livekit-plugins-elevenlabs 0.8.1__py3-none-any.whl → 1.0.0__py3-none-any.whl

livekit/plugins/elevenlabs/__init__.py

@@ -13,7 +13,7 @@
 # limitations under the License.
 
 from .models import TTSEncoding, TTSModels
-from .tts import DEFAULT_VOICE, TTS, Voice, VoiceSettings
+from .tts import DEFAULT_VOICE_ID, TTS, Voice, VoiceSettings
 from .version import __version__
 
 __all__ = [
@@ -22,7 +22,7 @@ __all__ = [
     "VoiceSettings",
     "TTSEncoding",
     "TTSModels",
-    "DEFAULT_VOICE",
+    "DEFAULT_VOICE_ID",
     "__version__",
 ]
 

livekit/plugins/elevenlabs/models.py

@@ -10,4 +10,12 @@ TTSModels = Literal[
     "eleven_flash_v2",
 ]
 
-TTSEncoding = Literal["mp3_44100",]
+TTSEncoding = Literal[
+    "mp3_22050_32",
+    "mp3_44100",
+    "mp3_44100_32",
+    "mp3_44100_64",
+    "mp3_44100_96",
+    "mp3_44100_128",
+    "mp3_44100_192",
+]

livekit/plugins/elevenlabs/tts.py

@@ -21,9 +21,10 @@ import json
 import os
 import weakref
 from dataclasses import dataclass
-from typing import Any, List, Optional
+from typing import Any
 
 import aiohttp
+
 from livekit.agents import (
     APIConnectionError,
     APIConnectOptions,
@@ -33,11 +34,19 @@ from livekit.agents import (
     tts,
     utils,
 )
+from livekit.agents.types import (
+    DEFAULT_API_CONNECT_OPTIONS,
+    NOT_GIVEN,
+    NotGivenOr,
+)
+from livekit.agents.utils import is_given
 
 from .log import logger
 from .models import TTSEncoding, TTSModels
 
-_DefaultEncoding: TTSEncoding = "mp3_44100"
+# by default, use 22.05kHz sample rate at 32kbps
+# in our testing,  reduce TTFB by about ~110ms
+_DefaultEncoding: TTSEncoding = "mp3_22050_32"
 
 
 def _sample_rate_from_format(output_format: TTSEncoding) -> int:
@@ -49,9 +58,9 @@ def _sample_rate_from_format(output_format: TTSEncoding) -> int:
 class VoiceSettings:
     stability: float  # [0.0 - 1.0]
     similarity_boost: float  # [0.0 - 1.0]
-    style: float | None = None  # [0.0 - 1.0]
-    speed: float | None = 1.0  # [0.8 - 1.2]
-    use_speaker_boost: bool | None = False
+    style: NotGivenOr[float] = NOT_GIVEN  # [0.0 - 1.0]
+    speed: NotGivenOr[float] = NOT_GIVEN  # [0.8 - 1.2]
+    use_speaker_boost: NotGivenOr[bool] = NOT_GIVEN
 
 
 @dataclass
@@ -59,22 +68,9 @@ class Voice:
     id: str
     name: str
     category: str
-    settings: VoiceSettings | None = None
-
-
-DEFAULT_VOICE = Voice(
-    id="EXAVITQu4vr4xnSDxMaL",
-    name="Bella",
-    category="premade",
-    settings=VoiceSettings(
-        stability=0.71,
-        speed=1.0,
-        similarity_boost=0.5,
-        style=0.0,
-        use_speaker_boost=True,
-    ),
-)
 
+
+DEFAULT_VOICE_ID = "EXAVITQu4vr4xnSDxMaL"
 API_BASE_URL_V1 = "https://api.elevenlabs.io/v1"
 AUTHORIZATION_HEADER = "xi-api-key"
 WS_INACTIVITY_TIMEOUT = 300
@@ -83,13 +79,14 @@ WS_INACTIVITY_TIMEOUT = 300
 @dataclass
 class _TTSOptions:
     api_key: str
-    voice: Voice
+    voice_id: str
+    voice_settings: NotGivenOr[VoiceSettings]
     model: TTSModels | str
-    language: str | None
+    language: NotGivenOr[str]
     base_url: str
     encoding: TTSEncoding
     sample_rate: int
-    streaming_latency: int
+    streaming_latency: NotGivenOr[int]
     word_tokenizer: tokenize.WordTokenizer
     chunk_length_schedule: list[int]
     enable_ssml_parsing: bool
@@ -100,68 +97,70 @@ class TTS(tts.TTS):
     def __init__(
         self,
         *,
-        voice: Voice = DEFAULT_VOICE,
+        voice_id: str = DEFAULT_VOICE_ID,
+        voice_settings: NotGivenOr[VoiceSettings] = NOT_GIVEN,
         model: TTSModels | str = "eleven_flash_v2_5",
-        api_key: str | None = None,
-        base_url: str | None = None,
-        streaming_latency: int = 0,
+        encoding: NotGivenOr[TTSEncoding] = NOT_GIVEN,
+        api_key: NotGivenOr[str] = NOT_GIVEN,
+        base_url: NotGivenOr[str] = NOT_GIVEN,
+        streaming_latency: NotGivenOr[int] = NOT_GIVEN,
         inactivity_timeout: int = WS_INACTIVITY_TIMEOUT,
-        word_tokenizer: Optional[tokenize.WordTokenizer] = None,
+        word_tokenizer: NotGivenOr[tokenize.WordTokenizer] = NOT_GIVEN,
         enable_ssml_parsing: bool = False,
-        chunk_length_schedule: list[int] = [80, 120, 200, 260],  # range is [50, 500]
+        chunk_length_schedule: NotGivenOr[list[int]] = NOT_GIVEN,  # range is [50, 500]
         http_session: aiohttp.ClientSession | None = None,
-        # deprecated
-        model_id: TTSModels | str | None = None,
-        language: str | None = None,
+        language: NotGivenOr[str] = NOT_GIVEN,
     ) -> None:
         """
         Create a new instance of ElevenLabs TTS.
 
         Args:
-            voice (Voice): Voice configuration. Defaults to `DEFAULT_VOICE`.
+            voice_id (str): Voice ID. Defaults to `DEFAULT_VOICE_ID`.
+            voice_settings (NotGivenOr[VoiceSettings]): Voice settings.
             model (TTSModels | str): TTS model to use. Defaults to "eleven_turbo_v2_5".
-            api_key (str | None): ElevenLabs API key. Can be set via argument or `ELEVEN_API_KEY` environment variable.
-            base_url (str | None): Custom base URL for the API. Optional.
-            streaming_latency (int): Optimize for streaming latency, defaults to 0 - disabled. 4 for max latency optimizations. deprecated
+            api_key (NotGivenOr[str]): ElevenLabs API key. Can be set via argument or `ELEVEN_API_KEY` environment variable.
+            base_url (NotGivenOr[str]): Custom base URL for the API. Optional.
+            streaming_latency (NotGivenOr[int]): Optimize for streaming latency, defaults to 0 - disabled. 4 for max latency optimizations. deprecated
             inactivity_timeout (int): Inactivity timeout in seconds for the websocket connection. Defaults to 300.
-            word_tokenizer (tokenize.WordTokenizer): Tokenizer for processing text. Defaults to basic WordTokenizer.
+            word_tokenizer (NotGivenOr[tokenize.WordTokenizer]): Tokenizer for processing text. Defaults to basic WordTokenizer.
             enable_ssml_parsing (bool): Enable SSML parsing for input text. Defaults to False.
-            chunk_length_schedule (list[int]): Schedule for chunk lengths, ranging from 50 to 500. Defaults to [80, 120, 200, 260].
+            chunk_length_schedule (NotGivenOr[list[int]]): Schedule for chunk lengths, ranging from 50 to 500. Defaults to [80, 120, 200, 260].
             http_session (aiohttp.ClientSession | None): Custom HTTP session for API requests. Optional.
-            language (str | None): Language code for the TTS model, as of 10/24/24 only valid for "eleven_turbo_v2_5". Optional.
-        """
+            language (NotGivenOr[str]): Language code for the TTS model, as of 10/24/24 only valid for "eleven_turbo_v2_5".
+        """  # noqa: E501
+
+        if not is_given(chunk_length_schedule):
+            chunk_length_schedule = [80, 120, 200, 260]
+
+        if not is_given(encoding):
+            encoding = _DefaultEncoding
 
         super().__init__(
             capabilities=tts.TTSCapabilities(
                 streaming=True,
             ),
-            sample_rate=_sample_rate_from_format(_DefaultEncoding),
+            sample_rate=_sample_rate_from_format(encoding),
             num_channels=1,
         )
 
-        if model_id is not None:
-            logger.warning(
-                "model_id is deprecated and will be removed in 1.5.0, use model instead",
-            )
-            model = model_id
-
-        api_key = api_key or os.environ.get("ELEVEN_API_KEY")
-        if not api_key:
+        elevenlabs_api_key = api_key if is_given(api_key) else os.environ.get("ELEVEN_API_KEY")
+        if not elevenlabs_api_key:
             raise ValueError(
-                "ElevenLabs API key is required, either as argument or set ELEVEN_API_KEY environmental variable"
+                "ElevenLabs API key is required, either as argument or set ELEVEN_API_KEY environmental variable"  # noqa: E501
             )
 
-        if word_tokenizer is None:
+        if not is_given(word_tokenizer):
             word_tokenizer = tokenize.basic.WordTokenizer(
                 ignore_punctuation=False  # punctuation can help for intonation
             )
 
         self._opts = _TTSOptions(
-            voice=voice,
+            voice_id=voice_id,
+            voice_settings=voice_settings,
             model=model,
-            api_key=api_key,
-            base_url=base_url or API_BASE_URL_V1,
-            encoding=_DefaultEncoding,
+            api_key=elevenlabs_api_key,
+            base_url=base_url if is_given(base_url) else API_BASE_URL_V1,
+            encoding=encoding,
             sample_rate=self.sample_rate,
             streaming_latency=streaming_latency,
             word_tokenizer=word_tokenizer,
@@ -179,7 +178,7 @@ class TTS(tts.TTS):
 
         return self._session
 
-    async def list_voices(self) -> List[Voice]:
+    async def list_voices(self) -> list[Voice]:
         async with self._ensure_session().get(
             f"{self._opts.base_url}/voices",
             headers={AUTHORIZATION_HEADER: self._opts.api_key},
@@ -189,26 +188,33 @@ class TTS(tts.TTS):
     def update_options(
         self,
         *,
-        voice: Voice = DEFAULT_VOICE,
-        model: TTSModels | str = "eleven_turbo_v2_5",
-        language: str | None = None,
+        voice_id: NotGivenOr[str] = NOT_GIVEN,
+        voice_settings: NotGivenOr[VoiceSettings] = NOT_GIVEN,
+        model: NotGivenOr[TTSModels | str] = NOT_GIVEN,
+        language: NotGivenOr[str] = NOT_GIVEN,
     ) -> None:
         """
         Args:
-            voice (Voice): Voice configuration. Defaults to `DEFAULT_VOICE`.
-            model (TTSModels | str): TTS model to use. Defaults to "eleven_turbo_v2_5".
-            language (str | None): Language code for the TTS model. Optional.
+            voice_id (NotGivenOr[str]): Voice ID.
+            voice_settings (NotGivenOr[VoiceSettings]): Voice settings.
+            model (NotGivenOr[TTSModels | str]): TTS model to use.
+            language (NotGivenOr[str]): Language code for the TTS model.
         """
-        self._opts.model = model or self._opts.model
-        self._opts.voice = voice or self._opts.voice
-        self._opts.language = language or self._opts.language
+        if is_given(model):
+            self._opts.model = model
+        if is_given(voice_id):
+            self._opts.voice_id = voice_id
+        if is_given(voice_settings):
+            self._opts.voice_settings = voice_settings
+        if is_given(language):
+            self._opts.language = language
 
     def synthesize(
         self,
         text: str,
         *,
-        conn_options: Optional[APIConnectOptions] = None,
-    ) -> "ChunkedStream":
+        conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS,
+    ) -> ChunkedStream:
         return ChunkedStream(
             tts=self,
             input_text=text,
@@ -218,8 +224,8 @@ class TTS(tts.TTS):
         )
 
     def stream(
-        self, *, conn_options: Optional[APIConnectOptions] = None
-    ) -> "SynthesizeStream":
+        self, *, conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS
+    ) -> SynthesizeStream:
         stream = SynthesizeStream(
             tts=self,
             conn_options=conn_options,
@@ -245,7 +251,7 @@ class ChunkedStream(tts.ChunkedStream):
         tts: TTS,
         input_text: str,
         opts: _TTSOptions,
-        conn_options: Optional[APIConnectOptions] = None,
+        conn_options: APIConnectOptions,
         session: aiohttp.ClientSession,
     ) -> None:
         super().__init__(tts=tts, input_text=input_text, conn_options=conn_options)
@@ -254,8 +260,8 @@ class ChunkedStream(tts.ChunkedStream):
     async def _run(self) -> None:
         request_id = utils.shortuuid()
         voice_settings = (
-            _strip_nones(dataclasses.asdict(self._opts.voice.settings))
-            if self._opts.voice.settings
+            _strip_nones(dataclasses.asdict(self._opts.voice_settings))
+            if is_given(self._opts.voice_settings)
             else None
         )
         data = {
@@ -322,7 +328,7 @@ class SynthesizeStream(tts.SynthesizeStream):
         tts: TTS,
         session: aiohttp.ClientSession,
         opts: _TTSOptions,
-        conn_options: Optional[APIConnectOptions] = None,
+        conn_options: APIConnectOptions,
     ):
         super().__init__(tts=tts, conn_options=conn_options)
         self._opts, self._session = opts, session
@@ -392,15 +398,13 @@ class SynthesizeStream(tts.SynthesizeStream):
         )
 
         # 11labs protocol expects the first message to be an "init msg"
-        init_pkt = dict(
-            text=" ",
-            voice_settings=_strip_nones(dataclasses.asdict(self._opts.voice.settings))
-            if self._opts.voice.settings
+        init_pkt = {
+            "text": " ",
+            "voice_settings": _strip_nones(dataclasses.asdict(self._opts.voice_settings))
+            if is_given(self._opts.voice_settings)
             else None,
-            generation_config=dict(
-                chunk_length_schedule=self._opts.chunk_length_schedule
-            ),
-        )
+            "generation_config": {"chunk_length_schedule": self._opts.chunk_length_schedule},
+        }
         await ws_conn.send_str(json.dumps(init_pkt))
         eos_sent = False
 
@@ -423,14 +427,14 @@ class SynthesizeStream(tts.SynthesizeStream):
                     else:
                         continue
 
-                data_pkt = dict(text=f"{text} ")  # must always end with a space
+                data_pkt = {"text": f"{text} "}  # must always end with a space
                 self._mark_started()
                 await ws_conn.send_str(json.dumps(data_pkt))
             if xml_content:
                 logger.warning("11labs stream ended with incomplete xml content")
 
             # no more token, mark eos
-            eos_pkt = dict(text="")
+            eos_pkt = {"text": ""}
             await ws_conn.send_str(json.dumps(eos_pkt))
             eos_sent = True
 
@@ -460,7 +464,7 @@ class SynthesizeStream(tts.SynthesizeStream):
                 ):
                     if not eos_sent:
                         raise APIStatusError(
-                            "11labs connection closed unexpectedly, not all tokens have been consumed",
+                            "11labs connection closed unexpectedly, not all tokens have been consumed",  # noqa: E501
                             request_id=request_id,
                         )
                     return
@@ -520,40 +524,39 @@ class SynthesizeStream(tts.SynthesizeStream):
 
 
 def _dict_to_voices_list(data: dict[str, Any]):
-    voices: List[Voice] = []
+    voices: list[Voice] = []
     for voice in data["voices"]:
         voices.append(
             Voice(
                 id=voice["voice_id"],
                 name=voice["name"],
                 category=voice["category"],
-                settings=None,
             )
         )
     return voices
 
 
 def _strip_nones(data: dict[str, Any]):
-    return {k: v for k, v in data.items() if v is not None}
+    return {k: v for k, v in data.items() if is_given(v) and v is not None}
 
 
 def _synthesize_url(opts: _TTSOptions) -> str:
     base_url = opts.base_url
-    voice_id = opts.voice.id
+    voice_id = opts.voice_id
     model_id = opts.model
     output_format = opts.encoding
     url = (
         f"{base_url}/text-to-speech/{voice_id}/stream?"
         f"model_id={model_id}&output_format={output_format}"
     )
-    if opts.streaming_latency:
+    if is_given(opts.streaming_latency):
         url += f"&optimize_streaming_latency={opts.streaming_latency}"
     return url
 
 
 def _stream_url(opts: _TTSOptions) -> str:
     base_url = opts.base_url
-    voice_id = opts.voice.id
+    voice_id = opts.voice_id
     model_id = opts.model
     output_format = opts.encoding
     enable_ssml = str(opts.enable_ssml_parsing).lower()
@@ -564,8 +567,8 @@ def _stream_url(opts: _TTSOptions) -> str:
         f"model_id={model_id}&output_format={output_format}&"
         f"enable_ssml_parsing={enable_ssml}&inactivity_timeout={inactivity_timeout}"
     )
-    if language is not None:
+    if is_given(language):
         url += f"&language_code={language}"
-    if opts.streaming_latency:
+    if is_given(opts.streaming_latency):
         url += f"&optimize_streaming_latency={opts.streaming_latency}"
     return url

livekit/plugins/elevenlabs/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "0.8.1"
+__version__ = "1.0.0"

{livekit_plugins_elevenlabs-0.8.1.dist-info → livekit_plugins_elevenlabs-1.0.0.dist-info}/METADATA

@@ -1,35 +1,25 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: livekit-plugins-elevenlabs
-Version: 0.8.1
+Version: 1.0.0
 Summary: Agent Framework plugin for voice synthesis with ElevenLabs' API.
-Home-page: https://github.com/livekit/agents
-License: Apache-2.0
 Project-URL: Documentation, https://docs.livekit.io
 Project-URL: Website, https://livekit.io/
 Project-URL: Source, https://github.com/livekit/agents
-Keywords: webrtc,realtime,audio,video,livekit,elevenlabs
+Author-email: LiveKit <hello@livekit.io>
+License-Expression: Apache-2.0
+Keywords: audio,elevenlabs,livekit,realtime,video,webrtc
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: Apache Software License
-Classifier: Topic :: Multimedia :: Sound/Audio
-Classifier: Topic :: Multimedia :: Video
-Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Multimedia :: Sound/Audio
+Classifier: Topic :: Multimedia :: Video
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.0
+Requires-Dist: livekit-agents[codecs]>=1.0.0
 Description-Content-Type: text/markdown
-Requires-Dist: livekit-agents[codecs]<1.0.0,>=0.12.16
-Dynamic: classifier
-Dynamic: description
-Dynamic: description-content-type
-Dynamic: home-page
-Dynamic: keywords
-Dynamic: license
-Dynamic: project-url
-Dynamic: requires-dist
-Dynamic: requires-python
-Dynamic: summary
 
 # LiveKit Plugins Elevenlabs
 

livekit_plugins_elevenlabs-1.0.0.dist-info/RECORD

@@ -0,0 +1,9 @@
+livekit/plugins/elevenlabs/__init__.py,sha256=Va24UYTuuosmRuTcuzd_DIHYQOgV-wSYKJIXmOSB2Go,1255
+livekit/plugins/elevenlabs/log.py,sha256=hIuXqDsEB5GBa7rQY3z4Uqi1oCqc_lRmCHZEmXz0LHw,73
+livekit/plugins/elevenlabs/models.py,sha256=p_wHEz15bdsNEqwzN831ysm70PNWQ-xeN__BKvGPZxA,401
+livekit/plugins/elevenlabs/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+livekit/plugins/elevenlabs/tts.py,sha256=gs9p4TwBAYX3vlsNn2XQ-oyPNUGcuvgix8K7vChRMmc,19985
+livekit/plugins/elevenlabs/version.py,sha256=nW89L_U9N4ukT3wAO3BeTqOaa87zLUOsEFz8TkiKIP8,600
+livekit_plugins_elevenlabs-1.0.0.dist-info/METADATA,sha256=EL7wso-EPaWpWwQ5OtxwDaIueFvHrSBEy7PPCigZ8SI,1312
+livekit_plugins_elevenlabs-1.0.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+livekit_plugins_elevenlabs-1.0.0.dist-info/RECORD,,

{livekit_plugins_elevenlabs-0.8.1.dist-info → livekit_plugins_elevenlabs-1.0.0.dist-info}/WHEEL

@@ -1,5 +1,4 @@
 Wheel-Version: 1.0
-Generator: setuptools (76.1.0)
+Generator: hatchling 1.27.0
 Root-Is-Purelib: true
 Tag: py3-none-any
-

livekit_plugins_elevenlabs-0.8.1.dist-info/RECORD

@@ -1,10 +0,0 @@
-livekit/plugins/elevenlabs/__init__.py,sha256=YZVadomFq3JWiZN6GWXJbuE4vaNNWq1CmdH25du8qwg,1249
-livekit/plugins/elevenlabs/log.py,sha256=hIuXqDsEB5GBa7rQY3z4Uqi1oCqc_lRmCHZEmXz0LHw,73
-livekit/plugins/elevenlabs/models.py,sha256=nB43wLS1ilzS7IxLYVSQxBjKPnbiPl4AHpHAOlG2i00,273
-livekit/plugins/elevenlabs/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-livekit/plugins/elevenlabs/tts.py,sha256=-w8IeAvyQER4PLraajJz6OWDufvKpD_fPM8oPsYtX9s,19335
-livekit/plugins/elevenlabs/version.py,sha256=PoHw-_DNE2B5SpeoQ-r6HSfVmbDgYuGamg0dN2jhayQ,600
-livekit_plugins_elevenlabs-0.8.1.dist-info/METADATA,sha256=l8gbEDr8EsedqYQiqBhx6K9XwAdTtnQWVCxmlyjVG9w,1529
-livekit_plugins_elevenlabs-0.8.1.dist-info/WHEEL,sha256=beeZ86-EfXScwlR_HKu4SllMC9wUEj_8Z_4FJ3egI2w,91
-livekit_plugins_elevenlabs-0.8.1.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
-livekit_plugins_elevenlabs-0.8.1.dist-info/RECORD,,

livekit_plugins_elevenlabs-0.8.1.dist-info/top_level.txt

@@ -1 +0,0 @@
-livekit