If you find a typo or a small error in the documentation you can fix it using GitHub’s online web editor. Generally speaking, we will be following https://help.github.com/en/articles/editing-files-in-another-users-repository
Find the file you want to edit under one of the folders above.
Click the pencil icon in the upper-right corner. If this is your first contribution to the Dataverse Project, the hover text over the pencil icon will say “Fork this project and edit this file”.
Make changes to the file and preview them.
In the Commit changes box, enter a description of the changes you have made and click Propose file change.
Under the Write tab, delete the long welcome message and write a few words about what you fixed.
Click Create Pull Request.
That’s it! Thank you for your contribution! Your pull request will be added manually to the main Dataverse Project board at https://github.com/orgs/IQSS/projects/2 and will go through code review and QA before it is merged into the “develop” branch. Along the way, developers might suggest changes or make them on your behalf. Once your pull request has been merged you will be listed as a contributor at https://github.com/IQSS/dataverse/graphs/contributors
Please see https://github.com/IQSS/dataverse/pull/5857 for an example of a quick fix that was merged (the “Files changed” tab shows how a typo was fixed).
If you would like to read more about the Dataverse Project’s use of GitHub, please see the Version Control section. For bug fixes and features we request that you create an issue before making a pull request but this is not at all necessary for quick fixes to the documentation.
The Dataverse guides are written using Sphinx (http://sphinx-doc.org). We recommend installing Sphinx and building the guides locally so you can get an accurate preview of your changes.
First, make a fork of https://github.com/IQSS/dataverse and clone your fork locally. Then change to the
Create a Python virtual environment, activate it, then install dependencies:
python3 -m venv venv
pip install -r requirements.txt
In some parts of the documentation, graphs are rendered as images using the Sphinx GraphViz extension.
Building the guides requires the
dot executable from GraphViz.
To edit the existing documentation:
Create a branch (see How to Make a Pull Request).
doc/sphinx-guides/sourceyou will find the .rst files that correspond to http://guides.dataverse.org.
Using your preferred text editor, open and edit the necessary files, or create new ones.
Once you are done, open a terminal, change directories to
doc/sphinx-guides, activate (or reactivate) your Python virtual environment, and build the guides.
After Sphinx is done processing the files you should notice that the
html folder in
doc/sphinx-guides/build directory has been updated.
You can click on the files in the
html folder to preview the changes.
Now you can make a commit with the changes to your own fork in GitHub and submit a pull request. See How to Make a Pull Request.
Every non-index page should use the following code to display a table of contents of internal sub-headings:
.. contents:: |toctitle| :local:
This code should be placed below any introductory text/images and directly above the first subheading, much like a Wikipedia page.
A good documentation is just like a website enhanced and upgraded by adding high quality and self-explanatory images. Often images depict a lot of written text in a simple manner. Within our Sphinx docs, you can add them in two ways: a) add a PNG image directly and include or b) use inline description languages like GraphViz (current only option).
While PNGs in the git repo can be linked directly via URL, Sphinx-generated images do not need a manual step and might provide higher visual quality. Especially in terms of quality of content, generated images can be extendend and improved by a textbased and reviewable commit, without needing raw data or source files and no diff around.
NOTE: When adding ReStructured Text (RST) cross references, use the hyphen character (
-) as the word separator for the cross reference label. For example,
my-reference-label would be the preferred label for a cross reference as opposed to, for example,
For installations hosting their own copies of the guides, note that as each version of the Dataverse Software is released, there is an updated version of the guides released with it. Google and other search engines index all versions, which may confuse users who discover your guides in the search results as to which version they should be looking at. When learning about your installation from the search results, it is best to be viewing the latest version.
In order to make it clear to the crawlers that we only want the latest version discoverable in their search results, we suggest adding this to your
User-agent: * Allow: /en/latest/ Disallow: /en/