livekit-plugins-cartesia 0.4.3__py3-none-any.whl → 0.4.6__py3-none-any.whl

livekit/plugins/cartesia/models.py

@@ -10,7 +10,7 @@ TTSEncoding = Literal[
 
 TTSModels = Literal["sonic-english", "sonic-multilingual"]
 TTSLanguages = Literal["en", "es", "fr", "de", "pt", "zh", "ja"]
-TTSDefaultVoiceId = "c2ac25f9-ecc4-4f56-9095-651354df60c0"
+TTSDefaultVoiceId = "794f9389-aac1-45b6-b726-9d9369183238"
 TTSVoiceSpeed = Literal["fastest", "fast", "normal", "slow", "slowest"]
 TTSVoiceEmotion = Literal[
     "anger:lowest",

livekit/plugins/cartesia/tts.py

@@ -24,7 +24,9 @@ from typing import Any
 import aiohttp
 from livekit import rtc
 from livekit.agents import (
+    DEFAULT_API_CONNECT_OPTIONS,
     APIConnectionError,
+    APIConnectOptions,
     APIStatusError,
     APITimeoutError,
     tokenize,
@@ -65,7 +67,7 @@ class TTS(tts.TTS):
     def __init__(
         self,
         *,
-        model: TTSModels | str = "sonic-english",
+        model: TTSModels | str = "sonic",
         language: str = "en",
         encoding: TTSEncoding = "pcm_s16le",
         voice: str | list[float] = TTSDefaultVoiceId,
@@ -149,23 +151,47 @@ class TTS(tts.TTS):
         if emotion is not None:
             self._opts.emotion = emotion
 
-    def synthesize(self, text: str) -> "ChunkedStream":
-        return ChunkedStream(self, text, self._opts, self._ensure_session())
+    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,
+            session=self._ensure_session(),
+        )
 
-    def stream(self) -> "SynthesizeStream":
-        return SynthesizeStream(self, self._opts, self._ensure_session())
+    def stream(
+        self, *, conn_options: APIConnectOptions = DEFAULT_API_CONNECT_OPTIONS
+    ) -> "SynthesizeStream":
+        return SynthesizeStream(
+            tts=self,
+            conn_options=conn_options,
+            opts=self._opts,
+            session=self._ensure_session(),
+        )
 
 
 class ChunkedStream(tts.ChunkedStream):
     """Synthesize chunked text using the bytes endpoint"""
 
     def __init__(
-        self, tts: TTS, text: str, opts: _TTSOptions, session: aiohttp.ClientSession
+        self,
+        *,
+        tts: TTS,
+        input_text: str,
+        conn_options: APIConnectOptions,
+        opts: _TTSOptions,
+        session: aiohttp.ClientSession,
     ) -> None:
-        super().__init__(tts, text)
+        super().__init__(tts=tts, input_text=input_text, conn_options=conn_options)
         self._opts, self._session = opts, session
 
-    async def _main_task(self) -> None:
+    async def _run(self) -> None:
         request_id = utils.shortuuid()
         bstream = utils.audio.AudioByteStream(
             sample_rate=self._opts.sample_rate, num_channels=NUM_CHANNELS
@@ -184,6 +210,10 @@ class ChunkedStream(tts.ChunkedStream):
                 "https://api.cartesia.ai/tts/bytes",
                 headers=headers,
                 json=json,
+                timeout=aiohttp.ClientTimeout(
+                    total=30,
+                    sock_connect=self._conn_options.timeout,
+                ),
             ) as resp:
                 resp.raise_for_status()
                 async for data, _ in resp.content.iter_chunks():
@@ -215,47 +245,22 @@ class ChunkedStream(tts.ChunkedStream):
 class SynthesizeStream(tts.SynthesizeStream):
     def __init__(
         self,
+        *,
         tts: TTS,
+        conn_options: APIConnectOptions,
         opts: _TTSOptions,
         session: aiohttp.ClientSession,
     ):
-        super().__init__(tts)
+        super().__init__(tts=tts, conn_options=conn_options)
         self._opts, self._session = opts, session
         self._sent_tokenizer_stream = tokenize.basic.SentenceTokenizer(
             min_sentence_len=BUFFERED_WORDS_COUNT
         ).stream()
 
-    @utils.log_exceptions(logger=logger)
-    async def _main_task(self) -> None:
-        retry_count = 0
-        max_retry = 3
-        while self._input_ch.qsize() or not self._input_ch.closed:
-            try:
-                url = f"wss://api.cartesia.ai/tts/websocket?api_key={self._opts.api_key}&cartesia_version={API_VERSION}"
-                ws = await self._session.ws_connect(url)
-                retry_count = 0  # connected successfully, reset the retry_count
-
-                await self._run_ws(ws)
-            except Exception as e:
-                if retry_count >= max_retry:
-                    logger.exception(
-                        f"failed to connect to Cartesia after {max_retry} tries"
-                    )
-                    break
-
-                retry_delay = min(retry_count * 2, 10)  # max 10s
-                retry_count += 1
-
-                logger.warning(
-                    f"Cartesia connection failed, retrying in {retry_delay}s",
-                    exc_info=e,
-                )
-                await asyncio.sleep(retry_delay)
-
-    async def _run_ws(self, ws: aiohttp.ClientWebSocketResponse) -> None:
+    async def _run(self) -> None:
         request_id = utils.shortuuid()
 
-        async def sentence_stream_task():
+        async def _sentence_stream_task(ws: aiohttp.ClientWebSocketResponse):
             base_pkt = _to_cartesia_options(self._opts)
             async for ev in self._sent_tokenizer_stream:
                 token_pkt = base_pkt.copy()
@@ -270,7 +275,7 @@ class SynthesizeStream(tts.SynthesizeStream):
             end_pkt["continue"] = False
             await ws.send_str(json.dumps(end_pkt))
 
-        async def input_task():
+        async def _input_task():
             async for data in self._input_ch:
                 if isinstance(data, self._FlushSentinel):
                     self._sent_tokenizer_stream.flush()
@@ -278,7 +283,7 @@ class SynthesizeStream(tts.SynthesizeStream):
                 self._sent_tokenizer_stream.push_text(data)
             self._sent_tokenizer_stream.end_input()
 
-        async def recv_task():
+        async def _recv_task(ws: aiohttp.ClientWebSocketResponse):
             audio_bstream = utils.audio.AudioByteStream(
                 sample_rate=self._opts.sample_rate,
                 num_channels=NUM_CHANNELS,
@@ -307,7 +312,10 @@ class SynthesizeStream(tts.SynthesizeStream):
                     aiohttp.WSMsgType.CLOSE,
                     aiohttp.WSMsgType.CLOSING,
                 ):
-                    raise Exception("Cartesia connection closed unexpectedly")
+                    raise APIStatusError(
+                        "Cartesia connection closed unexpectedly",
+                        request_id=request_id,
+                    )
 
                 if msg.type != aiohttp.WSMsgType.TEXT:
                     logger.warning("unexpected Cartesia message type %s", msg.type)
@@ -335,16 +343,28 @@ class SynthesizeStream(tts.SynthesizeStream):
                 else:
                     logger.error("unexpected Cartesia message %s", data)
 
-        tasks = [
-            asyncio.create_task(input_task()),
-            asyncio.create_task(sentence_stream_task()),
-            asyncio.create_task(recv_task()),
-        ]
+        url = f"wss://api.cartesia.ai/tts/websocket?api_key={self._opts.api_key}&cartesia_version={API_VERSION}"
+
+        ws: aiohttp.ClientWebSocketResponse | None = None
 
         try:
-            await asyncio.gather(*tasks)
+            ws = await asyncio.wait_for(
+                self._session.ws_connect(url), self._conn_options.timeout
+            )
+
+            tasks = [
+                asyncio.create_task(_input_task()),
+                asyncio.create_task(_sentence_stream_task(ws)),
+                asyncio.create_task(_recv_task(ws)),
+            ]
+
+            try:
+                await asyncio.gather(*tasks)
+            finally:
+                await utils.aio.gracefully_cancel(*tasks)
         finally:
-            await utils.aio.gracefully_cancel(*tasks)
+            if ws is not None:
+                await ws.close()
 
 
 def _to_cartesia_options(opts: _TTSOptions) -> dict[str, Any]:

livekit/plugins/cartesia/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "0.4.3"
+__version__ = "0.4.6"

{livekit_plugins_cartesia-0.4.3.dist-info → livekit_plugins_cartesia-0.4.6.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: livekit-plugins-cartesia
-Version: 0.4.3
+Version: 0.4.6
 Summary: LiveKit Agents Plugin for Cartesia
 Home-page: https://github.com/livekit/agents
 License: Apache-2.0
@@ -19,7 +19,17 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3 :: Only
 Requires-Python: >=3.9.0
 Description-Content-Type: text/markdown
-Requires-Dist: livekit-agents >=0.11
+Requires-Dist: livekit-agents>=0.12.3
+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 Cartesia
 

livekit_plugins_cartesia-0.4.6.dist-info/RECORD

@@ -0,0 +1,10 @@
+livekit/plugins/cartesia/__init__.py,sha256=UTa6Q7IxhRBCwPftowHEUDvmBg99J_UjGS_yxTzKD7g,1095
+livekit/plugins/cartesia/log.py,sha256=4Mnhjng_DU1dIWP9IWjIQGZ67EV3LnQhWMWCHVudJbo,71
+livekit/plugins/cartesia/models.py,sha256=56CJgo7my-w-vpedir_ImV_aqKASeLihE5DbcCCgGJI,950
+livekit/plugins/cartesia/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+livekit/plugins/cartesia/tts.py,sha256=M4Bxxu0cTyhA3QahgFubZcpREENw9C5GZjaepN4tPmc,13869
+livekit/plugins/cartesia/version.py,sha256=D_Vma_D7upABLhXhjK1P5nahb5uh4zpuIiqYDq9sgJ0,600
+livekit_plugins_cartesia-0.4.6.dist-info/METADATA,sha256=HGMhsernokDVPky0lA5zT6eNgK9MHHbiFi5eVoFKsPA,1462
+livekit_plugins_cartesia-0.4.6.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
+livekit_plugins_cartesia-0.4.6.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
+livekit_plugins_cartesia-0.4.6.dist-info/RECORD,,

{livekit_plugins_cartesia-0.4.3.dist-info → livekit_plugins_cartesia-0.4.6.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.3.0)
+Generator: setuptools (75.8.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

livekit_plugins_cartesia-0.4.3.dist-info/RECORD

@@ -1,10 +0,0 @@
-livekit/plugins/cartesia/__init__.py,sha256=UTa6Q7IxhRBCwPftowHEUDvmBg99J_UjGS_yxTzKD7g,1095
-livekit/plugins/cartesia/log.py,sha256=4Mnhjng_DU1dIWP9IWjIQGZ67EV3LnQhWMWCHVudJbo,71
-livekit/plugins/cartesia/models.py,sha256=fOO276Vzw3OkDUWUVcw7PH95ctFy38rj3q9I6_mYQ7M,950
-livekit/plugins/cartesia/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-livekit/plugins/cartesia/tts.py,sha256=2xwWOIjwLDOF4TbHlDibrZpUju9If8WrNpHQ2JMuBC0,13533
-livekit/plugins/cartesia/version.py,sha256=u7PSD5TBbPRIhE8vJkBVJzq_eGqYfg6RP5c3VKNlKGk,600
-livekit_plugins_cartesia-0.4.3.dist-info/METADATA,sha256=w5q0oz6rdHDL5cxAyT5hWbHqhZnOPnZYGl3aUKsr3z4,1246
-livekit_plugins_cartesia-0.4.3.dist-info/WHEEL,sha256=P9jw-gEje8ByB7_hXoICnHtVCrEwMQh-630tKvQWehc,91
-livekit_plugins_cartesia-0.4.3.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
-livekit_plugins_cartesia-0.4.3.dist-info/RECORD,,