chore: address CodeRabbit review feedback

1. Remove fork-specific issue URL placeholders (upstream-ready)
2. Add consistency assertions to LargeList test:
   - offset == 0 check
   - content verification (bytes embedded)
3. Add offset == 0 check to Nifti test for consistency

Author: The-Obstacle-Is-The-Way

src/datasets/table.py

@@ -2125,7 +2125,6 @@ def embed_array_storage(array: pa.Array, feature: "FeatureType", token_per_repo_
     # When ds.shard() or ds.select() creates a sliced view, array.values returns
     # values with internal offset references that can cause PyArrow's C++ layer
     # to crash when processing nested types like Sequence(Nifti()).
-    # See: https://github.com/huggingface/datasets/issues/XXXX
     if pa.types.is_list(array.type) or pa.types.is_large_list(array.type):
         if hasattr(array, "offset") and array.offset > 0:
             array = pa.concat_arrays([array])

tests/features/test_embed_storage_sliced.py

@@ -1,7 +1,7 @@
 """Tests for embed_array_storage with sliced/sharded arrays.
 
-Regression tests for https://github.com/huggingface/datasets/issues/XXXX
-(SIGKILL in embed_array_storage when processing sliced/sharded Arrow tables)
+Regression tests for SIGKILL crash when processing sliced/sharded Arrow tables
+with nested types like Sequence(Nifti()) or Sequence(Image()).
 """
 
 import pyarrow as pa
@@ -81,6 +81,8 @@ def test_embed_array_storage_sliced_list_nifti(self, shared_datadir):
         # This should NOT crash with SIGKILL
         embedded = embed_array_storage(sliced, List(Nifti()))
 
+        # The fix should make the result contiguous (offset = 0)
+        assert embedded.offset == 0, "Result should be contiguous after fix"
         assert len(embedded) == 2
         # Verify bytes were embedded
         assert embedded[0].as_py()[0]["bytes"] is not None
@@ -111,4 +113,10 @@ def test_embed_array_storage_sliced_large_list(self, shared_datadir):
         # This should NOT crash with SIGKILL
         embedded = embed_array_storage(sliced, LargeList(Image()))
 
+        # The fix should make the result contiguous (offset = 0)
+        assert embedded.offset == 0, "Result should be contiguous after fix"
         assert len(embedded) == 2
+        # Item 0 of sliced = Item 1 of original (has 2 images)
+        assert len(embedded[0].as_py()) == 2
+        # Verify bytes were embedded
+        assert embedded[0].as_py()[0]["bytes"] is not None