Adapt Learning courses

Adapt Learning builds responsive e-learning courses. ulams supports Adapt in two ways (“paths”); only the first is ready for authors today.

Path A: upload the SCORM export (works today)

1. In the Adapt authoring tool or framework, build the course with the spoor extension (adapt-contrib-spoor), which exports it as a SCORM package.
2. Upload the .zip in Courses → SCORMs. ulams recognises Adapt packages from their files and labels them Adapt in the Authoring tool column.
3. Put the package’s SCO in a lesson with a SCORM topic, or use it as the whole course.

From there it behaves exactly like any other SCORM package: same upload checks, same player, same completion tracking. See SCORM packages.

Path B: Adapt JSON sources

Path B keeps the Adapt course source itself in ulams, as versioned JSON (course, config, contentObjects, articles, blocks, components), and builds it into a SCORM package with a separate build service (ADR 0013). The built package goes through Path A, so there is still no separate Adapt topic type.

It is a feature flag, off by default. An operator turns it on with ADAPT_SOURCE_ENABLED=true and starts the build service (see Configuration). While it is off, Courses → Adapt shows a notice that Adapt sources are not enabled. You need the adapt_manage permission (given to admin).

The Adapt library

Courses → Adapt (/courses/adapt) lists sources with their title, status, current version, the version last built and the SCORM package it produced.

Status	Meaning
draft	Saved, not built yet, or changed since the last build
building	A build is queued or running. The list and the editor refresh every 3 seconds
built	The built version was imported as a SCORM package
failed	The build failed. The error is shown in the row and in the editor

1. New Adapt source: paste the JSON or upload a .json file, and optionally give a title (it defaults to the course title). The structure is checked first; errors are listed with their path, for example /components/3/_component.
2. Open the source with Edit to change the JSON. Save as new version (with an optional change note) keeps the previous versions; open an older one with Open in the version list to start from it.
3. Build turns the current version into a SCORM package. The editor asks you to save changes first.
4. When the status is built, the editor shows the package number. Add a SCORM topic to a lesson and select that package (see SCORM packages).

Deleting a source does not delete the SCORM package that was built from it.

Note

The version list comes from GET /api/admin/adapt/{id}/versions; the other endpoints are listed in API endpoints.

Configuration

For operators. Path A needs none of this.

Variable	Default	Meaning
ADAPT_SOURCE_ENABLED	false	Turns Path B on
ADAPT_BUILDER_URL	http://adapt-builder:8080	Where the API reaches the build service
ADAPT_BUILDER_TOKEN	none	Shared secret for calls to the build service; must match the service
ADAPT_BUILDER_TIMEOUT	300	Seconds the API waits for one build call
ADAPT_MAX_SOURCE_KB	4096	Largest source JSON accepted
ADAPT_QUEUE_CONNECTION, ADAPT_QUEUE	<driver>-builder, builder	Where the build job runs

The build runs as a job on the builder queue (job timeout 900 s, retry_after above it), so a worker for that queue must be running: see Queues and the scheduler. The build service is a separate GPL-3.0 program in api/adapt-builder (about 30 to 90 seconds of CPU per build). In the development stack start it with docker compose --profile adapt up -d adapt-builder from api/. Its image is not among the images published to GHCR, so a production install builds it from api/adapt-builder (see its README) and runs it where only the API can reach it.