Package org.springframework.ai.openai.metadata

Class OpenAiAudioTranscriptionResponseMetadata

java.lang.Object

org.springframework.ai.model.MutableResponseMetadata

org.springframework.ai.audio.transcription.AudioTranscriptionResponseMetadata

org.springframework.ai.openai.metadata.OpenAiAudioTranscriptionResponseMetadata

All Implemented Interfaces:

ResponseMetadata

public class OpenAiAudioTranscriptionResponseMetadata
extends AudioTranscriptionResponseMetadata

Audio transcription metadata implementation for OpenAI using the OpenAI Java SDK.

duration, usage and segment/word-level timestamps are only returned by the OpenAI API for the verbose_json and diarized_json response_format values; for other formats these accessors return null.

Author:

Christian Tzolov

Constructor Summary

Constructors

Constructor	Description
OpenAiAudioTranscriptionResponseMetadata()
OpenAiAudioTranscriptionResponseMetadata(@Nullable Double duration, @Nullable String language, @Nullable Object usage, @Nullable List<?> segments, @Nullable List<com.openai.models.audio.transcriptions.TranscriptionWord> words)

Method Summary

Modifier and Type	Method	Description
@Nullable Double	getDuration()	Returns the audio duration in seconds, or null if not returned by the API.
@Nullable String	getLanguage()	Returns the detected input language, or null if not returned by the API.
@Nullable List<?>	getSegments()	Returns the segment-level transcription breakdown, or null if not returned by the API.
@Nullable Object	getUsage()	Returns the provider-native usage object for this transcription, or null if not returned by the API.
@Nullable List<com.openai.models.audio.transcriptions.TranscriptionWord>	getWords()	Returns the word-level timestamps, or null if not returned by the API.
String	toString()

Methods inherited from class org.springframework.ai.model.MutableResponseMetadata

clear, computeIfAbsent, containsKey, entrySet, get, getOrDefault, getRawMap, getRequired, isEmpty, keySet, put, remove

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface org.springframework.ai.model.ResponseMetadata

getOrDefault

Constructor Details

OpenAiAudioTranscriptionResponseMetadata

public OpenAiAudioTranscriptionResponseMetadata()

OpenAiAudioTranscriptionResponseMetadata

public OpenAiAudioTranscriptionResponseMetadata(@Nullable Double duration,
 @Nullable String language,
 @Nullable Object usage,
 @Nullable List<?> segments,
 @Nullable List<com.openai.models.audio.transcriptions.TranscriptionWord> words)

Method Details

getDuration

public @Nullable Double getDuration()

Returns the audio duration in seconds, or null if not returned by the API.

getLanguage

public @Nullable String getLanguage()

Returns the detected input language, or null if not returned by the API.

getUsage

public @Nullable Object getUsage()

Returns the provider-native usage object for this transcription, or null if not returned by the API. The concrete type depends on the response format requested: TranscriptionVerbose.Usage for verbose_json, TranscriptionDiarized.Usage for diarized_json, or Transcription.Usage otherwise.

getSegments

public @Nullable List<?> getSegments()

Returns the segment-level transcription breakdown, or null if not returned by the API. Elements are TranscriptionSegment for verbose_json, or TranscriptionDiarizedSegment (carrying per-speaker attribution) for diarized_json.

getWords

public @Nullable List<com.openai.models.audio.transcriptions.TranscriptionWord> getWords()

Returns the word-level timestamps, or null if not returned by the API. Only populated for verbose_json when word-level timestamp_granularities were requested.

toString

public String toString()

Overrides:

toString in class Object