sarvamai 0.1.23a2__py3-none-any.whl → 0.1.23a4__py3-none-any.whl

sarvamai/speech_to_text/raw_client.py

@@ -1,6 +1,5 @@
 # This file was auto-generated by Fern from our API Definition.
 
-import json
 import typing
 from json.decoder import JSONDecodeError
 
@@ -8,7 +7,6 @@ from .. import core
 from ..core.api_error import ApiError
 from ..core.client_wrapper import AsyncClientWrapper, SyncClientWrapper
 from ..core.http_response import AsyncHttpResponse, HttpResponse
-from ..core.jsonable_encoder import jsonable_encoder
 from ..core.pydantic_utilities import parse_obj_as
 from ..core.request_options import RequestOptions
 from ..errors.bad_request_error import BadRequestError
@@ -18,6 +16,7 @@ from ..errors.service_unavailable_error import ServiceUnavailableError
 from ..errors.too_many_requests_error import TooManyRequestsError
 from ..errors.unprocessable_entity_error import UnprocessableEntityError
 from ..types.input_audio_codec import InputAudioCodec
+from ..types.mode import Mode
 from ..types.speech_to_text_language import SpeechToTextLanguage
 from ..types.speech_to_text_model import SpeechToTextModel
 from ..types.speech_to_text_response import SpeechToTextResponse
@@ -37,6 +36,7 @@ class RawSpeechToTextClient:
         *,
         file: core.File,
         model: typing.Optional[SpeechToTextModel] = OMIT,
+        mode: typing.Optional[Mode] = OMIT,
         language_code: typing.Optional[SpeechToTextLanguage] = OMIT,
         input_audio_codec: typing.Optional[InputAudioCodec] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
@@ -63,7 +63,18 @@ class RawSpeechToTextClient:
 
         model : typing.Optional[SpeechToTextModel]
             Specifies the model to use for speech-to-text conversion.
-            Note:- Default model is `saarika:v2.5`
+            - **saarika:v2.5** (default): Standard transcription model
+            - **saarika:v3**: Advanced transcription model
+            - **saaras:v3**: Advanced model with multiple output modes
+
+        mode : typing.Optional[Mode]
+            Mode of operation. **Only applicable when using saaras:v3 model.**
+            - **transcribe** (default): Standard transcription
+            - **translate**: Translation to English
+            - **indic-en**: Indic to English translation
+            - **verbatim**: Exact transcription
+            - **translit**: Transliteration to Latin script
+            - **codemix**: Code-mixed output
 
         language_code : typing.Optional[SpeechToTextLanguage]
             Specifies the language of the input audio.
@@ -86,7 +97,8 @@ class RawSpeechToTextClient:
             base_url=self._client_wrapper.get_environment().base,
             method="POST",
             data={
-                "model": json.dumps(jsonable_encoder(model)),
+                "model": model,
+                "mode": mode,
                 "language_code": language_code,
                 "input_audio_codec": input_audio_codec,
             },
@@ -111,9 +123,9 @@ class RawSpeechToTextClient:
                 raise BadRequestError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -122,9 +134,9 @@ class RawSpeechToTextClient:
                 raise ForbiddenError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -133,9 +145,9 @@ class RawSpeechToTextClient:
                 raise UnprocessableEntityError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -144,9 +156,9 @@ class RawSpeechToTextClient:
                 raise TooManyRequestsError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -155,9 +167,9 @@ class RawSpeechToTextClient:
                 raise InternalServerError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -166,9 +178,9 @@ class RawSpeechToTextClient:
                 raise ServiceUnavailableError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -230,7 +242,7 @@ class RawSpeechToTextClient:
             method="POST",
             data={
                 "prompt": prompt,
-                "model": json.dumps(jsonable_encoder(model)),
+                "model": model,
                 "input_audio_codec": input_audio_codec,
             },
             files={
@@ -254,9 +266,9 @@ class RawSpeechToTextClient:
                 raise BadRequestError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -265,9 +277,9 @@ class RawSpeechToTextClient:
                 raise ForbiddenError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -276,9 +288,9 @@ class RawSpeechToTextClient:
                 raise UnprocessableEntityError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -287,9 +299,9 @@ class RawSpeechToTextClient:
                 raise TooManyRequestsError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -298,9 +310,9 @@ class RawSpeechToTextClient:
                 raise InternalServerError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -309,9 +321,9 @@ class RawSpeechToTextClient:
                 raise ServiceUnavailableError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -331,6 +343,7 @@ class AsyncRawSpeechToTextClient:
         *,
         file: core.File,
         model: typing.Optional[SpeechToTextModel] = OMIT,
+        mode: typing.Optional[Mode] = OMIT,
         language_code: typing.Optional[SpeechToTextLanguage] = OMIT,
         input_audio_codec: typing.Optional[InputAudioCodec] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
@@ -357,7 +370,18 @@ class AsyncRawSpeechToTextClient:
 
         model : typing.Optional[SpeechToTextModel]
             Specifies the model to use for speech-to-text conversion.
-            Note:- Default model is `saarika:v2.5`
+            - **saarika:v2.5** (default): Standard transcription model
+            - **saarika:v3**: Advanced transcription model
+            - **saaras:v3**: Advanced model with multiple output modes
+
+        mode : typing.Optional[Mode]
+            Mode of operation. **Only applicable when using saaras:v3 model.**
+            - **transcribe** (default): Standard transcription
+            - **translate**: Translation to English
+            - **indic-en**: Indic to English translation
+            - **verbatim**: Exact transcription
+            - **translit**: Transliteration to Latin script
+            - **codemix**: Code-mixed output
 
         language_code : typing.Optional[SpeechToTextLanguage]
             Specifies the language of the input audio.
@@ -380,7 +404,8 @@ class AsyncRawSpeechToTextClient:
             base_url=self._client_wrapper.get_environment().base,
             method="POST",
             data={
-                "model": json.dumps(jsonable_encoder(model)),
+                "model": model,
+                "mode": mode,
                 "language_code": language_code,
                 "input_audio_codec": input_audio_codec,
             },
@@ -405,9 +430,9 @@ class AsyncRawSpeechToTextClient:
                 raise BadRequestError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -416,9 +441,9 @@ class AsyncRawSpeechToTextClient:
                 raise ForbiddenError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -427,9 +452,9 @@ class AsyncRawSpeechToTextClient:
                 raise UnprocessableEntityError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -438,9 +463,9 @@ class AsyncRawSpeechToTextClient:
                 raise TooManyRequestsError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -449,9 +474,9 @@ class AsyncRawSpeechToTextClient:
                 raise InternalServerError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -460,9 +485,9 @@ class AsyncRawSpeechToTextClient:
                 raise ServiceUnavailableError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -524,7 +549,7 @@ class AsyncRawSpeechToTextClient:
             method="POST",
             data={
                 "prompt": prompt,
-                "model": json.dumps(jsonable_encoder(model)),
+                "model": model,
                 "input_audio_codec": input_audio_codec,
             },
             files={
@@ -548,9 +573,9 @@ class AsyncRawSpeechToTextClient:
                 raise BadRequestError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -559,9 +584,9 @@ class AsyncRawSpeechToTextClient:
                 raise ForbiddenError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -570,9 +595,9 @@ class AsyncRawSpeechToTextClient:
                 raise UnprocessableEntityError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -581,9 +606,9 @@ class AsyncRawSpeechToTextClient:
                 raise TooManyRequestsError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -592,9 +617,9 @@ class AsyncRawSpeechToTextClient:
                 raise InternalServerError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),
@@ -603,9 +628,9 @@ class AsyncRawSpeechToTextClient:
                 raise ServiceUnavailableError(
                     headers=dict(_response.headers),
                     body=typing.cast(
-                        typing.Any,
+                        typing.Optional[typing.Any],
                         parse_obj_as(
-                            type_=typing.Any,  # type: ignore
+                            type_=typing.Optional[typing.Any],  # type: ignore
                             object_=_response.json(),
                         ),
                     ),

sarvamai/speech_to_text_job/client.py

@@ -12,6 +12,7 @@ from ..types.files_upload_response import FilesUploadResponse
 from ..types.job_status_v_1_response import JobStatusV1Response
 from ..types.speech_to_text_model import SpeechToTextModel
 from ..types.speech_to_text_language import SpeechToTextLanguage
+from ..types.mode import Mode
 from .raw_client import AsyncRawSpeechToTextJobClient, RawSpeechToTextJobClient
 from .job import AsyncSpeechToTextJob, SpeechToTextJob
 
@@ -72,7 +73,9 @@ class SpeechToTextJobClient:
         )
         """
         _response = self._raw_client.initialise(
-            job_parameters=job_parameters, callback=callback, request_options=request_options
+            job_parameters=job_parameters,
+            callback=callback,
+            request_options=request_options,
         )
         return _response.data
 
@@ -145,11 +148,17 @@ class SpeechToTextJobClient:
             job_id="job_id",
         )
         """
-        _response = self._raw_client.start(job_id, ptu_id=ptu_id, request_options=request_options)
+        _response = self._raw_client.start(
+            job_id, ptu_id=ptu_id, request_options=request_options
+        )
         return _response.data
 
     def get_upload_links(
-        self, *, job_id: str, files: typing.Sequence[str], request_options: typing.Optional[RequestOptions] = None
+        self,
+        *,
+        job_id: str,
+        files: typing.Sequence[str],
+        request_options: typing.Optional[RequestOptions] = None,
     ) -> FilesUploadResponse:
         """
         Start a speech to text bulk job V1
@@ -180,11 +189,17 @@ class SpeechToTextJobClient:
             files=["files"],
         )
         """
-        _response = self._raw_client.get_upload_links(job_id=job_id, files=files, request_options=request_options)
+        _response = self._raw_client.get_upload_links(
+            job_id=job_id, files=files, request_options=request_options
+        )
         return _response.data
 
     def get_download_links(
-        self, *, job_id: str, files: typing.Sequence[str], request_options: typing.Optional[RequestOptions] = None
+        self,
+        *,
+        job_id: str,
+        files: typing.Sequence[str],
+        request_options: typing.Optional[RequestOptions] = None,
     ) -> FilesDownloadResponse:
         """
         Start a speech to text bulk job V1
@@ -215,12 +230,15 @@ class SpeechToTextJobClient:
             files=["files"],
         )
         """
-        _response = self._raw_client.get_download_links(job_id=job_id, files=files, request_options=request_options)
+        _response = self._raw_client.get_download_links(
+            job_id=job_id, files=files, request_options=request_options
+        )
         return _response.data
 
     def create_job(
         self,
         model: SpeechToTextModel = "saarika:v2.5",
+        mode: typing.Optional[Mode] = None,
         with_diarization: bool = False,
         with_timestamps: bool = False,
         language_code: typing.Optional[SpeechToTextLanguage] = None,
@@ -236,6 +254,10 @@ class SpeechToTextJobClient:
         model : SpeechToTextModel, default="saarika:v2.5"
             The model to use for transcription.
 
+        mode : typing.Optional[Mode], default=None
+            Mode of operation. Only applicable for saaras:v3 model.
+            Options: transcribe, translate, indic-en, verbatim, translit, codemix
+
         with_diarization : typing.Optional[bool], default=False
             Whether to enable speaker diarization (distinguishing who said what).
 
@@ -244,7 +266,7 @@ class SpeechToTextJobClient:
 
         language_code : typing.Optional[SpeechToTextLanguage], default=None
             The language code of the input audio (e.g., "hi-IN", "bn-IN").
-            
+
         num_speakers : typing.Optional[int], default=None
             The number of distinct speakers in the audio, if known.
 
@@ -263,6 +285,7 @@ class SpeechToTextJobClient:
             job_parameters=SpeechToTextJobParametersParams(
                 language_code=language_code,
                 model=model,
+                mode=mode,  # type: ignore[typeddict-item]
                 num_speakers=num_speakers,  # type: ignore[typeddict-item]
                 with_diarization=with_diarization,
                 with_timestamps=with_timestamps,
@@ -350,7 +373,9 @@ class AsyncSpeechToTextJobClient:
         asyncio.run(main())
         """
         _response = await self._raw_client.initialise(
-            job_parameters=job_parameters, callback=callback, request_options=request_options
+            job_parameters=job_parameters,
+            callback=callback,
+            request_options=request_options,
         )
         return _response.data
 
@@ -392,7 +417,9 @@ class AsyncSpeechToTextJobClient:
 
         asyncio.run(main())
         """
-        _response = await self._raw_client.get_status(job_id, request_options=request_options)
+        _response = await self._raw_client.get_status(
+            job_id, request_options=request_options
+        )
         return _response.data
 
     async def start(
@@ -439,11 +466,17 @@ class AsyncSpeechToTextJobClient:
 
         asyncio.run(main())
         """
-        _response = await self._raw_client.start(job_id, ptu_id=ptu_id, request_options=request_options)
+        _response = await self._raw_client.start(
+            job_id, ptu_id=ptu_id, request_options=request_options
+        )
         return _response.data
 
     async def get_upload_links(
-        self, *, job_id: str, files: typing.Sequence[str], request_options: typing.Optional[RequestOptions] = None
+        self,
+        *,
+        job_id: str,
+        files: typing.Sequence[str],
+        request_options: typing.Optional[RequestOptions] = None,
     ) -> FilesUploadResponse:
         """
         Start a speech to text bulk job V1
@@ -482,11 +515,17 @@ class AsyncSpeechToTextJobClient:
 
         asyncio.run(main())
         """
-        _response = await self._raw_client.get_upload_links(job_id=job_id, files=files, request_options=request_options)
+        _response = await self._raw_client.get_upload_links(
+            job_id=job_id, files=files, request_options=request_options
+        )
         return _response.data
 
     async def get_download_links(
-        self, *, job_id: str, files: typing.Sequence[str], request_options: typing.Optional[RequestOptions] = None
+        self,
+        *,
+        job_id: str,
+        files: typing.Sequence[str],
+        request_options: typing.Optional[RequestOptions] = None,
     ) -> FilesDownloadResponse:
         """
         Start a speech to text bulk job V1
@@ -533,6 +572,7 @@ class AsyncSpeechToTextJobClient:
     async def create_job(
         self,
         model: SpeechToTextModel = "saarika:v2.5",
+        mode: typing.Optional[Mode] = None,
         with_diarization: bool = False,
         with_timestamps: bool = False,
         language_code: typing.Optional[SpeechToTextLanguage] = None,
@@ -548,6 +588,10 @@ class AsyncSpeechToTextJobClient:
         model : SpeechToTextModel, default="saarika:v2.5"
             The model to use for transcription.
 
+        mode : typing.Optional[Mode], default=None
+            Mode of operation. Only applicable for saaras:v3 model.
+            Options: transcribe, translate, indic-en, verbatim, translit, codemix
+
         with_diarization : typing.Optional[bool], default=False
             Whether to enable speaker diarization (distinguishing who said what).
 
@@ -556,8 +600,8 @@ class AsyncSpeechToTextJobClient:
 
         language_code : typing.Optional[SpeechToTextLanguage], default=None
             The language code of the input audio (e.g., "hi-IN", "bn-IN").
-            
-        num_speakers : typing.Optional[int], default=None
+
+        num_speakers : typing.Optional[int] = None
             The number of distinct speakers in the audio, if known.
 
         callback : typing.Optional[BulkJobCallbackParams], default=OMIT
@@ -575,6 +619,7 @@ class AsyncSpeechToTextJobClient:
             job_parameters=SpeechToTextJobParametersParams(
                 language_code=language_code,
                 model=model,
+                mode=mode,  # type: ignore[typeddict-item]
                 with_diarization=with_diarization,
                 with_timestamps=with_timestamps,
                 num_speakers=num_speakers,  # type: ignore[typeddict-item]