livekit-plugins-cartesia 0.1.0__py3-none-any.whl → 0.2.0__py3-none-any.whl

livekit/plugins/cartesia/models.py

@@ -9,39 +9,6 @@ TTSEncoding = Literal[
 ]
 
 
-TTSModels = Literal["upbeat-moon"]
-
-
-# fmt: off
-# Barbershop Man in upbeat-moon
-TTSDefaultVoiceEmbedding: list[float] = [
-    -0.033633083, 0.072083704, -0.01807767, -0.083488315, -0.04407617, 0.0022592682, 0.070505895,
-    0.023946615, -0.04788024, -0.06388413, -0.0716355, -0.0022612812, -0.0053448505, -0.07848381,
-    0.0348162, -0.053745482, -0.092399485, -0.02950225, 0.028591828, -0.10556894, 0.023313355,
-    0.06224387, 0.0362463, 0.029258432, 0.10769641, 0.043595582, -0.058543224, -0.080402784,
-    -0.0953816, -0.008988032, -0.0028981369, -0.004752721, -0.20742874, 0.058907595, 0.08813939,
-    -0.06192675, 0.099082634, -0.09661578, -0.0077761724, -0.013982456, -0.025798267, 0.04467142,
-    0.026222011, 0.023023574, 0.011227064, -0.17462021, -0.09880612, -0.1521035, -0.060464993,
-    -0.04735665, -0.09725187, -0.006127679, 0.15818526, -0.039493002, -0.067719474, 0.0066190436,
-    -0.10636633, 0.17073768, -0.051717706, 0.03186961, -0.020547207, -0.02244247, 0.013196935,
-    -0.06431055, -0.115360335, 0.016918058, -0.033195216, 0.11255181, 0.020366343, -0.041032124,
-    0.08780918, -0.040567942, 0.057276532, 0.05848221, -0.077479474, -0.073524915, -0.01913317,
-    -0.029291833, 0.11210393, -0.09859328, 0.2152541, -0.022976823, 0.028627992, -0.039598297,
-    0.041829932, -0.05593181, -0.06444655, -0.018057477, -0.008098263, 0.05994528, 0.10430693,
-    -0.13121894, -0.06512868, -0.026126215, 0.046727825, -0.17180993, -0.10577226, -0.08610466,
-    0.008862588, 0.09547498, -0.010965332, -0.061217085, -0.038954042, 0.019930292, -0.017192135,
-    0.007296275, 0.03273872, 0.04389937, -0.056483064, 0.003420891, -0.10319067, -0.015706042,
-    0.1308774, -0.0018035866, -0.03582506, 0.077131025, 0.013398928, 0.003188886, 0.12039741,
-    -0.033974767, 0.06899378, -0.059775922, -0.026934423, 0.028482193, 0.100996524, 0.004498743,
-    -0.02291186, 0.078752205, -0.0063796206, 0.04206536, 0.05721349, 0.06290694, 0.06130212,
-    0.096969016, -0.057664312, -0.16727506, -0.035220966, 0.090760484, 0.010039947, 0.06513242,
-    0.011055657, -0.004258431, -0.08316792, -0.15650468, -0.076931365, 0.11385587, -0.038372636,
-    0.015648656, -0.12029895, -0.06604956, 0.009441591, -0.11912808, 0.013378132, 0.029525978,
-    -0.0056742397, -0.0075976513, 0.019999338, -0.05521377, -0.07650746, -0.017710293, -0.033986397,
-    -0.047768556, 0.13857274, 0.099290825, 0.11736938, 0.017834296, -0.07140237, -0.052047748,
-    -0.06398965, -0.037033975, -0.061061256, -0.03330076, -0.024472248, -0.059656, 0.05359946,
-    -0.043915518, -0.086325996, 0.14189173, 0.021086395, 0.02945159, 0.1029604, 0.018490415,
-    -0.028736332, -0.025272416, -0.06082937, -0.031339463, -0.0007249595, 0.025595888, 0.007144545,
-    -0.16938712, -0.1160664, -0.0654145,
-]
-# fmt: on
+TTSModels = Literal["sonic-english", "sonic-multilingual"]
+TTSLanguages = Literal["en", "es", "fr", "de", "pt", "zh", "ja"]
+TTSDefaultVoiceId = "248be419-c632-4f23-adf1-5324ed7dbf1d"

livekit/plugins/cartesia/tts.py

@@ -14,18 +14,14 @@
 
 from __future__ import annotations
 
-import asyncio
-import contextlib
 import os
 from dataclasses import dataclass
-from typing import Optional
 
 import aiohttp
-from livekit import rtc
 from livekit.agents import tts, utils
 
 from .log import logger
-from .models import TTSDefaultVoiceEmbedding, TTSEncoding, TTSModels
+from .models import TTSDefaultVoiceId, TTSEncoding, TTSModels
 
 API_AUTH_HEADER = "X-API-Key"
 API_VERSION_HEADER = "Cartesia-Version"
@@ -39,21 +35,23 @@ class _TTSOptions:
     sample_rate: int
     voice: str | list[float]
     api_key: str
+    language: str
 
 
 class TTS(tts.TTS):
     def __init__(
         self,
         *,
-        model: TTSModels = "upbeat-moon",
+        model: TTSModels = "sonic-english",
+        language: str = "en",
         encoding: TTSEncoding = "pcm_s16le",
-        voice: str | list[float] = TTSDefaultVoiceEmbedding,
+        voice: str | list[float] = TTSDefaultVoiceId,
         sample_rate: int = 24000,
         api_key: str | None = None,
         http_session: aiohttp.ClientSession | None = None,
     ) -> None:
         super().__init__(
-            streaming_supported=False,
+            capabilities=tts.TTSCapabilities(streaming=False),
             sample_rate=sample_rate,
             num_channels=1,
         )
@@ -64,6 +62,7 @@ class TTS(tts.TTS):
 
         self._opts = _TTSOptions(
             model=model,
+            language=language,
             encoding=encoding,
             sample_rate=sample_rate,
             voice=voice,
@@ -73,14 +72,11 @@ class TTS(tts.TTS):
 
     def _ensure_session(self) -> aiohttp.ClientSession:
         if not self._session:
-            self._session = utils.http_session()
+            self._session = utils.http_context.http_session()
 
         return self._session
 
-    def synthesize(
-        self,
-        text: str,
-    ) -> "ChunkedStream":
+    def synthesize(self, text: str) -> "ChunkedStream":
         return ChunkedStream(text, self._opts, self._ensure_session())
 
 
@@ -88,14 +84,17 @@ class ChunkedStream(tts.ChunkedStream):
     def __init__(
         self, text: str, opts: _TTSOptions, session: aiohttp.ClientSession
     ) -> None:
-        self._opts = opts
-        self._text = text
-        self._session = session
-        self._main_task: asyncio.Task | None = None
-        self._queue = asyncio.Queue[Optional[tts.SynthesizedAudio]]()
+        super().__init__()
+        self._text, self._opts, self._session = text, opts, session
 
     @utils.log_exceptions(logger=logger)
-    async def _run(self):
+    async def _main_task(self):
+        bstream = utils.audio.AudioByteStream(
+            sample_rate=self._opts.sample_rate, num_channels=1
+        )
+        request_id = utils.shortuuid()
+        segment_id = utils.shortuuid()
+
         voice = {}
         if isinstance(self._opts.voice, str):
             voice["mode"] = "id"
@@ -104,77 +103,37 @@ class ChunkedStream(tts.ChunkedStream):
             voice["mode"] = "embedding"
             voice["embedding"] = self._opts.voice
 
-        try:
-            async with self._session.post(
-                "https://api.cartesia.ai/tts/bytes",
-                headers={
-                    API_AUTH_HEADER: f"{self._opts.api_key}",
-                    API_VERSION_HEADER: API_VERSION,
-                },
-                json={
-                    "model_id": self._opts.model,
-                    "transcript": self._text,
-                    "voice": voice,
-                    "output_format": {
-                        "container": "raw",
-                        "encoding": self._opts.encoding,
-                        "sample_rate": self._opts.sample_rate,
-                    },
-                },
-            ) as resp:
-                bytes_per_frame = (self._opts.sample_rate // 100) * 2
-                buf = bytearray()
-
-                async for data, _ in resp.content.iter_chunks():
-                    buf.extend(data)
-
-                    while len(buf) >= bytes_per_frame:
-                        frame_data = buf[:bytes_per_frame]
-                        buf = buf[bytes_per_frame:]
-
-                        self._queue.put_nowait(
-                            tts.SynthesizedAudio(
-                                text=self._text,
-                                data=rtc.AudioFrame(
-                                    data=frame_data,
-                                    sample_rate=self._opts.sample_rate,
-                                    num_channels=1,
-                                    samples_per_channel=len(frame_data) // 2,
-                                ),
-                            )
-                        )
-
-                # send any remaining data
-                if len(buf) > 0:
-                    self._queue.put_nowait(
+        data = {
+            "model_id": self._opts.model,
+            "transcript": self._text,
+            "voice": voice,
+            "output_format": {
+                "container": "raw",
+                "encoding": self._opts.encoding,
+                "sample_rate": self._opts.sample_rate,
+            },
+            "language": self._opts.language,
+        }
+
+        async with self._session.post(
+            "https://api.cartesia.ai/tts/bytes",
+            headers={
+                API_AUTH_HEADER: f"{self._opts.api_key}",
+                API_VERSION_HEADER: API_VERSION,
+            },
+            json=data,
+        ) as resp:
+            async for data, _ in resp.content.iter_chunks():
+                for frame in bstream.write(data):
+                    self._event_ch.send_nowait(
                         tts.SynthesizedAudio(
-                            text=self._text,
-                            data=rtc.AudioFrame(
-                                data=buf,
-                                sample_rate=self._opts.sample_rate,
-                                num_channels=1,
-                                samples_per_channel=len(buf) // 2,
-                            ),
+                            request_id=request_id, segment_id=segment_id, frame=frame
                         )
                     )
 
-        finally:
-            self._queue.put_nowait(None)
-
-    async def __anext__(self) -> tts.SynthesizedAudio:
-        if not self._main_task:
-            self._main_task = asyncio.create_task(self._run())
-
-        frame = await self._queue.get()
-        if frame is None:
-            raise StopAsyncIteration
-
-        return frame
-
-    async def aclose(self) -> None:
-        if not self._main_task:
-            return
-
-        self._main_task.cancel()
-        with contextlib.suppress(asyncio.CancelledError):
-            await self._main_task
+            for frame in bstream.flush():
+                self._event_ch.send_nowait(
+                    tts.SynthesizedAudio(
+                        request_id=request_id, segment_id=segment_id, frame=frame
+                    )
+                )

livekit/plugins/cartesia/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "0.1.0"
+__version__ = "0.2.0"

{livekit_plugins_cartesia-0.1.0.dist-info → livekit_plugins_cartesia-0.2.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: livekit-plugins-cartesia
-Version: 0.1.0
+Version: 0.2.0
 Summary: LiveKit Agents Plugin for Cartesia
 Home-page: https://github.com/livekit/agents
 License: Apache-2.0
@@ -14,14 +14,12 @@ Classifier: Topic :: Multimedia :: Sound/Audio
 Classifier: Topic :: Multimedia :: Video
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.8
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3 :: Only
-Requires-Python: >=3.8.0
+Requires-Python: >=3.9.0
 Description-Content-Type: text/markdown
-Requires-Dist: livekit ~=0.11
-Requires-Dist: livekit-agents ~=0.7
+Requires-Dist: livekit-agents >=0.8.0.dev0
 
 # LiveKit Plugins Cartesia
 

livekit_plugins_cartesia-0.2.0.dist-info/RECORD

@@ -0,0 +1,10 @@
+livekit/plugins/cartesia/__init__.py,sha256=_a8u7qqya1pjZTV19gNOpMKTO7ccAVZAeCukiDKAG-U,937
+livekit/plugins/cartesia/log.py,sha256=4Mnhjng_DU1dIWP9IWjIQGZ67EV3LnQhWMWCHVudJbo,71
+livekit/plugins/cartesia/models.py,sha256=06S-Z-M90kB-kEOQsQk70xfQUD-TztU4ZIU_AfAyUMc,335
+livekit/plugins/cartesia/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+livekit/plugins/cartesia/tts.py,sha256=S5BMSVtsbNI_c2PpgyFK6wvleudmJZLTUt3ZmGNKlRI,4319
+livekit/plugins/cartesia/version.py,sha256=cLFCdnm5S21CiJ5UJBcqfRvvFkCQ8p6M5fFUJVJkEiM,600
+livekit_plugins_cartesia-0.2.0.dist-info/METADATA,sha256=z0oS7uWDrLjMFWqYxGNJD9SENd5W5sdNwogpe-tTeA4,1252
+livekit_plugins_cartesia-0.2.0.dist-info/WHEEL,sha256=Wyh-_nZ0DJYolHNn1_hMa4lM7uDedD_RGVwbmTjyItk,91
+livekit_plugins_cartesia-0.2.0.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
+livekit_plugins_cartesia-0.2.0.dist-info/RECORD,,

{livekit_plugins_cartesia-0.1.0.dist-info → livekit_plugins_cartesia-0.2.0.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.43.0)
+Generator: setuptools (71.1.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

livekit_plugins_cartesia-0.1.0.dist-info/RECORD

@@ -1,10 +0,0 @@
-livekit/plugins/cartesia/__init__.py,sha256=_a8u7qqya1pjZTV19gNOpMKTO7ccAVZAeCukiDKAG-U,937
-livekit/plugins/cartesia/log.py,sha256=4Mnhjng_DU1dIWP9IWjIQGZ67EV3LnQhWMWCHVudJbo,71
-livekit/plugins/cartesia/models.py,sha256=Qhl51ZScuB61bEzN1tBlHMuHO_kCXSzuVOicYa16EL8,2922
-livekit/plugins/cartesia/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-livekit/plugins/cartesia/tts.py,sha256=16BneZFQQsS-lB9Ug1HYj4QW7-VnNdpTJ0CW5A1b9EU,5725
-livekit/plugins/cartesia/version.py,sha256=vQH9cItKAVYAmrLbOntkbLqmxrUZrPiKb1TjkZ8jRKQ,600
-livekit_plugins_cartesia-0.1.0.dist-info/METADATA,sha256=lpj60lrnFHBzooiT3f3Dtt74J0vq9ZvHfxqCb7YCBmg,1325
-livekit_plugins_cartesia-0.1.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-livekit_plugins_cartesia-0.1.0.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
-livekit_plugins_cartesia-0.1.0.dist-info/RECORD,,