Struct TranscriptionLoader 

pub struct TranscriptionLoader { /* private fields */ }

Loader that handles media files via sidecar subtitle detection and whisper-apr-based speech-to-text transcription.

When a media file has a sidecar subtitle (.srt or .vtt) adjacent to it, the subtitle is loaded directly. Otherwise, the audio is decoded and transcribed using the whisper-apr Whisper ASR engine.

Example

use aprender_rag::loader::transcription::{TranscriptionLoader, TranscriptionConfig};
use aprender_rag::loader::LoaderRegistry;

let mut registry = LoaderRegistry::new();
registry.register(Box::new(TranscriptionLoader::with_defaults()));
// Now the registry handles .mp4, .wav, etc. via sidecar detection

Implementations

impl TranscriptionLoader

pub fn new(config: TranscriptionConfig) -> Self

Create a new transcription loader with the given configuration.

If config.model_path is set, loads the whisper-apr model eagerly. Otherwise, transcription of files without sidecars will fail gracefully.

pub fn with_defaults() -> Self

Create a loader with default configuration (no model loaded).

pub fn config(&self) -> &TranscriptionConfig

Access the transcription configuration.

pub fn has_model(&self) -> bool

Check if a Whisper model is loaded and ready for transcription.

Trait Implementations

impl Debug for TranscriptionLoader

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl DocumentLoader for TranscriptionLoader

fn supported_extensions(&self) -> Vec<&str>

File extensions this loader handles (lowercase, without dot).

fn load(&self, path: &Path) -> Result<Document>

Load a file and produce a Document.

fn can_load(&self, path: &Path) -> bool

Returns true if this loader can handle the given path.