Enhance voice configuration with idle prompt features and new TTS settings

- Added idle prompt timeout, maximum count, and text to multiple voice configuration files to improve user interaction during idle periods.
- Updated greeting mode to 'fastgpt_opener' in relevant configurations for a more dynamic greeting experience.
- Introduced a new voice configuration file for xfyun TTS, including detailed service settings and parameters.
- Refactored the pipeline to handle idle prompts and user turn events, ensuring smoother interaction flow.
- Adjusted the VAD and turn configurations to accommodate new idle prompt features.

src/voice/config.py

@@ -67,6 +67,12 @@ class VADConfig:
 class TurnConfig:
     vad: VADConfig = field(default_factory=VADConfig)
     user_speech_timeout_sec: float = 1.0
+    idle_prompt_timeout_sec: float = 0.0
+    idle_prompt_max_count: int = 1
+    idle_prompt_text: str = (
+        "我先停在这里。你可以继续说你的想法，"
+        "或者让我根据刚才的内容帮你整理下一步。"
+    )
     interruption_min_chars: int = 3
     interruption_use_interim: bool = True
     interruption_short_replies: list[str] = field(
@@ -209,8 +215,10 @@ def config_from_dict(data: dict) -> EngineConfig:
     agent = _dict(data.get("agent"))
     if agent.get("greeting") == "":
         agent["greeting"] = None
-    if agent.get("greeting_mode") not in (None, "generated", "fixed", "off"):
-        raise ValueError("agent.greeting_mode must be one of: generated, fixed, off")
+    if agent.get("greeting_mode") not in (None, "generated", "fixed", "off", "fastgpt_opener"):
+        raise ValueError(
+            "agent.greeting_mode must be one of: generated, fixed, off, fastgpt_opener"
+        )
     response_state = ResponseStateConfig(**_dict(agent.pop("response_state")))
     if response_state.max_prefix_chars < 1:
         raise ValueError("agent.response_state.max_prefix_chars must be greater than 0")
@@ -231,6 +239,10 @@ def config_from_dict(data: dict) -> EngineConfig:
         llm["app_id"] = None
     if not isinstance(llm.get("variables"), dict):
         llm["variables"] = {}
+    if agent.get("greeting_mode") == "fastgpt_opener" and llm["provider"] != "fastgpt":
+        raise ValueError(
+            "agent.greeting_mode='fastgpt_opener' requires services.llm.provider='fastgpt'"
+        )
 
     turn = _dict(data.get("turn"))
     vad = _dict(turn.get("vad"))
@@ -244,6 +256,15 @@ def config_from_dict(data: dict) -> EngineConfig:
             user_speech_timeout_sec=float(
                 turn.get("user_speech_timeout_sec", TurnConfig().user_speech_timeout_sec)
             ),
+            idle_prompt_timeout_sec=float(
+                turn.get("idle_prompt_timeout_sec", TurnConfig().idle_prompt_timeout_sec)
+            ),
+            idle_prompt_max_count=int(
+                turn.get("idle_prompt_max_count", TurnConfig().idle_prompt_max_count)
+            ),
+            idle_prompt_text=str(
+                turn.get("idle_prompt_text", TurnConfig().idle_prompt_text)
+            ),
             interruption_min_chars=int(
                 turn.get("interruption_min_chars", TurnConfig().interruption_min_chars)
             ),