Improve DatasetInfo loading error message

FileNotFoundError doesn't make much sense, because there can be different causes why the file cannot be loaded.

PiperOrigin-RevId: 736818093

Author: tomvdw

tensorflow_datasets/core/dataset_info.py

@@ -702,20 +702,12 @@ def read_from_directory(self, dataset_info_dir: epath.PathLike) -> None:
         be the root directory of a specific dataset version.
 
     Raises:
-      FileNotFoundError: If the dataset_info.json can't be found.
+      DatasetInfoFileError: If the dataset_info.json can't be read.
     """
     logging.info("Load dataset info from %s", dataset_info_dir)
 
     # Load the metadata from disk
-    try:
-      parsed_proto = read_from_json(dataset_info_path(dataset_info_dir))
-    except Exception as e:
-      raise FileNotFoundError(
-          "Tried to load `DatasetInfo` from a directory which does not exist or"
-          " does not contain `dataset_info.json`. Please delete the directory "
-          f"`{dataset_info_dir}`  if you are trying to re-generate the "
-          "dataset."
-      ) from e
+    parsed_proto = read_from_json(dataset_info_path(dataset_info_dir))
 
     if str(self.version) != parsed_proto.version:
       raise AssertionError(
@@ -1128,12 +1120,14 @@ def read_from_json(path: epath.PathLike) -> dataset_info_pb2.DatasetInfo:
     the DatasetInfo proto.
 
   Raises:
-    FileNotFoundError: If the builder_dir does not exist.
+    DatasetInfoFileError: If the dataset info file cannot be read.
   """
   try:
     json_str = epath.Path(path).read_text()
   except OSError as e:
-    raise FileNotFoundError(f"Could not load dataset info from {path}") from e
+    raise DatasetInfoFileError(
+        f"Could not read dataset info from {path}"
+    ) from e
   # Parse it back into a proto.
   parsed_proto = json_format.Parse(json_str, dataset_info_pb2.DatasetInfo())
   return parsed_proto
@@ -1151,7 +1145,7 @@ def read_proto_from_builder_dir(
     The DatasetInfo proto as read from the builder dir.
 
   Raises:
-    FileNotFoundError: If the builder_dir does not exist.
+    DatasetInfoFileError: If the dataset info file cannot be read.
   """
   builder_dir = epath.Path(builder_dir).expanduser()
   info_path = builder_dir / constants.DATASET_INFO_FILENAME
@@ -1173,8 +1167,7 @@ def read_full_proto_from_builder_dir(
     dir.
 
   Raises:
-    FileNotFoundError: If the builder_dir does not exist or it doesn't contain
-    dataset_info.json.
+    DatasetInfoFileError: If the dataset info file cannot be read.
   """
   builder_dir = epath.Path(builder_dir).expanduser()
   info_path = builder_dir / constants.DATASET_INFO_FILENAME
@@ -1513,3 +1506,12 @@ def save_metadata(self, data_dir):
           self[key] = json.load(f)
     self._tempdir.rmtree()
     super(BeamMetadataDict, self).save_metadata(data_dir)
+
+
+class DatasetInfoFileError(OSError):
+  """Raised when the dataset info file cannot be read.
+
+  We use a custom exception rather than native exceptions, because different
+  backend of etils.epath will throw different exceptions. This exception
+  catches them all.
+  """

tensorflow_datasets/core/dataset_info_test.py

@@ -97,7 +97,8 @@ def tearDownClass(cls):
   def test_non_existent_dir(self):
     info = dataset_info.DatasetInfo(builder=self._builder)
     with self.assertRaisesWithPredicateMatch(
-        FileNotFoundError, "from a directory which does not exist"
+        dataset_info.DatasetInfoFileError,
+        "Could not read dataset info from",
     ):
       info.read_from_directory(_NON_EXISTENT_DIR)
 

tensorflow_datasets/core/read_only_builder_test.py

@@ -28,6 +28,7 @@
 from tensorflow_datasets import testing
 from tensorflow_datasets.core import constants
 from tensorflow_datasets.core import dataset_builder
+from tensorflow_datasets.core import dataset_info as dataset_info_lib
 from tensorflow_datasets.core import dataset_utils
 from tensorflow_datasets.core import load
 from tensorflow_datasets.core import proto
@@ -275,7 +276,10 @@ def test_restore_blocked_versions(
 
 
 def test_builder_from_directory_dir_not_exists(tmp_path: pathlib.Path):
-  with pytest.raises(FileNotFoundError, match='Could not load dataset info'):
+  with pytest.raises(
+      dataset_info_lib.DatasetInfoFileError,
+      match='Could not read dataset info from',
+  ):
     read_only_builder.builder_from_directory(tmp_path)
 
 

tensorflow_datasets/core/sequential_writer.py

@@ -256,7 +256,7 @@ def __init__(
               f'Trying to append a dataset with name {ds_info.name}'
               f' to an existing dataset with name {self._ds_info.name}'
           )
-      except FileNotFoundError:
+      except dataset_info.DatasetInfoFileError:
         self._ds_info.set_file_format(
             file_format=self._file_format,
             # if it was set, we want this to fail to warn the user