This repository holds the source for the Software Carpentry web site. Lessons are not stored in this repository: please see the lessons page for links to their repositories.
Software Carpentry is an open source/open access project, and we welcome contributions of all kinds. By contributing, you are agreeing that Software Carpentry may redistribute your work under these licenses.
Rebuilding the web site locally to check changes requires:
We use Jekyll because it's what GitHub uses; we use Python because most of our volunteers speak it. The Python packages we depend on can be installed using:
$ pip install -r requirements.txt
We try to use the same MarkDown interpreters as GitHub does for consistency. On OS X, we suggest you use a recent Ruby to get access to these. If you don't have Homebrew or MacPorts installed, here's a quick recipe to get started using HomeBrew:
ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
brew install ruby
gem install jekyll
gem install RedCloth
The commands used to rebuild the website are stored in Makefile
.
- Type
make
on its own to see a list of all available commands. - Type
make clean
to get rid of generated files. - Type
make cache
to rebuild cached information about workshops and repositories. You only need to do this once a day or so, since that information changes relatively infrequently, but you do need to do it at least once before you can preview the website. (Depending on your machine, this takes 15-30 seconds.) - Type
make site
to build everything in_site
for testing. (Depending on your machine, this takes about 10-15 seconds.) You can then open_site/index.html
and other pages in_site
to see what your changes will look like.
Note: Disqus comments will not load properly, since you'll be on your machine rather than our server.
The simplest way to create a new blog post is to mail the text to Greg Wilson, who will format it and post it for you. If you'd like to submit your post as a pull request, then:
-
Fork this repository on GitHub, and clone that to your desktop.
-
Create a file
blog/YYYY/MM/post-name.html
, whereYYYY
is a four-digit year andMM
is a two-digit month. (The directoriesYYYY
andYYYY/MM
will already exist unless you're the first person to blog this month or this year.) -
Fill in the metadata at the top of the file (or if you have copied an existing blog post, edit the values):
layout
must beblog
.author
must be the author's name. (It does not need to be quoted.)title
is the post title.date
the date of the post (in YYYY-MM-DD format).time
must be an HH:MM clock time, such as 09:30. Every post must have a unique timestamp so that posts can be ordered.category
must be a list of category identifiers, e.g.,["Euphoric State University", "Assessment"]
You can use an empty list if you want, and we'll fill in categories for you.
-
Use HTML to write the post.
- Use
<!-- start excerpt -->
and<!-- end excerpt -->
to mark a paragraph or two at the start as the excerpt to show in feed readers. - If you need to refer to our email address, it is
{{page.contact}}
. - If you need to another post, or something else on the site, use
{{page.root}}/path/to/file
.
- Use
-
Please add any images your blog post needs to the same blog/YYYY/MM directory as the post itself. Please use lower-case names without special characters for image files.
-
When you're satisfied with your post,
git add path/to/post
andgit commit -m "Adding a blog post about something or other"
will commit it to your local copy (on your laptop). -
git push origin master
will push it to your clone on GitHub (assuming you've added your fork on GitHub as a remote calledorigin
). -
Go to GitHub and issue a pull request from your clone to
swcarpentry/site
, then assign it togvwilson
for proof-reading.
If you have set up a GitHub website for a repository,
and would like it listed on our website,
all you need to do is add one line to the file config/workshop_urls.yml
and send us a pull request
(or if you'd rather, just mail Greg
and he'll add it for you).
Please add the URL of the GitHub repository,
not the website itself,
i.e.,
add https://github.com/someone/yyyy-mm-dd-site
rather than http://someone.github.io/yyyy-mm-dd
.
make site
(and its partners make dev
and make install
) do the following:
-
Run
bin/preprocess.py
to create the./_config.yml
file required by Jekyll and the_includes/recent_blog_posts.html
file containing excerpts of recent blog posts. This tool collects metadata about blog posts and workshops and combines it with static configuration information. -
Run Jekyll to build the web site.
-
Run
bin/make_rss_feed.py
to generate the RSS feed filefeed.xml
. -
Run
bin/make_calendar.py
to generate the ICal calendar fileworkshops.ics
. -
Copy
./_htaccess
to create the.htaccess
file for the web site.
bin/preprocess.py
needs two generated files to work properly:
-
_workshop_cache.yml
: stores information about recent and upcoming workshops. This file is created bybin/get_workshop_info.py
. -
_dashboard_cache.yml
: stores information about Software Carpentry's GitHub projects. This file is created bybin/make-dashboard.py
.
The tool used to rebuild the dashboard cache requires credentials in order to run, since use of GitHub's API is throttled for unauthenticated users and programs. If you wish to regenerate these files:
-
Get a GitHub API token.
-
Save it in a file called
git-token.txt
in the root directory of this site. (This file is ignored by Git, since tokens should not be shared between people.)
After that,
you can run make cache
to re-create the two files.
Please do not commit them.
Software Carpentry uses Open Badges to recognize people's skills and accomplishments. To create badges, you must install PyPNG module:
$ pip install pypng
Use bin/badge-create.py
to create a new badge, e.g.:
$ python bin/badge-create.py username email instructor
The badges scripts in bin
should be compatible with Python2 and Python3.
To bake the badge we use bin/badgebakery.py
which was provided by the Open Badge Team.