Raise FileNotFoundError when passing data_files that don't exist (#6155)

* raise filenotfound

* only raise if no magic char in pattern

* add test

* use has_magic

Author: lhoestq

src/datasets/data_files.py

@@ -1,6 +1,7 @@
 import os
 import re
 from functools import partial
+from glob import has_magic
 from pathlib import Path, PurePath
 from typing import Callable, Dict, List, Optional, Set, Tuple, Union
 
@@ -583,7 +584,8 @@ def from_patterns(
                     )
                 )
             except FileNotFoundError:
-                pass
+                if not has_magic(pattern):
+                    raise
         origin_metadata = _get_origin_metadata(data_files, download_config=download_config)
         return cls(data_files, origin_metadata)
 

tests/test_data_files.py

@@ -380,6 +380,11 @@ def test_DataFilesList_from_patterns_locally_with_extra_files(complex_data_dir,
     assert len(data_files_list.origin_metadata) == 2
 
 
+def test_DataFilesList_from_patterns_raises_FileNotFoundError(complex_data_dir):
+    with pytest.raises(FileNotFoundError):
+        DataFilesList.from_patterns(["file_that_doesnt_exist.txt"], complex_data_dir)
+
+
 @pytest.mark.parametrize("pattern", _TEST_PATTERNS)
 def test_DataFilesDict_from_patterns_in_dataset_repository(
     hub_dataset_repo_path, hub_dataset_repo_patterns_results, pattern