livekit-plugins-hume 1.0.17__py3-none-any.whl

livekit/plugins/hume/__init__.py

@@ -0,0 +1,56 @@
+# Copyright 2023 LiveKit, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import annotations
+
+__version__ = "1.0.0"
+
+# make imports available
+from hume.tts import (
+    Format,
+    PostedContext,
+    PostedUtterance,
+    PostedUtteranceVoiceWithId,
+    PostedUtteranceVoiceWithName,
+)
+from livekit.agents import Plugin
+
+from .tts import TTS
+
+# all exports
+__all__ = [
+    "TTS",
+    "Format",
+    "PostedUtterance",
+    "PostedContext",
+    "PostedUtteranceVoiceWithName",
+    "PostedUtteranceVoiceWithId",
+]
+
+
+class HumeAIPlugin(Plugin):
+    def __init__(self) -> None:
+        super().__init__(__name__, __version__, __package__)
+
+
+Plugin.register_plugin(HumeAIPlugin())
+
+# Cleanup docs of unexported modules
+_module = dir()
+NOT_IN_ALL = [m for m in _module if m not in __all__]
+
+__pdoc__ = {}
+
+for n in NOT_IN_ALL:
+    __pdoc__[n] = False

livekit/plugins/hume/log.py

@@ -0,0 +1,3 @@
+import logging
+
+logger = logging.getLogger("livekit.plugins.hume")

livekit/plugins/hume/py.typed

@@ -0,0 +1 @@
+ 

livekit/plugins/hume/tts.py

@@ -0,0 +1,297 @@
+# Copyright 2023 LiveKit, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import annotations
+
+import asyncio
+import base64
+import os
+from dataclasses import dataclass
+
+import aiohttp
+
+from hume import AsyncHumeClient
+from hume.tts import Format, FormatWav, PostedContext, PostedUtterance, PostedUtteranceVoiceWithName
+from livekit.agents import (
+    APIConnectionError,
+    APIConnectOptions,
+    APITimeoutError,
+    tokenize,
+    tts,
+    utils,
+)
+from livekit.agents.types import (
+    DEFAULT_API_CONNECT_OPTIONS,
+    NOT_GIVEN,
+    NotGivenOr,
+)
+from livekit.agents.utils import is_given
+
+# Default audio settings
+DEFAULT_SAMPLE_RATE = 24000
+DEFAULT_NUM_CHANNELS = 1
+
+# Default TTS settings
+DEFAULT_VOICE = PostedUtteranceVoiceWithName(name="Colton Rivers", provider="HUME_AI")
+
+# text is required in PostedUtterance but it is declared as an empty string
+# it will be overwritten when input tokens are received
+DEFAULT_UTTERANCE = PostedUtterance(
+    voice=DEFAULT_VOICE, speed=1, trailing_silence=0.35, description="", text=""
+)
+
+
+@dataclass
+class _TTSOptions:
+    """TTS options for Hume API"""
+
+    api_key: str
+    utterance_options: PostedUtterance
+    context: PostedContext | None
+    format: Format
+    sample_rate: int
+    split_utterances: bool
+    strip_headers: bool
+    num_generations: int
+    instant_mode: bool
+    word_tokenizer: tokenize.WordTokenizer
+
+
+class TTS(tts.TTS):
+    def __init__(
+        self,
+        *,
+        utterance_options: NotGivenOr[PostedUtterance] = NOT_GIVEN,
+        context: NotGivenOr[PostedContext] = NOT_GIVEN,
+        format: NotGivenOr[Format] = NOT_GIVEN,
+        split_utterances: bool = False,
+        num_generations: int = 1,
+        instant_mode: bool = False,
+        strip_headers: bool = True,
+        api_key: NotGivenOr[str] = NOT_GIVEN,
+        word_tokenizer: tokenize.WordTokenizer | None = None,
+        http_session: aiohttp.ClientSession | None = None,
+        sample_rate: int = 24000,
+    ) -> None:
+        """Initialize the Hume TTS client.
+
+        See https://dev.hume.ai/reference/text-to-speech-tts/synthesize-json-streaming for API doc
+
+        Args:
+            utterance_options (NotGivenOr[PostedUtterance]): Default options for utterances,
+                including description, voice, and delivery controls.
+            context (NotGivenOr[PostedContext]): Utterances to use as context for generating
+                consistent speech style and prosody across multiple requests.
+            format (NotGivenOr[Format]): Specifies the output audio file format (WAV, MP3 or PCM).
+                Defaults to WAV format.
+            split_utterances (bool): Controls how audio output is segmented in the response.
+                When enabled (True), input utterances are split into natural-sounding segments.
+                When disabled (False), maintains one-to-one mapping between input and output.
+                Defaults to False.
+            num_generations (int): Number of generations of the audio to produce.
+                Must be between 1 and 5. Defaults to 1.
+            instant_mode (bool): Enables ultra-low latency streaming, reducing time to first chunk.
+                Recommended for real-time applications. Only for streaming endpoints.
+                With this enabled, requests incur 10% higher cost. Defaults to False.
+            strip_headers (bool): If enabled, the audio for all the chunks of a generation.
+                Once concatenated together, will constitute a single audio file.
+                If disabled, each chunk’s audio will be its own audio file, each with its headers.
+            api_key (NotGivenOr[str]): Hume API key for authentication. If not provided,
+                will attempt to read from HUME_API_KEY environment variable.
+            word_tokenizer (tokenize.WordTokenizer | None): Custom word tokenizer to use for text.
+                If None, a basic word tokenizer will be used.
+            http_session (aiohttp.ClientSession | None): Optional HTTP session for API requests.
+                If None, a new session will be created.
+            sample_rate (int): Audio sample rate in Hz. Defaults to 24000.
+        """
+
+        super().__init__(
+            capabilities=tts.TTSCapabilities(
+                streaming=False,
+            ),
+            sample_rate=sample_rate,
+            num_channels=DEFAULT_NUM_CHANNELS,
+        )
+
+        self._api_key = api_key if is_given(api_key) else os.environ.get("HUME_API_KEY")
+        if not self._api_key:
+            raise ValueError(
+                "Hume API key is required, either as argument or set HUME_API_KEY env variable"
+            )
+
+        if not word_tokenizer:
+            word_tokenizer = tokenize.basic.WordTokenizer(ignore_punctuation=False)
+
+        self._opts = _TTSOptions(
+            utterance_options=utterance_options
+            if is_given(utterance_options)
+            else DEFAULT_UTTERANCE,
+            context=context if is_given(context) else None,
+            format=format if is_given(format) else FormatWav(),
+            api_key=self._api_key,
+            sample_rate=self.sample_rate,
+            split_utterances=split_utterances,
+            num_generations=num_generations,
+            strip_headers=strip_headers,
+            instant_mode=instant_mode,
+            word_tokenizer=word_tokenizer,
+        )
+
+        self._client = AsyncHumeClient(api_key=self._api_key)
+        self._session = http_session
+
+    def _ensure_session(self) -> aiohttp.ClientSession:
+        if not self._session:
+            self._session = utils.http_context.http_session()
+        return self._session
+
+    def update_options(
+        self,
+        *,
+        utterance_options: NotGivenOr[PostedUtterance] = NOT_GIVEN,
+        context: NotGivenOr[PostedContext] = NOT_GIVEN,
+        format: NotGivenOr[Format] = NOT_GIVEN,
+        split_utterances: NotGivenOr[bool] = NOT_GIVEN,
+        num_generations: NotGivenOr[int] = NOT_GIVEN,
+        instant_mode: NotGivenOr[bool] = NOT_GIVEN,
+        strip_headers: NotGivenOr[bool] = NOT_GIVEN,
+    ) -> None:
+        """Update TTS options for synthesizing speech.
+
+        Args:
+            utterance_options (NotGivenOr[PostedUtterance]): Options for utterances,
+                including text, description, voice, and additional controls.
+            context (Optional[PostedContext]): Utterances to use as context for generating
+                consistent speech style and prosody across multiple requests.
+            format (NotGivenOr[Format]): Specifies the output audio file format (WAV, MP3 or PCM).
+            split_utterances (NotGivenOr[bool]): Controls how audio output is segmented.
+                When True, utterances are split into natural-sounding segments.
+                When False, maintains one-to-one mapping between input and output.
+            num_generations (NotGivenOr[int]): Number of speech generations to produce (1-5).
+            instant_mode (NotGivenOr[bool]): Enables ultra-low latency streaming.
+                Reduces time to first audio chunk, recommended for real-time applications.
+                Note: Incurs 10% higher cost when enabled.
+            strip_headers (NotGivenOr[bool]): If enabled, the audio for the chunks of a generation.
+                Once concatenated together, will constitute a single audio file.
+                If disabled, each chunk’s audio will be its own audio file, each with its headers.
+        """
+
+        if is_given(utterance_options):
+            # text is required in PostedUtterance but it is declared as an empty string
+            # it will be overwritten when input tokens are received
+            self._opts.utterance_options = PostedUtterance(
+                description=utterance_options.description if utterance_options.description else "",
+                voice=utterance_options.voice if utterance_options.voice else DEFAULT_VOICE,
+                speed=utterance_options.speed if utterance_options.speed else 1,
+                trailing_silence=utterance_options.trailing_silence
+                if utterance_options.trailing_silence
+                else 0.35,
+                text="",
+            )
+        if is_given(format):
+            self._opts.format = format
+        if is_given(context):
+            self._opts.context = context
+        if is_given(split_utterances):
+            self._opts.split_utterances = split_utterances
+        if is_given(num_generations):
+            self._opts.num_generations = num_generations
+        if is_given(instant_mode):
+            self._opts.instant_mode = instant_mode
+        if is_given(strip_headers):
+            self._opts.strip_headers = strip_headers
+
+    def synthesize(
+        self,
+        text: str,
+        *,
+        conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS,
+    ) -> ChunkedStream:
+        return ChunkedStream(
+            tts=self,
+            input_text=text,
+            conn_options=conn_options,
+            opts=self._opts,
+        )
+
+
+class ChunkedStream(tts.ChunkedStream):
+    """Stream for Hume TTS JSON streaming API."""
+
+    def __init__(
+        self,
+        *,
+        tts: TTS,
+        input_text: str,
+        opts: _TTSOptions,
+        conn_options: APIConnectOptions,
+    ) -> None:
+        super().__init__(tts=tts, input_text=input_text, conn_options=conn_options)
+        self._opts = opts
+        self._client = tts._client
+
+    async def _run(self) -> None:
+        request_id = utils.shortuuid()
+
+        decoder = utils.codecs.AudioStreamDecoder(
+            sample_rate=self._opts.sample_rate,
+            num_channels=DEFAULT_NUM_CHANNELS,
+        )
+
+        decode_task: asyncio.Task | None = None
+
+        try:
+
+            async def _decode_loop():
+                try:
+                    async for chunk in self._client.tts.synthesize_json_streaming(
+                        utterances=[
+                            PostedUtterance(
+                                text=self._input_text,
+                                description=self._opts.utterance_options.description,
+                                voice=self._opts.utterance_options.voice,
+                                speed=self._opts.utterance_options.speed,
+                                trailing_silence=self._opts.utterance_options.trailing_silence,
+                            )
+                        ],
+                        context=self._opts.context,
+                        format=self._opts.format,
+                        num_generations=self._opts.num_generations,
+                        split_utterances=self._opts.split_utterances,
+                        instant_mode=self._opts.instant_mode,
+                        strip_headers=self._opts.strip_headers,
+                    ):
+                        decoder.push(base64.b64decode(chunk.audio))
+
+                finally:
+                    decoder.end_input()
+
+            decode_task = asyncio.create_task(_decode_loop())
+            emitter = tts.SynthesizedAudioEmitter(
+                event_ch=self._event_ch,
+                request_id=request_id,
+            )
+            async for frame in decoder:
+                emitter.push(frame)
+
+            emitter.flush()
+
+        except asyncio.TimeoutError:
+            raise APITimeoutError() from None
+        except Exception as e:
+            raise APIConnectionError() from e
+        finally:
+            if decode_task:
+                await utils.aio.gracefully_cancel(decode_task)
+            await decoder.aclose()

livekit/plugins/hume/version.py

@@ -0,0 +1,15 @@
+# Copyright 2024 LiveKit, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+__version__ = "1.0.17"

livekit_plugins_hume-1.0.17.dist-info/METADATA

@@ -0,0 +1,34 @@
+Metadata-Version: 2.4
+Name: livekit-plugins-hume
+Version: 1.0.17
+Summary: Hume TTS plugin for LiveKit agents
+Project-URL: Documentation, https://docs.livekit.io
+Project-URL: Website, https://livekit.io/
+Project-URL: Source, https://github.com/livekit/agents
+Author-email: LiveKit <info@livekit.io>
+License-Expression: Apache-2.0
+Keywords: Hume,HumeAI,Octave,audio,livekit,realtime,webrtc
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Multimedia :: Sound/Audio
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.9.0
+Requires-Dist: aiohttp>=3.8.0
+Requires-Dist: hume
+Requires-Dist: livekit-agents>=1.0.17
+Description-Content-Type: text/markdown
+
+# LiveKit Plugins Hume AI TTS
+
+LiveKit Agents Framework plugin for [Hume](https://www.hume.ai/) Text-to-Speech API.
+
+## Installation
+
+```bash
+pip install livekit-plugins-hume
+```
+
+You will need an API Key from Hume, it can be set as an environment variable: `HUME_API_KEY`. You can get it from [here](https://platform.hume.ai/settings/keys)

livekit_plugins_hume-1.0.17.dist-info/RECORD

@@ -0,0 +1,9 @@
+livekit/plugins/hume/__init__.py,sha256=CdEjcQRVL3dBso4xBL-zOgCESSqwH0Xdb01VT35P8u0,1362
+livekit/plugins/hume/log.py,sha256=TwpK1FOwgD6Jb0A2nl-9nIgi0q5qWo9HGDrDuV_2g0g,67
+livekit/plugins/hume/models.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+livekit/plugins/hume/py.typed,sha256=Nqnn8clbgv-5l0PgxcTOldg8mkMKrFn4TvPL-rYUUGg,1
+livekit/plugins/hume/tts.py,sha256=aVlp-PebRsIily2mcsCewuZzcgHKwzbBSYwHcFnSo0w,12029
+livekit/plugins/hume/version.py,sha256=oT9vgJC1WR2E9D9qKy-VZ5neWTTotVE-IZcSbmiQP98,601
+livekit_plugins_hume-1.0.17.dist-info/METADATA,sha256=EpRs_Biq7BWbNk8P-COP1Sgm0LqZiMd6L1Zp--oDsN8,1251
+livekit_plugins_hume-1.0.17.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+livekit_plugins_hume-1.0.17.dist-info/RECORD,,

livekit_plugins_hume-1.0.17.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.27.0
+Root-Is-Purelib: true
+Tag: py3-none-any