Module loader 

Document loading abstraction for pluggable file format support.

The DocumentLoader trait decouples file format handling from the RAG pipeline. Built-in loaders handle text (.txt, .md) and subtitle (.srt, .vtt) formats. Third parties can implement DocumentLoader for any format.

The LoaderRegistry dispatches loading to the appropriate loader based on file extension, with support for sidecar subtitle files adjacent to media files.

Example

use aprender_rag::loader::LoaderRegistry;
use std::path::Path;

let registry = LoaderRegistry::new();
let extensions = registry.supported_extensions();
assert!(extensions.contains(&"txt"));
assert!(extensions.contains(&"srt"));

Re-exports

pub use transcription::TranscriptionLoader;

Modules

transcription

Feature-gated transcription loader using whisper-apr for speech-to-text.

Structs

ImageLoader

Loads image files by extracting text via Tesseract OCR.

LoaderRegistry

Registry that dispatches file loading to the appropriate DocumentLoader.

SubtitleLoader

Loads subtitle files and produces Documents with timestamp metadata.

TextLoader

Loads plain text and Markdown files.

Traits

DocumentLoader

Abstraction for loading files of any format into Documents.