experiment_video_dataset import ExperimentVideoDataset
 from boiling_learning.utils.pathutils import PathLike, resolve
 
 
-class Case(ImageDataset):
+class Case(ExperimentVideoDataset):
     def __init__(
         self,
         path: PathLike,
@@ -47,7 +47,7 @@ def set_video_data_from_file(
         self,
         data_path: Optional[PathLike] = None,
         remove_absent: bool = False,
-        keys: ImageDataset.VideoDataKeys = ImageDataset.VideoDataKeys(),
+        keys: ExperimentVideoDataset.VideoDataKeys = ExperimentVideoDataset.VideoDataKeys(),
     ) -> None:
         super().set_video_data_from_file(
             data_path or self.video_data_path,

boiling_learning/preprocessing/experiment_video_dataset.py

@@ -13,7 +13,7 @@
 from boiling_learning.utils.pathutils import PathLike, resolve
 
 
-class ImageDataset(KeyedSet[str, ExperimentVideo]):
+class ExperimentVideoDataset(KeyedSet[str, ExperimentVideo]):
     @dataclass(frozen=True)
     class VideoDataKeys(ExperimentVideo.VideoDataKeys):
         name: str = 'name'
@@ -36,11 +36,11 @@ def name(self) -> str:
     @classmethod
     def make_union(
         cls,
-        *others: ImageDataset,
+        *others: ExperimentVideoDataset,
         namer: Callable[[List[str]], str] = '+'.join,
-    ) -> ImageDataset:
+    ) -> ExperimentVideoDataset:
         name = namer([other.name for other in others])
-        image_dataset = ImageDataset(name)
+        image_dataset = ExperimentVideoDataset(name)
         image_dataset.update(*others)
         return image_dataset
 
@@ -92,13 +92,13 @@ def set_video_data_from_file(
         self.set_video_data(video_data, remove_absent=remove_absent)
 
 
-@json.encode.instance(ImageDataset)
-def _encode_image_dataset(obj: ImageDataset) -> List[json.JSONDataType]:
+@json.encode.instance(ExperimentVideoDataset)
+def _encode_image_dataset(obj: ExperimentVideoDataset) -> List[json.JSONDataType]:
     return json.serialize(sorted(obj, key=_get_experiment_video_name))
 
 
-@describe.instance(ImageDataset)
-def _describe_image_dataset(obj: ImageDataset) -> List[Path]:
+@describe.instance(ExperimentVideoDataset)
+def _describe_image_dataset(obj: ExperimentVideoDataset) -> List[Path]:
     return describe(sorted(obj, key=_get_experiment_video_name))
 
 

boiling_learning/scripts/interact_frames.py

@@ -6,7 +6,7 @@
 # from ipywidgets import interact, widgets
 # from skimage.transform import downscale_local_mean as downscale
 
-# from boiling_learning.preprocessing.image_datasets import ImageDataset
+# from boiling_learning.preprocessing.image_datasets import ExperimentVideoDataset
 # from boiling_learning.preprocessing.experiment_video import ExperimentVideo
 
 # try:
@@ -34,7 +34,7 @@
 
 
 # def _interact_dataset_frames(
-#     datasets: Tuple[ImageDataset, ...], imshow: Callable[[np.ndarray], Any]
+#     datasets: Tuple[ExperimentVideoDataset, ...], imshow: Callable[[np.ndarray], Any]
 # ) -> None:
 #     datasets_options = [(dataset.name, dataset) for dataset in datasets]
 #     default_dataset = datasets_options[0][1]
@@ -74,7 +74,7 @@
 
 #     experiment_videos_widget.observe(update_max_index, 'value')
 
-#     def show_frames(dataset: ImageDataset, ev: ExperimentVideo, idx: int) -> None:
+#     def show_frames(dataset: ExperimentVideoDataset, ev: ExperimentVideo, idx: int) -> None:
 #         imshow(ev[idx])
 
 #     interact(
@@ -85,7 +85,7 @@
 #     )
 
 
-# def main(datasets: Iterable[ImageDataset], colab_backend: bool = False) -> None:
+# def main(datasets: Iterable[ExperimentVideoDataset], colab_backend: bool = False) -> None:
 #     datasets = tuple(datasets)
 
 #     imshow = (

boiling_learning/scripts/load_dataset_tree.py

@@ -3,16 +3,16 @@
 from loguru import logger
 
 from boiling_learning.preprocessing.experiment_video import ExperimentVideo
-from boiling_learning.preprocessing.experiment_video_dataset import ImageDataset
+from boiling_learning.preprocessing.experiment_video_dataset import ExperimentVideoDataset
 from boiling_learning.utils.pathutils import PathLike, resolve
 
 
-def main(datapath: PathLike) -> List[ImageDataset]:
+def main(datapath: PathLike) -> List[ExperimentVideoDataset]:
     logger.info(f'Loading cases from {datapath}')
 
     datapath = resolve(datapath)
 
-    datasets: List[ImageDataset] = []
+    datasets: List[ExperimentVideoDataset] = []
     for casedir in datapath.iterdir():
         logger.debug(f'Searching for subcases in {casedir}')
         if not casedir.is_dir():
@@ -27,7 +27,7 @@ def main(datapath: PathLike) -> List[ImageDataset]:
 
             subcase = subcasedir.name
 
-            dataset = ImageDataset(f'{case}:{subcase}')
+            dataset = ExperimentVideoDataset(f'{case}:{subcase}')
             for testdir in subcasedir.iterdir():
                 logger.debug(f'Searching for videos in {testdir}')
 

boiling_learning/scripts/set_condensation_datasets_data.py

@@ -7,7 +7,7 @@
 from loguru import logger
 
 from boiling_learning.preprocessing.experiment_video import ExperimentVideo
-from boiling_learning.preprocessing.experiment_video_dataset import ImageDataset
+from boiling_learning.preprocessing.experiment_video_dataset import ExperimentVideoDataset
 from boiling_learning.preprocessing.video import Video
 from boiling_learning.scripts.utils.setting_data import check_experiment_video_dataframe_indices
 from boiling_learning.utils.frozendicts import frozendict
@@ -23,7 +23,9 @@
 _TIMEDELTA_PATTERN = re.compile(r'(?P<h>\d{2}):(?P<min>\d{2}):(?P<s>\d{2})')
 
 
-def main(datasets: Iterable[ImageDataset], dataspecpath: PathLike) -> Tuple[ImageDataset, ...]:
+def main(
+    datasets: Iterable[ExperimentVideoDataset], dataspecpath: PathLike
+) -> Tuple[ExperimentVideoDataset, ...]:
     logger.info('Setting condensation data')
 
     datasets = tuple(datasets)
@@ -110,7 +112,7 @@ def _parse_timedelta(s: Optional[str]) -> Optional[timedelta]:
     return timedelta(hours=int(m['h']), minutes=int(m['min']), seconds=int(m['s']))
 
 
-def _make_dataframe(dataset: ImageDataset) -> None:
+def _make_dataframe(dataset: ExperimentVideoDataset) -> None:
     missing: List[ExperimentVideo] = []
     for ev in dataset:
         try:
@@ -131,8 +133,10 @@ def _make_dataframe(dataset: ImageDataset) -> None:
         ev.save_df(overwrite=False)
 
 
-def _group_datasets(datasets: Iterable[ImageDataset]) -> Tuple[ImageDataset, ...]:
-    datasets_dict: Dict[str, ImageDataset] = {}
+def _group_datasets(
+    datasets: Iterable[ExperimentVideoDataset],
+) -> Tuple[ExperimentVideoDataset, ...]:
+    datasets_dict: Dict[str, ExperimentVideoDataset] = {}
     for dataset in datasets:
         logger.debug(f'Reading condensation dataset {dataset.name}')
 
@@ -145,7 +149,7 @@ def _group_datasets(datasets: Iterable[ImageDataset]) -> Tuple[ImageDataset, ...
             dataset_name = f'{case}:{subcase}'
 
             if dataset_name not in datasets_dict:
-                datasets_dict[dataset_name] = ImageDataset(dataset_name)
+                datasets_dict[dataset_name] = ExperimentVideoDataset(dataset_name)
 
             datasets_dict[dataset_name].add(ev)
 

main.py

@@ -60,7 +60,7 @@
     strategy_scope,
 )
 from boiling_learning.preprocessing.experiment_video import ExperimentVideo
-from boiling_learning.preprocessing.experiment_video_dataset import ImageDataset
+from boiling_learning.preprocessing.experiment_video_dataset import ExperimentVideoDataset
 from boiling_learning.preprocessing.hdf5 import get_frame_from_hdf5, video_to_hdf5
 from boiling_learning.preprocessing.transformers import DictTransformer, Transformer
 from boiling_learning.preprocessing.video import PimsVideo, VideoFrame
@@ -180,7 +180,7 @@ def items(self) -> ItemsView[str, Any]:
     fps_cache_path=Path('.cache', 'fps'),
     end_frame_index_cache_path=Path('.cache', 'end_frame_index'),
 )
-condensation_datasets_merged_all = ImageDataset.make_union(*condensation_datasets_merged)
+condensation_datasets_merged_all = ExperimentVideoDataset.make_union(*condensation_datasets_merged)
 
 boiling_preprocessors, boiling_augmentors = make_boiling_processors.main(
     direct_visualization=True,
@@ -253,15 +253,15 @@ def sliceable_dataset_from_video_and_transformers(
 
 @dataclass(frozen=True)
 class GetImageDatasetParams:
-    image_dataset: ImageDataset
+    image_dataset: ExperimentVideoDataset
     transformers: List[Transformer[VideoFrame, VideoFrame]]
     splits: DatasetSplits
     target: Optional[str] = None
     dataset_size: Optional[Union[int, Fraction]] = None
 
 
 def _get_image_dataset(
-    image_dataset: ImageDataset,
+    image_dataset: ExperimentVideoDataset,
     transformers: List[Transformer[VideoFrame, VideoFrame]],
     splits: DatasetSplits,
     dataset_size: Optional[Union[int, Fraction]] = None,