feat: TTS 服务功能完善（任务管理、配置优化、客户端实现）

backend-single/tts-service/app.py

@@ -1,5 +1,6 @@
 import subprocess
 from pathlib import Path
+from typing import Optional
 
 from fastapi import FastAPI
 from pydantic import BaseModel, Field
@@ -16,6 +17,42 @@ class SynthesizeRequest(BaseModel):
     text: str = Field(min_length=1, max_length=5000)
     voice: str = "default_zh_female"
     outputPath: str
+    speechRate: Optional[float] = Field(default=0.92, ge=0.6, le=1.4)
+    pitch: Optional[float] = Field(default=0.0, ge=-20.0, le=20.0)
+    emotion: Optional[str] = "story"
+
+
+def clamp(value: float, minimum: float, maximum: float) -> float:
+    return max(minimum, min(maximum, value))
+
+
+def resolve_piper_args(request: SynthesizeRequest) -> list[str]:
+    speech_rate = clamp(float(request.speechRate or 0.92), 0.6, 1.4)
+    emotion = (request.emotion or "story").lower()
+    length_scale = round(1.0 / speech_rate, 2)
+    sentence_silence = 0.46
+    noise_scale = 0.64
+    noise_w = 0.72
+
+    if emotion in {"calm", "soft", "warm"}:
+        sentence_silence = 0.5
+        noise_scale = 0.58
+        noise_w = 0.68
+    elif emotion in {"story", "narration", "expressive"}:
+        sentence_silence = 0.48
+        noise_scale = 0.68
+        noise_w = 0.76
+
+    return [
+        "--sentence-silence",
+        str(sentence_silence),
+        "--length_scale",
+        str(length_scale),
+        "--noise_scale",
+        str(noise_scale),
+        "--noise_w",
+        str(noise_w),
+    ]
 
 
 @app.get("/health")
@@ -47,8 +84,7 @@ def synthesize(request: SynthesizeRequest):
                 str(PIPER_CONFIG),
                 "--output_file",
                 str(output),
-                "--sentence-silence",
-                "0.35",
+                *resolve_piper_args(request),
             ],
             input=request.text,
             text=True,