dv-pipecat-ai 0.0.85.dev837__tar.gz → 0.0.85.dev841__tar.gz

{dv_pipecat_ai-0.0.85.dev837 → dv_pipecat_ai-0.0.85.dev841}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.85.dev837
+Version: 0.0.85.dev841
 Summary: An open source framework for voice (and multimodal) assistants
 License-Expression: BSD-2-Clause
 Project-URL: Source, https://github.com/pipecat-ai/pipecat
@@ -26,6 +26,7 @@ Requires-Dist: numpy<3,>=1.26.4
 Requires-Dist: Pillow<12,>=11.1.0
 Requires-Dist: protobuf~=5.29.3
 Requires-Dist: pydantic<3,>=2.10.6
+Requires-Dist: PyJWT<3,>=2.8.0
 Requires-Dist: pyloudnorm~=0.1.1
 Requires-Dist: resampy~=0.4.3
 Requires-Dist: soxr~=0.5.0

{dv_pipecat_ai-0.0.85.dev837 → dv_pipecat_ai-0.0.85.dev841}/pyproject.toml

@@ -31,6 +31,7 @@ dependencies = [
     "Pillow>=11.1.0,<12",
     "protobuf~=5.29.3",
     "pydantic>=2.10.6,<3",
+    "PyJWT>=2.8.0,<3",
     "pyloudnorm~=0.1.1",
     "resampy~=0.4.3",
     "soxr~=0.5.0",

{dv_pipecat_ai-0.0.85.dev837 → dv_pipecat_ai-0.0.85.dev841}/src/dv_pipecat_ai.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dv-pipecat-ai
-Version: 0.0.85.dev837
+Version: 0.0.85.dev841
 Summary: An open source framework for voice (and multimodal) assistants
 License-Expression: BSD-2-Clause
 Project-URL: Source, https://github.com/pipecat-ai/pipecat
@@ -26,6 +26,7 @@ Requires-Dist: numpy<3,>=1.26.4
 Requires-Dist: Pillow<12,>=11.1.0
 Requires-Dist: protobuf~=5.29.3
 Requires-Dist: pydantic<3,>=2.10.6
+Requires-Dist: PyJWT<3,>=2.8.0
 Requires-Dist: pyloudnorm~=0.1.1
 Requires-Dist: resampy~=0.4.3
 Requires-Dist: soxr~=0.5.0

{dv_pipecat_ai-0.0.85.dev837 → dv_pipecat_ai-0.0.85.dev841}/src/dv_pipecat_ai.egg-info/requires.txt

@@ -8,6 +8,7 @@ numpy<3,>=1.26.4
 Pillow<12,>=11.1.0
 protobuf~=5.29.3
 pydantic<3,>=2.10.6
+PyJWT<3,>=2.8.0
 pyloudnorm~=0.1.1
 resampy~=0.4.3
 soxr~=0.5.0

{dv_pipecat_ai-0.0.85.dev837 → dv_pipecat_ai-0.0.85.dev841}/src/pipecat/services/soniox/stt.py

@@ -49,6 +49,33 @@ END_TOKEN = "<end>"
 FINALIZED_TOKEN = "<fin>"
 
 
+class SonioxContextGeneralItem(BaseModel):
+    """Represents a key-value pair for structured general context information."""
+
+    key: str
+    value: str
+
+
+class SonioxContextTranslationTerm(BaseModel):
+    """Represents a custom translation mapping for ambiguous or domain-specific terms."""
+
+    source: str
+    target: str
+
+
+class SonioxContextObject(BaseModel):
+    """Context object for models with context_version 2, for Soniox stt-rt-v3-preview and higher.
+
+    Learn more about context in the documentation:
+    https://soniox.com/docs/stt/concepts/context
+    """
+
+    general: Optional[List[SonioxContextGeneralItem]] = None
+    text: Optional[str] = None
+    terms: Optional[List[str]] = None
+    translation_terms: Optional[List[SonioxContextTranslationTerm]] = None
+
+
 class SonioxInputParams(BaseModel):
     """Real-time transcription settings.
 
@@ -60,9 +87,9 @@ class SonioxInputParams(BaseModel):
         audio_format: Audio format to use for transcription.
         num_channels: Number of channels to use for transcription.
         language_hints: List of language hints to use for transcription.
-        context: Customization for transcription.
-        enable_non_final_tokens: Whether to enable non-final tokens. If false, only final tokens will be returned.
-        max_non_final_tokens_duration_ms: Maximum duration of non-final tokens.
+        context: Customization for transcription. String for models with context_version 1 and ContextObject for models with context_version 2.
+        enable_speaker_diarization: Whether to enable speaker diarization. Tokens are annotated with speaker IDs.
+        enable_language_identification: Whether to enable language identification. Tokens are annotated with language IDs.
         client_reference_id: Client reference ID to use for transcription.
     """
 
@@ -72,10 +99,10 @@ class SonioxInputParams(BaseModel):
     num_channels: Optional[int] = 1
 
     language_hints: Optional[List[Language]] = None
-    context: Optional[str] = None
+    context: Optional[SonioxContextObject | str] = None
 
-    enable_non_final_tokens: Optional[bool] = True
-    max_non_final_tokens_duration_ms: Optional[int] = None
+    enable_speaker_diarization: Optional[bool] = False
+    enable_language_identification: Optional[bool] = False
 
     client_reference_id: Optional[str] = None
 
@@ -173,6 +200,10 @@ class SonioxSTTService(STTService):
         # Either one or the other is required.
         enable_endpoint_detection = not self._vad_force_turn_endpoint
 
+        context = self._params.context
+        if isinstance(context, SonioxContextObject):
+            context = context.model_dump()
+
         # Send the initial configuration message.
         config = {
             "api_key": self._api_key,
@@ -182,9 +213,9 @@ class SonioxSTTService(STTService):
             "enable_endpoint_detection": enable_endpoint_detection,
             "sample_rate": self.sample_rate,
             "language_hints": _prepare_language_hints(self._params.language_hints),
-            "context": self._params.context,
-            "enable_non_final_tokens": self._params.enable_non_final_tokens,
-            "max_non_final_tokens_duration_ms": self._params.max_non_final_tokens_duration_ms,
+            "context": context,
+            "enable_speaker_diarization": self._params.enable_speaker_diarization,
+            "enable_language_identification": self._params.enable_language_identification,
             "client_reference_id": self._params.client_reference_id,
         }
 

{dv_pipecat_ai-0.0.85.dev837 → dv_pipecat_ai-0.0.85.dev841}/src/pipecat/services/vistaar/llm.py

@@ -10,9 +10,17 @@ from typing import Any, AsyncGenerator, Dict, List, Optional
 from urllib.parse import urlencode
 
 import httpx
+import jwt
 from loguru import logger
 from pydantic import BaseModel, Field
 
+try:
+    import redis.asyncio as redis
+    REDIS_AVAILABLE = True
+except ImportError:
+    REDIS_AVAILABLE = False
+    redis = None
+
 from pipecat.frames.frames import (
     CancelFrame,
     EndFrame,
@@ -55,7 +63,9 @@ class VistaarLLMService(LLMService):
         Parameters:
             source_lang: Source language code (e.g., 'mr' for Marathi, 'hi' for Hindi).
             target_lang: Target language code for responses.
-            session_id: Session ID for maintaining conversation context.
+            session_id: Session ID for maintaining conversation context (also used for JWT caching).
+            pre_query_response_phrases: List of phrases to say while waiting for response.
+            phone_number: Phone number for JWT subject claim.
             extra: Additional model-specific parameters
         """
 
@@ -63,6 +73,7 @@ class VistaarLLMService(LLMService):
         target_lang: Optional[str] = Field(default="mr")
         session_id: Optional[str] = Field(default=None)
         pre_query_response_phrases: Optional[List[str]] = Field(default_factory=list)
+        phone_number: Optional[str] = Field(default="UNKNOWN")
         extra: Optional[Dict[str, Any]] = Field(default_factory=dict)
 
     def __init__(
@@ -72,6 +83,9 @@ class VistaarLLMService(LLMService):
         params: Optional[InputParams] = None,
         timeout: float = 30.0,
         interim_timeout: float = 5.0,
+        redis_client: Optional[Any] = None,  # redis.Redis type
+        jwt_private_key: Optional[str] = None,
+        jwt_token_expiry: int = 3600,
         **kwargs,
     ):
         """Initialize Vistaar LLM service.
@@ -81,6 +95,9 @@ class VistaarLLMService(LLMService):
             params: Input parameters for model configuration and behavior.
             timeout: Request timeout in seconds. Defaults to 30.0 seconds.
             interim_timeout: Time in seconds before sending interim message. Defaults to 5.0 seconds.
+            redis_client: Optional Redis client for JWT token caching.
+            jwt_private_key: Optional RSA private key in PEM format for JWT signing.
+            jwt_token_expiry: JWT token expiry time in seconds. Defaults to 3600 (1 hour).
             **kwargs: Additional arguments passed to the parent LLMService.
         """
         super().__init__(**kwargs)
@@ -95,6 +112,16 @@ class VistaarLLMService(LLMService):
         self._extra = params.extra if isinstance(params.extra, dict) else {}
         self._timeout = timeout
         self._interim_timeout = interim_timeout
+        self._phone_number = params.phone_number
+
+        # JWT authentication setup
+        self._redis_client = redis_client
+        self._jwt_private_key = jwt_private_key
+        self._jwt_token_expiry = jwt_token_expiry
+        self._jwt_issuer = "voice-provider"
+
+        if self._jwt_private_key and not self._redis_client:
+            logger.warning("JWT private key provided but no Redis client for caching. JWT auth will regenerate tokens on each request.")
 
         # Create an async HTTP client
         self._client = httpx.AsyncClient(timeout=httpx.Timeout(self._timeout), verify=False)
@@ -112,6 +139,53 @@ class VistaarLLMService(LLMService):
             f"Vistaar LLM initialized - Base URL: {self._base_url}, Session ID: {self._session_id}, Source Lang: {self._source_lang}, Target Lang: {self._target_lang}, Timeout: {self._timeout}s"
         )
 
+    async def _get_jwt_token(self) -> Optional[str]:
+        """Generate or retrieve a cached JWT token.
+
+        Returns:
+            JWT token string or None if JWT auth is not configured.
+        """
+        if not self._jwt_private_key:
+            return None
+
+        # Try to get from Redis cache if available
+        if self._redis_client and self._session_id:
+            redis_key = f"vistaar_jwt:{self._session_id}"
+            try:
+                cached_token = await self._redis_client.get(redis_key)
+                if cached_token:
+                    logger.debug(f"Retrieved JWT token from Redis cache for session_id: {self._session_id}")
+                    return cached_token.decode('utf-8') if isinstance(cached_token, bytes) else cached_token
+            except Exception as e:
+                logger.warning(f"Redis cache retrieval failed: {e}. Generating new token.")
+
+        # Generate new token
+        current_time = int(time.time())
+        payload = {
+            "sub": self._phone_number,  # Subject identifier (phone number)
+            "iss": self._jwt_issuer,    # Issuer
+            "iat": current_time,         # Issued at timestamp
+            "exp": current_time + self._jwt_token_expiry  # Expiration timestamp
+        }
+
+        token = jwt.encode(payload, self._jwt_private_key, algorithm="RS256")
+        logger.info(f"Generated new JWT token for {self._phone_number}, expires in {self._jwt_token_expiry}s")
+
+        # Cache in Redis if available
+        if self._redis_client and self._session_id:
+            redis_key = f"vistaar_jwt:{self._session_id}"
+            try:
+                await self._redis_client.setex(
+                    redis_key,
+                    self._jwt_token_expiry,
+                    token
+                )
+                logger.debug(f"Cached JWT token in Redis for session_id: {self._session_id} with {self._jwt_token_expiry}s TTL")
+            except Exception as e:
+                logger.warning(f"Redis cache storage failed: {e}. Continuing without cache.")
+
+        return token
+
     async def _extract_messages_to_query(self, context: OpenAILLMContext) -> str:
         """Extract only the last user message from context.
 
@@ -259,9 +333,20 @@ class VistaarLLMService(LLMService):
         self._interim_in_progress = False
         self._interim_completion_event.clear()  # Reset the event for new request
 
+        # Prepare headers with JWT authentication if configured
+        headers = {}
+        try:
+            jwt_token = await self._get_jwt_token()
+            if jwt_token:
+                headers["Authorization"] = f"Bearer {jwt_token}"
+                logger.debug(f"Added JWT authentication header for session_id: {self._session_id}")
+        except Exception as e:
+            logger.error(f"Failed to generate JWT token: {e}")
+            raise
+
         try:
             # Use httpx to handle SSE streaming
-            async with self._client.stream("GET", url) as response:
+            async with self._client.stream("GET", url, headers=headers) as response:
                 self._current_response = response  # Store for potential cancellation
                 response.raise_for_status()