cartesia 1.0.11__tar.gz → 1.0.13__tar.gz

{cartesia-1.0.11 → cartesia-1.0.13}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: cartesia
-Version: 1.0.11
+Version: 1.0.13
 Summary: The official Python library for the Cartesia API.
 Home-page: 
 Author: Cartesia, Inc.
@@ -73,6 +73,11 @@ print("The embedding for", voice["name"], "is", voice["embedding"])
 # Clone a voice using filepath
 cloned_voice_embedding = client.voices.clone(filepath="path/to/voice")
 
+# Mix voices together
+mixed_voice_embedding = client.voices.mix(
+    [{ "id": "voice_id_1", "weight": 0.5 }, { "id": "voice_id_2", "weight": 0.25 }, { "id": "voice_id_3", "weight": 0.25 }]
+)
+
 # Create a new voice
 new_voice = client.voices.create(
     name="New Voice",
@@ -504,6 +509,7 @@ You can enhance the voice output by adjusting the `speed` and `emotion` paramete
 
 Speed Options:
 - `slowest`, `slow`, `normal`, `fast`, `fastest`
+- Float values between -1.0 and 1.0, where -1.0 is the slowest speed and 1.0 is the fastest speed.
 
 Emotion Options:
 Use a list of tags in the format `emotion_name:level` where:

{cartesia-1.0.11 → cartesia-1.0.13}/README.md

@@ -56,6 +56,11 @@ print("The embedding for", voice["name"], "is", voice["embedding"])
 # Clone a voice using filepath
 cloned_voice_embedding = client.voices.clone(filepath="path/to/voice")
 
+# Mix voices together
+mixed_voice_embedding = client.voices.mix(
+    [{ "id": "voice_id_1", "weight": 0.5 }, { "id": "voice_id_2", "weight": 0.25 }, { "id": "voice_id_3", "weight": 0.25 }]
+)
+
 # Create a new voice
 new_voice = client.voices.create(
     name="New Voice",
@@ -487,6 +492,7 @@ You can enhance the voice output by adjusting the `speed` and `emotion` paramete
 
 Speed Options:
 - `slowest`, `slow`, `normal`, `fast`, `fastest`
+- Float values between -1.0 and 1.0, where -1.0 is the slowest speed and 1.0 is the fastest speed.
 
 Emotion Options:
 Use a list of tags in the format `emotion_name:level` where:

{cartesia-1.0.11 → cartesia-1.0.13}/cartesia/client.py

@@ -32,7 +32,6 @@ except ImportError:
     IS_WEBSOCKET_SYNC_AVAILABLE = False
 
 from iterators import TimeoutIterator
-from websockets.sync.client import connect
 
 from cartesia._types import (
     DeprecatedOutputFormatMapping,
@@ -50,6 +49,7 @@ DEFAULT_BASE_URL = "api.cartesia.ai"
 DEFAULT_CARTESIA_VERSION = "2024-06-10"  # latest version
 DEFAULT_TIMEOUT = 30  # seconds
 DEFAULT_NUM_CONNECTIONS = 10  # connections per client
+DEFAULT_VOICE_EMBEDDING = [1.0] * 192  # Default voice embedding is a 192 sized float array
 
 BACKOFF_FACTOR = 1
 MAX_RETRIES = 3
@@ -261,6 +261,40 @@ class Voices(Resource):
 
         return response.json()
 
+    def mix(self, voices: List[Dict[str, Union[str, float]]]) -> List[float]:
+        """Mix multiple voices together.
+
+        Args:
+            voices: A list of dictionaries, each containing either:
+                        - 'id': The ID of an existing voice
+                        - 'embedding': A voice embedding
+                    AND
+                        - 'weight': The weight of the voice in the mix (0.0 to 1.0)
+
+        Returns:
+            The embedding of the mixed voice as a list of floats.
+
+        Raises:
+            ValueError: If the request fails or if the input is invalid.
+        """
+        url = f"{self._http_url()}/voices/mix"
+
+        if not voices or not isinstance(voices, list):
+            raise ValueError("voices must be a non-empty list")
+
+        response = httpx.post(
+            url,
+            headers=self.headers,
+            json={"voices": voices},
+            timeout=self.timeout,
+        )
+
+        if not response.is_success:
+            raise ValueError(f"Failed to mix voices. Error: {response.text}")
+
+        result = response.json()
+        return result["embedding"]
+
 
 class _TTSContext:
     """Manage a single context over a WebSocket.
@@ -857,15 +891,17 @@ class TTS(Resource):
         if voice_id is None and voice_embedding is None:
             raise ValueError("Either voice_id or voice_embedding must be specified.")
 
-        if voice_id is not None and voice_embedding is not None:
-            raise ValueError("Only one of voice_id or voice_embedding should be specified.")
+        voice = {}
+
+        if voice_id is not None:
+            voice["id"] = voice_id
+
+        if voice_embedding is not None:
+            voice["embedding"] = voice_embedding
 
-        if voice_id:
-            voice = {"mode": "id", "id": voice_id}
-        else:
-            voice = {"mode": "embedding", "embedding": voice_embedding}
         if experimental_voice_controls is not None:
             voice["__experimental_controls"] = experimental_voice_controls
+
         return voice
 
 
@@ -1129,7 +1165,7 @@ class _AsyncTTSContext:
             model_id=DEFAULT_MODEL_ID,
             transcript="",
             output_format=TTS.get_output_format("raw_pcm_f32le_44100"),
-            voice_id="a0e99841-438c-4a64-b679-ae501e7d6091",  # Default voice ID since it's a required input for now
+            voice_embedding=DEFAULT_VOICE_EMBEDDING,  # Default voice embedding since it's a required input for now.
             context_id=self._context_id,
             continue_=False,
         )

cartesia-1.0.13/cartesia/version.py

@@ -0,0 +1 @@
+__version__ = "1.0.13"

{cartesia-1.0.11 → cartesia-1.0.13}/cartesia.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: cartesia
-Version: 1.0.11
+Version: 1.0.13
 Summary: The official Python library for the Cartesia API.
 Home-page: 
 Author: Cartesia, Inc.
@@ -73,6 +73,11 @@ print("The embedding for", voice["name"], "is", voice["embedding"])
 # Clone a voice using filepath
 cloned_voice_embedding = client.voices.clone(filepath="path/to/voice")
 
+# Mix voices together
+mixed_voice_embedding = client.voices.mix(
+    [{ "id": "voice_id_1", "weight": 0.5 }, { "id": "voice_id_2", "weight": 0.25 }, { "id": "voice_id_3", "weight": 0.25 }]
+)
+
 # Create a new voice
 new_voice = client.voices.create(
     name="New Voice",
@@ -504,6 +509,7 @@ You can enhance the voice output by adjusting the `speed` and `emotion` paramete
 
 Speed Options:
 - `slowest`, `slow`, `normal`, `fast`, `fastest`
+- Float values between -1.0 and 1.0, where -1.0 is the slowest speed and 1.0 is the fastest speed.
 
 Emotion Options:
 Use a list of tags in the format `emotion_name:level` where: