sarvamai 0.1.5a3__py3-none-any.whl → 0.1.5a5__py3-none-any.whl

sarvamai/__init__.py

@@ -3,6 +3,9 @@
 # isort: skip_file
 
 from .types import (
+    AudioData,
+    AudioDataEncoding,
+    AudioMessage,
     ChatCompletionRequestAssistantMessage,
     ChatCompletionRequestMessage,
     ChatCompletionRequestMessage_Assistant,
@@ -17,9 +20,12 @@ from .types import (
     DiarizedEntry,
     DiarizedTranscript,
     ErrorCode,
+    ErrorData,
     ErrorDetails,
     ErrorMessage,
+    EventsData,
     FinishReason,
+    Format,
     LanguageIdentificationResponse,
     NumeralsFormat,
     ReasoningEffort,
@@ -29,6 +35,9 @@ from .types import (
     SpeechToTextLanguage,
     SpeechToTextModel,
     SpeechToTextResponse,
+    SpeechToTextStreamingResponse,
+    SpeechToTextStreamingResponseData,
+    SpeechToTextStreamingResponseType,
     SpeechToTextTranslateLanguage,
     SpeechToTextTranslateModel,
     SpeechToTextTranslateResponse,
@@ -39,14 +48,16 @@ from .types import (
     TextToSpeechResponse,
     TextToSpeechSpeaker,
     TimestampsModel,
+    TranscriptionData,
+    TranscriptionMetrics,
     TranslateMode,
     TranslateModel,
-    TranslatePostprocessing,
     TranslateSourceLanguage,
     TranslateSpeakerGender,
     TranslateTargetLanguage,
     TranslationResponse,
     TranslatiterateTargetLanguage,
+    TransliterateMode,
     TransliterateSourceLanguage,
     TransliterationResponse,
 )
@@ -58,10 +69,12 @@ from .errors import (
     TooManyRequestsError,
     UnprocessableEntityError,
 )
-from . import chat, speech_to_text, text, text_to_speech
+from . import chat, speech_to_text, speech_to_text_streaming, text, text_to_speech
 from .client import AsyncSarvamAI, SarvamAI
 from .environment import SarvamAIEnvironment
 from .requests import (
+    AudioDataParams,
+    AudioMessageParams,
     ChatCompletionRequestAssistantMessageParams,
     ChatCompletionRequestMessageParams,
     ChatCompletionRequestMessage_AssistantParams,
@@ -75,21 +88,33 @@ from .requests import (
     CreateChatCompletionResponseParams,
     DiarizedEntryParams,
     DiarizedTranscriptParams,
+    ErrorDataParams,
     ErrorDetailsParams,
     ErrorMessageParams,
+    EventsDataParams,
     LanguageIdentificationResponseParams,
     SpeechToTextResponseParams,
+    SpeechToTextStreamingResponseDataParams,
+    SpeechToTextStreamingResponseParams,
     SpeechToTextTranslateResponseParams,
     StopConfigurationParams,
     TextToSpeechResponseParams,
     TimestampsModelParams,
+    TranscriptionDataParams,
+    TranscriptionMetricsParams,
     TranslationResponseParams,
     TransliterationResponseParams,
 )
+from .speech_to_text_streaming import SpeechToTextStreamingLanguageCode, SpeechToTextStreamingModel
 from .version import __version__
 
 __all__ = [
     "AsyncSarvamAI",
+    "AudioData",
+    "AudioDataEncoding",
+    "AudioDataParams",
+    "AudioMessage",
+    "AudioMessageParams",
     "BadRequestError",
     "ChatCompletionRequestAssistantMessage",
     "ChatCompletionRequestAssistantMessageParams",
@@ -118,12 +143,17 @@ __all__ = [
     "DiarizedTranscript",
     "DiarizedTranscriptParams",
     "ErrorCode",
+    "ErrorData",
+    "ErrorDataParams",
     "ErrorDetails",
     "ErrorDetailsParams",
     "ErrorMessage",
     "ErrorMessageParams",
+    "EventsData",
+    "EventsDataParams",
     "FinishReason",
     "ForbiddenError",
+    "Format",
     "InternalServerError",
     "LanguageIdentificationResponse",
     "LanguageIdentificationResponseParams",
@@ -139,6 +169,13 @@ __all__ = [
     "SpeechToTextModel",
     "SpeechToTextResponse",
     "SpeechToTextResponseParams",
+    "SpeechToTextStreamingLanguageCode",
+    "SpeechToTextStreamingModel",
+    "SpeechToTextStreamingResponse",
+    "SpeechToTextStreamingResponseData",
+    "SpeechToTextStreamingResponseDataParams",
+    "SpeechToTextStreamingResponseParams",
+    "SpeechToTextStreamingResponseType",
     "SpeechToTextTranslateLanguage",
     "SpeechToTextTranslateModel",
     "SpeechToTextTranslateResponse",
@@ -154,15 +191,19 @@ __all__ = [
     "TimestampsModel",
     "TimestampsModelParams",
     "TooManyRequestsError",
+    "TranscriptionData",
+    "TranscriptionDataParams",
+    "TranscriptionMetrics",
+    "TranscriptionMetricsParams",
     "TranslateMode",
     "TranslateModel",
-    "TranslatePostprocessing",
     "TranslateSourceLanguage",
     "TranslateSpeakerGender",
     "TranslateTargetLanguage",
     "TranslationResponse",
     "TranslationResponseParams",
     "TranslatiterateTargetLanguage",
+    "TransliterateMode",
     "TransliterateSourceLanguage",
     "TransliterationResponse",
     "TransliterationResponseParams",
@@ -170,6 +211,7 @@ __all__ = [
     "__version__",
     "chat",
     "speech_to_text",
+    "speech_to_text_streaming",
     "text",
     "text_to_speech",
 ]

sarvamai/chat/client.py

@@ -83,7 +83,6 @@ class ChatClient:
         seed : typing.Optional[int]
             This feature is in Beta.
             If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same `seed` and parameters should return the same result.
-            Determinism is not guaranteed, and you should refer to the `system_fingerprint` response parameter to monitor changes in the backend.
 
         frequency_penalty : typing.Optional[float]
             Number between -2.0 and 2.0. Positive values penalize new tokens based on
@@ -204,7 +203,6 @@ class AsyncChatClient:
         seed : typing.Optional[int]
             This feature is in Beta.
             If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same `seed` and parameters should return the same result.
-            Determinism is not guaranteed, and you should refer to the `system_fingerprint` response parameter to monitor changes in the backend.
 
         frequency_penalty : typing.Optional[float]
             Number between -2.0 and 2.0. Positive values penalize new tokens based on

sarvamai/chat/raw_client.py

@@ -81,7 +81,6 @@ class RawChatClient:
         seed : typing.Optional[int]
             This feature is in Beta.
             If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same `seed` and parameters should return the same result.
-            Determinism is not guaranteed, and you should refer to the `system_fingerprint` response parameter to monitor changes in the backend.
 
         frequency_penalty : typing.Optional[float]
             Number between -2.0 and 2.0. Positive values penalize new tokens based on
@@ -262,7 +261,6 @@ class AsyncRawChatClient:
         seed : typing.Optional[int]
             This feature is in Beta.
             If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same `seed` and parameters should return the same result.
-            Determinism is not guaranteed, and you should refer to the `system_fingerprint` response parameter to monitor changes in the backend.
 
         frequency_penalty : typing.Optional[float]
             Number between -2.0 and 2.0. Positive values penalize new tokens based on

sarvamai/client.py

@@ -9,6 +9,7 @@ from .core.api_error import ApiError
 from .core.client_wrapper import AsyncClientWrapper, SyncClientWrapper
 from .environment import SarvamAIEnvironment
 from .speech_to_text.client import AsyncSpeechToTextClient, SpeechToTextClient
+from .speech_to_text_streaming.client import AsyncSpeechToTextStreamingClient, SpeechToTextStreamingClient
 from .text.client import AsyncTextClient, TextClient
 from .text_to_speech.client import AsyncTextToSpeechClient, TextToSpeechClient
 
@@ -77,6 +78,7 @@ class SarvamAI:
         self.speech_to_text = SpeechToTextClient(client_wrapper=self._client_wrapper)
         self.text_to_speech = TextToSpeechClient(client_wrapper=self._client_wrapper)
         self.chat = ChatClient(client_wrapper=self._client_wrapper)
+        self.speech_to_text_streaming = SpeechToTextStreamingClient(client_wrapper=self._client_wrapper)
 
 
 class AsyncSarvamAI:
@@ -143,3 +145,4 @@ class AsyncSarvamAI:
         self.speech_to_text = AsyncSpeechToTextClient(client_wrapper=self._client_wrapper)
         self.text_to_speech = AsyncTextToSpeechClient(client_wrapper=self._client_wrapper)
         self.chat = AsyncChatClient(client_wrapper=self._client_wrapper)
+        self.speech_to_text_streaming = AsyncSpeechToTextStreamingClient(client_wrapper=self._client_wrapper)

sarvamai/core/client_wrapper.py

@@ -17,10 +17,10 @@ class BaseClientWrapper:
 
     def get_headers(self) -> typing.Dict[str, str]:
         headers: typing.Dict[str, str] = {
-            "User-Agent": "sarvamai/0.1.5a3",
+            "User-Agent": "sarvamai/0.1.5a5",
             "X-Fern-Language": "Python",
             "X-Fern-SDK-Name": "sarvamai",
-            "X-Fern-SDK-Version": "0.1.5a3",
+            "X-Fern-SDK-Version": "0.1.5a5",
         }
         headers["api-subscription-key"] = self.api_subscription_key
         return headers

sarvamai/requests/__init__.py

@@ -2,6 +2,8 @@
 
 # isort: skip_file
 
+from .audio_data import AudioDataParams
+from .audio_message import AudioMessageParams
 from .chat_completion_request_assistant_message import ChatCompletionRequestAssistantMessageParams
 from .chat_completion_request_message import (
     ChatCompletionRequestMessageParams,
@@ -17,18 +19,26 @@ from .completion_usage import CompletionUsageParams
 from .create_chat_completion_response import CreateChatCompletionResponseParams
 from .diarized_entry import DiarizedEntryParams
 from .diarized_transcript import DiarizedTranscriptParams
+from .error_data import ErrorDataParams
 from .error_details import ErrorDetailsParams
 from .error_message import ErrorMessageParams
+from .events_data import EventsDataParams
 from .language_identification_response import LanguageIdentificationResponseParams
 from .speech_to_text_response import SpeechToTextResponseParams
+from .speech_to_text_streaming_response import SpeechToTextStreamingResponseParams
+from .speech_to_text_streaming_response_data import SpeechToTextStreamingResponseDataParams
 from .speech_to_text_translate_response import SpeechToTextTranslateResponseParams
 from .stop_configuration import StopConfigurationParams
 from .text_to_speech_response import TextToSpeechResponseParams
 from .timestamps_model import TimestampsModelParams
+from .transcription_data import TranscriptionDataParams
+from .transcription_metrics import TranscriptionMetricsParams
 from .translation_response import TranslationResponseParams
 from .transliteration_response import TransliterationResponseParams
 
 __all__ = [
+    "AudioDataParams",
+    "AudioMessageParams",
     "ChatCompletionRequestAssistantMessageParams",
     "ChatCompletionRequestMessageParams",
     "ChatCompletionRequestMessage_AssistantParams",
@@ -42,14 +52,20 @@ __all__ = [
     "CreateChatCompletionResponseParams",
     "DiarizedEntryParams",
     "DiarizedTranscriptParams",
+    "ErrorDataParams",
     "ErrorDetailsParams",
     "ErrorMessageParams",
+    "EventsDataParams",
     "LanguageIdentificationResponseParams",
     "SpeechToTextResponseParams",
+    "SpeechToTextStreamingResponseDataParams",
+    "SpeechToTextStreamingResponseParams",
     "SpeechToTextTranslateResponseParams",
     "StopConfigurationParams",
     "TextToSpeechResponseParams",
     "TimestampsModelParams",
+    "TranscriptionDataParams",
+    "TranscriptionMetricsParams",
     "TranslationResponseParams",
     "TransliterationResponseParams",
 ]

sarvamai/requests/audio_data.py

@@ -0,0 +1,21 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing_extensions
+from ..types.audio_data_encoding import AudioDataEncoding
+
+
+class AudioDataParams(typing_extensions.TypedDict):
+    data: str
+    """
+    Base64 encoded audio data
+    """
+
+    sample_rate: int
+    """
+    Audio sample rate in Hz (16kHz preferred, 8kHz least preferred)
+    """
+
+    encoding: AudioDataEncoding
+    """
+    Audio encoding format
+    """

sarvamai/requests/audio_message.py

@@ -0,0 +1,8 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing_extensions
+from .audio_data import AudioDataParams
+
+
+class AudioMessageParams(typing_extensions.TypedDict):
+    audio: AudioDataParams

sarvamai/requests/error_data.py

@@ -0,0 +1,15 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing_extensions
+
+
+class ErrorDataParams(typing_extensions.TypedDict):
+    error: str
+    """
+    Error message
+    """
+
+    code: str
+    """
+    Error code
+    """

sarvamai/requests/events_data.py

@@ -0,0 +1,17 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import datetime as dt
+
+import typing_extensions
+
+
+class EventsDataParams(typing_extensions.TypedDict):
+    event_type: str
+    """
+    Type of event
+    """
+
+    timestamp: dt.datetime
+    """
+    Event timestamp
+    """

sarvamai/requests/speech_to_text_streaming_response.py

@@ -0,0 +1,10 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing_extensions
+from ..types.speech_to_text_streaming_response_type import SpeechToTextStreamingResponseType
+from .speech_to_text_streaming_response_data import SpeechToTextStreamingResponseDataParams
+
+
+class SpeechToTextStreamingResponseParams(typing_extensions.TypedDict):
+    type: SpeechToTextStreamingResponseType
+    data: SpeechToTextStreamingResponseDataParams

sarvamai/requests/speech_to_text_streaming_response_data.py

@@ -0,0 +1,9 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+from .error_data import ErrorDataParams
+from .events_data import EventsDataParams
+from .transcription_data import TranscriptionDataParams
+
+SpeechToTextStreamingResponseDataParams = typing.Union[TranscriptionDataParams, ErrorDataParams, EventsDataParams]

sarvamai/requests/transcription_data.py

@@ -0,0 +1,35 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+import typing_extensions
+from .transcription_metrics import TranscriptionMetricsParams
+
+
+class TranscriptionDataParams(typing_extensions.TypedDict):
+    request_id: str
+    """
+    Unique identifier for the request
+    """
+
+    transcript: str
+    """
+    Transcript of the provided speech
+    """
+
+    timestamps: typing_extensions.NotRequired[typing.Dict[str, typing.Optional[typing.Any]]]
+    """
+    Timestamp information (if available)
+    """
+
+    diarized_transcript: typing_extensions.NotRequired[typing.Dict[str, typing.Optional[typing.Any]]]
+    """
+    Diarized transcript of the provided speech
+    """
+
+    language_code: typing_extensions.NotRequired[str]
+    """
+    BCP-47 code of detected language
+    """
+
+    metrics: TranscriptionMetricsParams

sarvamai/requests/transcription_metrics.py

@@ -0,0 +1,15 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing_extensions
+
+
+class TranscriptionMetricsParams(typing_extensions.TypedDict):
+    audio_duration: float
+    """
+    Duration of processed audio in seconds
+    """
+
+    processing_latency: float
+    """
+    Processing latency in seconds
+    """

sarvamai/speech_to_text_streaming/__init__.py

@@ -0,0 +1,7 @@
+# This file was auto-generated by Fern from our API Definition.
+
+# isort: skip_file
+
+from .types import SpeechToTextStreamingLanguageCode, SpeechToTextStreamingModel
+
+__all__ = ["SpeechToTextStreamingLanguageCode", "SpeechToTextStreamingModel"]

sarvamai/speech_to_text_streaming/client.py

@@ -0,0 +1,189 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+from contextlib import asynccontextmanager, contextmanager
+
+import httpx
+import websockets
+import websockets.sync.client as websockets_sync_client
+from ..core.api_error import ApiError
+from ..core.client_wrapper import AsyncClientWrapper, SyncClientWrapper
+from ..core.request_options import RequestOptions
+from .raw_client import AsyncRawSpeechToTextStreamingClient, RawSpeechToTextStreamingClient
+from .socket_client import AsyncSpeechToTextStreamingSocketClient, SpeechToTextStreamingSocketClient
+from .types.speech_to_text_streaming_language_code import SpeechToTextStreamingLanguageCode
+from .types.speech_to_text_streaming_model import SpeechToTextStreamingModel
+
+
+class SpeechToTextStreamingClient:
+    def __init__(self, *, client_wrapper: SyncClientWrapper):
+        self._raw_client = RawSpeechToTextStreamingClient(client_wrapper=client_wrapper)
+
+    @property
+    def with_raw_response(self) -> RawSpeechToTextStreamingClient:
+        """
+        Retrieves a raw implementation of this client that returns raw responses.
+
+        Returns
+        -------
+        RawSpeechToTextStreamingClient
+        """
+        return self._raw_client
+
+    @contextmanager
+    def connect(
+        self,
+        *,
+        language_code: SpeechToTextStreamingLanguageCode,
+        model: SpeechToTextStreamingModel,
+        high_vad_sensitivity: typing.Optional[str] = None,
+        vad_signals: typing.Optional[str] = None,
+        api_subscription_key: typing.Optional[str] = None,
+        request_options: typing.Optional[RequestOptions] = None,
+    ) -> typing.Iterator[SpeechToTextStreamingSocketClient]:
+        """
+        WebSocket channel for real-time speech to text streaming
+
+        Parameters
+        ----------
+        language_code : SpeechToTextStreamingLanguageCode
+            Language code for speech recognition
+
+        model : SpeechToTextStreamingModel
+            Speech to text model to use
+
+        high_vad_sensitivity : typing.Optional[str]
+            Enable high VAD (Voice Activity Detection) sensitivity
+
+        vad_signals : typing.Optional[str]
+            Enable VAD signals in response
+
+        api_subscription_key : typing.Optional[str]
+            API subscription key for authentication
+
+        request_options : typing.Optional[RequestOptions]
+            Request-specific configuration.
+
+        Returns
+        -------
+        SpeechToTextStreamingSocketClient
+        """
+        ws_url = self._raw_client._client_wrapper.get_environment().production + "/speech-to-text/ws"
+        query_params = httpx.QueryParams()
+        if language_code is not None:
+            query_params = query_params.add("language-code", language_code)
+        if model is not None:
+            query_params = query_params.add("model", model)
+        if high_vad_sensitivity is not None:
+            query_params = query_params.add("high_vad_sensitivity", high_vad_sensitivity)
+        if vad_signals is not None:
+            query_params = query_params.add("vad_signals", vad_signals)
+        ws_url = ws_url + f"?{query_params}"
+        headers = self._raw_client._client_wrapper.get_headers()
+        if api_subscription_key is not None:
+            headers["Api-Subscription-Key"] = str(api_subscription_key)
+        if request_options and "additional_headers" in request_options:
+            headers.update(request_options["additional_headers"])
+        try:
+            with websockets_sync_client.connect(ws_url, additional_headers=headers) as protocol:
+                yield SpeechToTextStreamingSocketClient(websocket=protocol)
+        except websockets.exceptions.InvalidStatusCode as exc:
+            status_code: int = exc.status_code
+            if status_code == 401:
+                raise ApiError(
+                    status_code=status_code,
+                    headers=dict(headers),
+                    body="Websocket initialized with invalid credentials.",
+                )
+            raise ApiError(
+                status_code=status_code,
+                headers=dict(headers),
+                body="Unexpected error when initializing websocket connection.",
+            )
+
+
+class AsyncSpeechToTextStreamingClient:
+    def __init__(self, *, client_wrapper: AsyncClientWrapper):
+        self._raw_client = AsyncRawSpeechToTextStreamingClient(client_wrapper=client_wrapper)
+
+    @property
+    def with_raw_response(self) -> AsyncRawSpeechToTextStreamingClient:
+        """
+        Retrieves a raw implementation of this client that returns raw responses.
+
+        Returns
+        -------
+        AsyncRawSpeechToTextStreamingClient
+        """
+        return self._raw_client
+
+    @asynccontextmanager
+    async def connect(
+        self,
+        *,
+        language_code: SpeechToTextStreamingLanguageCode,
+        model: SpeechToTextStreamingModel,
+        high_vad_sensitivity: typing.Optional[str] = None,
+        vad_signals: typing.Optional[str] = None,
+        api_subscription_key: typing.Optional[str] = None,
+        request_options: typing.Optional[RequestOptions] = None,
+    ) -> typing.AsyncIterator[AsyncSpeechToTextStreamingSocketClient]:
+        """
+        WebSocket channel for real-time speech to text streaming
+
+        Parameters
+        ----------
+        language_code : SpeechToTextStreamingLanguageCode
+            Language code for speech recognition
+
+        model : SpeechToTextStreamingModel
+            Speech to text model to use
+
+        high_vad_sensitivity : typing.Optional[str]
+            Enable high VAD (Voice Activity Detection) sensitivity
+
+        vad_signals : typing.Optional[str]
+            Enable VAD signals in response
+
+        api_subscription_key : typing.Optional[str]
+            API subscription key for authentication
+
+        request_options : typing.Optional[RequestOptions]
+            Request-specific configuration.
+
+        Returns
+        -------
+        AsyncSpeechToTextStreamingSocketClient
+        """
+        ws_url = self._raw_client._client_wrapper.get_environment().production + "/speech-to-text/ws"
+        query_params = httpx.QueryParams()
+        if language_code is not None:
+            query_params = query_params.add("language-code", language_code)
+        if model is not None:
+            query_params = query_params.add("model", model)
+        if high_vad_sensitivity is not None:
+            query_params = query_params.add("high_vad_sensitivity", high_vad_sensitivity)
+        if vad_signals is not None:
+            query_params = query_params.add("vad_signals", vad_signals)
+        ws_url = ws_url + f"?{query_params}"
+        headers = self._raw_client._client_wrapper.get_headers()
+        if api_subscription_key is not None:
+            headers["Api-Subscription-Key"] = str(api_subscription_key)
+        if request_options and "additional_headers" in request_options:
+            headers.update(request_options["additional_headers"])
+        try:
+            async with websockets.connect(ws_url, extra_headers=headers) as protocol:
+                yield AsyncSpeechToTextStreamingSocketClient(websocket=protocol)
+        except websockets.exceptions.InvalidStatusCode as exc:
+            status_code: int = exc.status_code
+            if status_code == 401:
+                raise ApiError(
+                    status_code=status_code,
+                    headers=dict(headers),
+                    body="Websocket initialized with invalid credentials.",
+                )
+            raise ApiError(
+                status_code=status_code,
+                headers=dict(headers),
+                body="Unexpected error when initializing websocket connection.",
+            )