This is the repository that is used to create docs.graylog.org running at readthedocs. After changes are pushed to the specific branches, the new pages are built automatically. When the build runs without error, the new pages are available after some minutes.
The typical workflow to make changes to the documentation and preview them before you push the changes to this repository needs the following preparation.
- clone the repository (to your workplace)
- initialize the virtual environment for python
- install the required python packages
After the above is done, changes can be made and previewed with the following
- create git branch for your changes
- change to the virtual environment
- make the modifications/additions
- run the build and check for errors
- push the branch to GitHub and create a pull request
Now a review of the changes is needed and, finally, it will be merged by the reviewer into the branch/version of the documentation you created the pull request for. Should the change be available in other versions of the documentation this should be written in the pull request that the reviewer can push this to the specific versions too.
We really recommend to use the docker-compose environment for writing documentation. Please only use all other provided options when you know how to handle errors on your own!
- git
- python (including pip)
- virtualenv (
pip install virtualenv)
- virtualenv (
- make
- browser (to preview)
It is very likely that you already have all needed software available. If not we recommend homebrew for Mac and the package manager of the Linux distribution you are using.
Multiple options are given to deal with python and different versions on a Mac, the recomment and failsafe way is to use pyenv to manage python environments.
"The basic premise of all Python development is to never use the system Python. You do not want the Mac OS X 'default Python' to be 'python3.' You want to never care about default Python." (Moshe Zadka)
# brew install pyenv
# pyenv install 3.7.5
# pyenv global 3.7.3
# echo -e 'if command -v pyenv 1>/dev/null 2>&1; then\n eval "$(pyenv init -)"\nfi' >> ~/.zshrc
# echo -e 'if command -v pyenv 1>/dev/null 2>&1; then\n eval "$(pyenv init -)"\nfi' >> ~/.bash_profile
The Version for python might need to be adjusted, at the time of writing this was the most recent version of python. In addition check for other aliases that might do something with python.
Clone the repository to your workbench:
# git clone https://github.com/Graylog2/documentation.git
Create and enter the python virtual environment:
# cd documentation
# virtualenv -p python3 .
# source ./bin/activate
Install Sphinx, the theme we are using, and sphinx-autobuild:
# pip install -r requirements.txt
For Windows, we recommend chocolatey to install the requirements. If that is not possible to use, download and install the software from the project webpages.
This can not be a complete guide, but this might give you some guidance.
Installation the needed software using chocolatey. Github Desktop and Notepad+ are optional, but a good addition. The commands need to be run in your administrator Powershell.
# choco install python
# choco install git
# choco install make
# choco install github-desktop
# choco install notepadplus
Use the Github Desktop application to clone the repository, by default this can be found in C:\Users\$USERNAME\Documents\GitHub\$REPONAME. Or you run git clone https://github.com/Graylog2/documentation.git in your Workbench directory.
Open Powershell at this location and enable your user to run scripts:
# Set-ExecutionPolicy -Scope CurrentUser -ExecutionPolicy Bypass -Force;
Use the python package manager pip to install virtulenv. Create the virtual environment, change into the environment and install all requirements:
# pip install virtualenv
# virtualenv .
# .\Scripts\activate
# pip install -r requirements.txt
Once the above is done you are prepared to contribute to the documentation and preview the work live in your local browser. See the daily usage chapter.
To ease up contributions to the documentation, you can also use Docker to create a local environment for the documentation. The image is built locally and uses the current base dir of this repository for creating and serving the documentation. This way, there's no need to handle virtual environments. Just build the image and run.
docker build -t graylog/documentation -f Dockerfile .
docker run -it -d --rm -v `pwd`:/web -u $(id -u):$(id -g) -p 8000:8000 --name graylog-documentation graylog/documentation
If unsure how to run what command with docker, the simple script docker_run.sh was created. This runs on Mac and Linux (maybe Windows when bash is available). The simplification is done in a way that you just need to run the command. If necessary it will build the image, and start the docker container. If the container is up and running it will ask if the container should be stopped. In addition it will be ask if the image should be removed too.
If unsure how to operate the docker image - use this script.
When docker-compose is installed, you could use this too - the provided docker-compose.yml includes all settings and information that are needed to build and run the docker container. All well known docker-compose commands can be used.
docker-compose up
To rebuild the image, for example to include the new pip requirements the build command is needed.
docker-compose build
How to work with git, create branches and push them will not be covered in this document. Change into the prepared directory, update the sources (git pull), change into the virtual python environment (Linux/Mac source ./bin/active, Windows .\Scripts\activate) and start making the changes.
Build the static documentation and open them in your browser:
# make html
# open _build/html/index.html
Build the documentation and automatically build them on any change:
# make livehtml
# open http://127.0.0.1:8000/
Once you've run make html or make livehtml reStructuredText syntax errors will not be displayed.
It's recommended to run make clean && make html to check for any errors before creating a PR.
When updating an image you will need to change its name, otherwise the CDN cache will not immediately deliver the new image.
If you don't want to change the name, just add a suffix like _v2.
You might update your python version and the virtualenv did not work anymore for some reasons. The linked gist can guide you out of that situation:
https://gist.github.com/jalogisch/96ebccd9a050a2d1c489a2c6149e549d
Do not run it blindly - check if the used paths are matching your local system!