update code
Browse files
app.py
CHANGED
@@ -56,40 +56,63 @@ class AudioProcessor:
|
|
56 |
|
57 |
def _process_with_ffmpeg(self, audio_input):
|
58 |
# Your existing FFmpeg processing
|
59 |
-
pass
|
60 |
-
|
61 |
-
def _process_with_sounddevice(self, audio_input):
|
62 |
-
# Process using sounddevice
|
63 |
-
duration = 5 # seconds
|
64 |
-
print(f"Recording with sounddevice (rate={self.sample_rate})...")
|
65 |
-
audio_data = sd.rec(int(duration * self.sample_rate),
|
66 |
-
samplerate=self.sample_rate,
|
67 |
-
channels=1)
|
68 |
-
sd.wait()
|
69 |
-
return (audio_data.flatten(), self.sample_rate)
|
70 |
-
|
71 |
-
def _process_with_librosa(self, audio_input):
|
72 |
-
# Process using librosa
|
73 |
-
import librosa
|
74 |
if isinstance(audio_input, tuple):
|
75 |
return audio_input
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
# Handle other input types
|
80 |
with tempfile.NamedTemporaryFile(suffix='.wav', delete=False) as tmp:
|
81 |
tmp.write(audio_input)
|
82 |
tmp.flush()
|
83 |
-
|
|
|
|
|
|
|
|
|
84 |
os.unlink(tmp.name)
|
85 |
-
return (
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
86 |
|
87 |
def _process_fallback(self, audio_input):
|
88 |
-
# Simple numpy fallback
|
89 |
if isinstance(audio_input, tuple):
|
90 |
return audio_input
|
91 |
-
return (np.random.random(16000), 16000 #
|
92 |
-
|
93 |
# Modified Interface with Audio Debugging
|
94 |
def create_debug_interface():
|
95 |
audio_processor = AudioProcessor()
|
|
|
56 |
|
57 |
def _process_with_ffmpeg(self, audio_input):
|
58 |
# Your existing FFmpeg processing
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
59 |
if isinstance(audio_input, tuple):
|
60 |
return audio_input
|
61 |
+
try:
|
62 |
+
import ffmpeg
|
63 |
+
# Process audio file with ffmpeg
|
|
|
64 |
with tempfile.NamedTemporaryFile(suffix='.wav', delete=False) as tmp:
|
65 |
tmp.write(audio_input)
|
66 |
tmp.flush()
|
67 |
+
out, _ = (
|
68 |
+
ffmpeg.input(tmp.name)
|
69 |
+
.output('pipe:', format='f32le', ac=1, ar=self.sample_rate)
|
70 |
+
.run(capture_stdout=True)
|
71 |
+
)
|
72 |
os.unlink(tmp.name)
|
73 |
+
return (np.frombuffer(out, dtype=np.float32), self.sample_rate)
|
74 |
+
except Exception as e:
|
75 |
+
raise Exception(f"FFmpeg processing failed: {str(e)}")
|
76 |
+
|
77 |
+
def _process_with_sounddevice(self, audio_input):
|
78 |
+
# Process using sounddevice
|
79 |
+
if isinstance(audio_input, tuple):
|
80 |
+
return audio_input
|
81 |
+
try:
|
82 |
+
duration = 5 # seconds
|
83 |
+
print(f"Recording with sounddevice (rate={self.sample_rate})...")
|
84 |
+
audio_data = sd.rec(int(duration * self.sample_rate),
|
85 |
+
samplerate=self.sample_rate,
|
86 |
+
channels=1)
|
87 |
+
sd.wait()
|
88 |
+
return (audio_data.flatten(), self.sample_rate)
|
89 |
+
except Exception as e:
|
90 |
+
raise Exception(f"Sounddevice processing failed: {str(e)}")
|
91 |
+
|
92 |
+
def _process_with_librosa(self, audio_input):
|
93 |
+
# Process using librosa
|
94 |
+
try:
|
95 |
+
import librosa
|
96 |
+
if isinstance(audio_input, tuple):
|
97 |
+
return audio_input
|
98 |
+
elif isinstance(audio_input, str):
|
99 |
+
return librosa.load(audio_input, sr=self.sample_rate)
|
100 |
+
else:
|
101 |
+
# Handle other input types
|
102 |
+
with tempfile.NamedTemporaryFile(suffix='.wav', delete=False) as tmp:
|
103 |
+
tmp.write(audio_input)
|
104 |
+
tmp.flush()
|
105 |
+
data, sr = librosa.load(tmp.name, sr=self.sample_rate)
|
106 |
+
os.unlink(tmp.name)
|
107 |
+
return (data, sr)
|
108 |
+
except Exception as e:
|
109 |
+
raise Exception(f"Librosa processing failed: {str(e)}")
|
110 |
|
111 |
def _process_fallback(self, audio_input):
|
112 |
+
# Simple numpy fallback with proper error handling
|
113 |
if isinstance(audio_input, tuple):
|
114 |
return audio_input
|
115 |
+
return (np.random.random(16000), 16000) # Now properly closed
|
|
|
116 |
# Modified Interface with Audio Debugging
|
117 |
def create_debug_interface():
|
118 |
audio_processor = AudioProcessor()
|