Merge pull request coqui-ai#2236 from coqui-ai/data_checker

Add ability to read audio from paths via util.audio

.github/workflows/lint.yml

@@ -18,7 +18,9 @@ jobs:
           python-version: ${{ matrix.pyver }}
       - name: Run training unittests
         run: |
+          sudo apt-get install -y --no-install-recommends libopusfile0 libopus-dev libopusfile-dev
           ./ci_scripts/train-unittests.sh
+
   pre-commit-checks:
     name: "Lin|Pre-commit checks"
     runs-on: ubuntu-20.04

Dockerfile.build

@@ -19,6 +19,7 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
         libbz2-dev \
         libboost-all-dev \
         libgsm1-dev \
+        libopusfile0 \
         libltdl-dev \
         liblzma-dev \
         libmagic-dev \

Dockerfile.train

@@ -36,6 +36,7 @@ RUN apt-get update && \
         git \
         wget \
         libopus0 \
+        libvorbisfile3 \
         libopusfile0 \
         libsndfile1 \
         sox \

tests/test_audio_utils.py

@@ -0,0 +1,54 @@
+import unittest
+from argparse import Namespace
+from pathlib import Path
+
+from coqui_stt_training.util import audio
+
+
+def from_here(path):
+    here = Path(__file__)
+    return here.parent / path
+
+
+class TestValidateReadAudio(unittest.TestCase):
+    def test_flac(self):
+        audio_path = from_here("../data/smoke_test/LDC93S1.flac")
+        audio_type = audio.get_loadable_audio_type_from_extension(audio_path.suffix)
+        result = audio.read_audio(audio_type, str(audio_path))
+        self.assertIsInstance(result[0], audio.AudioFormat)
+
+    def test_wav(self):
+        audio_path = from_here("../data/smoke_test/LDC93S1.wav")
+        audio_type = audio.get_loadable_audio_type_from_extension(audio_path.suffix)
+        result = audio.read_audio(audio_type, str(audio_path))
+        self.assertIsInstance(result[0], audio.AudioFormat)
+
+    def test_ogg_opus(self):
+        audio_path = from_here("../data/smoke_test/LDC93S1.opus")
+        audio_type = audio.get_loadable_audio_type_from_extension(audio_path.suffix)
+        result = audio.read_audio(audio_type, str(audio_path))
+        self.assertIsInstance(result[0], audio.AudioFormat)
+
+
+class TestValidateReadDuration(unittest.TestCase):
+    def test_flac(self):
+        audio_path = from_here("../data/smoke_test/LDC93S1.flac")
+        audio_type = audio.get_loadable_audio_type_from_extension(audio_path.suffix)
+        result = audio.read_duration(audio_type, str(audio_path))
+        self.assertIsInstance(result, float)
+
+    def test_wav(self):
+        audio_path = from_here("../data/smoke_test/LDC93S1.wav")
+        audio_type = audio.get_loadable_audio_type_from_extension(audio_path.suffix)
+        result = audio.read_duration(audio_type, str(audio_path))
+        self.assertIsInstance(result, float)
+
+    def test_ogg_opus(self):
+        audio_path = from_here("../data/smoke_test/LDC93S1.opus")
+        audio_type = audio.get_loadable_audio_type_from_extension(audio_path.suffix)
+        result = audio.read_duration(audio_type, str(audio_path))
+        self.assertIsInstance(result, float)
+
+
+if __name__ == "__main__":
+    unittest.main()