Merge pull request #2536 from gladiaio/PLA-38-missing-config-parameters

Gladia - add missing config parameters

CHANGELOG.md

@@ -42,6 +42,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   sending it through the transport. This makes sending DTMF generic across all
   output transports.
 
+- Added new config parameters to `GladiaSTTService`. 
+  - PreProcessingConfig > `audio_enhancer` to enhance audio quality.
+  - CustomVocabularyItem > `pronunciations` and `language` to specify special pronunciations and in which language it will be pronounced.
+
 ## Changed
 
 - `pipecat.frames.frames.KeypadEntry` is deprecated and has been moved to

src/pipecat/services/gladia/config.py

@@ -29,9 +29,11 @@ class PreProcessingConfig(BaseModel):
     """Configuration for audio pre-processing options.
 
     Parameters:
+        audio_enhancer: Apply pre-processing to the audio stream to enhance quality
         speech_threshold: Sensitivity for speech detection (0-1)
     """
 
+    audio_enhancer: Optional[bool] = None
     speech_threshold: Optional[float] = None
 
 
@@ -41,10 +43,14 @@ class CustomVocabularyItem(BaseModel):
     Parameters:
         value: The vocabulary word or phrase
         intensity: The bias intensity for this vocabulary item (0-1)
+        pronunciations: The pronunciations used in the transcription.
+        language: Specify the language in which it will be pronounced when sound comparison occurs. Default to transcription language.
     """
 
     value: str
     intensity: float
+    pronunciations: Optional[List[str]] = None
+    language: Optional[str] = None
 
 
 class CustomVocabularyConfig(BaseModel):
@@ -170,7 +176,7 @@ class GladiaInputParams(BaseModel):
     channels: Optional[int] = 1
     custom_metadata: Optional[Dict[str, Any]] = None
     endpointing: Optional[float] = None
-    maximum_duration_without_endpointing: Optional[int] = 10
+    maximum_duration_without_endpointing: Optional[int] = 5
     language: Optional[Language] = None  # Deprecated
     language_config: Optional[LanguageConfig] = None
     pre_processing: Optional[PreProcessingConfig] = None

src/pipecat/services/gladia/stt.py

@@ -18,6 +18,7 @@ from typing import Any, AsyncGenerator, Dict, Literal, Optional
 import aiohttp
 from loguru import logger
 
+from pipecat import __version__ as pipecat_version
 from pipecat.frames.frames import (
     CancelFrame,
     EndFrame,
@@ -303,8 +304,8 @@ class GladiaSTTService(STTService):
         }
 
         # Add custom_metadata if provided
-        if self._params.custom_metadata:
-            settings["custom_metadata"] = self._params.custom_metadata
+        settings["custom_metadata"] = dict(self._params.custom_metadata or {})
+        settings["custom_metadata"]["pipecat"] = pipecat_version
 
         # Add endpointing parameters if provided
         if self._params.endpointing is not None:
@@ -430,6 +431,7 @@ class GladiaSTTService(STTService):
                     response = await self._setup_gladia(settings)
                     self._session_url = response["url"]
                     self._reconnection_attempts = 0
+                    logger.info(f"Session URL : {self._session_url}")
 
                 # Connect with automatic reconnection
                 async with websocket_connect(self._session_url) as websocket: