cartesia 1.0.10__tar.gz → 1.0.11__tar.gz

{cartesia-1.0.10 → cartesia-1.0.11}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: cartesia
-Version: 1.0.10
+Version: 1.0.11
 Summary: The official Python library for the Cartesia API.
 Home-page: 
 Author: Cartesia, Inc.

{cartesia-1.0.10 → cartesia-1.0.11}/cartesia/client.py

@@ -294,6 +294,7 @@ class _TTSContext:
         context_id: Optional[str] = None,
         duration: Optional[int] = None,
         language: Optional[str] = None,
+        add_timestamps: bool = False,
         _experimental_voice_controls: Optional[VoiceControls] = None,
     ) -> Generator[bytes, None, None]:
         """Send audio generation requests to the WebSocket and yield responses.
@@ -307,6 +308,7 @@ class _TTSContext:
             context_id: The context ID to use for the request. If not specified, a random context ID will be generated.
             duration: The duration of the audio in seconds.
             language: The language code for the audio request. This can only be used with `model_id = sonic-multilingual`
+            add_timestamps: Whether to return word-level timestamps.
             _experimental_voice_controls: Experimental voice controls for controlling speed and emotion.
                 Note: This is an experimental feature and may change rapidly in future releases.
 
@@ -341,6 +343,7 @@ class _TTSContext:
             },
             "context_id": self._context_id,
             "language": language,
+            "add_timestamps": add_timestamps,
         }
 
         if duration is not None:

cartesia-1.0.11/cartesia/version.py

@@ -0,0 +1 @@
+__version__ = "1.0.11"

{cartesia-1.0.10 → cartesia-1.0.11}/cartesia.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: cartesia
-Version: 1.0.10
+Version: 1.0.11
 Summary: The official Python library for the Cartesia API.
 Home-page: 
 Author: Cartesia, Inc.

{cartesia-1.0.10 → cartesia-1.0.11}/tests/test_tts.py

@@ -190,7 +190,6 @@ def test_websocket_send_timestamps(resources: _Resources, stream: bool):
     
     ws.close()
 
-        
 @pytest.mark.parametrize("_experimental_voice_controls", [None, EXPERIMENTAL_VOICE_CONTROLS, EXPERIMENTAL_VOICE_CONTROLS_2])
 def test_sse_send_context_manager(resources: _Resources, _experimental_voice_controls: VoiceControls):
     logger.info("Testing SSE send context manager")
@@ -456,6 +455,28 @@ def test_sync_continuation_websocket_context_send():
             assert isinstance(out["audio"], bytes)
     finally:
         ws.close()
+        
+def test_sync_context_send_timestamps(resources: _Resources):
+    logger.info("Testing WebSocket send")
+    client = resources.client
+    transcripts = ["Hello, world!", "I'\''m generating audio on Cartesia."]
+
+    ws = client.tts.websocket()
+    ctx = ws.context()
+    output_generate = ctx.send(transcript=chunk_generator(transcripts), voice_id=SAMPLE_VOICE_ID, output_format={
+        "container": "raw",
+        "encoding": "pcm_f32le",
+        "sample_rate": 44100
+    }, model_id=DEFAULT_MODEL_ID, add_timestamps=True)
+
+    has_wordtimestamps = False
+    for out in output_generate:
+        has_wordtimestamps |= "word_timestamps" in out
+        _validate_schema(out)
+
+    assert has_wordtimestamps, "No word timestamps found"
+    
+    ws.close()
     
 @pytest.mark.asyncio
 async def test_continuation_websocket_context_send():

cartesia-1.0.10/cartesia/version.py

@@ -1 +0,0 @@
-__version__ = "1.0.10"