feat: add speech_threshold

GitOrigin-RevId: e4fdb960f375bd01c3d79f3bc85bc346fcaf9368

Author: dmccrystal

assemblyai/types.py

@@ -413,6 +413,9 @@ class RawTranscriptionConfig(BaseModel):
         - Dutch
     """
 
+    speech_threshold: Optional[float]
+    "Reject audio files that contain less than this fraction of speech. Valid values are in the range [0,1] inclusive"
+
     class Config:
         extra = Extra.allow
 
@@ -451,6 +454,7 @@ def __init__(
         summary_type: Optional[SummarizationType] = None,
         auto_highlights: Optional[bool] = None,
         language_detection: Optional[bool] = None,
+        speech_threshold: Optional[float] = None,
         raw_transcription_config: Optional[RawTranscriptionConfig] = None,
     ) -> None:
         """
@@ -485,6 +489,7 @@ def __init__(
             summary_type: The summarization type to use in case `summarization` is enabled
             auto_highlights: Detect important phrases and words in your transcription text.
             language_detection: Identify the dominant language that’s spoken in an audio file, and route the file to the appropriate model for the detected language.
+            speech_threshold: Reject audio files that contain less than this fraction of speech. Valid values are in the range [0,1] inclusive
             raw_transcription_config: Create the config from a `RawTranscriptionConfig`
         """
         self._raw_transcription_config = raw_transcription_config
@@ -529,6 +534,7 @@ def __init__(
         )
         self.auto_highlights = auto_highlights
         self.language_detection = language_detection
+        self.speech_threshold = speech_threshold
 
     @property
     def raw(self) -> RawTranscriptionConfig:
@@ -854,6 +860,21 @@ def language_detection(self, enable: Optional[bool]) -> None:
 
         self._raw_transcription_config.language_detection = enable
 
+    @property
+    def speech_threshold(self) -> Optional[float]:
+        "Returns the current speech threshold."
+
+        return self._raw_transcription_config.speech_threshold
+
+    @speech_threshold.setter
+    def speech_threshold(self, threshold: Optional[float]) -> None:
+        "Reject audio files that contain less than this fraction of speech. Valid values are in the range [0,1] inclusive"
+
+        if threshold is not None and (threshold < 0 or threshold > 1):
+            raise ValueError("speech_threshold must be between 0 and 1 (inclusive).")
+
+        self._raw_transcription_config.speech_threshold = threshold
+
     # endregion
 
     # region: Convenience (helper) methods
@@ -1434,6 +1455,9 @@ class BaseTranscript(BaseModel):
         - Dutch
     """
 
+    speech_threshold: Optional[float]
+    "Reject audio files that contain less than this fraction of speech. Valid values are in the range [0,1] inclusive"
+
 
 class TranscriptRequest(BaseTranscript):
     """

setup.py

@@ -7,19 +7,19 @@
 
 setup(
     name="assemblyai",
-    version="0.11.0",
+    version="0.7.1",
     description="AssemblyAI Python SDK",
     author="AssemblyAI",
     author_email="engineering.sdk@assemblyai.com",
     packages=find_packages(),
     install_requires=[
         "httpx>=0.19.0",
         "pydantic>=1.7.0",
-        "typing-extensions>=3.7,<4.6",
-        "websockets>=11.0",
+        "typing-extensions",
+        "websockets>=10.0",
     ],
     extras_require={
-        "extras": ["pyaudio>=0.2.13"],
+        "extras": ["pyaudio"],
     },
     classifiers=[
         "Development Status :: 3 - Alpha",

tests/unit/factories.py

@@ -90,6 +90,7 @@ class Meta:
     summary_type = None
     auto_highlights = False
     language_detection = False
+    speech_threshold = None
 
 
 class TranscriptCompletedResponseFactory(BaseTranscriptFactory):

tests/unit/test_config.py

@@ -18,3 +18,14 @@ def test_configuration_are_none_by_default():
             pytest.fail(
                 f"Configuration field {name} is {value} and not None by default."
             )
+
+
+def test_speech_threshold_fails_if_outside_range():
+    """
+    Tests that an exception is raised if the value for speech_threshold is outside the range of [0, 1].
+    """
+
+    with pytest.raises(ValueError, match="speech_threshold"):
+        aai.TranscriptionConfig(speech_threshold=1.5)
+    with pytest.raises(ValueError, match="speech_threshold"):
+        aai.TranscriptionConfig(speech_threshold=-0.5)

tests/unit/test_transcript_group.py

@@ -1,5 +1,6 @@
 import uuid
 
+import assemblyai as aai
 import assemblyai as aai
 from tests.unit import factories
 

tox.ini

@@ -16,15 +16,14 @@ deps =
     pydantic1.9: pydantic>=1.9.0,<1.10.0
     pydantic1.8: pydantic>=1.8.0,<1.9.0
     pydantic1.7: pydantic>=1.7.0,<1.8.0
-    typing-extensions: typing-extensions>=3.7,<4.6
+    typing-extensions: typing-extensions>=3.7
     # extra dependencies
-    pyaudiolatest: pyaudio
+    pyaudio: pyaudio
     pyaudio0.2: pyaudio>=0.2.13,<0.3.0
     # test dependencies
     pytest
     pytest-httpx
     pytest-xdist
-    pytest-mock
     pytest-cov
     factory-boy
 allowlist_externals = pytest