NuShell scripts for managing Canvas LMS instances
# Get all published courses that use Zoom
> canvas courses list --published=true
| insert using_zoom? {
not (
canvas courses tabs
| where label == "Zoom" and hidden == false
| is-empty
)
}
| select id name using_zoom?
╭───┬────┬────────────┬─────────────╮
│ # │ id │ name │ using_zoom? │
├───┼────┼────────────┼─────────────┤
│ 0 │ 1 │ Admin Test │ true │
│ 1 │ 2 │ Test 2 │ false │
╰───┴────┴────────────┴─────────────╯
# Get all manually enrolled (non-SIS) users in courses
> canvas courses list
| insert manual_enrollments {
canvas enrollments list
| where sis_import_id == null
| flatten user --all
| select role name user_id
}
| flatten manual_enrollments --all
| select id name manual_enrollments_name user_id role
| rename course_id course user user_id role
╭───┬───────────┬────────────┬────────────┬─────────┬───────────────────╮
│ # │ course_id │ course │ user │ user_id │ role │
├───┼───────────┼────────────┼────────────┼─────────┼───────────────────┤
│ 0 │ 1 │ Admin Test │ Admin Alex │ 5 │ TeacherEnrollment │
│ 1 │ 2 │ Test 2 │ Student │ 6 │ StudentEnrollment │
╰───┴───────────┴────────────┴────────────┴─────────┴───────────────────╯After installing NuShell, add this repository to your $env.NU_LIB_DIRS. Next, in Canvas,
generate a personal access token.
Then, define the following environment variables in your NuShell environment:
let-env CANVAS_URL = "https://yourschool.instructure.com" # No trailing slash
let-env CANVAS_TOKEN = "YOUR-CANVAS-ACCESS-TOKEN"
let-env CANVAS_ROOT_ACCOUNT_ID = 1 # Optional, set this if you're self-hosting Canvas or have an odd account structurePersonally, I have each Canvas instance I manage in its own file that I can use. For example, for a production
environment, I define a URL and token in prod.nu inside of an export-env block.
Once your environment is configured, just use canvas.nu from a NuShell prompt. Run canvas my user to verify
the script is working correctly.
> canvas my user
╭──────────────────┬─────────────────────────────────────────────────────────────╮
│ id │ 1 │
│ name │ Site Admin │
│ created_at │ a week ago │
│ sortable_name │ site-admin@localhost │
│ short_name │ site-admin@localhost │
│ sis_user_id │ │
│ integration_id │ │
│ sis_import_id │ │
│ login_id │ site-admin@localhost │
│ avatar_url │ http://canvas.instructure.com/images/messages/avatar-50.png │
│ email │ site-admin@localhost │
│ locale │ │
│ effective_locale │ en │
│ │ ╭─────────────────────────────┬───────╮ │
│ permissions │ │ can_update_name │ true │ │
│ │ │ can_update_avatar │ false │ │
│ │ │ limit_parent_app_web_access │ false │ │
│ │ ╰─────────────────────────────┴───────╯ │
╰──────────────────┴─────────────────────────────────────────────────────────────╯Managing multiple Canvas LMS instances can be time-consuming, especially when it comes to tasks that involve bulk operations, such as adding users or courses. While there has been some limited effort to develop shell scripts that interact with the Canvas API, there is a lack of comprehensive command-line tools to manage Canvas LMS instances. For anyone familiar with other excellent tools, such as gam, this toolbox aims to provide a similar level of comfort and flexibility.
The main goal of the toolbox is to provide a simple and efficient way to perform common, low-level Canvas LMS management tasks, such as creating and deleting courses, managing users and groups, and performing course enrollment operations. By using the shell, admins can quickly and easily automate routine tasks, reducing the time and effort required to manage Canvas LMS instances.
At the moment, this project is a thin wrapper around the limited bits of the Canvas API that I've bothered to implement. Currently, the criteria for what API endpoints have been implemented is those that I need for my day-to-day work. As this project matures, I hope to include more high-level workflows.
NuShell is awesome.
But seriously, being able to easily pipe and filter data is extremely useful for the bulk tasks I run daily. For the Canvas tasks I've automated in the past, I mostly used python and UCF's wonderful canvasapi library. The examples above consisted of about 10 minutes of work each and are extremely readable whereas my equivalent python scripts would likely not be as succient and would have taken about an hour to implement.