Add unit test for LibriMix dataset (#2659)

Summary:
Besides the unit test, the PR also addresses these issues:
- The original `LibriMix` dataset only supports "min" mode, which means the audio length is the minimum of all clean sources. It is default for source separation task. Users may also want to use "max" mode which allows for end-to-end separation and recognition. The PR adds ``mode`` argument to let users decide which dataset they want to use.
- If the task is ``"enh_both"``, the target is the audios in ``mix_clean`` instead of separate clean sources. The PR fixes it to use ``mix_clean`` as target.

Pull Request resolved: https://github.com/pytorch/audio/pull/2659

Reviewed By: carolineechen

Differential Revision: D40229227

Pulled By: nateanl

fbshipit-source-id: fc07e0d88a245e1367656d3767cf98168a799235

test/torchaudio_unittest/datasets/librimix_test.py

+import os
+
+from parameterized import parameterized
+
+from torchaudio.datasets import LibriMix
+from torchaudio_unittest.common_utils import get_whitenoise, save_wav, TempDirMixin, TorchaudioTestCase
+
+_SAMPLE_RATE = 8000
+_TASKS_TO_MIXTURE = {
+    "sep_clean": "mix_clean",
+    "enh_single": "mix_single",
+    "enh_both": "mix_both",
+    "sep_noisy": "mix_both",
+}
+
+
+def _save_wav(filepath: str, seed: int):
+    wav = get_whitenoise(
+        sample_rate=_SAMPLE_RATE,
+        duration=0.01,
+        n_channels=1,
+        seed=seed,
+    )
+    save_wav(filepath, wav, _SAMPLE_RATE)
+    return wav
+
+
+def get_mock_dataset(root_dir: str, num_speaker: int):
+    """
+    root_dir: directory to the mocked dataset
+    """
+    mocked_data = []
+    seed = 0
+    base_dir = os.path.join(root_dir, f"Libri{num_speaker}Mix", "wav8k", "min", "train-360")
+    os.makedirs(base_dir, exist_ok=True)
+    for utterance_id in range(10):
+        filename = f"{utterance_id}.wav"
+        task_outputs = {}
+        for task in _TASKS_TO_MIXTURE:
+            # create mixture folder. The folder names depends on the task.
+            mixture_folder = _TASKS_TO_MIXTURE[task]
+            mixture_dir = os.path.join(base_dir, mixture_folder)
+            os.makedirs(mixture_dir, exist_ok=True)
+            mixture_path = os.path.join(mixture_dir, filename)
+            mixture = _save_wav(mixture_path, seed)
+            sources = []
+            if task == "enh_both":
+                sources = [task_outputs["sep_clean"][1]]
+            else:
+                for speaker_id in range(num_speaker):
+                    source_dir = os.path.join(base_dir, f"s{speaker_id+1}")
+                    os.makedirs(source_dir, exist_ok=True)
+                    source_path = os.path.join(source_dir, filename)
+                    if os.path.exists(source_path):
+                        sources = task_outputs["sep_clean"][2]
+                        break
+                    else:
+                        source = _save_wav(source_path, seed)
+                        sources.append(source)
+                        seed += 1
+            task_outputs[task] = (_SAMPLE_RATE, mixture, sources)
+        mocked_data.append(task_outputs)
+
+    return mocked_data
+
+
+class TestLibriMix(TempDirMixin, TorchaudioTestCase):
+    backend = "default"
+
+    root_dir = None
+    samples_2spk = []
+    samples_3spk = []
+
+    @classmethod
+    def setUpClass(cls):
+        cls.root_dir = cls.get_base_temp_dir()
+        cls.samples_2spk = get_mock_dataset(cls.root_dir, 2)
+        cls.samples_3spk = get_mock_dataset(cls.root_dir, 3)
+
+    def _test_librimix(self, dataset, samples, task):
+        num_samples = 0
+        for i, (sample_rate, mixture, sources) in enumerate(dataset):
+            assert sample_rate == samples[i][task][0]
+            self.assertEqual(mixture, samples[i][task][1])
+            assert len(sources) == len(samples[i][task][2])
+            for j in range(len(sources)):
+                self.assertEqual(sources[j], samples[i][task][2][j])
+            num_samples += 1
+
+        assert num_samples == len(samples)
+
+    @parameterized.expand([("sep_clean"), ("enh_single",), ("enh_both",), ("sep_noisy",)])
+    def test_librimix_2speaker(self, task):
+        dataset = LibriMix(self.root_dir, num_speakers=2, sample_rate=_SAMPLE_RATE, task=task)
+        self._test_librimix(dataset, self.samples_2spk, task)
+
+    @parameterized.expand([("sep_clean"), ("enh_single",), ("enh_both",), ("sep_noisy",)])
+    def test_librimix_3speaker(self, task):
+        dataset = LibriMix(self.root_dir, num_speakers=3, sample_rate=_SAMPLE_RATE, task=task)
+        self._test_librimix(dataset, self.samples_3spk, task)

torchaudio/datasets/librimix.py

@@ -7,6 +7,13 @@ from torch.utils.data import Dataset
 
 SampleType = Tuple[int, torch.Tensor, List[torch.Tensor]]
 
+_TASKS_TO_MIXTURE = {
+    "sep_clean": "mix_clean",
+    "enh_single": "mix_single",
+    "enh_both": "mix_both",
+    "sep_noisy": "mix_both",
+}
+
 
 class LibriMix(Dataset):
     r"""*LibriMix* :cite:`cosentino2020librimix` dataset.
@@ -19,12 +26,17 @@ class LibriMix(Dataset):
         num_speakers (int, optional): The number of speakers, which determines the directories
             to traverse. The Dataset will traverse ``s1`` to ``sN`` directories to collect
             N source audios. (Default: 2)
-        sample_rate (int, optional): sample rate of audio files. The ``sample_rate`` determines
+        sample_rate (int, optional): Sample rate of audio files. The ``sample_rate`` determines
             which subdirectory the audio are fetched. If any of the audio has a different sample
             rate, raises ``ValueError``. Options: [8000, 16000] (Default: 8000)
-        task (str, optional): the task of LibriMix.
+        task (str, optional): The task of LibriMix.
             Options: [``"enh_single"``, ``"enh_both"``, ``"sep_clean"``, ``"sep_noisy"``]
             (Default: ``"sep_clean"``)
+        mode (str, optional): The mode when creating the mixture. If set to ``"min"``, the lengths of mixture
+            and sources are the minimum length of all sources. If set to ``"max"``, the lengths of mixture and
+            sources are zero padded to the maximum length of all sources.
+            Options: [``"min"``, ``"max"``]
+            (Default: ``"min"``)
 
     Note:
         The LibriMix dataset needs to be manually generated. Please check https://github.com/JorisCos/LibriMix
@@ -37,20 +49,26 @@ class LibriMix(Dataset):
         num_speakers: int = 2,
         sample_rate: int = 8000,
         task: str = "sep_clean",
+        mode: str = "min",
     ):
         self.root = Path(root) / f"Libri{num_speakers}Mix"
+        if mode not in ["max", "min"]:
+            raise ValueError(f'Expect ``mode`` to be one in ["min", "max"]. Found {mode}.')
         if sample_rate == 8000:
-            self.root = self.root / "wav8k/min" / subset
+            self.root = self.root / "wav8k" / mode / subset
         elif sample_rate == 16000:
-            self.root = self.root / "wav16k/min" / subset
+            self.root = self.root / "wav16k" / mode / subset
         else:
             raise ValueError(f"Unsupported sample rate. Found {sample_rate}.")
         self.sample_rate = sample_rate
         self.task = task
-        self.mix_dir = (self.root / f"mix_{task.split('_')[1]}").resolve()
+        self.mix_dir = (self.root / _TASKS_TO_MIXTURE[task]).resolve()
+        if task == "enh_both":
+            self.src_dirs = [(self.root / "mix_clean")]
+        else:
             self.src_dirs = [(self.root / f"s{i+1}").resolve() for i in range(num_speakers)]
 
-        self.files = [p.name for p in self.mix_dir.glob("*wav")]
+        self.files = [p.name for p in self.mix_dir.glob("*.wav")]
         self.files.sort()
 
     def _load_audio(self, path) -> torch.Tensor:
@@ -80,6 +98,7 @@ class LibriMix(Dataset):
 
         Args:
             key (int): The index of the sample to be loaded
+
         Returns:
             Tuple of the following items;