sarvamai 0.1.19a2__tar.gz → 0.1.19a4__tar.gz

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sarvamai
-Version: 0.1.19a2
+Version: 0.1.19a4
 Summary: 
 Requires-Python: >=3.8,<4.0
 Classifier: Intended Audience :: Developers

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/pyproject.toml

@@ -3,7 +3,7 @@ name = "sarvamai"
 
 [tool.poetry]
 name = "sarvamai"
-version = "0.1.19a2"
+version = "0.1.19a4"
 description = ""
 readme = "README.md"
 authors = []

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/src/sarvamai/core/client_wrapper.py

@@ -23,10 +23,10 @@ class BaseClientWrapper:
 
     def get_headers(self) -> typing.Dict[str, str]:
         headers: typing.Dict[str, str] = {
-            "User-Agent": "sarvamai/0.1.19a2",
+            "User-Agent": "sarvamai/0.1.19a4",
             "X-Fern-Language": "Python",
             "X-Fern-SDK-Name": "sarvamai",
-            "X-Fern-SDK-Version": "0.1.19a2",
+            "X-Fern-SDK-Version": "0.1.19a4",
             **(self.get_custom_headers() or {}),
         }
         headers["api-subscription-key"] = self.api_subscription_key

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/src/sarvamai/requests/config_message.py

@@ -13,5 +13,5 @@ class ConfigMessageParams(typing_extensions.TypedDict):
 
     prompt: typing_extensions.NotRequired[str]
     """
-    Prompt for ASR model to improve transcription accuracy
+    Prompt for ASR model to improve transcription accuracy.
     """

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/src/sarvamai/requests/stt_flush_signal.py

@@ -14,8 +14,3 @@ class SttFlushSignalParams(typing_extensions.TypedDict):
     """
     Type identifier for flush signal
     """
-
-    reason: typing_extensions.NotRequired[str]
-    """
-    Optional reason for flushing (e.g., "end_of_segment", "manual_flush")
-    """

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/src/sarvamai/speech_to_text_translate_streaming/socket_client.py

@@ -10,9 +10,14 @@ from ..core.pydantic_utilities import parse_obj_as
 from ..types.audio_data import AudioData
 from ..types.audio_message import AudioMessage
 from ..types.config_message import ConfigMessage
-from ..types.speech_to_text_translate_streaming_response import SpeechToTextTranslateStreamingResponse
+from ..types.speech_to_text_translate_streaming_response import (
+    SpeechToTextTranslateStreamingResponse,
+)
+from ..types.stt_flush_signal import SttFlushSignal
 
-SpeechToTextTranslateStreamingSocketClientResponse = typing.Union[SpeechToTextTranslateStreamingResponse]
+SpeechToTextTranslateStreamingSocketClientResponse = typing.Union[
+    SpeechToTextTranslateStreamingResponse
+]
 
 
 class AsyncSpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
@@ -38,7 +43,11 @@ class AsyncSpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         self._emit(EventType.OPEN, None)
         try:
             async for raw_message in self._websocket:
-                raw_message = json.loads(raw_message) if isinstance(raw_message, str) else raw_message
+                raw_message = (
+                    json.loads(raw_message)
+                    if isinstance(raw_message, str)
+                    else raw_message
+                )
                 parsed = parse_obj_as(SpeechToTextTranslateStreamingSocketClientResponse, raw_message)  # type: ignore
                 self._emit(EventType.MESSAGE, parsed)
         except websockets.WebSocketException as exc:
@@ -46,7 +55,9 @@ class AsyncSpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         finally:
             self._emit(EventType.CLOSE, None)
 
-    async def translate(self, audio: str, encoding: str = "audio/wav", sample_rate: int = 16000):
+    async def translate(
+        self, audio: str, encoding: str = "audio/wav", sample_rate: int = 16000
+    ):
         """
         Sends audio translation request to the server.
         :param audio: Base64 encoded audio data
@@ -54,7 +65,9 @@ class AsyncSpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         :param sample_rate (Optional): Audio sample rate in Hz (default is 16000)
         """
         return await self._send_speech_to_text_translate_streaming_audio_message(
-            message=AudioMessage(audio=AudioData(data=audio, encoding=encoding, sample_rate=sample_rate))
+            message=AudioMessage(
+                audio=AudioData(data=audio, encoding=encoding, sample_rate=sample_rate)
+            )
         )
 
     async def set_prompt(self, prompt: str) -> None:
@@ -66,15 +79,31 @@ class AsyncSpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         message = ConfigMessage(prompt=prompt)
         await self._send_config_message(message)
 
+    async def flush(self, reason: typing.Optional[str] = None) -> None:
+        """
+        Signal to flush the audio buffer and force finalize partial
+        transcriptions and translations. Use this to force processing of any
+        remaining audio that hasn't been transcribed and translated yet.
+
+        :param reason: Optional reason for flushing (e.g., "end_of_segment",
+                      "manual_flush")
+        """
+        message = SttFlushSignal(reason=reason)
+        await self._send_model(message)
+
     async def recv(self) -> SpeechToTextTranslateStreamingSocketClientResponse:
         """
         Receive a message from the websocket connection.
         """
         data = await self._websocket.recv()
         data = json.loads(data) if isinstance(data, str) else data
-        return parse_obj_as(SpeechToTextTranslateStreamingSocketClientResponse, data)  # type: ignore
+        return parse_obj_as(
+            SpeechToTextTranslateStreamingSocketClientResponse, data
+        )  # type: ignore
 
-    async def _send_speech_to_text_translate_streaming_audio_message(self, message: AudioMessage) -> None:
+    async def _send_speech_to_text_translate_streaming_audio_message(
+        self, message: AudioMessage
+    ) -> None:
         """
         Send a message to the websocket connection.
         The message will be sent as a AudioMessage.
@@ -126,7 +155,11 @@ class SpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         self._emit(EventType.OPEN, None)
         try:
             for raw_message in self._websocket:
-                raw_message = json.loads(raw_message) if isinstance(raw_message, str) else raw_message
+                raw_message = (
+                    json.loads(raw_message)
+                    if isinstance(raw_message, str)
+                    else raw_message
+                )
                 parsed = parse_obj_as(SpeechToTextTranslateStreamingSocketClientResponse, raw_message)  # type: ignore
                 self._emit(EventType.MESSAGE, parsed)
         except websockets.WebSocketException as exc:
@@ -134,7 +167,9 @@ class SpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         finally:
             self._emit(EventType.CLOSE, None)
 
-    def translate(self, audio: str, encoding: str = "audio/wav", sample_rate: int = 16000):
+    def translate(
+        self, audio: str, encoding: str = "audio/wav", sample_rate: int = 16000
+    ):
         """
         Sends audio translation request to the server.
         :param audio: Base64 encoded audio data
@@ -142,7 +177,9 @@ class SpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         :param sample_rate: Audio sample rate in Hz (default is 16000)
         """
         return self._send_speech_to_text_translate_streaming_audio_message(
-            message=AudioMessage(audio=AudioData(data=audio, encoding=encoding, sample_rate=sample_rate))
+            message=AudioMessage(
+                audio=AudioData(data=audio, encoding=encoding, sample_rate=sample_rate)
+            )
         )
 
     def set_prompt(self, prompt: str) -> None:
@@ -154,6 +191,18 @@ class SpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         message = ConfigMessage(prompt=prompt)
         self._send_config_message(message)
 
+    def flush(self, reason: typing.Optional[str] = None) -> None:
+        """
+        Signal to flush the audio buffer and force finalize partial transcriptions and translations.
+        Use this to force processing of any remaining audio that hasn't been
+        transcribed and translated yet.
+
+        :param reason: Optional reason for flushing (e.g., "end_of_segment",
+                      "manual_flush")
+        """
+        message = SttFlushSignal(reason=reason)
+        self._send_model(message)
+
     def recv(self) -> SpeechToTextTranslateStreamingSocketClientResponse:
         """
         Receive a message from the websocket connection.
@@ -169,7 +218,9 @@ class SpeechToTextTranslateStreamingSocketClient(EventEmitterMixin):
         """
         self._send_model(message)
 
-    def _send_speech_to_text_translate_streaming_audio_message(self, message: AudioMessage) -> None:
+    def _send_speech_to_text_translate_streaming_audio_message(
+        self, message: AudioMessage
+    ) -> None:
         """
         Send a message to the websocket connection.
         The message will be sent as a AudioMessage.

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/src/sarvamai/types/config_message.py

@@ -14,7 +14,7 @@ class ConfigMessage(UniversalBaseModel):
 
     prompt: typing.Optional[str] = pydantic.Field(default=None)
     """
-    Prompt for ASR model to improve transcription accuracy
+    Prompt for ASR model to improve transcription accuracy.
     """
 
     if IS_PYDANTIC_V2:

{sarvamai-0.1.19a2 → sarvamai-0.1.19a4}/src/sarvamai/types/stt_flush_signal.py

@@ -16,11 +16,6 @@ class SttFlushSignal(UniversalBaseModel):
     Type identifier for flush signal
     """
 
-    reason: typing.Optional[str] = pydantic.Field(default=None)
-    """
-    Optional reason for flushing (e.g., "end_of_segment", "manual_flush")
-    """
-
     if IS_PYDANTIC_V2:
         model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow", frozen=True)  # type: ignore # Pydantic v2
     else: