Danbyte plugin template

A starter for building a Danbyte plugin. Click “Use this template” (or clone it), rename a few things, and you have a working plugin that adds a model, a REST API, a monitoring check kind, an automation provider, and server-driven UI — with no changes to Danbyte core and no frontend build.

Danbyte plugins are trusted, in-process Python packages (NetBox-style): an operator installs the package, lists it in the PLUGINS setting, and restarts. See the full guide: Danbyte docs → Architecture → Plugins.

---

What's in here

danbyte_myplugin/
├── apps.py            # DanbytePluginConfig — plugin metadata
├── danbyte_plugin.py  # THE registration entry point (autodiscovered)
├── models.py          # a sample tenant-scoped model (Gadget)
├── serializers.py     # DRF serializer (tags + custom fields)
├── viewsets.py        # tenant-scoped, RBAC default-closed CRUD
├── api_urls.py        # mounted at /api/plugins/<slug>/
├── checks.py          # a sample monitoring check kind
└── migrations/
pyproject.toml         # packaging (hatchling)
.github/workflows/     # publish to PyPI on release (Trusted Publishing)

Rename it (once)

Pick a name — e.g. danbyte_acme. Then:

1. Rename the danbyte_myplugin/ folder → danbyte_acme/.
2. In pyproject.toml: name = "danbyte-acme" and packages = ["danbyte_acme"].
3. In apps.py: name = "danbyte_acme", and pick a slug (e.g. acme).
4. Search-and-replace myplugin → acme and danbyte_myplugin → danbyte_acme across the package (slugs, URLs, the migration's model, audit_type).
5. Rename the sample Gadget model to yours (or delete it and add your own).

Develop against a Danbyte checkout

Your plugin imports from the host (core, api, auth_api, plugins, danbyte_checks), so develop it inside a Danbyte environment:

# in your Danbyte checkout, with its virtualenv active
pip install -e /path/to/danbyte-acme          # editable install
echo 'PLUGINS=danbyte_acme' >> .env           # enable it
python manage.py makemigrations danbyte_acme  # generate real migrations
python manage.py migrate
python manage.py runserver                    # (or restart your services)

Open Settings → Deployment → Plugins & services — your plugin appears as loaded, and its nav item + pages show up (subject to RBAC). Grant yourself the object-type permission (e.g. gadget) to see and edit rows.

What you can register (danbyte_plugin.py)

You want…	Call / mixin
A model in RBAC + import/export + eventing	register_object_type("danbyte_acme.Gadget", "Gadgets", "Plugins")
Custom fields on it	mix in core.models.CustomFieldsMixin
Tags on it	mix in core.models.TaggableMixin
It targetable by object-reference custom fields	register_reference_model(ReferenceModel(...))
Change-log / audit	register_audited_model("danbyte_acme.Gadget")
A REST API	ship api_urls.py (auto-mounted at /api/plugins/<slug>/)
An automation runner	register_automation_provider(kind, runner)
An import source / notification channel	register_import_source(...) / register_notification_channel(...)
A monitoring check kind	@danbyte_checks.base.register on a Checker
Sidebar nav / pages / dashboard panels	register_nav_item / register_page / register_dashboard_panel

Rules that matter

• apps.py and __init__.py must be import-safe — no model imports at the top level (Danbyte reads your metadata before the app registry exists).
• Every model uses a UUID primary key and a tenant FK; scope all querysets to the active tenant (the TenantScopedViewSet base does this).
• New endpoints are default-closed — registering the object type is what makes yours require a <model>.* grant. Never trust a client-supplied id to belong to the current tenant.

Publish to PyPI

The included workflow publishes on a GitHub Release using Trusted Publishing (OIDC — no token stored):

1. On PyPI: create the project (or reserve the name), then Publishing → add a GitHub publisher: owner your-org, repo danbyte-acme, workflow publish.yml, environment pypi.
2. In GitHub: create an environment named pypi.
3. Cut a GitHub Release (tag e.g. v0.1.0). The workflow builds and uploads.

Prefer a token? Set a PYPI_API_TOKEN repo secret and pass it to the publish step (see the comment in .github/workflows/publish.yml). Never commit a token.

Users then install your plugin with pip install danbyte-acme, add it to PLUGINS, and apply (migrate + restart).

License

Apache-2.0, matching Danbyte. Change it to suit your plugin.