remove comments which were moved to github

rising/loading/container.py

@@ -39,8 +39,6 @@ def split_by_index(self, split: SplitType) -> None:
         for key, idx in split.items():
             self._dset[key] = self._dataset.get_subset(idx)
 
-    # TODO: Shouldn"t the kfold methods instead yield the current datasets
-    #  instead of the whole cointainer?
     def kfold_by_index(self, splits: typing.Iterable[SplitType]):
         """
         Produces kfold splits based on the given indices.

rising/loading/dataset.py

@@ -9,6 +9,7 @@
 
 from torch.utils.data import Dataset as TorchDset
 from rising.loading.debug_mode import get_debug_mode
+from rising import AbstractMixin
 from torch.multiprocessing import Pool
 
 
@@ -328,10 +329,7 @@ def __len__(self) -> int:
         return len(self.data)
 
 
-# TODO: Maybe we should add the dataset baseclass as baseclass of this as well
-#  (since it should just extend it and still have all the other dataset
-#  functionalities)?
-class IDManager:
+class IDManager(AbstractMixin):
     def __init__(self, id_key: str, cache_ids: bool = True, **kwargs):
         """
         Helper class to add additional functionality to Datasets
@@ -442,8 +440,6 @@ def __init__(self, data_path, load_fn, id_key, cache_ids=True,
         **kwargs :
             additional keyword arguments
         """
-        # TODO: Shouldn't we call the baseclasses explicitly here? with super
-        #  it is not clear, which baseclass is actually called
         super().__init__(data_path=data_path, load_fn=load_fn, id_key=id_key,
                          cache_ids=cache_ids, **kwargs)
 
@@ -467,7 +463,5 @@ def __init__(self, data_path, load_fn, id_key, cache_ids=True,
         **kwargs :
             additional keyword arguments
         """
-        # TODO: Shouldn't we call the baseclasses explicitly here? with super
-        #  it is not clear, which baseclass is actually called
         super().__init__(data_path=data_path, load_fn=load_fn, id_key=id_key,
                          cache_ids=cache_ids, **kwargs)

rising/loading/splitter.py

@@ -155,10 +155,6 @@ def index_split_grouped(self, groups_key: str = "id", **kwargs) -> SplitType:
         --------
         Shuffling cannot be deactivated
         """
-        # TODO: maybe we should implement a single split function, which
-        #  handles random, stratificated and grouped splitting? This would not
-        #  be hard at all (based on the first look at sklearn internals) and
-        #  would remove some code duolication in here
         split_dict = {}
         groups = [d[groups_key] for d in self._dataset]