feat: support arbitrary attributes for speak provider

Author: naomi-lgbt

deepgram/__init__.py

@@ -342,7 +342,6 @@
     Listen,
     ListenProvider,
     Speak,
-    SpeakProvider,
     Header,
     Item,
     Properties,

deepgram/client.py

@@ -356,7 +356,6 @@
     Listen,
     ListenProvider,
     Speak,
-    SpeakProvider,
     Header,
     Item,
     Properties,

deepgram/clients/__init__.py

@@ -365,7 +365,6 @@
     Listen,
     ListenProvider,
     Speak,
-    SpeakProvider,
     Header,
     Item,
     Properties,

deepgram/clients/agent/__init__.py

@@ -40,7 +40,6 @@
     Listen,
     ListenProvider,
     Speak,
-    SpeakProvider,
     Header,
     Item,
     Properties,

deepgram/clients/agent/client.py

@@ -39,7 +39,6 @@
     Listen as LatestListen,
     ListenProvider as LatestListenProvider,
     Speak as LatestSpeak,
-    SpeakProvider as LatestSpeakProvider,
     Header as LatestHeader,
     Item as LatestItem,
     Properties as LatestProperties,
@@ -87,7 +86,6 @@
 Listen = LatestListen
 ListenProvider = LatestListenProvider
 Speak = LatestSpeak
-SpeakProvider = LatestSpeakProvider
 Header = LatestHeader
 Item = LatestItem
 Properties = LatestProperties

deepgram/clients/agent/v1/__init__.py

@@ -44,7 +44,6 @@
     Listen,
     ListenProvider,
     Speak,
-    SpeakProvider,
     Header,
     Item,
     Properties,

deepgram/clients/agent/v1/websocket/__init__.py

@@ -35,7 +35,6 @@
     Listen,
     ListenProvider,
     Speak,
-    SpeakProvider,
     Header,
     Item,
     Properties,

deepgram/clients/agent/v1/websocket/options.py

@@ -2,7 +2,7 @@
 # Use of this source code is governed by a MIT license that can be found in the LICENSE file.
 # SPDX-License-Identifier: MIT
 
-from typing import List, Optional, Union, Any, Tuple
+from typing import List, Optional, Union, Any, Tuple, Dict
 import logging
 
 from dataclasses import dataclass, field
@@ -167,52 +167,6 @@ class ThinkProvider(BaseResponse):
     )
 
 
-@dataclass
-class SpeakProvider(BaseResponse):
-    """
-    This class defines the provider for the Speak model.
-    """
-
-    type: Optional[str] = field(default="deepgram")
-    """
-    Deepgram OR OpenAI model to use.
-    """
-    model: Optional[str] = field(
-        default="aura-2-thalia-en",
-        metadata=dataclass_config(exclude=lambda f: f is None),
-    )
-    """
-    ElevenLabs or Cartesia model to use.
-    """
-    model_id: Optional[str] = field(
-        default=None, metadata=dataclass_config(exclude=lambda f: f is None)
-    )
-    """
-    Cartesia voice configuration.
-    """
-    voice: Optional[CartesiaVoice] = field(
-        default=None, metadata=dataclass_config(exclude=lambda f: f is None)
-    )
-    """
-    Cartesia language.
-    """
-    language: Optional[str] = field(
-        default=None, metadata=dataclass_config(exclude=lambda f: f is None)
-    )
-    """
-    ElevenLabs language.
-    """
-    language_code: Optional[str] = field(
-        default=None, metadata=dataclass_config(exclude=lambda f: f is None)
-    )
-
-    def __getitem__(self, key):
-        _dict = self.to_dict()
-        if "voice" in _dict and isinstance(_dict["voice"], dict):
-            _dict["voice"] = CartesiaVoice.from_dict(_dict["voice"])
-        return _dict[key]
-
-
 @dataclass
 class Think(BaseResponse):
     """
@@ -264,15 +218,26 @@ class Speak(BaseResponse):
     This class defines any configuration settings for the Speak model.
     """
 
-    provider: SpeakProvider = field(default_factory=SpeakProvider)
+    provider: dict = field(default_factory=dict)
     endpoint: Optional[Endpoint] = field(
         default=None, metadata=dataclass_config(exclude=lambda f: f is None)
     )
 
+    def __post_init__(self):
+        # Allow attribute-style access to provider dict
+        class AttrDict(dict):
+            def __getattr__(self, name):
+                try:
+                    return self[name]
+                except KeyError:
+                    raise AttributeError(name)
+            def __setattr__(self, name, value):
+                self[name] = value
+        if not isinstance(self.provider, AttrDict):
+            self.provider = AttrDict(self.provider)
+
     def __getitem__(self, key):
         _dict = self.to_dict()
-        if "provider" in _dict and isinstance(_dict["provider"], dict):
-            _dict["provider"] = SpeakProvider.from_dict(_dict["provider"])
         if "endpoint" in _dict and isinstance(_dict["endpoint"], dict):
             _dict["endpoint"] = Endpoint.from_dict(_dict["endpoint"])
         return _dict[key]

examples/agent/arbitrary_keys/main.py

@@ -0,0 +1,100 @@
+# Copyright 2025 Deepgram SDK contributors. All Rights Reserved.
+# Use of this source code is governed by a MIT license that can be found in the LICENSE file.
+# SPDX-License-Identifier: MIT
+
+# Import dependencies and set up the main function
+import requests
+import wave
+import io
+import time
+import os
+import json
+import threading
+from datetime import datetime
+
+from deepgram import (
+    DeepgramClient,
+    DeepgramClientOptions,
+    AgentWebSocketEvents,
+    AgentKeepAlive,
+)
+from deepgram.clients.agent.v1.websocket.options import SettingsOptions
+
+def main():
+    try:
+        # Initialize the Voice Agent
+        api_key = os.getenv("DEEPGRAM_API_KEY")
+        if not api_key:
+            raise ValueError("DEEPGRAM_API_KEY environment variable is not set")
+        print(f"API Key found:")
+
+        # Initialize Deepgram client
+        config = DeepgramClientOptions(
+            options={
+                "keepalive": "true",
+                # "speaker_playback": "true",
+            },
+        )
+        deepgram = DeepgramClient(api_key, config)
+        connection = deepgram.agent.websocket.v("1")
+        print("Created WebSocket connection...")
+
+        # 4. Configure the Agent
+        options = SettingsOptions()
+        # Audio input configuration
+        options.audio.input.encoding = "linear16"
+        options.audio.input.sample_rate = 24000
+        # Audio output configuration
+        options.audio.output.encoding = "linear16"
+        options.audio.output.sample_rate = 24000
+        options.audio.output.container = "wav"
+        # Agent configuration
+        options.agent.language = "en"
+        options.agent.listen.provider.type = "deepgram"
+        options.agent.listen.provider.model = "nova-3"
+        options.agent.think.provider.type = "open_ai"
+        options.agent.think.provider.model = "gpt-4o-mini"
+        options.agent.think.prompt = "You are a friendly AI assistant."
+        options.agent.speak.provider.type = "deepgram"
+        options.agent.speak.provider.model = "aura-2-thalia-en"
+        options.agent.greeting = "Hello! How can I help you today?"
+        options.agent.speak.provider.arbitrary_key = "test"
+
+        def on_welcome(self, welcome, **kwargs):
+            print(f"Welcome message received: {welcome}")
+            with open("chatlog.txt", 'a') as chatlog:
+                chatlog.write(f"Welcome message: {welcome}\n")
+
+        def on_settings_applied(self, settings_applied, **kwargs):
+            print(f"Settings applied: {settings_applied}")
+            with open("chatlog.txt", 'a') as chatlog:
+                chatlog.write(f"Settings applied: {settings_applied}\n")
+
+        def on_error(self, error, **kwargs):
+            print(f"Error received: {error}")
+            with open("chatlog.txt", 'a') as chatlog:
+                chatlog.write(f"Error: {error}\n")
+
+        # Register handlers
+        connection.on(AgentWebSocketEvents.Welcome, on_welcome)
+        connection.on(AgentWebSocketEvents.SettingsApplied, on_settings_applied)
+        connection.on(AgentWebSocketEvents.Error, on_error)
+        print("Event handlers registered")
+
+        # Start the connection
+        print("Starting WebSocket connection...")
+        print(options)
+        if not connection.start(options):
+            print("Failed to start connection")
+            return
+        print("WebSocket connection started successfully")
+
+        # Cleanup
+        connection.finish()
+        print("Finished")
+
+    except Exception as e:
+        print(f"Error: {str(e)}")
+
+if __name__ == "__main__":
+    main()