modifications

torchaudio/datasets/__init__.py

@@ -6,7 +6,7 @@
 from .gtzan import GTZAN
 from .librilight_limited import LibriLightLimited
 from .librimix import LibriMix
-from .librispeech import LIBRISPEECH, LIBRISPEECHBase
+from .librispeech import LIBRISPEECH, LibriSpeechBase
 from .libritts import LIBRITTS
 from .ljspeech import LJSPEECH
 from .musdb_hq import MUSDB_HQ
@@ -21,7 +21,7 @@
 __all__ = [
     "COMMONVOICE",
     "LIBRISPEECH",
-    "LIBRISPEECHBase",
+    "LibriSpeechBase",
     "LibriLightLimited",
     "SPEECHCOMMANDS",
     "VCTK_092",

torchaudio/datasets/librispeech.py

@@ -44,19 +44,19 @@ def download_librispeech(root, url):
 
 
 def get_librispeech_metadata(
-    fileid: str, path: str, ext_audio: str, ext_txt: str
+    fileid: str, root: str, subset: str, ext_audio: str, ext_txt: str
 ) -> Tuple[str, int, str, int, int, int]:
     speaker_id, chapter_id, utterance_id = fileid.split("-")
 
     # Get audio path and sample rate
     fileid_audio = f"{speaker_id}-{chapter_id}-{utterance_id}"
-    filepath = os.path.join(speaker_id, chapter_id, f"{fileid_audio}{ext_audio}")
-    full_path = os.path.join(path, filepath)
+    filepath = os.path.join(subset, speaker_id, chapter_id, f"{fileid_audio}{ext_audio}")
+    full_path = os.path.join(root, filepath)
     sample_rate = torchaudio.info(full_path).sample_rate
 
     # Load text
     file_text = f"{speaker_id}-{chapter_id}{ext_txt}"
-    file_text = os.path.join(path, speaker_id, chapter_id, file_text)
+    file_text = os.path.join(root, subset, speaker_id, chapter_id, file_text)
     with open(file_text) as ft:
         for line in ft:
             fileid_text, transcript = line.strip().split(" ", 1)
@@ -79,11 +79,14 @@ def get_librispeech_metadata(
 def load_librispeech_item(
     fileid: str, path: str, ext_audio: str, ext_txt: str
 ) -> Tuple[Tensor, int, str, int, int, int]:
+    path = os.path.normpath(path)
+    root = os.path.dirname(path)
+    subset = os.path.basename(path)
+
     filepath, sample_rate, transcript, speaker_id, chapter_id, utterance_id = get_librispeech_metadata(
-        fileid, path, ext_audio, ext_txt
+        fileid, root, subset, ext_audio, ext_txt
     )
-    full_path = os.path.join(path, filepath)
-    waveform, _ = torchaudio.load(full_path)
+    waveform, _ = torchaudio.load(os.path.join(root, filepath))
 
     return (
         waveform,
@@ -95,47 +98,33 @@ def load_librispeech_item(
     )
 
 
-class LIBRISPEECHBase(Dataset):
+class LibriSpeechBase(Dataset):
     """Create a Dataset for *LibriSpeech* [:footcite:`7178964`].
 
     Args:
-        root (str or Path): Path to the directory where the dataset is found or downloaded.
-        url (str, optional): The URL to download the dataset from,
-            or the type of the dataset to dowload.
-            Allowed type values are ``"dev-clean"``, ``"dev-other"``, ``"test-clean"``,
+        root (str or Path): Path to the directory where the dataset is found.
+        subset (str, optional): Subset of LibriSpeech to use.
+            Valid options: ``"dev-clean"``, ``"dev-other"``, ``"test-clean"``,
             ``"test-other"``, ``"train-clean-100"``, ``"train-clean-360"`` and
             ``"train-other-500"``. (default: ``"train-clean-100"``)
-        folder_in_archive (str, optional):
-            The top-level directory of the dataset. (default: ``"LibriSpeech"``)
-        download (bool, optional):
-            Whether to download the dataset if it is not found at root path. (default: ``False``).
     """
 
     _ext_txt = ".trans.txt"
     _ext_audio = ".flac"
 
-    def __init__(
-        self,
-        root: Union[str, Path],
-        url: str = URL,
-        folder_in_archive: str = FOLDER_IN_ARCHIVE,
-        download: bool = False,
-    ) -> None:
-        if url not in _DATA_SUBSETS:
-            raise ValueError(f"Invalid url '{url}' given; please provide one of {_DATA_SUBSETS}.")
+    def __init__(self, root: Union[str, Path], subset: str = "train-clean-100") -> None:
+        if subset not in _DATA_SUBSETS:
+            raise ValueError(f"Invalid subset '{subset}' given; please provide one of {_DATA_SUBSETS}.")
 
         root = os.fspath(root)
-        self._path = os.path.join(root, folder_in_archive, url)
+        self._root = root
+        self._subset = subset
+        subset_path = os.path.join(root, subset)
 
-        if not os.path.isdir(self._path):
-            if download:
-                download_librispeech(root, url)
-            else:
-                raise RuntimeError(
-                    f"Dataset not found at {self._path}. Please set `download=True` to download the dataset."
-                )
+        if not os.path.isdir(subset_path):
+            raise RuntimeError(f"Dataset not found at {subset_path}.")
 
-        self._walker = sorted(str(p.stem) for p in Path(self._path).glob("*/*/*" + self._ext_audio))
+        self._walker = sorted(str(p.stem) for p in Path(subset_path).glob("*/*/*" + self._ext_audio))
 
     def __getitem__(self, n: int) -> Tuple[Tensor, int, str, int, int, int]:
         """Load the n-th sample from the dataset.
@@ -148,13 +137,13 @@ def __getitem__(self, n: int) -> Tuple[Tensor, int, str, int, int, int]:
             ``(filepath, sample_rate, transcript, speaker_id, chapter_id, utterance_id)``
         """
         fileid = self._walker[n]
-        return get_librispeech_metadata(fileid, self._path, self._ext_audio, self._ext_txt)
+        return get_librispeech_metadata(fileid, self._root, self._subset, self._ext_audio, self._ext_txt)
 
     def __len__(self) -> int:
         return len(self._walker)
 
 
-class LIBRISPEECH(LIBRISPEECHBase):
+class LIBRISPEECH(LibriSpeechBase):
     """Create a Dataset for *LibriSpeech* [:footcite:`7178964`].
 
     Args:
@@ -177,7 +166,13 @@ def __init__(
         folder_in_archive: str = FOLDER_IN_ARCHIVE,
         download: bool = False,
     ) -> None:
-        super().__init__(root)
+        archive = os.path.join(root, folder_in_archive)
+        self._path = os.path.join(archive, url)
+
+        if not os.path.isdir(self._path) and download:
+            download_librispeech(archive, url)
+
+        super().__init__(root=root, subset=url)
 
     def __getitem__(self, n: int) -> Tuple[Tensor, int, str, int, int, int]:
         """Load the n-th sample from the dataset.