Upload tts-server.py

tts-server.py

@@ -10,9 +10,8 @@ from typing import Optional, List, Dict, Any, AsyncGenerator
 
 import aiohttp
 import websockets
-from fastapi import FastAPI, HTTPException, Header, Request, BackgroundTasks, Depends
+from fastapi import FastAPI, HTTPException, Header, Request, BackgroundTasks
 from fastapi.responses import StreamingResponse, JSONResponse
-from fastapi.security import HTTPBearer, HTTPAuthorizationCredentials
 from pydantic import BaseModel, Field
 import uvicorn
 from contextlib import asynccontextmanager
@@ -33,7 +32,6 @@ PORT = 1547
 HOST = "0.0.0.0"
 MODELS_FILE = "models.json"
 COOKIE_DIR = "cookie"  # Directory to store cookie txt files
-AUTH_PASSWORD = os.getenv("PASSWORD", "sk-wei123")
 
 # Initialize FastAPI
 @asynccontextmanager
@@ -52,12 +50,6 @@ async def lifespan(app: FastAPI):
     yield
 
 app = FastAPI(title="Doubao TTS OpenAI API Server", lifespan=lifespan)
-security = HTTPBearer()
-
-async def verify_token(credentials: HTTPAuthorizationCredentials = Depends(security)):
-    if credentials.credentials != AUTH_PASSWORD:
-        raise HTTPException(status_code=401, detail="Invalid authentication token")
-    return credentials.credentials
 
 # ==========================================
 # Cookie Manager
@@ -244,7 +236,7 @@ class DoubaoTTS:
         except Exception as e:
             logger.error(f"Failed to save models.json: {e}")
 
-    async def stream_audio(self, text: str, voice: str, format: str = "aac", speed: float = 1.0, pitch: float = 1.0) -> AsyncGenerator[bytes, None]:
+    async def stream_audio(self, text: str, voice: str, speed: float = 1.0, pitch: float = 1.0) -> AsyncGenerator[bytes, None]:
         """Connect to WebSocket and yield audio chunks with retry logic."""
         
         # Map OpenAI speed (0.25 - 4.0) to Doubao rate (-100 to 100)
@@ -262,7 +254,7 @@ class DoubaoTTS:
                 return
 
             params = self._get_common_params()
-            ws_url = f"{self.ws_url}?format={format}&speaker={voice}&speech_rate={doubao_rate}&pitch={doubao_pitch}{params}"
+            ws_url = f"{self.ws_url}?format=aac&speaker={voice}&speech_rate={doubao_rate}&pitch={doubao_pitch}{params}"
             
             headers = {
                 "Cookie": cookie,
@@ -271,9 +263,6 @@ class DoubaoTTS:
             }
 
             try:
-                # Try to use extra_headers, if it fails, it might be a version issue or specific environment quirk
-                # But standard websockets library uses extra_headers.
-                # Let's try to be safe and use the standard way.
                 async with websockets.connect(ws_url, extra_headers=headers) as ws:
                     # Send Text Event
                     msg = {
@@ -348,7 +337,7 @@ async def check_speech_endpoint():
     return {"status": "ok", "message": "Speech endpoint is ready"}
 
 @app.post("/v1/audio/speech")
-async def create_speech(req: OpenAIRequest, token: str = Depends(verify_token)):
+async def create_speech(req: OpenAIRequest):
     """
     OpenAI-compatible speech generation endpoint.
     """
@@ -359,15 +348,8 @@ async def create_speech(req: OpenAIRequest, token: str = Depends(verify_token)):
     if req.response_format == "mp3":
         media_type = "audio/mpeg"
     
-    # Determine format to request from Doubao
-    target_format = "aac"
-    if req.response_format == "mp3":
-        target_format = "mp3"
-    elif req.response_format == "pcm":
-        target_format = "pcm"
-
     return StreamingResponse(
-        engine.stream_audio(req.input, req.voice, target_format, req.speed, req.pitch),
+        engine.stream_audio(req.input, req.voice, req.speed, req.pitch),
         media_type=media_type
     )