Update app.py

app.py

@@ -7,8 +7,11 @@ import torch
 from gtts import gTTS
 import os
 import yt_dlp
-import re
-
+#import re
+import io
+import numpy as np 
+import scipy.io.wavfile as wav 
+#from fastapi.responses import JSONResponse 
 
 hf_token = os.getenv("HF_TOKEN")
 app = FastAPI()
@@ -75,19 +78,15 @@ async def chat(request: ChatRequest):
 
 # Endpoint voice chat + TTS
 """@app.post("/voice_chat")
-    async def voice_chat(file: UploadFile = File(...)):
+async def voice_chat(file: UploadFile = File(...)):
     file_location = f"temp_{file.filename}"
     with open(file_location, "wb") as f:
         f.write(await file.read())
 
     result = whisper_model.transcribe(file_location, language="vi")
     user_text = result["text"]
-    os.remove(file_location)"""
-import io
-import numpy as np 
-import scipy.io.wavfile as wav 
-from fastapi import FastAPI, Request
-from fastapi.responses import JSONResponse 
+    os.remove(file_location)
+"""
 @app.post("/voice_chat")
 async def voice_chat(request: Request): 
     # Đọc dữ liệu âm thanh thô từ ESP32 gửi lên
@@ -97,14 +96,15 @@ async def voice_chat(request: Request):
     audio_np = np.frombuffer(raw_audio, dtype=np.int16)
     # Chuyển thành file WAV trong bộ nhớ
     wav_io = io.BytesIO()
-    wav.write(wav_io, sample_rate, audio_np) wav_io.seek(0) 
+    wav.write(wav_io, sample_rate, audio_np)
+    wav_io.seek(0) 
     # Lưu file WAV tạm để dùng với Whisper 
     with open("temp_audio.wav", "wb") as f: f.write(wav_io.read()) 
     # Gọi Whisper để chuyển âm thanh thành văn bản 
     import whisper model = whisper.load_model("base") 
     result = model.transcribe("temp_audio.wav", language="vi") 
     user_text = result["text"] 
-    return JSONResponse(content={"text": user_text})
+    #return JSONResponse(content={"text": user_text})
 # Kiểm tra yêu cầu mở nhạc
     if any(kw in user_text.lower() for kw in ["nghe nhạc", "mở bài hát", "bài hát", "bài"]):
         song_name = extract_song_name(user_text)
@@ -129,7 +129,7 @@ async def voice_chat(request: Request):
     return {
         "user_text": user_text,
         "response": response_text,
-        "audio_url": f"/get_audio"
+        "audio_url": "/get_audio"
     }
 
 # Endpoint trả về file âm thanh