This scripts generates files and meta-data for the device catalog from various sources.
Currently, it uses the Zephyr repository as the only source.
It outputs files into
vendor/zephyrdocs/firmware/hardwarestatic/img
which can then be re-used by Docusaurus pages and in the future other Golioth systems, too.
You'll need node and bash to run it. It's only been tested with bash.
There is a build-device-catalog node.js script available on the project level. Running it from the
repository root is the preferred execution method for generating the device catalog:
npm run build-device-catalog
You can also execute the individual scripts from that node.js script.
If you want to further extend the scripts, the following notes should help you get started.
Fetch the Zephyr repository into the vendor/zephyr directory by running ./get-zephyr-boards.sh from the repository root.
This will perform a sparse checkout of the main branch, or update the already checked out repository.
The sparse checkout is set to only the boards directory of the Zephyr repo, but is still ~430 MB download.
You'll need Linux shell to execute the .sh script.
The index.js script will process and transform the available data sources into a folder structure in the docs/firmware/hardware and static/img.
We have number of different support levels for boards, indicating how well they are integrated with our platform.
To add support level annotation for a board
- look up the architecture of the board
- look up the Board ID of the board
- edit
scripts/device-catalog/support.json - under the specific architecture and Board ID, add the support level
The support level values are
unverified= board with unknown compatibility with Goliothcommunity= community members verified the board work with Goliothverified= Golioth team verified the board works at some point in timequickstart= Golioth team provides support and periodically tests compatibility