nautobot-nvdatamodels 0.17.0

A Nautobot app that provides tools for modeling NVIDIA products and resources, such as NVLink domains and compute clusters.

NVDataModels Nautobot App

Features

This app provides the tools necessary to model NVIDIA components in Nautobot.

This app provides the following models:

• NVLinkDomain
• NVLinkDomainMembership
• ResourceBlock
• ResourceBlockMembership

Installation

Once the project is mature, installing this app will follow the standard process for installing any Nautobot app:

1. Add nautobot-app-nvdatamodels to Nautobot's pyproject.toml
2. Add nautobot_nvdatamodels to PLUGINS in nautobot_config.py

Currently, there are no plugin configuration options to add to PLUGINS_CONFIG in nautobot_config.py.

Development Environment Setup

Prerequisites

This project requires Docker, Docker Compose, and uv.

On macOS, use homebrew to install Docker, Docker Compose, and uv:

brew install --cask docker
brew install docker-compose uv

Setup

1. Clone this project locally.

2. Install dependencies.

   uv sync --all-extras
   

3. Build image.

   invoke build
   

4. Start containers.

   invoke start
   

The first time you start the containers with a fresh database, it will take 1-2 minutes for the Nautobot container to come up. Check progress by tailing the logs:

invoke logs -f  # <Ctrl-C> to stop

To enable automatic linting and formatting on file save, install the ruff VSCode plugin.

Once it's up, access the development server at http://localhost:8080.

Tips

Changes to code will auto-reload the worker for quick iteration. Tail the worker's logs for easier debugging:

invoke logs -f -s worker  # <Ctrl-C> to stop

If you're working with jobs, it's convenient to enable all jobs (which are disabled by default):

invoke enable-jobs  # Idempotently enables all disabled jobs, if any

If you're developing a job and need to re-run it often for testing, consider this approach:

1. Run the job once in the GUI.

2. Copy the JobResult's UUID.

3. Re-run the job with the same parameters from the CLI:

   invoke rerun-job <job-result-uuid>
   

This is not only faster than the GUI-only method, but also gives you the option to feed debug output directly into AI agents, if you use an AI-assisted IDE like Cursor.

It's often helpful to populate Nautobot with dummy data for testing:

invoke cli
nautobot-server generate_test_data

See all available commands with:

invoke --help