TranscriptionSession Class Reference

Mutable per-session transcription interface. More...

#include <transcriptionengine.h>

Public Member Functions
	TranscriptionSession (const TranscriptionSession &)=delete
TranscriptionSession &	operator= (const TranscriptionSession &)=delete
	TranscriptionSession (TranscriptionSession &&)=delete
TranscriptionSession &	operator= (TranscriptionSession &&)=delete
virtual QString	backendName () const =0
	Returns the backend identifier for this live session.
virtual bool	warmup (RuntimeError *error=nullptr)=0
	Performs optional backend warmup for this session.
virtual TranscriptUpdate	pushAudioChunk (const AudioChunk &chunk)=0
	Ingests one normalized audio chunk into the live session.
virtual TranscriptUpdate	finish ()=0
	Flushes the current utterance and emits any final transcript events.
virtual TranscriptUpdate	cancel ()=0
	Requests cooperative cancellation of any active decode.

Detailed Description

Mutable per-session transcription interface.

Sessions own backend state that may be warmed up, reused, and kept isolated per thread or request flow.

Definition at line 58 of file transcriptionengine.h.

Member Function Documentation

backendName()

virtual QString TranscriptionSession::backendName ( ) const

pure virtual

Returns the backend identifier for this live session.

Returns

Short backend name used for logs and diagnostics.

Implemented in CpuReferenceTranscriber, and WhisperCppTranscriber.

cancel()

virtual TranscriptUpdate TranscriptionSession::cancel ( )

pure virtual

Requests cooperative cancellation of any active decode.

Implementations should stop in-flight backend work best-effort without using thread interruption.

Implemented in CpuReferenceTranscriber, and WhisperCppTranscriber.

finish()

virtual TranscriptUpdate TranscriptionSession::finish ( )

pure virtual

Flushes the current utterance and emits any final transcript events.

Returns

Final transcript events or a structured runtime failure.

Implemented in CpuReferenceTranscriber, and WhisperCppTranscriber.

pushAudioChunk()

virtual TranscriptUpdate TranscriptionSession::pushAudioChunk ( const AudioChunk &  chunk )

pure virtual

Ingests one normalized audio chunk into the live session.

Parameters

chunk

Product-owned normalized audio chunk.

Returns

Any transcript events emitted while ingesting the chunk.

Implemented in CpuReferenceTranscriber, and WhisperCppTranscriber.

warmup()

virtual bool TranscriptionSession::warmup ( RuntimeError *  error = nullptr )

pure virtual

Performs optional backend warmup for this session.

Parameters

error

Optional destination for a structured failure reason.

Returns

true if the session is ready for transcription, otherwise false.

Implemented in CpuReferenceTranscriber, and WhisperCppTranscriber.

---

The documentation for this class was generated from the following file:

• src/transcription/transcriptionengine.h