🐛 fix backend resolution

spleeter/audio/__init__.py

@@ -12,6 +12,11 @@
 
 from enum import Enum
 
+# pyright: reportMissingImports=false
+# pylint: disable=import-error
+import tensorflow as tf
+# pylint: enable=import-error
+
 __email__ = 'spleeter@deezer.com'
 __author__ = 'Deezer Research'
 __license__ = 'MIT License'
@@ -34,3 +39,12 @@ class STFTBackend(str, Enum):
     AUTO: str = 'auto'
     TENSORFLOW: str = 'tensorflow'
     LIBROSA: str = 'librosa'
+
+    def resolve(cls: type, backend: str) -> str:
+        if backend not in cls.__members__.items():
+            raise ValueError(f'Unsupported backend {backend}')
+        if backend == cls.AUTO:
+            if len(tf.config.list_physical_devices('GPU')):
+                return cls.TENSORFLOW
+            return STFTBackend.LIBROSA
+        return backend

spleeter/separator.py

@@ -129,8 +129,7 @@ class Separator(object):
         else:
             self._pool = None
         self._tasks = []
-        # NOTE: provide type check here ?
-        self._params['stft_backend'] = stft_backend
+        self._params['stft_backend'] = STFTBackend.resolve(stft_backend)
         self._data_generator = DataGenerator()
 
     def __del__(self) -> None:
@@ -333,11 +332,6 @@ class Separator(object):
                     (Optional) string describing the waveform (e.g. filename).
         """
         backend: str = self._params['stft_backend']
-        if backend == STFTBackend.AUTO:
-            if len(tf.config.list_physical_devices('GPU')):
-                backend = STFTBackend.TENSORFLOW
-            else:
-                backend = STFTBackend.LIBROSA
         if backend == STFTBackend.TENSORFLOW:
             return self._separate_tensorflow(waveform, audio_descriptor)
         elif backend == STFTBackend.LIBROSA: