sarvamai 0.1.23a3__py3-none-any.whl → 0.1.23a5__py3-none-any.whl

sarvamai/speech_to_text_streaming/__init__.py

@@ -2,48 +2,20 @@
 
 # isort: skip_file
 
-import typing
-from importlib import import_module
-
-if typing.TYPE_CHECKING:
-    from .types import (
-        SpeechToTextStreamingFlushSignal,
-        SpeechToTextStreamingHighVadSensitivity,
-        SpeechToTextStreamingLanguageCode,
-        SpeechToTextStreamingVadSignals,
-    )
-_dynamic_imports: typing.Dict[str, str] = {
-    "SpeechToTextStreamingFlushSignal": ".types",
-    "SpeechToTextStreamingHighVadSensitivity": ".types",
-    "SpeechToTextStreamingLanguageCode": ".types",
-    "SpeechToTextStreamingVadSignals": ".types",
-}
-
-
-def __getattr__(attr_name: str) -> typing.Any:
-    module_name = _dynamic_imports.get(attr_name)
-    if module_name is None:
-        raise AttributeError(f"No {attr_name} found in _dynamic_imports for module name -> {__name__}")
-    try:
-        module = import_module(module_name, __package__)
-        if module_name == f".{attr_name}":
-            return module
-        else:
-            return getattr(module, attr_name)
-    except ImportError as e:
-        raise ImportError(f"Failed to import {attr_name} from {module_name}: {e}") from e
-    except AttributeError as e:
-        raise AttributeError(f"Failed to get {attr_name} from {module_name}: {e}") from e
-
-
-def __dir__():
-    lazy_attrs = list(_dynamic_imports.keys())
-    return sorted(lazy_attrs)
-
+from .types import (
+    SpeechToTextStreamingFlushSignal,
+    SpeechToTextStreamingHighVadSensitivity,
+    SpeechToTextStreamingLanguageCode,
+    SpeechToTextStreamingMode,
+    SpeechToTextStreamingModel,
+    SpeechToTextStreamingVadSignals,
+)
 
 __all__ = [
     "SpeechToTextStreamingFlushSignal",
     "SpeechToTextStreamingHighVadSensitivity",
     "SpeechToTextStreamingLanguageCode",
+    "SpeechToTextStreamingMode",
+    "SpeechToTextStreamingModel",
     "SpeechToTextStreamingVadSignals",
 ]

sarvamai/speech_to_text_streaming/client.py

@@ -14,6 +14,8 @@ from .socket_client import AsyncSpeechToTextStreamingSocketClient, SpeechToTextS
 from .types.speech_to_text_streaming_flush_signal import SpeechToTextStreamingFlushSignal
 from .types.speech_to_text_streaming_high_vad_sensitivity import SpeechToTextStreamingHighVadSensitivity
 from .types.speech_to_text_streaming_language_code import SpeechToTextStreamingLanguageCode
+from .types.speech_to_text_streaming_mode import SpeechToTextStreamingMode
+from .types.speech_to_text_streaming_model import SpeechToTextStreamingModel
 from .types.speech_to_text_streaming_vad_signals import SpeechToTextStreamingVadSignals
 
 try:
@@ -42,7 +44,8 @@ class SpeechToTextStreamingClient:
         self,
         *,
         language_code: SpeechToTextStreamingLanguageCode,
-        model: typing.Optional[typing.Literal["saarika:v2.5"]] = None,
+        model: typing.Optional[SpeechToTextStreamingModel] = None,
+        mode: typing.Optional[SpeechToTextStreamingMode] = None,
         sample_rate: typing.Optional[str] = None,
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
@@ -60,10 +63,47 @@ class SpeechToTextStreamingClient:
         Parameters
         ----------
         language_code : SpeechToTextStreamingLanguageCode
-            Language code for speech recognition
+            Specifies the language of the input audio in BCP-47 format.
 
-        model : typing.Optional[typing.Literal["saarika:v2.5"]]
-            Speech to text model to use
+            **Available Options:**
+            - `hi-IN`: Hindi
+            - `bn-IN`: Bengali
+            - `gu-IN`: Gujarati
+            - `kn-IN`: Kannada
+            - `ml-IN`: Malayalam
+            - `mr-IN`: Marathi
+            - `od-IN`: Odia
+            - `pa-IN`: Punjabi
+            - `ta-IN`: Tamil
+            - `te-IN`: Telugu
+            - `en-IN`: English
+
+        model : typing.Optional[SpeechToTextStreamingModel]
+            Specifies the model to use for speech-to-text conversion.
+
+            - **saarika:v2.5** (default): Transcribes audio in the spoken language.
+
+            - **saaras:v3**: State-of-the-art model with flexible output formats. Supports multiple modes via the `mode` parameter: transcribe, translate, verbatim, translit, codemix.
+
+        mode : typing.Optional[SpeechToTextStreamingMode]
+            Mode of operation. **Only applicable when using saaras:v3 model.**
+
+            Example audio: 'मेरा फोन नंबर है 9840950950'
+
+            - **transcribe** (default): Standard transcription in the original language with proper formatting and number normalization.
+              - Output: `मेरा फोन नंबर है 9840950950`
+
+            - **translate**: Translates speech from any supported Indic language to English.
+              - Output: `My phone number is 9840950950`
+
+            - **verbatim**: Exact word-for-word transcription without normalization, preserving filler words and spoken numbers as-is.
+              - Output: `मेरा फोन नंबर है नौ आठ चार zero नौ पांच zero नौ पांच zero`
+
+            - **translit**: Romanization - Transliterates speech to Latin/Roman script only.
+              - Output: `mera phone number hai 9840950950`
+
+            - **codemix**: Code-mixed text with English words in English and Indic words in native script.
+              - Output: `मेरा phone number है 9840950950`
 
         sample_rate : typing.Optional[str]
             Audio sample rate for the WebSocket connection. When specified as a connection parameter, only 16kHz and 8kHz are supported. 8kHz is only available via this connection parameter. If not specified, defaults to 16kHz.
@@ -93,6 +133,8 @@ class SpeechToTextStreamingClient:
             query_params = query_params.add("language-code", language_code)
         if model is not None:
             query_params = query_params.add("model", model)
+        if mode is not None:
+            query_params = query_params.add("mode", mode)
         if sample_rate is not None:
             query_params = query_params.add("sample_rate", sample_rate)
         if high_vad_sensitivity is not None:
@@ -145,7 +187,8 @@ class AsyncSpeechToTextStreamingClient:
         self,
         *,
         language_code: SpeechToTextStreamingLanguageCode,
-        model: typing.Optional[typing.Literal["saarika:v2.5"]] = None,
+        model: typing.Optional[SpeechToTextStreamingModel] = None,
+        mode: typing.Optional[SpeechToTextStreamingMode] = None,
         sample_rate: typing.Optional[str] = None,
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
@@ -163,10 +206,47 @@ class AsyncSpeechToTextStreamingClient:
         Parameters
         ----------
         language_code : SpeechToTextStreamingLanguageCode
-            Language code for speech recognition
+            Specifies the language of the input audio in BCP-47 format.
+
+            **Available Options:**
+            - `hi-IN`: Hindi
+            - `bn-IN`: Bengali
+            - `gu-IN`: Gujarati
+            - `kn-IN`: Kannada
+            - `ml-IN`: Malayalam
+            - `mr-IN`: Marathi
+            - `od-IN`: Odia
+            - `pa-IN`: Punjabi
+            - `ta-IN`: Tamil
+            - `te-IN`: Telugu
+            - `en-IN`: English
+
+        model : typing.Optional[SpeechToTextStreamingModel]
+            Specifies the model to use for speech-to-text conversion.
+
+            - **saarika:v2.5** (default): Transcribes audio in the spoken language.
+
+            - **saaras:v3**: State-of-the-art model with flexible output formats. Supports multiple modes via the `mode` parameter: transcribe, translate, verbatim, translit, codemix.
+
+        mode : typing.Optional[SpeechToTextStreamingMode]
+            Mode of operation. **Only applicable when using saaras:v3 model.**
+
+            Example audio: 'मेरा फोन नंबर है 9840950950'
+
+            - **transcribe** (default): Standard transcription in the original language with proper formatting and number normalization.
+              - Output: `मेरा फोन नंबर है 9840950950`
+
+            - **translate**: Translates speech from any supported Indic language to English.
+              - Output: `My phone number is 9840950950`
+
+            - **verbatim**: Exact word-for-word transcription without normalization, preserving filler words and spoken numbers as-is.
+              - Output: `मेरा फोन नंबर है नौ आठ चार zero नौ पांच zero नौ पांच zero`
+
+            - **translit**: Romanization - Transliterates speech to Latin/Roman script only.
+              - Output: `mera phone number hai 9840950950`
 
-        model : typing.Optional[typing.Literal["saarika:v2.5"]]
-            Speech to text model to use
+            - **codemix**: Code-mixed text with English words in English and Indic words in native script.
+              - Output: `मेरा phone number है 9840950950`
 
         sample_rate : typing.Optional[str]
             Audio sample rate for the WebSocket connection. When specified as a connection parameter, only 16kHz and 8kHz are supported. 8kHz is only available via this connection parameter. If not specified, defaults to 16kHz.
@@ -196,6 +276,8 @@ class AsyncSpeechToTextStreamingClient:
             query_params = query_params.add("language-code", language_code)
         if model is not None:
             query_params = query_params.add("model", model)
+        if mode is not None:
+            query_params = query_params.add("mode", mode)
         if sample_rate is not None:
             query_params = query_params.add("sample_rate", sample_rate)
         if high_vad_sensitivity is not None:

sarvamai/speech_to_text_streaming/raw_client.py

@@ -13,6 +13,8 @@ from .socket_client import AsyncSpeechToTextStreamingSocketClient, SpeechToTextS
 from .types.speech_to_text_streaming_flush_signal import SpeechToTextStreamingFlushSignal
 from .types.speech_to_text_streaming_high_vad_sensitivity import SpeechToTextStreamingHighVadSensitivity
 from .types.speech_to_text_streaming_language_code import SpeechToTextStreamingLanguageCode
+from .types.speech_to_text_streaming_mode import SpeechToTextStreamingMode
+from .types.speech_to_text_streaming_model import SpeechToTextStreamingModel
 from .types.speech_to_text_streaming_vad_signals import SpeechToTextStreamingVadSignals
 
 try:
@@ -30,7 +32,8 @@ class RawSpeechToTextStreamingClient:
         self,
         *,
         language_code: SpeechToTextStreamingLanguageCode,
-        model: typing.Optional[typing.Literal["saarika:v2.5"]] = None,
+        model: typing.Optional[SpeechToTextStreamingModel] = None,
+        mode: typing.Optional[SpeechToTextStreamingMode] = None,
         sample_rate: typing.Optional[str] = None,
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
@@ -48,10 +51,47 @@ class RawSpeechToTextStreamingClient:
         Parameters
         ----------
         language_code : SpeechToTextStreamingLanguageCode
-            Language code for speech recognition
+            Specifies the language of the input audio in BCP-47 format.
 
-        model : typing.Optional[typing.Literal["saarika:v2.5"]]
-            Speech to text model to use
+            **Available Options:**
+            - `hi-IN`: Hindi
+            - `bn-IN`: Bengali
+            - `gu-IN`: Gujarati
+            - `kn-IN`: Kannada
+            - `ml-IN`: Malayalam
+            - `mr-IN`: Marathi
+            - `od-IN`: Odia
+            - `pa-IN`: Punjabi
+            - `ta-IN`: Tamil
+            - `te-IN`: Telugu
+            - `en-IN`: English
+
+        model : typing.Optional[SpeechToTextStreamingModel]
+            Specifies the model to use for speech-to-text conversion.
+
+            - **saarika:v2.5** (default): Transcribes audio in the spoken language.
+
+            - **saaras:v3**: State-of-the-art model with flexible output formats. Supports multiple modes via the `mode` parameter: transcribe, translate, verbatim, translit, codemix.
+
+        mode : typing.Optional[SpeechToTextStreamingMode]
+            Mode of operation. **Only applicable when using saaras:v3 model.**
+
+            Example audio: 'मेरा फोन नंबर है 9840950950'
+
+            - **transcribe** (default): Standard transcription in the original language with proper formatting and number normalization.
+              - Output: `मेरा फोन नंबर है 9840950950`
+
+            - **translate**: Translates speech from any supported Indic language to English.
+              - Output: `My phone number is 9840950950`
+
+            - **verbatim**: Exact word-for-word transcription without normalization, preserving filler words and spoken numbers as-is.
+              - Output: `मेरा फोन नंबर है नौ आठ चार zero नौ पांच zero नौ पांच zero`
+
+            - **translit**: Romanization - Transliterates speech to Latin/Roman script only.
+              - Output: `mera phone number hai 9840950950`
+
+            - **codemix**: Code-mixed text with English words in English and Indic words in native script.
+              - Output: `मेरा phone number है 9840950950`
 
         sample_rate : typing.Optional[str]
             Audio sample rate for the WebSocket connection. When specified as a connection parameter, only 16kHz and 8kHz are supported. 8kHz is only available via this connection parameter. If not specified, defaults to 16kHz.
@@ -81,6 +121,8 @@ class RawSpeechToTextStreamingClient:
             query_params = query_params.add("language-code", language_code)
         if model is not None:
             query_params = query_params.add("model", model)
+        if mode is not None:
+            query_params = query_params.add("mode", mode)
         if sample_rate is not None:
             query_params = query_params.add("sample_rate", sample_rate)
         if high_vad_sensitivity is not None:
@@ -122,7 +164,8 @@ class AsyncRawSpeechToTextStreamingClient:
         self,
         *,
         language_code: SpeechToTextStreamingLanguageCode,
-        model: typing.Optional[typing.Literal["saarika:v2.5"]] = None,
+        model: typing.Optional[SpeechToTextStreamingModel] = None,
+        mode: typing.Optional[SpeechToTextStreamingMode] = None,
         sample_rate: typing.Optional[str] = None,
         high_vad_sensitivity: typing.Optional[SpeechToTextStreamingHighVadSensitivity] = None,
         vad_signals: typing.Optional[SpeechToTextStreamingVadSignals] = None,
@@ -140,10 +183,47 @@ class AsyncRawSpeechToTextStreamingClient:
         Parameters
         ----------
         language_code : SpeechToTextStreamingLanguageCode
-            Language code for speech recognition
+            Specifies the language of the input audio in BCP-47 format.
+
+            **Available Options:**
+            - `hi-IN`: Hindi
+            - `bn-IN`: Bengali
+            - `gu-IN`: Gujarati
+            - `kn-IN`: Kannada
+            - `ml-IN`: Malayalam
+            - `mr-IN`: Marathi
+            - `od-IN`: Odia
+            - `pa-IN`: Punjabi
+            - `ta-IN`: Tamil
+            - `te-IN`: Telugu
+            - `en-IN`: English
+
+        model : typing.Optional[SpeechToTextStreamingModel]
+            Specifies the model to use for speech-to-text conversion.
+
+            - **saarika:v2.5** (default): Transcribes audio in the spoken language.
+
+            - **saaras:v3**: State-of-the-art model with flexible output formats. Supports multiple modes via the `mode` parameter: transcribe, translate, verbatim, translit, codemix.
+
+        mode : typing.Optional[SpeechToTextStreamingMode]
+            Mode of operation. **Only applicable when using saaras:v3 model.**
+
+            Example audio: 'मेरा फोन नंबर है 9840950950'
+
+            - **transcribe** (default): Standard transcription in the original language with proper formatting and number normalization.
+              - Output: `मेरा फोन नंबर है 9840950950`
+
+            - **translate**: Translates speech from any supported Indic language to English.
+              - Output: `My phone number is 9840950950`
+
+            - **verbatim**: Exact word-for-word transcription without normalization, preserving filler words and spoken numbers as-is.
+              - Output: `मेरा फोन नंबर है नौ आठ चार zero नौ पांच zero नौ पांच zero`
+
+            - **translit**: Romanization - Transliterates speech to Latin/Roman script only.
+              - Output: `mera phone number hai 9840950950`
 
-        model : typing.Optional[typing.Literal["saarika:v2.5"]]
-            Speech to text model to use
+            - **codemix**: Code-mixed text with English words in English and Indic words in native script.
+              - Output: `मेरा phone number है 9840950950`
 
         sample_rate : typing.Optional[str]
             Audio sample rate for the WebSocket connection. When specified as a connection parameter, only 16kHz and 8kHz are supported. 8kHz is only available via this connection parameter. If not specified, defaults to 16kHz.
@@ -173,6 +253,8 @@ class AsyncRawSpeechToTextStreamingClient:
             query_params = query_params.add("language-code", language_code)
         if model is not None:
             query_params = query_params.add("model", model)
+        if mode is not None:
+            query_params = query_params.add("mode", mode)
         if sample_rate is not None:
             query_params = query_params.add("sample_rate", sample_rate)
         if high_vad_sensitivity is not None:

sarvamai/speech_to_text_streaming/types/__init__.py

@@ -2,46 +2,18 @@
 
 # isort: skip_file
 
-import typing
-from importlib import import_module
-
-if typing.TYPE_CHECKING:
-    from .speech_to_text_streaming_flush_signal import SpeechToTextStreamingFlushSignal
-    from .speech_to_text_streaming_high_vad_sensitivity import SpeechToTextStreamingHighVadSensitivity
-    from .speech_to_text_streaming_language_code import SpeechToTextStreamingLanguageCode
-    from .speech_to_text_streaming_vad_signals import SpeechToTextStreamingVadSignals
-_dynamic_imports: typing.Dict[str, str] = {
-    "SpeechToTextStreamingFlushSignal": ".speech_to_text_streaming_flush_signal",
-    "SpeechToTextStreamingHighVadSensitivity": ".speech_to_text_streaming_high_vad_sensitivity",
-    "SpeechToTextStreamingLanguageCode": ".speech_to_text_streaming_language_code",
-    "SpeechToTextStreamingVadSignals": ".speech_to_text_streaming_vad_signals",
-}
-
-
-def __getattr__(attr_name: str) -> typing.Any:
-    module_name = _dynamic_imports.get(attr_name)
-    if module_name is None:
-        raise AttributeError(f"No {attr_name} found in _dynamic_imports for module name -> {__name__}")
-    try:
-        module = import_module(module_name, __package__)
-        if module_name == f".{attr_name}":
-            return module
-        else:
-            return getattr(module, attr_name)
-    except ImportError as e:
-        raise ImportError(f"Failed to import {attr_name} from {module_name}: {e}") from e
-    except AttributeError as e:
-        raise AttributeError(f"Failed to get {attr_name} from {module_name}: {e}") from e
-
-
-def __dir__():
-    lazy_attrs = list(_dynamic_imports.keys())
-    return sorted(lazy_attrs)
-
+from .speech_to_text_streaming_flush_signal import SpeechToTextStreamingFlushSignal
+from .speech_to_text_streaming_high_vad_sensitivity import SpeechToTextStreamingHighVadSensitivity
+from .speech_to_text_streaming_language_code import SpeechToTextStreamingLanguageCode
+from .speech_to_text_streaming_mode import SpeechToTextStreamingMode
+from .speech_to_text_streaming_model import SpeechToTextStreamingModel
+from .speech_to_text_streaming_vad_signals import SpeechToTextStreamingVadSignals
 
 __all__ = [
     "SpeechToTextStreamingFlushSignal",
     "SpeechToTextStreamingHighVadSensitivity",
     "SpeechToTextStreamingLanguageCode",
+    "SpeechToTextStreamingMode",
+    "SpeechToTextStreamingModel",
     "SpeechToTextStreamingVadSignals",
 ]

sarvamai/speech_to_text_streaming/types/speech_to_text_streaming_mode.py

@@ -0,0 +1,7 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+SpeechToTextStreamingMode = typing.Union[
+    typing.Literal["transcribe", "translate", "verbatim", "translit", "codemix"], typing.Any
+]

sarvamai/speech_to_text_streaming/types/speech_to_text_streaming_model.py

@@ -0,0 +1,5 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+SpeechToTextStreamingModel = typing.Union[typing.Literal["saarika:v2.5", "saaras:v3"], typing.Any]