# Importing Storyline and Rise modules

How Storyline 360 and Rise 360 modules are read into one model and converted into Elabora modules, with a report of every loss (beta, D47). Part of the [architecture](../architecture.md), the contract every change follows.

A module made with Storyline 360 or Rise 360 becomes an Elabora module: its blocks, layers, events and variables, its media copied, a draft plan, and its report. Elabora never traces a slide: it reads it (decision D47).

| Part | Role |
|---|---|
| `@elabora/core` `import/model.ts` | One model for every source: sections of pages, items in the source's terms (a text with its fonts and size, an image, a question…, or a `fallback` or `unsupported` feature), boxes on a slide, start times, hidden items, layers, triggers with conditions and actions, variables |
| `import/storyline-source.ts` | The `.story` (an OPC package of XML parts): `story/story.xml` (scenes, variables, quizzes, the media registry, matched to `story/media/` by MD5, else by size), each slide part (`loc` boxes, `zOrder`, plain and `Document` texts, pictures, buttons, questions and their `scoreData`, layers and feedback layers, `trigLst` triggers and `trigCond` conditions) |
| `import/storyline-published.ts` | A publication's `html5/data/js/` (`parseProvidedData`: the `globalProvideData` wrapping): `data.js` (scenes, slides, variables, questions with their right answers, assets), `frame.js` (scene titles, notes), each slide (layers, objects with their boxes, `vartext` as HTML or blocks of spans, `imagelib`, events and action groups, the timeline's show times); questions as D63's activities: matching and drag and drop (the right answer's pairs) as a sort in pairs, sequence as a sort in order, fill in the blank and word bank as fill-ins, Likert as reflections on a scale; the player's own triggers and variables left out |
| `import/rise.ts` | A Rise export's course JSON (`findRiseCourse`: the ways a page embeds it: 2017 `var data`, 2018 `window.course`, 2021–2023 `window.courseData`, 2025 `deserialize(…)`, or a locale file named by the page, `readRiseLocale`; and the 2026 `runtime-data.js`, `readRiseRuntime`), its lessons, sections, blocks and quiz lessons, a scenario block read as a dialogue (D61); D63's blocks: statements A–D, quotes (author, portrait, on a picture), image presentations (Rise's crop as the focal point), galleries, two columns, dividers, panels and tabs (an item's picture; its embed a link, its video after the panels), flip cards, process and timeline as steps, labeled graphics as interactive pictures (markers in percent), sorting (unscored categories, twelve cards a sort), matching and fill in the blank as activities, charts as D60's, 2017's map and sorting lessons; a Continue that waits for the block above opens once explored (once answered after knowledge checks only); media found by their crushed key, src or key |
| `import/layout.ts` | A slide's reading order: a recursive XY cut (bands, then columns), or the author's focus order when a reader gives one; its visual title; its backgrounds |
| `import/convert.ts` | The model as files (`course.json`, `plan.json`, `screens/*.json`), the media to copy, the vetted fonts to add, and the report; `choices` take the report's decisions otherwise. A reader gives D60's and D63's blocks as `block` items, their props as written and their media with where each goes and what its absence does (the prop left out, the entry gone, or the item's `otherwise`); assets never take a name a sync service's conflict copy would have |
| `import/report.ts` | The report's shape (`elabora-import-report/1`) and words, in French or English |
| `@elabora/node` `import/source.ts` | A ZIP (`ZipReader`) or a folder read alike: paths, small texts read whole, media copied by streams and hashed in chunks; nothing outside the source is reachable |
| `import/index.ts` | `detectImport(file)` (by content: `story/story.xml`, `html5/data/js/data.js`, a Rise player page or a 2026 `runtime-data.js`, the one nearest the root first, so a Rise course that holds a Storyline block's publication, or an older export in a folder, is imported as the Rise course; an Elabora archive or folder told apart), `importCourse({ file, dir?, … })`: the module written into `dir`, or built in the library's staging folder and moved into place, its fonts added as `elabora font add` does (journaled `add_font` changes), `import-report.json`, then validated; `brief` and `defaultLanguage` (the app's « Nouveau module », D51), `replace: <project id>` (imported again in its place, history and sources kept, `replaceWithStaged`) |

- **Inputs**: a `.story`; a publication, a folder or its zip (the course may sit in a folder of the zip), or its `story.html`; a Rise export, web, SCORM or xAPI, zipped or not. A Storyline block inside a Rise export is reported, not imported.
- **Storyline's source (`.story`)**: its matching, sequence, fill in the blank, word bank and Likert questions stay fallbacks (their instruction in a text): the format of those questions in the source is not read yet.
- **What is never taken**: Articulate's code, media whose original path is in Articulate's folders (the Content Library, its textures) and Storyline characters (a note marks the place), the source's font files (fonts without a web licence Elabora knows become vetted fonts, a decision per family), the author's JavaScript.
- **The CLI**: `elabora import <file> [dir] [--title] [--id] [--lang] [--choose <decision>=<choice>,…] [--json]`. No MCP tool: the MCP server works on one course and importing makes another; agents run the command. The app calls `importCourse` from « Nouveau module » and shows the report (its sentence, its decisions with their choices; `choices` runs it again).