openspeechapi 0.1.0__py3-none-any.whl

openspeech/engine_registry.yaml

@@ -0,0 +1,347 @@
+# Engine Registry — lists all available engines with basic metadata.
+# Provider code supplies default_settings and field_options at runtime.
+# This file can be updated from a remote URL in the future.
+version: 1
+
+engines:
+  # ---- Cloud STT ----
+  - name: openai-stt
+    vendor: openai
+    provider: openai-stt
+    type: stt
+    category: cloud
+    display_name: "OpenAI STT"
+    description: "OpenAI Whisper API (also compatible with Groq, SiliconFlow)"
+    default_alias: openai_stt
+    default_exec_mode: remote
+    pip_extras: [openai]
+
+  - name: deepgram-stt
+    vendor: deepgram
+    provider: deepgram
+    type: stt
+    category: cloud
+    display_name: "Deepgram STT"
+    description: "Deepgram Nova STT — fast, accurate, word-level timestamps"
+    default_alias: deepgram_stt
+    default_exec_mode: remote
+    pip_extras: [deepgram]
+
+  - name: elevenlabs-stt
+    vendor: elevenlabs
+    provider: elevenlabs-stt
+    type: stt
+    category: cloud
+    display_name: "ElevenLabs STT"
+    description: "ElevenLabs Scribe Speech-to-Text"
+    default_alias: elevenlabs_stt
+    default_exec_mode: remote
+    pip_extras: [elevenlabs-stt]
+
+  - name: deepgram-tts
+    vendor: deepgram
+    provider: deepgram-tts
+    type: tts
+    category: cloud
+    display_name: "Deepgram TTS"
+    description: "Deepgram Aura TTS — fast, natural text-to-speech"
+    default_alias: deepgram_tts
+    default_exec_mode: remote
+
+  - name: google-stt
+    vendor: google
+    provider: google-stt
+    type: stt
+    category: cloud
+    display_name: "Google STT"
+    description: "Google Cloud Speech-to-Text v1"
+    default_alias: google_stt
+    default_exec_mode: remote
+
+  - name: azure-stt
+    vendor: azure
+    provider: azure-stt
+    type: stt
+    category: cloud
+    display_name: "Azure STT"
+    description: "Azure Cognitive Services Speech-to-Text"
+    default_alias: azure_stt
+    default_exec_mode: remote
+
+  - name: assemblyai-stt
+    vendor: assemblyai
+    provider: assemblyai-stt
+    type: stt
+    category: cloud
+    display_name: "AssemblyAI STT"
+    description: "AssemblyAI STT — async upload + poll"
+    default_alias: assemblyai_stt
+    default_exec_mode: remote
+
+  - name: volcengine-stt
+    vendor: volcengine
+    provider: volcengine-stt
+    type: stt
+    category: cloud
+    display_name: "Volcengine STT"
+    description: "Volcengine (火山引擎) Speech-to-Text"
+    default_alias: volcengine_stt
+    default_exec_mode: remote
+
+  - name: alibaba-stt
+    vendor: alibaba
+    provider: alibaba-stt
+    type: stt
+    category: cloud
+    display_name: "Alibaba STT"
+    description: "Alibaba Cloud (阿里云百炼) STT — OpenAI-compatible"
+    default_alias: alibaba_stt
+    default_exec_mode: remote
+
+  - name: tencent-stt
+    vendor: tencent
+    provider: tencent-stt
+    type: stt
+    category: cloud
+    display_name: "Tencent STT"
+    description: "Tencent Cloud (腾讯云) Speech-to-Text"
+    default_alias: tencent_stt
+    default_exec_mode: remote
+
+  - name: baidu-stt
+    vendor: baidu
+    provider: baidu-stt
+    type: stt
+    category: cloud
+    display_name: "Baidu STT"
+    description: "Baidu (百度) Speech-to-Text"
+    default_alias: baidu_stt
+    default_exec_mode: remote
+
+  - name: iflytek-stt
+    vendor: iflytek
+    provider: iflytek-stt
+    type: stt
+    category: cloud
+    display_name: "iFlytek STT"
+    description: "iFlytek (科大讯飞) Speech-to-Text"
+    default_alias: iflytek_stt
+    default_exec_mode: remote
+    pip_extras: [iflytek-stt]
+
+  # ---- Cloud TTS ----
+  - name: openai-tts
+    vendor: openai
+    provider: openai-tts
+    type: tts
+    category: cloud
+    display_name: "OpenAI TTS"
+    description: "OpenAI TTS API (also compatible with Groq, SiliconFlow, Minimax)"
+    default_alias: openai_tts
+    default_exec_mode: remote
+    pip_extras: [openai]
+
+  - name: elevenlabs-tts
+    vendor: elevenlabs
+    provider: elevenlabs
+    type: tts
+    category: cloud
+    display_name: "ElevenLabs TTS"
+    description: "ElevenLabs TTS — high quality, voice cloning"
+    default_alias: elevenlabs
+    default_exec_mode: remote
+    pip_extras: [elevenlabs-tts]
+
+  - name: minimax-tts
+    vendor: minimax
+    provider: minimax
+    type: tts
+    category: cloud
+    display_name: "MiniMax TTS"
+    description: "MiniMax TTS — 300+ voices, OpenAI-compatible"
+    default_alias: minimax
+    default_exec_mode: remote
+
+  - name: google-tts
+    vendor: google
+    provider: google-tts
+    type: tts
+    category: cloud
+    display_name: "Google TTS"
+    description: "Google Cloud Text-to-Speech"
+    default_alias: google_tts
+    default_exec_mode: remote
+
+  - name: azure-tts
+    vendor: azure
+    provider: azure-tts
+    type: tts
+    category: cloud
+    display_name: "Azure TTS"
+    description: "Azure Cognitive Services Text-to-Speech"
+    default_alias: azure_tts
+    default_exec_mode: remote
+
+  - name: volcengine-tts
+    vendor: volcengine
+    provider: volcengine-tts
+    type: tts
+    category: cloud
+    display_name: "Volcengine TTS"
+    description: "Volcengine (火山引擎) Text-to-Speech"
+    default_alias: volcengine_tts
+    default_exec_mode: remote
+
+  - name: alibaba-tts
+    vendor: alibaba
+    provider: alibaba-tts
+    type: tts
+    category: cloud
+    display_name: "Alibaba TTS"
+    description: "Alibaba Cloud (阿里云百炼) TTS — OpenAI-compatible"
+    default_alias: alibaba_tts
+    default_exec_mode: remote
+
+  - name: tencent-tts
+    vendor: tencent
+    provider: tencent-tts
+    type: tts
+    category: cloud
+    display_name: "Tencent TTS"
+    description: "Tencent Cloud (腾讯云) Text-to-Speech"
+    default_alias: tencent_tts
+    default_exec_mode: remote
+
+  - name: baidu-tts
+    vendor: baidu
+    provider: baidu-tts
+    type: tts
+    category: cloud
+    display_name: "Baidu TTS"
+    description: "Baidu (百度) Text-to-Speech"
+    default_alias: baidu_tts
+    default_exec_mode: remote
+
+  - name: iflytek-tts
+    vendor: iflytek
+    provider: iflytek-tts
+    type: tts
+    category: cloud
+    display_name: "iFlytek TTS"
+    description: "iFlytek (科大讯飞) Text-to-Speech"
+    default_alias: iflytek_tts
+    default_exec_mode: remote
+    pip_extras: [iflytek-tts]
+
+  # ---- Local STT ----
+  - name: faster-whisper
+    provider: faster-whisper
+    type: stt
+    category: local
+    display_name: "Faster Whisper"
+    description: "Faster Whisper — CTranslate2-based, fast local STT"
+    default_alias: stt_fw
+    default_exec_mode: subprocess
+    pip_extras: [faster-whisper]
+
+  - name: whisper
+    provider: whisper
+    type: stt
+    category: local
+    display_name: "Whisper"
+    description: "OpenAI Whisper — original local STT model"
+    default_alias: stt_whisper
+    default_exec_mode: subprocess
+    pip_extras: [whisper]
+
+  - name: whisperlivekit
+    provider: whisperlivekit-stt
+    type: stt
+    category: local
+    display_name: "WhisperLiveKit"
+    description: "WhisperLiveKit — streaming STT with MLX backend"
+    default_alias: stt_wlk
+    default_exec_mode: local
+    pip_extras: [whisperlivekit]
+
+  - name: sherpa-onnx
+    provider: sherpa-onnx-stt
+    type: stt
+    category: local
+    display_name: "Sherpa ONNX"
+    description: "Sherpa-ONNX — lightweight streaming STT"
+    default_alias: stt_sherpa
+    default_exec_mode: local
+
+  # ---- Local TTS ----
+  - name: fish-speech
+    provider: fish-speech
+    type: tts
+    category: local
+    display_name: "Fish Speech"
+    description: "Fish Speech — high quality local TTS"
+    default_alias: fish_speech
+    default_exec_mode: local
+    pip_extras: [fish-speech]
+
+  # ---- macOS Native ----
+  - name: macos-stt
+    provider: macos-stt
+    type: stt
+    category: native
+    display_name: "macOS STT"
+    description: "macOS SFSpeechRecognizer — built-in speech recognition"
+    default_alias: macos_stt
+    default_exec_mode: in_process
+    platforms: [darwin]
+
+  - name: macos-tts
+    provider: macos-say
+    type: tts
+    category: native
+    display_name: "macOS TTS"
+    description: "macOS say command — built-in text-to-speech"
+    default_alias: macos_tts
+    default_exec_mode: in_process
+    platforms: [darwin]
+
+  # ---- Windows Native ----
+  - name: windows-stt
+    provider: windows-stt
+    type: stt
+    category: native
+    display_name: "Windows STT"
+    description: "Windows System.Speech — built-in speech recognition"
+    default_alias: windows_stt
+    default_exec_mode: in_process
+    platforms: [win32]
+
+  - name: windows-tts
+    provider: windows-tts
+    type: tts
+    category: native
+    display_name: "Windows TTS"
+    description: "Windows SAPI5 — built-in text-to-speech"
+    default_alias: windows_tts
+    default_exec_mode: in_process
+    platforms: [win32]
+    pip_extras: [windows-tts]
+
+  # ---- Cross-platform Native (auto-detect) ----
+  - name: native-stt
+    provider: native-stt
+    type: stt
+    category: native
+    display_name: "Native STT"
+    description: "Platform native STT (macOS SFSpeech / Windows System.Speech)"
+    default_alias: native_stt
+    default_exec_mode: in_process
+
+  - name: native-tts
+    provider: native-tts
+    type: tts
+    category: native
+    display_name: "Native TTS"
+    description: "Platform native TTS (macOS say / Windows SAPI5)"
+    default_alias: native_tts
+    default_exec_mode: in_process

openspeech/exceptions.py

@@ -0,0 +1,51 @@
+"""OpenSpeech exception hierarchy."""
+
+
+class OpenSpeechError(Exception):
+    """Base exception for all OpenSpeech errors."""
+
+
+class ProviderError(OpenSpeechError):
+    """Error from a provider invocation."""
+
+    def __init__(self, provider_name: str, original_type: str, message: str) -> None:
+        self.provider_name = provider_name
+        self.original_type = original_type
+        super().__init__(f"[{provider_name}] {original_type}: {message}")
+
+
+class ProviderCrashedError(OpenSpeechError):
+    """Provider subprocess crashed during invocation."""
+
+    def __init__(self, provider_name: str) -> None:
+        self.provider_name = provider_name
+        super().__init__(f"Provider '{provider_name}' crashed during invocation")
+
+
+class ProviderUnavailableError(OpenSpeechError):
+    """Provider exhausted restart attempts and is unavailable."""
+
+    def __init__(self, provider_name: str) -> None:
+        self.provider_name = provider_name
+        super().__init__(f"Provider '{provider_name}' is unavailable (max restarts exhausted)")
+
+
+class ConfigError(OpenSpeechError):
+    """Configuration loading or validation error."""
+
+
+class FanOutAllFailedError(OpenSpeechError):
+    """All providers in a fanout invocation failed."""
+
+    def __init__(self, errors: dict[str, Exception]) -> None:
+        self.errors = errors
+        names = ", ".join(errors.keys())
+        super().__init__(f"All fanout providers failed: {names}")
+
+
+class ProviderNotFoundError(OpenSpeechError):
+    """Provider not found in registry or dispatcher."""
+
+    def __init__(self, name: str) -> None:
+        self.name = name
+        super().__init__(f"Provider '{name}' not found")