fix bugs in test_all

tests/test_all.py

@@ -24,14 +24,10 @@
     'taper_asymmetric': True,
     'downsample_factor': 4,
     'padding': 'valid',
-    'DEVICE_compute': 'cpu',
-    'DEVICE_return': 'cpu',
-    'batch_size': 1000,
-    'return_complex': False,
+    'take_abs': True,
     'filters': None,
     'verbose': False,
     'plot_pref': False,
-    'progressBar': False,
 }
 
 
@@ -73,7 +69,7 @@ def test_peak_in_spectrogram_at_sine_wave_frequency(
     # Apply the VQT to this sine wave
     spectrogram, _, freqs = v(input_signal)
     # Convert freqs to a tensor for easier handling
-    freqs_tensor = torch.tensor(freqs, dtype=torch.float32)
+    freqs_tensor = torch.as_tensor(freqs, dtype=torch.float32)
     # Locate the peak in the spectrogram
     peak_index = torch.argmax(spectrogram[0], dim=0)  # Assuming the output shape is (n_channels, n_freq_bins, time_bins) 
     assert torch.all(peak_index == peak_index[0]), "Expected a single peak in the spectrogram"
@@ -114,13 +110,8 @@ def test_constant_signal_transformation():
     taper_asymmetric=st.booleans(),
     downsample_factor=st.integers(min_value=1, max_value=100),
     padding=st.sampled_from(['valid', 'same']),
-    # DEVICE_compute=st.sampled_from(['cpu', 'cuda']),
-    # DEVICE_return=st.sampled_from(['cpu', 'cuda']),
-    batch_size=st.integers(min_value=1, max_value=1000),
-    # return_complex=st.booleans(),
     # filters=st.none() | st.sampled_from([None]),
     # plot_pref=st.booleans(),
-    # progressBar=st.booleans(),
 
     n_channels=st.integers(min_value=1, max_value=100),
     n_dim=st.integers(min_value=1, max_value=2),
@@ -137,13 +128,9 @@ def test_vqt_params(
     taper_asymmetric,
     downsample_factor,
     padding,
-    # DEVICE_compute,
-    # DEVICE_return,
-    batch_size,
-    # return_complex,
+    # take_abs,
     # filters,
     # plot_pref,
-    # progressBar,
 
     n_channels,
     n_dim,
@@ -160,13 +147,9 @@ def test_vqt_params(
         'taper_asymmetric': taper_asymmetric,
         'downsample_factor': downsample_factor,
         'padding': padding,
-        # 'DEVICE_compute': DEVICE_compute,
-        # 'DEVICE_return': DEVICE_return,
-        'batch_size': batch_size,
-        # 'return_complex': return_complex,
+        # 'take_abs': take_abs,
         # 'filters': filters,
         # 'plot_pref': plot_pref,
-        # 'progressBar': progressBar,
     }
     params.update({k: v for k, v in params_vqt.items() if k not in params})
     v = vqt.VQT(**params)
@@ -182,9 +165,9 @@ def test_vqt_params(
     assert output is not None, "Output should not be None"
     # Check output shape
     assert output.shape[1] == params['n_freq_bins'], "VQT output shape does not match the number of frequency bins"
-    # Check all output is real if return_complex is False
-    if not params['return_complex']:
-        assert torch.all(torch.isreal(output)), "VQT output should be real if return_complex is False"
+    # Check all output is real if take_abs is True
+    if params['take_abs']:
+        assert torch.all(torch.isreal(output)), "VQT output should be real if take_abs is True"
 
 
 def test_vqt_filters():
@@ -202,7 +185,7 @@ def test_vqt_filters():
         taper_asymmetric=params['taper_asymmetric'],
         plot_pref=params['plot_pref']
     )
-    filters, wins = filters.numpy(), wins.numpy()
+    filters, freqs, wins = filters.numpy(), freqs.numpy(), wins.numpy()
 
     ## Filters, freqs, and wins should all have the following properties:
     ### They are not all zeros, not all ones, and do not contain NaNs or infinities