sayna-client 0.0.1__py3-none-any.whl

sayna_client/errors.py

@@ -0,0 +1,81 @@
+"""Custom exceptions for the Sayna SDK."""
+
+from typing import Any
+
+
+class SaynaError(Exception):
+    """Base error class for all Sayna SDK errors."""
+
+    def __init__(self, message: str) -> None:
+        """Initialize the error.
+
+        Args:
+            message: Error description
+        """
+        super().__init__(message)
+        self.message = message
+
+
+class SaynaNotConnectedError(SaynaError):
+    """Error raised when attempting to use the client before it's connected."""
+
+    def __init__(self, message: str = "Not connected to Sayna WebSocket") -> None:
+        """Initialize the error.
+
+        Args:
+            message: Error description
+        """
+        super().__init__(message)
+
+
+class SaynaNotReadyError(SaynaError):
+    """Error raised when attempting operations before the client is ready."""
+
+    def __init__(
+        self,
+        message: str = "Sayna voice providers are not ready. Wait for the connection to be established.",
+    ) -> None:
+        """Initialize the error.
+
+        Args:
+            message: Error description
+        """
+        super().__init__(message)
+
+
+class SaynaConnectionError(SaynaError):
+    """Error raised when WebSocket connection fails."""
+
+    def __init__(self, message: str, cause: Any = None) -> None:
+        """Initialize the error.
+
+        Args:
+            message: Error description
+            cause: The underlying exception that caused this error
+        """
+        super().__init__(message)
+        self.cause = cause
+
+
+class SaynaValidationError(SaynaError):
+    """Error raised when invalid parameters are provided."""
+
+    def __init__(self, message: str) -> None:
+        """Initialize the error.
+
+        Args:
+            message: Error description
+        """
+        super().__init__(message)
+
+
+class SaynaServerError(SaynaError):
+    """Error raised when the server returns an error."""
+
+    def __init__(self, message: str) -> None:
+        """Initialize the error.
+
+        Args:
+            message: Error description
+        """
+        super().__init__(message)

sayna_client/http_client.py

@@ -0,0 +1,169 @@
+"""Generic HTTP client for Sayna REST API calls."""
+
+import json
+from typing import Any, Optional
+
+import aiohttp
+
+from sayna_client.errors import SaynaServerError, SaynaValidationError
+
+
+class SaynaHttpClient:
+    """Generic HTTP client for making REST API requests to Sayna server."""
+
+    def __init__(self, base_url: str, api_key: Optional[str] = None) -> None:
+        """Initialize the HTTP client.
+
+        Args:
+            base_url: Base URL for the Sayna API (e.g., 'https://api.sayna.com')
+            api_key: Optional API key for authentication
+        """
+        self.base_url = base_url.rstrip("/")
+        self.api_key = api_key
+        self._session: Optional[aiohttp.ClientSession] = None
+
+    async def _ensure_session(self) -> aiohttp.ClientSession:
+        """Ensure an HTTP session exists."""
+        if self._session is None or self._session.closed:
+            headers = {"Content-Type": "application/json"}
+            if self.api_key:
+                headers["Authorization"] = f"Bearer {self.api_key}"
+
+            self._session = aiohttp.ClientSession(headers=headers)
+
+        return self._session
+
+    async def close(self) -> None:
+        """Close the HTTP session."""
+        if self._session and not self._session.closed:
+            await self._session.close()
+            self._session = None
+
+    async def get(
+        self,
+        endpoint: str,
+        params: Optional[dict[str, Any]] = None,
+    ) -> dict[str, Any]:
+        """Make a GET request.
+
+        Args:
+            endpoint: API endpoint path (e.g., '/voices')
+            params: Optional query parameters
+
+        Returns:
+            JSON response as a dictionary
+
+        Raises:
+            SaynaServerError: If the server returns an error
+            SaynaValidationError: If the request is invalid
+        """
+        session = await self._ensure_session()
+        url = f"{self.base_url}{endpoint}"
+
+        async with session.get(url, params=params) as response:
+            return await self._handle_response(response)
+
+    async def post(
+        self,
+        endpoint: str,
+        data: Optional[dict[str, Any]] = None,
+        json_data: Optional[dict[str, Any]] = None,
+    ) -> dict[str, Any]:
+        """Make a POST request.
+
+        Args:
+            endpoint: API endpoint path (e.g., '/speak')
+            data: Optional form data
+            json_data: Optional JSON payload
+
+        Returns:
+            JSON response as a dictionary
+
+        Raises:
+            SaynaServerError: If the server returns an error
+            SaynaValidationError: If the request is invalid
+        """
+        session = await self._ensure_session()
+        url = f"{self.base_url}{endpoint}"
+
+        async with session.post(url, data=data, json=json_data) as response:
+            return await self._handle_response(response)
+
+    async def post_binary(
+        self,
+        endpoint: str,
+        json_data: Optional[dict[str, Any]] = None,
+    ) -> tuple[bytes, dict[str, str]]:
+        """Make a POST request expecting binary response.
+
+        Args:
+            endpoint: API endpoint path (e.g., '/speak')
+            json_data: Optional JSON payload
+
+        Returns:
+            Tuple of (binary_data, response_headers)
+
+        Raises:
+            SaynaServerError: If the server returns an error
+            SaynaValidationError: If the request is invalid
+        """
+        session = await self._ensure_session()
+        url = f"{self.base_url}{endpoint}"
+
+        async with session.post(url, json=json_data) as response:
+            if response.status >= 400:
+                # Try to parse error message
+                try:
+                    error_data = await response.json()
+                    error_msg = error_data.get("error", f"HTTP {response.status}")
+                except Exception:
+                    error_msg = f"HTTP {response.status}: {response.reason}"
+
+                if response.status >= 500:
+                    raise SaynaServerError(error_msg)
+                else:
+                    raise SaynaValidationError(error_msg)
+
+            binary_data = await response.read()
+            headers = dict(response.headers)
+            return binary_data, headers
+
+    async def _handle_response(self, response: aiohttp.ClientResponse) -> dict[str, Any]:
+        """Handle HTTP response and raise appropriate errors.
+
+        Args:
+            response: aiohttp response object
+
+        Returns:
+            Parsed JSON response
+
+        Raises:
+            SaynaServerError: If the server returns a 5xx error
+            SaynaValidationError: If the request is invalid (4xx error)
+        """
+        if response.status >= 400:
+            try:
+                error_data = await response.json()
+                error_msg = error_data.get("error", f"HTTP {response.status}")
+            except Exception:
+                error_msg = f"HTTP {response.status}: {response.reason}"
+
+            if response.status >= 500:
+                raise SaynaServerError(error_msg)
+            else:
+                raise SaynaValidationError(error_msg)
+
+        try:
+            json_response: dict[str, Any] = await response.json()
+            return json_response
+        except json.JSONDecodeError as e:
+            raise SaynaServerError(f"Failed to decode JSON response: {e}") from e
+
+    async def __aenter__(self) -> "SaynaHttpClient":
+        """Async context manager entry."""
+        await self._ensure_session()
+        return self
+
+    async def __aexit__(self, exc_type: Any, exc_val: Any, exc_tb: Any) -> None:
+        """Async context manager exit."""
+        await self.close()

sayna_client/types.py

@@ -0,0 +1,267 @@
+"""Type definitions for the Sayna SDK."""
+
+from typing import Any, Literal, Optional, Union
+
+from pydantic import BaseModel, Field, RootModel
+
+
+class Pronunciation(BaseModel):
+    """Word pronunciation override for text-to-speech."""
+
+    word: str = Field(..., description="The word to be pronounced differently")
+    pronunciation: str = Field(..., description="Phonetic pronunciation or alternative spelling")
+
+
+class STTConfig(BaseModel):
+    """Speech-to-Text (STT) configuration options."""
+
+    provider: str = Field(..., description="The STT provider to use (e.g., 'deepgram', 'google')")
+    language: str = Field(..., description="Language code for speech recognition (e.g., 'en-US')")
+    sample_rate: int = Field(..., description="Audio sample rate in Hz (e.g., 16000, 44100)")
+    channels: int = Field(..., description="Number of audio channels (1 for mono, 2 for stereo)")
+    punctuation: bool = Field(..., description="Whether to include punctuation in transcriptions")
+    encoding: str = Field(..., description="Audio encoding format (e.g., 'linear16', 'opus')")
+    model: str = Field(..., description="STT model identifier to use")
+
+
+class TTSConfig(BaseModel):
+    """Text-to-Speech (TTS) configuration options."""
+
+    provider: str = Field(..., description="The TTS provider to use (e.g., 'elevenlabs', 'google')")
+    voice_id: str = Field(..., description="Voice identifier for the selected provider")
+    speaking_rate: float = Field(
+        ..., description="Speech rate multiplier (e.g., 1.0 for normal, 1.5 for faster)"
+    )
+    audio_format: str = Field(..., description="Audio format for TTS output (e.g., 'mp3', 'pcm')")
+    sample_rate: int = Field(..., description="Audio sample rate in Hz (e.g., 16000, 44100)")
+    connection_timeout: int = Field(..., description="Connection timeout in milliseconds")
+    request_timeout: int = Field(..., description="Request timeout in milliseconds")
+    model: str = Field(..., description="TTS model identifier to use")
+    pronunciations: list[Pronunciation] = Field(
+        default_factory=list, description="Custom pronunciation overrides"
+    )
+
+
+class LiveKitConfig(BaseModel):
+    """LiveKit room configuration for real-time communication."""
+
+    room_name: str = Field(..., description="LiveKit room name to join")
+    enable_recording: Optional[bool] = Field(
+        default=False, description="Whether to enable session recording"
+    )
+    recording_file_key: Optional[str] = Field(
+        default=None,
+        description="Storage key for the recording file (required when enable_recording is true)",
+    )
+    sayna_participant_identity: Optional[str] = Field(
+        default="sayna-ai", description="Identity assigned to the agent participant"
+    )
+    sayna_participant_name: Optional[str] = Field(
+        default="Sayna AI", description="Display name for the agent participant"
+    )
+    listen_participants: Optional[list[str]] = Field(
+        default_factory=list,
+        description="List of participant identities to monitor (empty = all participants)",
+    )
+
+
+# ============================================================================
+# Outgoing Messages (Client -> Server)
+# ============================================================================
+
+
+class ConfigMessage(BaseModel):
+    """Configuration message sent to initialize the Sayna WebSocket connection."""
+
+    type: Literal["config"] = "config"
+    audio: Optional[bool] = Field(default=True, description="Whether audio streaming is enabled")
+    stt_config: Optional[STTConfig] = Field(
+        default=None, description="Speech-to-text configuration (required when audio=true)"
+    )
+    tts_config: Optional[TTSConfig] = Field(
+        default=None, description="Text-to-speech configuration (required when audio=true)"
+    )
+    livekit: Optional[LiveKitConfig] = Field(
+        default=None, description="Optional LiveKit room configuration"
+    )
+
+
+class SpeakMessage(BaseModel):
+    """Message to request text-to-speech synthesis."""
+
+    type: Literal["speak"] = "speak"
+    text: str = Field(..., description="Text to synthesize")
+    flush: Optional[bool] = Field(
+        default=None, description="Whether to flush the TTS queue before speaking"
+    )
+    allow_interruption: Optional[bool] = Field(
+        default=None, description="Whether this speech can be interrupted"
+    )
+
+
+class ClearMessage(BaseModel):
+    """Message to clear the TTS queue."""
+
+    type: Literal["clear"] = "clear"
+
+
+class SendMessageMessage(BaseModel):
+    """Message to send data to the Sayna session."""
+
+    type: Literal["send_message"] = "send_message"
+    message: str = Field(..., description="Message content")
+    role: str = Field(..., description="Message role (e.g., 'user', 'assistant')")
+    topic: Optional[str] = Field(default=None, description="Optional topic identifier")
+    debug: Optional[dict[str, Any]] = Field(default=None, description="Optional debug metadata")
+
+
+# ============================================================================
+# Incoming Messages (Server -> Client)
+# ============================================================================
+
+
+class ReadyMessage(BaseModel):
+    """Message received when the Sayna connection is ready."""
+
+    type: Literal["ready"] = "ready"
+    livekit_room_name: Optional[str] = Field(
+        default=None, description="LiveKit room name (present only when LiveKit is enabled)"
+    )
+    livekit_url: str = Field(..., description="LiveKit WebSocket URL configured on the server")
+    sayna_participant_identity: Optional[str] = Field(
+        default=None,
+        description="Identity assigned to the agent participant when LiveKit is enabled",
+    )
+    sayna_participant_name: Optional[str] = Field(
+        default=None,
+        description="Display name assigned to the agent participant when LiveKit is enabled",
+    )
+
+
+class STTResultMessage(BaseModel):
+    """Speech-to-text transcription result."""
+
+    type: Literal["stt_result"] = "stt_result"
+    transcript: str = Field(..., description="Transcribed text")
+    is_final: bool = Field(..., description="Whether this is a final transcription")
+    is_speech_final: bool = Field(..., description="Whether speech has concluded")
+    confidence: float = Field(..., description="Transcription confidence score (0-1)")
+
+
+class ErrorMessage(BaseModel):
+    """Error message from the Sayna server."""
+
+    type: Literal["error"] = "error"
+    message: str = Field(..., description="Error description")
+
+
+class SaynaMessage(BaseModel):
+    """Message data from a Sayna session participant."""
+
+    message: Optional[str] = Field(default=None, description="Message content")
+    data: Optional[str] = Field(default=None, description="Additional data payload")
+    identity: str = Field(..., description="Participant identity")
+    topic: str = Field(..., description="Message topic")
+    room: str = Field(..., description="Room identifier")
+    timestamp: int = Field(..., description="Unix timestamp in milliseconds")
+
+
+class MessageMessage(BaseModel):
+    """Wrapper for participant messages."""
+
+    type: Literal["message"] = "message"
+    message: SaynaMessage = Field(..., description="The message data")
+
+
+class Participant(BaseModel):
+    """Information about a session participant."""
+
+    identity: str = Field(..., description="Unique participant identity")
+    name: Optional[str] = Field(default=None, description="Optional display name")
+    room: str = Field(..., description="Room identifier")
+    timestamp: int = Field(..., description="Unix timestamp in milliseconds")
+
+
+class ParticipantDisconnectedMessage(BaseModel):
+    """Message received when a participant disconnects."""
+
+    type: Literal["participant_disconnected"] = "participant_disconnected"
+    participant: Participant = Field(..., description="The disconnected participant")
+
+
+class TTSPlaybackCompleteMessage(BaseModel):
+    """Message received when the TTS playback is complete."""
+
+    type: Literal["tts_playback_complete"] = "tts_playback_complete"
+    timestamp: int = Field(..., description="Unix timestamp in milliseconds")
+
+
+# ============================================================================
+# REST API Types
+# ============================================================================
+
+
+class HealthResponse(BaseModel):
+    """Response from GET / health endpoint."""
+
+    status: str = Field(..., description="Health status (should be 'OK')")
+
+
+class VoiceDescriptor(BaseModel):
+    """Voice descriptor from a TTS provider."""
+
+    id: str = Field(..., description="Provider-specific identifier for the voice profile")
+    sample: str = Field(default="", description="URL to a preview audio sample")
+    name: str = Field(..., description="Human-readable name supplied by the provider")
+    accent: str = Field(default="Unknown", description="Detected accent associated with the voice")
+    gender: str = Field(
+        default="Unknown", description="Inferred gender label from provider metadata"
+    )
+    language: str = Field(default="Unknown", description="Primary language for synthesis")
+
+
+class VoicesResponse(RootModel[dict[str, list[VoiceDescriptor]]]):
+    """Response from GET /voices endpoint.
+
+    Dictionary where keys are provider names and values are lists of voice descriptors.
+    """
+
+    root: dict[str, list[VoiceDescriptor]]
+
+
+class LiveKitTokenRequest(BaseModel):
+    """Request body for POST /livekit/token."""
+
+    room_name: str = Field(..., description="LiveKit room to join or create")
+    participant_name: str = Field(..., description="Display name assigned to the participant")
+    participant_identity: str = Field(..., description="Unique identifier for the participant")
+
+
+class LiveKitTokenResponse(BaseModel):
+    """Response from POST /livekit/token."""
+
+    token: str = Field(..., description="JWT granting LiveKit permissions")
+    room_name: str = Field(..., description="Echo of the requested room")
+    participant_identity: str = Field(..., description="Echo of the requested identity")
+    livekit_url: str = Field(..., description="WebSocket endpoint for the LiveKit server")
+
+
+class SpeakRequest(BaseModel):
+    """Request body for POST /speak."""
+
+    text: str = Field(..., description="Text to convert to speech")
+    tts_config: TTSConfig = Field(..., description="Provider configuration without API credentials")
+
+
+# ============================================================================
+# Union Types
+# ============================================================================
+
+OutgoingMessage = Union[
+    ReadyMessage,
+    STTResultMessage,
+    ErrorMessage,
+    MessageMessage,
+    ParticipantDisconnectedMessage,
+    TTSPlaybackCompleteMessage,
+]