Sort list in test_get_available_shards remove flakiness in split_test.py

The TensorFlow Datasets Authors · The TensorFlow Datasets Authors · commit 4a4d3c72d364 · 2025-03-05T03:37:53.000-08:00
PiperOrigin-RevId: 733664768
diff --git a/tensorflow_datasets/core/splits_test.py b/tensorflow_datasets/core/splits_test.py
@@ -727,7 +727,7 @@ def test_get_available_shards(self):
             split='train', data_dir=os.fspath(tmp_dir), dataset_name='ds'
         ),
     )
-    self.assertEqual(
+    self.assertCountEqual(
         [train_shard1, train_shard_incorrect],
         split_info.get_available_shards(tmp_dir, strict_matching=False),
     )

Original file line number	Diff line number	Diff line change
`@@ -727,7 +727,7 @@ def test_get_available_shards(self):`
`727`	`727`	`split='train', data_dir=os.fspath(tmp_dir), dataset_name='ds'`
`728`	`728`	`),`
`729`	`729`	`)`
`730`		`- self.assertEqual(`
	`730`	`+ self.assertCountEqual(`
`731`	`731`	`[train_shard1, train_shard_incorrect],`
`732`	`732`	`split_info.get_available_shards(tmp_dir, strict_matching=False),`
`733`	`733`	`)`