webscout 8.3.5__py3-none-any.whl → 8.3.7__py3-none-any.whl

webscout/Provider/TTS/deepgram.py

@@ -1,156 +1,409 @@
-import time
-import requests
-import pathlib
-import base64
-import tempfile
-from io import BytesIO
-from webscout import exceptions
-from concurrent.futures import ThreadPoolExecutor, as_completed
-from webscout.litagent import LitAgent
-from . import utils
-from .base import BaseTTSProvider
-
-class DeepgramTTS(BaseTTSProvider):
-    """
-    Text-to-speech provider using the DeepgramTTS API.
-    """
-    # Request headers
-    headers: dict[str, str] = {
-        "User-Agent": LitAgent().random()
-    }
-    all_voices: dict[str, str] = {
-        "Asteria": "aura-asteria-en", "Arcas": "aura-arcas-en", "Luna": "aura-luna-en",
-        "Zeus": "aura-zeus-en", "Orpheus": "aura-orpheus-en", "Angus": "aura-angus-en",
-        "Athena": "aura-athena-en", "Helios": "aura-helios-en", "Hera": "aura-hera-en",
-        "Orion": "aura-orion-en", "Perseus": "aura-perseus-en", "Stella": "aura-stella-en"
-    }
-
-    def __init__(self, timeout: int = 20, proxies: dict = None):
-        """Initializes the DeepgramTTS TTS client."""
-        super().__init__()
-        self.session = requests.Session()
-        self.session.headers.update(self.headers)
-        if proxies:
-            self.session.proxies.update(proxies)
-        self.timeout = timeout
-
-    def tts(self, text: str, voice: str = "Brian", verbose: bool = True) -> str:
-        """
-        Converts text to speech using the DeepgramTTS API and saves it to a file.
-
-        Args:
-            text (str): The text to convert to speech
-            voice (str): The voice to use for TTS (default: "Brian")
-            verbose (bool): Whether to print progress messages (default: True)
-
-        Returns:
-            str: Path to the generated audio file
-
-        Raises:
-            AssertionError: If the specified voice is not available
-            requests.RequestException: If there's an error communicating with the API
-            RuntimeError: If there's an error processing the audio
-        """
-        assert (
-            voice in self.all_voices
-        ), f"Voice '{voice}' not one of [{', '.join(self.all_voices.keys())}]"
-
-        url = "https://deepgram.com/api/ttsAudioGeneration"
-        filename = pathlib.Path(tempfile.mktemp(suffix=".mp3", dir=self.temp_dir))
-
-        # Split text into sentences using the utils module
-        sentences = utils.split_sentences(text)
-        if verbose:
-            for index, sen in enumerate(sentences):
-                print(f"[debug] Sentence {index}: {sen}")
-
-        def generate_audio_for_chunk(part_text: str, part_number: int):
-            """
-            Generate audio for a single chunk of text.
-
-            Args:
-                part_text (str): The text chunk to convert
-                part_number (int): The chunk number for ordering
-
-            Returns:
-                tuple: (part_number, audio_data)
-
-            Raises:
-                requests.RequestException: If there's an API error
-            """
-            max_retries = 3
-            retry_count = 0
-
-            while retry_count < max_retries:
-                try:
-                    payload = {"text": part_text, "model": self.all_voices[voice]}
-                    response = self.session.post(
-                        url=url,
-                        headers=self.headers,
-                        json=payload,
-                        stream=True,
-                        timeout=self.timeout
-                    )
-                    response.raise_for_status()
-
-                    response_data = response.json().get('data')
-                    if response_data:
-                        audio_data = base64.b64decode(response_data)
-                        if verbose:
-                            print(f"[debug] Chunk {part_number} processed successfully")
-                        return part_number, audio_data
-
-                    if verbose:
-                        print(f"[debug] No data received for chunk {part_number}. Attempt {retry_count + 1}/{max_retries}")
-
-                except requests.RequestException as e:
-                    if verbose:
-                        print(f"[debug] Error processing chunk {part_number}: {str(e)}. Attempt {retry_count + 1}/{max_retries}")
-                    if retry_count == max_retries - 1:
-                        raise
-
-                retry_count += 1
-                time.sleep(1)
-
-            raise RuntimeError(f"Failed to generate audio for chunk {part_number} after {max_retries} attempts")
-
-        try:
-            # Using ThreadPoolExecutor to handle requests concurrently
-            with ThreadPoolExecutor() as executor:
-                futures = {
-                    executor.submit(generate_audio_for_chunk, sentence.strip(), chunk_num): chunk_num
-                    for chunk_num, sentence in enumerate(sentences, start=1)
-                }
-
-                # Dictionary to store results with order preserved
-                audio_chunks = {}
-
-                for future in as_completed(futures):
-                    chunk_num = futures[future]
-                    try:
-                        part_number, audio_data = future.result()
-                        audio_chunks[part_number] = audio_data
-                    except Exception as e:
-                        raise RuntimeError(f"Failed to generate audio for chunk {chunk_num}: {str(e)}")
-
-                # Combine all audio chunks in order
-                with open(filename, 'wb') as f:
-                    for chunk_num in sorted(audio_chunks.keys()):
-                        f.write(audio_chunks[chunk_num])
-
-                if verbose:
-                    print(f"[debug] Audio saved to {filename}")
-                return str(filename)
-
-        except Exception as e:
-            print(f"[debug] Failed to generate audio: {str(e)}") if verbose else None
-            raise RuntimeError(f"Failed to generate audio: {str(e)}")
-
-# Example usage
-if __name__ == "__main__":
-    deepgram = DeepgramTTS()
-    text = "This is a test of the DeepgramTTS text-to-speech API. It supports multiple sentences. Let's see how it works!"
-
-    print("[debug] Generating audio...")
-    audio_file = deepgram.tts(text, voice="Asteria")
-    print(f"Audio saved to: {audio_file}")
+##################################################################################
+##  Deepgram TTS Provider                                                      ##
+##################################################################################
+import time
+import requests
+import pathlib
+import base64
+import tempfile
+from io import BytesIO
+from webscout import exceptions
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from webscout.litagent import LitAgent
+try:
+    from . import utils
+    from .base import BaseTTSProvider
+except ImportError:
+    # Handle direct execution
+    import sys
+    import os
+    sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..', '..', '..'))
+    from webscout.Provider.TTS import utils
+    from webscout.Provider.TTS.base import BaseTTSProvider
+
+class DeepgramTTS(BaseTTSProvider):
+    """
+    Text-to-speech provider using the Deepgram API with OpenAI-compatible interface.
+    
+    This provider follows the OpenAI TTS API structure with support for:
+    - Multiple TTS models (gpt-4o-mini-tts, tts-1, tts-1-hd)
+    - Deepgram's Aura voices optimized for English
+    - Voice instructions for controlling speech aspects
+    - Multiple output formats
+    - Streaming support
+    """
+    
+    # Request headers
+    headers: dict[str, str] = {
+        "User-Agent": LitAgent().random()
+    }
+    
+    # Override supported models for Deepgram
+    SUPPORTED_MODELS = None
+    
+    # Deepgram Aura voices (mapped to OpenAI-compatible names)
+    SUPPORTED_VOICES = [
+        "asteria",   # aura-asteria-en - Clear and articulate female voice
+        "arcas",     # aura-arcas-en - Warm male voice
+        "luna",      # aura-luna-en - Gentle female voice
+        "zeus",      # aura-zeus-en - Authoritative male voice
+        "orpheus",   # aura-orpheus-en - Melodic male voice
+        "angus",     # aura-angus-en - Scottish-accented male voice
+        "athena",    # aura-athena-en - Professional female voice
+        "helios",    # aura-helios-en - Bright male voice
+        "hera",      # aura-hera-en - Mature female voice
+        "orion",     # aura-orion-en - Deep male voice
+        "perseus",   # aura-perseus-en - Young male voice
+        "stella"     # aura-stella-en - Friendly female voice
+    ]
+    
+    # Voice mapping for Deepgram API compatibility
+    voice_mapping = {
+        "asteria": "aura-asteria-en",
+        "arcas": "aura-arcas-en", 
+        "luna": "aura-luna-en",
+        "zeus": "aura-zeus-en",
+        "orpheus": "aura-orpheus-en",
+        "angus": "aura-angus-en",
+        "athena": "aura-athena-en",
+        "helios": "aura-helios-en",
+        "hera": "aura-hera-en",
+        "orion": "aura-orion-en",
+        "perseus": "aura-perseus-en",
+        "stella": "aura-stella-en"
+    }
+    
+    # Legacy voice mapping for backward compatibility
+    all_voices: dict[str, str] = {
+        "Asteria": "aura-asteria-en", "Arcas": "aura-arcas-en", "Luna": "aura-luna-en",
+        "Zeus": "aura-zeus-en", "Orpheus": "aura-orpheus-en", "Angus": "aura-angus-en",
+        "Athena": "aura-athena-en", "Helios": "aura-helios-en", "Hera": "aura-hera-en",
+        "Orion": "aura-orion-en", "Perseus": "aura-perseus-en", "Stella": "aura-stella-en"
+    }
+
+    def __init__(self, timeout: int = 20, proxies: dict = None):
+        """
+        Initialize the Deepgram TTS client.
+        
+        Args:
+            timeout (int): Request timeout in seconds
+            proxies (dict): Proxy configuration
+        """
+        super().__init__()
+        self.api_url = "https://deepgram.com/api/ttsAudioGeneration"
+        self.session = requests.Session()
+        self.session.headers.update(self.headers)
+        if proxies:
+            self.session.proxies.update(proxies)
+        self.timeout = timeout
+        # Override defaults for Deepgram
+        self.default_voice = "asteria"
+
+    def tts(
+        self, 
+        text: str, 
+        model: str = "gpt-4o-mini-tts",
+        voice: str = "asteria", 
+        response_format: str = "mp3",
+        instructions: str = None, 
+        verbose: bool = True
+    ) -> str:
+        """
+        Convert text to speech using Deepgram API with OpenAI-compatible parameters.
+
+        Args:
+            text (str): The text to convert to speech (max 10,000 characters)
+            model (str): The TTS model to use (gpt-4o-mini-tts, tts-1, tts-1-hd)
+            voice (str): The voice to use for TTS (asteria, arcas, luna, zeus, etc.)
+            response_format (str): Audio format (mp3, opus, aac, flac, wav, pcm)
+            instructions (str): Voice instructions (not used by Deepgram but kept for compatibility)
+            verbose (bool): Whether to print debug information
+
+        Returns:
+            str: Path to the generated audio file
+
+        Raises:
+            ValueError: If input parameters are invalid
+            exceptions.FailedToGenerateResponseError: If there is an error generating or saving the audio
+        """
+        # Validate input parameters
+        if not text or not isinstance(text, str):
+            raise ValueError("Input text must be a non-empty string")
+        if len(text) > 10000:
+            raise ValueError("Input text exceeds maximum allowed length of 10,000 characters")
+            
+        # Validate model, voice, and format using base class methods
+        model = self.validate_model(model)
+        voice = self.validate_voice(voice)
+        response_format = self.validate_format(response_format)
+        
+        # Map voice to Deepgram API format
+        deepgram_voice = self.voice_mapping.get(voice, voice)
+        
+        # Create temporary file with appropriate extension
+        file_extension = f".{response_format}" if response_format != "pcm" else ".wav"
+        filename = pathlib.Path(tempfile.mktemp(suffix=file_extension, dir=self.temp_dir))
+
+        # Split text into sentences using the utils module for better processing
+        sentences = utils.split_sentences(text)
+        if verbose:
+            print(f"[debug] Processing {len(sentences)} sentences")
+            print(f"[debug] Model: {model}")
+            print(f"[debug] Voice: {voice} -> {deepgram_voice}")
+            print(f"[debug] Format: {response_format}")
+
+        def generate_audio_for_chunk(part_text: str, part_number: int):
+            """
+            Generate audio for a single chunk of text.
+
+            Args:
+                part_text (str): The text chunk to convert
+                part_number (int): The chunk number for ordering
+
+            Returns:
+                tuple: (part_number, audio_data)
+
+            Raises:
+                requests.RequestException: If there's an API error
+            """
+            max_retries = 3
+            retry_count = 0
+
+            while retry_count < max_retries:
+                try:
+                    payload = {
+                        "text": part_text, 
+                        "model": deepgram_voice,
+                        # Add model parameter for future Deepgram API compatibility
+                        "tts_model": model
+                    }
+                    response = self.session.post(
+                        url=self.api_url,
+                        headers=self.headers,
+                        json=payload,
+                        stream=True,
+                        timeout=self.timeout
+                    )
+                    response.raise_for_status()
+
+                    response_data = response.json().get('data')
+                    if response_data:
+                        audio_data = base64.b64decode(response_data)
+                        if verbose:
+                            print(f"[debug] Chunk {part_number} processed successfully")
+                        return part_number, audio_data
+
+                    if verbose:
+                        print(f"[debug] No data received for chunk {part_number}. Attempt {retry_count + 1}/{max_retries}")
+
+                except requests.RequestException as e:
+                    if verbose:
+                        print(f"[debug] Error processing chunk {part_number}: {str(e)}. Attempt {retry_count + 1}/{max_retries}")
+                    if retry_count == max_retries - 1:
+                        raise exceptions.FailedToGenerateResponseError(f"Failed to generate audio for chunk {part_number}: {str(e)}")
+
+                retry_count += 1
+                time.sleep(1)
+
+            raise exceptions.FailedToGenerateResponseError(f"Failed to generate audio for chunk {part_number} after {max_retries} attempts")
+
+        try:
+            # Using ThreadPoolExecutor to handle requests concurrently
+            with ThreadPoolExecutor() as executor:
+                futures = {
+                    executor.submit(generate_audio_for_chunk, sentence.strip(), chunk_num): chunk_num
+                    for chunk_num, sentence in enumerate(sentences, start=1)
+                }
+
+                # Dictionary to store results with order preserved
+                audio_chunks = {}
+
+                for future in as_completed(futures):
+                    chunk_num = futures[future]
+                    try:
+                        part_number, audio_data = future.result()
+                        audio_chunks[part_number] = audio_data
+                    except Exception as e:
+                        raise exceptions.FailedToGenerateResponseError(f"Failed to generate audio for chunk {chunk_num}: {str(e)}")
+
+                # Combine all audio chunks in order
+                with open(filename, 'wb') as f:
+                    for chunk_num in sorted(audio_chunks.keys()):
+                        f.write(audio_chunks[chunk_num])
+
+                if verbose:
+                    print(f"[debug] Speech generated successfully")
+                    print(f"[debug] Audio saved to {filename}")
+                    
+                return str(filename)
+
+        except Exception as e:
+            if verbose:
+                print(f"[debug] Failed to generate audio: {str(e)}")
+            raise exceptions.FailedToGenerateResponseError(f"Failed to generate audio: {str(e)}")
+
+    def create_speech(
+        self,
+        input: str,
+        model: str = "gpt-4o-mini-tts", 
+        voice: str = "asteria",
+        response_format: str = "mp3",
+        instructions: str = None,
+        verbose: bool = False
+    ) -> str:
+        """
+        OpenAI-compatible speech creation interface.
+        
+        Args:
+            input (str): The text to convert to speech
+            model (str): The TTS model to use
+            voice (str): The voice to use
+            response_format (str): Audio format
+            instructions (str): Voice instructions (not used by Deepgram)
+            verbose (bool): Whether to print debug information
+            
+        Returns:
+            str: Path to the generated audio file
+        """
+        return self.tts(
+            text=input,
+            model=model,
+            voice=voice,
+            response_format=response_format,
+            instructions=instructions,
+            verbose=verbose
+        )
+
+    def with_streaming_response(self):
+        """
+        Return a streaming response context manager (OpenAI-compatible).
+        
+        Returns:
+            StreamingResponseContextManager: Context manager for streaming responses
+        """
+        return StreamingResponseContextManager(self)
+
+
+class StreamingResponseContextManager:
+    """
+    Context manager for streaming TTS responses (OpenAI-compatible).
+    """
+    
+    def __init__(self, tts_provider: DeepgramTTS):
+        self.tts_provider = tts_provider
+        self.audio_file = None
+    
+    def create(
+        self,
+        input: str,
+        model: str = "gpt-4o-mini-tts",
+        voice: str = "asteria", 
+        response_format: str = "mp3",
+        instructions: str = None
+    ):
+        """
+        Create speech with streaming capability.
+        
+        Args:
+            input (str): The text to convert to speech
+            model (str): The TTS model to use
+            voice (str): The voice to use
+            response_format (str): Audio format
+            instructions (str): Voice instructions
+            
+        Returns:
+            StreamingResponse: Streaming response object
+        """
+        self.audio_file = self.tts_provider.create_speech(
+            input=input,
+            model=model,
+            voice=voice,
+            response_format=response_format,
+            instructions=instructions
+        )
+        return StreamingResponse(self.audio_file)
+    
+    def __enter__(self):
+        return self
+    
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        pass
+
+
+class StreamingResponse:
+    """
+    Streaming response object for TTS audio (OpenAI-compatible).
+    """
+    
+    def __init__(self, audio_file: str):
+        self.audio_file = audio_file
+    
+    def __enter__(self):
+        """Enter the context manager."""
+        return self
+    
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        """Exit the context manager."""
+        pass
+    
+    def stream_to_file(self, file_path: str, chunk_size: int = 1024):
+        """
+        Stream audio content to a file.
+        
+        Args:
+            file_path (str): Destination file path
+            chunk_size (int): Size of chunks to read/write
+        """
+        import shutil
+        shutil.copy2(self.audio_file, file_path)
+    
+    def iter_bytes(self, chunk_size: int = 1024):
+        """
+        Iterate over audio bytes in chunks.
+        
+        Args:
+            chunk_size (int): Size of chunks to yield
+            
+        Yields:
+            bytes: Audio data chunks
+        """
+        with open(self.audio_file, 'rb') as f:
+            while chunk := f.read(chunk_size):
+                yield chunk
+
+
+# Example usage
+if __name__ == "__main__":
+    # Example usage demonstrating OpenAI-compatible interface
+    deepgram = DeepgramTTS()
+    
+    try:
+        # Basic usage
+        print("Testing basic speech generation...")
+        audio_file = deepgram.create_speech(
+            input="Today is a wonderful day to build something people love!",
+            model="gpt-4o-mini-tts",
+            voice="asteria",
+            instructions="Speak in a cheerful and positive tone."
+        )
+        print(f"Audio file generated: {audio_file}")
+        
+        # Streaming usage
+        print("\nTesting streaming response...")
+        with deepgram.with_streaming_response().create(
+            input="This is a streaming test with Deepgram Aura voices.",
+            voice="luna",
+            response_format="wav"
+        ) as response:
+            response.stream_to_file("deepgram_streaming_test.wav")
+            print("Streaming audio saved to deepgram_streaming_test.wav")
+            
+        # Legacy compatibility test
+        print("\nTesting legacy voice compatibility...")
+        legacy_audio = deepgram.tts("Testing legacy voice format.", voice="Asteria")
+        print(f"Legacy audio file generated: {legacy_audio}")
+        
+    except exceptions.FailedToGenerateResponseError as e:
+        print(f"Error: {e}")
+    except Exception as e:
+        print(f"Unexpected error: {e}")