# Hudra Architecture

## Overview

Hudra is a Manifest V3 browser extension that presents the existing DOM of an
HTML lesson.

It does not convert the lesson into a proprietary format.

## Slide detection

The current presentation model uses HTML headings:

- H1 identifies the lesson title
- H2 identifies the beginning of a new slide
- H3 identifies the beginning of a new slide
- H4 identifies a subsection within the current slide
- content following an H2 or H3 belongs to that slide until the next H2 or H3

Pages without H2 or H3 headings are presented as a single slide.

## Supported pages

Hudra starts only on secure QLearn pages at `qlearn.eq.edu.au` (including
subdomains) and Canvas pages on an `instructure.com` subdomain. A supported
page must also contain a recognised lesson-content container and an H1 lesson
title. Unsupported pages receive a temporary in-page notice and are not put
into presentation mode.

## Main components

### background.js

Handles extension-level behaviour and presentation launch logic.
Active session routing and window-restoration metadata is kept in
`chrome.storage.session` so presenter controls survive service-worker restarts.
Lesson content and slide previews are not persisted.

### controls.html

Provides the presenter dashboard interface.

### controls.css

Styles the presenter controls.

### controls.js

Manages presenter navigation, controls and audience communication.

### presenter.js

Creates and manages the audience presentation experience.

### page-detection.js

Contains the shared lesson-container and title detection used before launch
and when presentation mode starts.

### manifest.json

Defines the extension metadata, permissions and version.

## Design constraints

Hudra should avoid:

- cloning complex media unnecessarily
- fragile LMS-specific DOM assumptions
- large content-detection systems
- destructive modification of the original lesson

New media features should be implemented and tested independently.