NVMe-Spex is a tool for linting- and generating a structured representation of the data-structures contained in an NVMe specification document(https://nvmexpress.org/).
NVMe-Spex provides two front-ends to interact with the tool. The most basic interface to interact with NVMe-Spex is the command line tool as an alternative a web-application is also provided where the docx/html files can be uploaded and processed.
Please see the full documentation at: https://samsungds.github.io/Spex/
There are multiple options for installing NVME-Spex. The list of supported method is: Nix package, python package, docker image and installing nvme-spex from source.
NVME-Spex has libxml2(https://gitlab.gnome.org/GNOME/libxml2) as a dependency. In any other case than for Nix and docker installs it is necessary to install libxml2 manually.
For apt based systems such as debian or ubuntu the following commands will install libxml2:
sudo apt update
sudo apt install libxml2-dev
Open a terminal and issue the following command to install libxml2:
brew install libxml2
Given a copy of Nix, the environment will be automatically configured, and is guaranteed to match the environment in which we develop and for which CI testing is done.
See https://samsungds.github.io/Spex/setup/nix.html for details.
NOTE Nix works natively on Linux and MacOS, and on Windows via WSL. See the link above for notes on installing- and using Nix on all 3 platforms.
It is possible to install the nvme-spex tool directly from pypi.org via the pipx or pip tool. The nvme-spex tool is build with lxml and therefor has a none python dependency that needs to be installed which is libxml2. See instructions for you particular operation systems to install libxml2.
When the dependency is met is is possible to install nvme-spex with following command:
pipx install nvme-spex
NOTE that in this case, you will still need Python 3.11 or later and
libxml2.
NVMe-Spex is also packaged as a docker container image and can the newest release can be pulled and run by executing the following docker commands:
docker pull ghcr.io/samsungds/nvme-spex-webserver:latest
docker run -p 8000:8000 ghcr.io/samsungds/nvme-spex-webserver:latest
When the docker container is successfully running the web application can be accessed in the browser at http://localhost:8000 .
Note for Windows it is possible to install docker though the docker desktop(https://docs.docker.com/desktop/install/windows-install/) or the podman project (https://podman.io/docs/installation). We aware of docker desktops licensing model.
Clone repository with git:
git clone [email protected]:SamsungDS/Spex.git
cd spex
make build install
make build install build the package and installs Spex for your user in
a separate Python virtual environment - meaning the dependencies of Spex are
not impacted by other Python software you install. make uninstall removes
Spex, if installed.
NOTE that in this case, you will still need Python 3.11 or later and
libxml2.
See https://samsungds.github.io/Spex/user_guide/using_spex.html for details.
Note that targets like build, check, format and docs all expect to
operate in a properly configured environment.
To understand your options, see https://samsungds.github.io/Spex/setup/index.html
Finally, if you use nix, make dev will put you inside a properly configured
development environment.
This target runs all the tests that our regular CI run would to ensure the software is in a good state. If you make modifications, please run this before submitting.
This target automatically re-arranges package imports and reformats the code to comply with our coding standards. Please use this ahead of submitting changes.
This target builds the documentation locally. This is mostly useful when writing previewing changes to the documentation locally.
These targets respectively build and removes a source distribution python package. You most likely won't need or want this.