Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -17,72 +17,60 @@ except Exception as e:
|
|
17 |
else:
|
18 |
model_load_error = None
|
19 |
|
20 |
-
# ---
|
21 |
-
def predict_emotion(audio_file):
|
22 |
-
if classifier is None:
|
23 |
-
return {"error": f"Model load failed: {model_load_error}"}
|
24 |
-
if audio_file is None:
|
25 |
-
return {"error": "No audio input provided."}
|
26 |
-
|
27 |
-
try:
|
28 |
-
if isinstance(audio_file, str):
|
29 |
-
audio_path = audio_file
|
30 |
-
elif isinstance(audio_file, tuple):
|
31 |
-
sample_rate, audio_array = audio_file
|
32 |
-
temp_audio_path = "temp_audio.wav"
|
33 |
-
sf.write(temp_audio_path, audio_array, sample_rate)
|
34 |
-
audio_path = temp_audio_path
|
35 |
-
else:
|
36 |
-
return {"error": f"Unsupported input type: {type(audio_file)}"}
|
37 |
-
|
38 |
-
results = classifier(audio_path, top_k=5)
|
39 |
-
return {item['label']: round(item['score'], 3) for item in results}
|
40 |
-
except Exception as e:
|
41 |
-
return {"error": f"Prediction error: {str(e)}"}
|
42 |
-
finally:
|
43 |
-
if 'temp_audio_path' in locals() and os.path.exists(temp_audio_path):
|
44 |
-
os.remove(temp_audio_path)
|
45 |
-
|
46 |
-
# --- FastAPI App for Base64 API ---
|
47 |
app = FastAPI()
|
48 |
|
49 |
@app.post("/api/predict/")
|
50 |
async def predict_emotion_api(request: Request):
|
51 |
if classifier is None:
|
52 |
-
return JSONResponse(content={"error": f"Model
|
53 |
|
54 |
try:
|
55 |
body = await request.json()
|
56 |
-
|
57 |
-
|
58 |
-
return JSONResponse(content={"error": "Missing 'data' field with base64 audio."}, status_code=400)
|
59 |
|
60 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
61 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_file:
|
62 |
temp_file.write(audio_data)
|
63 |
temp_audio_path = temp_file.name
|
64 |
|
65 |
results = classifier(temp_audio_path, top_k=5)
|
66 |
-
os.unlink(temp_audio_path)
|
|
|
|
|
|
|
67 |
|
68 |
-
return {item['label']: round(item['score'], 3) for item in results}
|
69 |
except Exception as e:
|
70 |
-
return JSONResponse(content={"error": f"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
71 |
|
72 |
-
# --- Gradio UI ---
|
73 |
gradio_interface = gr.Interface(
|
74 |
-
fn=
|
75 |
inputs=gr.Audio(sources=["microphone", "upload"], type="filepath", label="Upload Audio or Record"),
|
76 |
outputs=gr.Label(num_top_classes=5, label="Emotion Predictions"),
|
77 |
title="Audio Emotion Detector",
|
78 |
-
description="
|
79 |
allow_flagging="never"
|
80 |
)
|
81 |
|
82 |
-
# --- Mount Gradio
|
83 |
-
|
84 |
-
|
85 |
-
# --- Launch for local/dev use only ---
|
86 |
-
if __name__ == "__main__":
|
87 |
-
gradio_interface.queue()
|
88 |
-
uvicorn.run(app, host="0.0.0.0", port=7860)
|
|
|
17 |
else:
|
18 |
model_load_error = None
|
19 |
|
20 |
+
# --- FastAPI App for a dedicated, robust API ---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
app = FastAPI()
|
22 |
|
23 |
@app.post("/api/predict/")
|
24 |
async def predict_emotion_api(request: Request):
|
25 |
if classifier is None:
|
26 |
+
return JSONResponse(content={"error": f"Model is not loaded: {model_load_error}"}, status_code=503)
|
27 |
|
28 |
try:
|
29 |
body = await request.json()
|
30 |
+
# The JS FileReader sends a string like "data:audio/wav;base64,AABBCC..."
|
31 |
+
base64_with_prefix = body.get("data")
|
|
|
32 |
|
33 |
+
if not base64_with_prefix:
|
34 |
+
return JSONResponse(content={"error": "Missing 'data' field in request body."}, status_code=400)
|
35 |
+
|
36 |
+
# Robustly strip the prefix to get the pure base64 data
|
37 |
+
try:
|
38 |
+
# Find the comma that separates the prefix from the data
|
39 |
+
header, encoded = base64_with_prefix.split(",", 1)
|
40 |
+
audio_data = base64.b64decode(encoded)
|
41 |
+
except (ValueError, TypeError):
|
42 |
+
return JSONResponse(content={"error": "Invalid base64 data format."}, status_code=400)
|
43 |
+
|
44 |
+
# Write to a temporary file for the pipeline
|
45 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_file:
|
46 |
temp_file.write(audio_data)
|
47 |
temp_audio_path = temp_file.name
|
48 |
|
49 |
results = classifier(temp_audio_path, top_k=5)
|
50 |
+
os.unlink(temp_audio_path) # Clean up the temp file
|
51 |
+
|
52 |
+
# Return a successful response
|
53 |
+
return JSONResponse(content={"data": results})
|
54 |
|
|
|
55 |
except Exception as e:
|
56 |
+
return JSONResponse(content={"error": f"Internal server error during prediction: {str(e)}"}, status_code=500)
|
57 |
+
|
58 |
+
# --- Gradio UI function (optional, for the direct Space page) ---
|
59 |
+
def gradio_predict_wrapper(audio_file):
|
60 |
+
# This is just for the UI on the Hugging Face page itself
|
61 |
+
if audio_file is None: return {"error": "Please provide an audio file."}
|
62 |
+
results = classifier(audio_file, top_k=5)
|
63 |
+
return {item['label']: round(item['score'], 3) for item in results}
|
64 |
|
|
|
65 |
gradio_interface = gr.Interface(
|
66 |
+
fn=gradio_predict_wrapper,
|
67 |
inputs=gr.Audio(sources=["microphone", "upload"], type="filepath", label="Upload Audio or Record"),
|
68 |
outputs=gr.Label(num_top_classes=5, label="Emotion Predictions"),
|
69 |
title="Audio Emotion Detector",
|
70 |
+
description="This UI is for direct demonstration. The primary API is at /api/predict/",
|
71 |
allow_flagging="never"
|
72 |
)
|
73 |
|
74 |
+
# --- Mount the Gradio UI onto the FastAPI app ---
|
75 |
+
# The API at /api/predict/ will work even if the UI is at a different path.
|
76 |
+
app = gr.mount_gradio_app(app, gradio_interface, path="/ui")
|
|
|
|
|
|
|
|