Skip to content

tool

AudioTranscriptionTool

Bases: BaseTool

A tool that can be used to generate transcriptions from input audio.

Source code in griptape/tools/audio_transcription/tool.py 
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
@define
class AudioTranscriptionTool(BaseTool):
    """A tool that can be used to generate transcriptions from input audio."""

    engine: AudioTranscriptionEngine = field(kw_only=True)
    audio_loader: AudioLoader = field(default=Factory(lambda: AudioLoader()), kw_only=True)

    @activity(
        config={
            "description": "This tool can be used to generate transcriptions of audio files on disk.",
            "schema": Schema({Literal("path", description="The paths to an audio file on disk."): str}),
        },
    )
    def transcribe_audio_from_disk(self, params: dict) -> TextArtifact | ErrorArtifact:
        audio_path = params["values"]["path"]

        audio_artifact = self.audio_loader.load(Path(audio_path).read_bytes())

        return self.engine.run(audio_artifact)

    @activity(
        config={
            "description": "This tool can be used to generate the transcription of an audio artifact in memory.",
            "schema": Schema({"schema": Schema({"memory_name": str, "artifact_namespace": str, "artifact_name": str})}),
        },
    )
    def transcribe_audio_from_memory(self, params: dict[str, Any]) -> TextArtifact | ErrorArtifact:
        memory = self.find_input_memory(params["values"]["memory_name"])
        artifact_namespace = params["values"]["artifact_namespace"]
        artifact_name = params["values"]["artifact_name"]

        if memory is None:
            return ErrorArtifact("memory not found")

        audio_artifact = cast(
            AudioArtifact,
            load_artifact_from_memory(memory, artifact_namespace, artifact_name, AudioArtifact),
        )

        return self.engine.run(audio_artifact)

audio_loader: AudioLoader = field(default=Factory(lambda: AudioLoader()), kw_only=True) class-attribute instance-attribute

engine: AudioTranscriptionEngine = field(kw_only=True) class-attribute instance-attribute

transcribe_audio_from_disk(params)

Source code in griptape/tools/audio_transcription/tool.py 
26
27
28
29
30
31
32
33
34
35
36
37
@activity(
    config={
        "description": "This tool can be used to generate transcriptions of audio files on disk.",
        "schema": Schema({Literal("path", description="The paths to an audio file on disk."): str}),
    },
)
def transcribe_audio_from_disk(self, params: dict) -> TextArtifact | ErrorArtifact:
    audio_path = params["values"]["path"]

    audio_artifact = self.audio_loader.load(Path(audio_path).read_bytes())

    return self.engine.run(audio_artifact)

transcribe_audio_from_memory(params)

Source code in griptape/tools/audio_transcription/tool.py 
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
@activity(
    config={
        "description": "This tool can be used to generate the transcription of an audio artifact in memory.",
        "schema": Schema({"schema": Schema({"memory_name": str, "artifact_namespace": str, "artifact_name": str})}),
    },
)
def transcribe_audio_from_memory(self, params: dict[str, Any]) -> TextArtifact | ErrorArtifact:
    memory = self.find_input_memory(params["values"]["memory_name"])
    artifact_namespace = params["values"]["artifact_namespace"]
    artifact_name = params["values"]["artifact_name"]

    if memory is None:
        return ErrorArtifact("memory not found")

    audio_artifact = cast(
        AudioArtifact,
        load_artifact_from_memory(memory, artifact_namespace, artifact_name, AudioArtifact),
    )

    return self.engine.run(audio_artifact)