| title |
|---|
README |
For tech people: Clone the repo, make the changes, push them back. Follow the conventions laid out just below.
For you other people: Read on.
Keep in mind:
- The production website is mostly purple and is called neicweb on GitHub.
- The development system
- is more orange (orange headings, orange stripe in the header)
- has "TESTING" in the window name
- and is called neicweb-devel on GitHub.
Howto:
- Prepare your pictures, pdfs and other media files. Please try to make your pictures "widescreen", or roughly 3:2, or 16:9. Consider making miniatures for very large pictures. Name the files so they match exactly what you want them to be on the server.
- Upload your media files to the /media/ folder. Click the "Upload files" button near the top, and follow instructions onscreen.
- Write the article text on GitHub. News article files live in the _posts/ folder:
- Click a news article that looks similar to what you want.
- Please feel free to click around and compare what you see with the end result on the actual web server.
- Click the little pencil top right to "Edit this page".
- Copy all the article text (click inside text box, Ctlr-A for select all, and Ctrl-C for copy).
- Go back to the _posts/ directory.
- Click the little plus next to the _posts/ folder title
- Paste the article contents that you previously copied into the large text area.
- Click the filename box, and add a title on the form:
- YYYY-MM-DD-some-title.wiki
- (notice it starts with eg. today's date, all letters are lowercase, and it has .wiki at the end)
- Edit the article so it looks like you want it, using mediawiki syntax.
- If the long lines disappear off the edge of the screen, select "Soft wrap" in the drop down menu in the top right corner of the text area.
- You can click the "Preview changes" button in the top left corner of the text area to check that your text is fine.
- When you are happy, go live with your new article on the web server using the green "Commit changes" button below the text area.
- Go through the steps above again, but for the production system. Paste the news article that you just wrote.
- Done.
Go here: here: https://github.com/neicnordic/neicweb
Navigate to each file you want to delete. Click the little trashcan icon to the right near the top. Follow instructions onscreen.
When you delete a news article, please delete the media files as well as the news article text.
- Media files here: https://github.com/neicnordic/neicweb/tree/master/media
- News article texts here: https://github.com/neicnordic/neicweb/tree/master/_posts
- NeIC's web pages are static HTML built from text files written in the Mediawiki markup language for easier formatting.
- The .wiki files contain the text/content of the web pages, while the graphical design and navigational menus are wrapped around the content automatically by template files and some Ruby scripts.
- The build system is Jekyll (a popular site/blog-generator written in Ruby).
- The "source code" for the web site is kept in a repository at GitHub https://github.com/neicnordic/nscweb for version control. You need to be added as collaborator to have write access to it.
- When changes are committed to the GitHub repository, GitHub talks to our web server, which checks out the source and rebuilds the site. Updates are typically visible on http://neic.nordforsk.org/ within a few minutes.
- The easiest way to update an existing web page or create a new one, is to log into the GitHub:s web page and do it there using their integrated text editor. This has a preview feature so you can easily verify that you got the formatting right before saving your changes.
- Images and other media cannot be added through the web browser, but you need to clone the repository locally, add the files to the neicweb/media/ directory, commit your changes, and push your commits back to github. You can use command line git for this, or you can install a GitHub graphical user interface program on your computer to do this, for example GitHub for Windows or GitHub for Mac. Note however that the GUI programs use the word Sync for all transactions with github.com, so when this manual tells you to update, pull, or push, just press the Sync button on the top right.
- The hard-core way of editing the web site is to clone the git repository to your own machine, update the relevant Markdown files and then commit the changes. Optionally, you can also install Jekyll yourself, compile the full site, and preview it locally.
NeIC's web site was remade in 2015. The current version is a modified copy of a solution developed by Peter Larson at NSC, and is based on the Jekyll static site compiler and the Pandoc universal document converter. The site, as served, is still just static HTML, in the sense that there is no on-the-fly generation of webpages from content stored in databases. Everything on the site gets generated once by Jekyll and then saved as HTML files on the web server. The prime advantage of this approach for the typical content editor is that it is possible to write pages in simple Mediawiki syntax, which NeIC does anyway for all its wikis and meeting minutes.
For the web designer, Jekyll offers a flexible templating solution. For the system administrator this setup has few and simple components, so there are less things that can fail, misbehave or need maintenance. For the data manager, this solution has virtually no risk of ever losing content, since no changes go live without passing through state of the art version control and backup on another continent. From a security point of view, the attack surface is very small.
The processing pipeline looks like this:
- The content you write as a .wiki file, is converted to HTML content by Pandoc.
- The HTML content is inserted into the web page templates.
- Some "dynamic" content, such as breadcrumbs and the menus, is computed by Jekyll and inserted into the templates.
- The final, complete, HTML files are stored on disk in a special output directory.
- The output directory is copied to the web server.
The structure of the web site is modelled on the award-winning gov.uk site with a hierarchical, progressive menu system of categories under which everything is placed. This "tree" of pages is built directly from the file system directory tree and mirrors it one-to-one. For example, to add a new item in the menu on the first page, one would create a new folder in the root directory of the web site source tree.
The first level of categories are displayed in the top menu, and these are:
- About: Staff, addresses, organization, history, and partners
- Activities: Information about our ongoing projects and operations.
- Events: A neic-events calendar webview, with curent and past events, like we have on the wiki.
- Publications: Public documents produced by NeIC.
- Skandinavisk: Page(s?) in Skandinavisk.
The web site source comes in the form of folder called "neicweb", which is a git repository, to allow version control. The actual folder structure in the web site source mirrors the visual structure:
- index.html (This is the front page)
- about/index.wiki (Corresponds to the "About" item in the top menu)
- about/team/index.wiki (The menu item "Team and Contact" under the "About" category).
- about/areas/bms/index.wiki
- ...
Some directories do not automatically turn up as web pages, because they do not have any index.* files in them, and this is just the way we want it. For example:
- assets/: Building blocks for the website; icons, scripts, style sheets, etc.
- media/: Media content; news post images, public documents, etc.
- neic2015/: Information on the NeIC conference in May 2015.
- news/: News posts are readily accesible from the many links on the front page and do not need another link in the top menu.
- _layouts/: web page templates. The overall graphical design like page headers, placement of content, and the column structure is defined here.
- _plugins/: contains Ruby scripts to customize the behavior of Jekyll to get certain NSC-specific features, like the navigational menus (generate_navigation.rb) and the breadcrumb links (breadcrumbs.rb).
- _posts/: news articles go here. They are automatically picked up by Jekyll and put in the right in the web site.
- If you do not have an account at GitHub, you first need to sign up to get an account. The sign-up process is straightforward, directly from the front page.
- Once you have the account, talk to your project leader or area coordinator to get added to the "neicnordic" team, so that you can edit the pages.
- Navigate to the repository url https://github.com/neicnordic/neicweb. You will be presented with a list of files and folders, which you can navigate through, just like on a computer. Remember, the names of folders correspond to the list of topics in the navigational menu on the web page.
- All pages are called index.wiki (or index.html in some cases). For example, the file /about/index.wiki contains the text displayed on the About page (http://neic.nordforsk.org/about/).
- To edit a file, click on the index.wiki file, this will show the contents of the file. Next step is to click on the pencil icon to open up the text editor.
- Make the changes you want.
- In the bottom of the editor page, there is a button called "Commit change". That is how you save changes. But before doing that, please write a commit message briefly describing what you did, and why you did it. Since all files are called "index.wiki", it is recommended to write which menu item you changed in the short commit message (the first field of text), e.g. "Updated team page", and if you are feeling ambitious, some further description in the extended commit message (the second field of text).
- First upload the images you need (see below).
- Find a page that looks similar to what you want.
- Copy the source code for it, and paste it into a new page.
- Edit it until you are happy with it.
- Save it.
If you need to create a completely new page, you are faced by a conundrum: you need to create folder, because the actual file is supposed to go into e.g. /about/operations/index.wiki but there is no button for making folders on Github. There is only a + sign button, which creates new file. The solution is to write the new folder as part of the file name. So in the about directory, you hit the + button, and write e.g. operations/index.wiki as the new file name. That will create a new folder called operations with an index.wiki file for the content, which you now can edit.
Let us look at an example page, /about/index.wiki:
= About NeIC = '''The Nordic e-Infrastructure Collaboration (NeIC) is an organisation that facilitates the development and operation of high-quality e-Infrastructure solutions in areas of joint Nordic interest. The NeIC is a distributed organisation consisting of technical experts from academic high-performance computing centres across the Nordic countries.''' The collaboration was launched in 2003 in response to a decision by the research funding agencies in the Nordic countries to contribute to a distributed computing infrastructure for analysing and storing the data generated by experiments at the [http://home.web.cern.ch/about/accelerators/large-hadron-collider Large Hadron Collider (LHC)] at [http://home.web.cern.ch/ CERN] (the European Organization for Nuclear Research, physicists and engineers). New, innovative services can be developed and deployed by bringing together the competencies of all the Nordic e-Infrastructure providers, and challengingthis pool of highly skilled IT experts with the needs of the Nordic user communities. This is precisely what the NeIC aims to do in key areas of common Nordic interest. By combining the expertise already found in the Nordic region, we can provide leverage for more efficient and responsive e-Infrastructure, thereby avoiding costly duplication of work in each individual country. An important focus of the NeIC in the coming years will be on innovation in new e-Infrastructure solutions, in addition to selected services such as the Nordic Tier-1 service for the WLCG community. The NeIC was established as an organisational unit under NordForsk in Oslo on 1 January 2012.
There are two sections in a page, a metadata section, and a content section. The metadata section is delimited by three minus signs in a row (---) before it begins, and the same where it ends. The metadata part contains instructions to Jekyll about how to build the page. Note that there are quotes around the metadata values, and a space between the colon and quote character. There are two crucial metadata fields:
- title: contains the title/heading of the page. This is somewhat confusing, because it is the title as it appears in the navigational menu which links to this page, not the actual title that shows up on the page.
- description: contains the text for the brief description that shows up in mouse-over tooltips in the navigational menu.
- menu: controls the order of this page in the menu. The lower the number the earlier it comes in the menu. Default is 100. Pages with the same number are ordered alphabetically. See more below of how to control the sorting order of the menus.
- hidden: controls whether this page should show up in automatically generated menus. Set to yes (without quotes) to make the page available online but not advertised (unless you specifically provide links to them yourself). See above for examples.
- layout controls the which page template is being used to render the page. This is set up to automatically give normal pages the "menu" layout while news articles get the "news" layout. You probably shouldn't change this.
<img class="smallpic" src="/media/warface.png">). However, Pandoc is not perfect, and thus sometimes gets this html stuff wrong. Should you encounter this, you can work around this by naming the file index.html instead, and writing the page in pure HTML.
The navigational menus are built by inspecting all the folders at that particular level in the file system tree. Under each folder, there will be an index.wiki file which contains a metadata section. That section is read and the title/description/sorting order for each page is collected and used to build the menu.
The news items are actually blog posts in the Jekyll system. They are stored in the _posts/ directory. The list of news items displayed on the front page is generated automatically from the list of file in the _posts/ directory. Thus, you are not supposed to edit the front page to add news items. There is a specific system for file names of the news items. If you look inside the folder you will see files named like this:
2014-04-11-mathlabseminar.wiki 2014-05-23-prace.wiki ...
The file names are supposed to highlight the date of publication, followed by a dash (minus sign), and then something descriptive, using dashes instead of spaces.
[YYYY-MM-DD]-[very-brief-description].md
The description is mostly for internal purposes. It does get reflected in the URL of the web page of the news post, but it is not used anywhere else. The actual title of post is read from the metadata section in the post.
The metadata of news item is similar to a normal page, with a few differences. Let's use an example (see it live):
= Engaged discussions at the Nordic accounting workshop 2014 in Uppsala = '''Lively discussions on hot topics as energy-, GPU/Accelerator- and Cloud accounting have been held at this year's Nordic accounting workshop.''' <a href="/media/2014-10-16-nordic-accounting-workshop.jpeg"> <img class="smallpic" src="/media/2014-10-16-nordic-accounting-workshop.jpeg"> </a> The participants unanimously agreed that such small dedicated meetings are very useful in getting the right people to talk to each other. Also the remote participants from Norway, Iceland and the UK could contribute to the meeting in a meaningful way. All slides and links are online on the [https://wiki.neic.no/wiki/Nordic_accounting_workshop_2014 wiki page]. The complete minutes will be made available during next week. The next come-together will be during the dedicated workshop session on accounting at the NeIC2015 conference near Helsinki in May 2015. Special thanks go to the local staff at Uppsala University for their great support.
Like normal pages, news articles also have two parts; the metadata part and the content part.
There are two crucial and one useful metadata fields:
- title: contains the title/heading of the news article. This is somewhat confusing, because it is the title as it appears in news article listings which links to this page, not the actual title that shows up on the page.
- summary contains the brief summary description that shows up in news article listings, like on the front page (this example has been trimmed down for brevity). Most often, this will be repeated in boldface as the first paragraph of the article.
- image optional. If given, this points out a path to a small (in kilobytes) image to show next to the article in news article lists. Note: be sure to use a small image (in kilobytes), as a news article list of many large images will be painful to load on a smartphone with bad reception. Take a moment to make absolutely sure of this, as the images will be resized (size on screen) automatically for aesthetic resons and for reasons of convenience, so it may not be immediately obviious to you if you get this wrong.
- layout controls the which page template is being used to render the page. This is set up to automatically give normal pages the "menu" layout while news articles get the "news" layout. You probably shouldn't change this.
Images and other media can be added through the web browser. Navigate to the directory where you want upload the files, click the "Upload files" button near the top, and follow instructions onscreen.
You can of course also clone the repository locally, add the files to the neicweb/media/ directory, commit your changes, and push your commits back to github. You can use command line git for this, or you can install a GitHub graphical user interface program on your computer to do this, for example GitHub for Windows or GitHub for Mac. Note however that the GUI programs use the word Sync for all transactions with github.com, so when this manual tells you to update, pull, or push, just press the Sync button on the top right.
To add your image to the page you can use an HTML img tag like so, to add the image to the page at exactly that spot on the page, in exactly the size it actually is:
<img src="/media/nancy-mini.jpeg">
To place a small image to the side with text flowing next to it like in most news articles, add class="smallpic" or class="smallpic-left" like so:
<img class="smallpic" src="/media/nancy-mini.jpeg"> <img class="smallpic-left" src="/media/nancy-mini.jpeg">
To let the reader click the image to enlarge it, upload two images and wrap the "img" tag in an "a" tag like so:
<a href="/media/nancy-large.jpeg"> <img class="smallpic" src="/media/nancy-mini.jpeg"> </a>
See the preceding section for a real example.
Use normal Mediawiki links, like so:
[[/media/cognitus.pdf|download the Cognitus report]] [http://www.nordforsk.org/files/strategy.pdf download the NordForsk strategy]
If you want to do development of the web site system itself, or simply preview pages, without them being pushed to the live web server, you need to set up a local Jekyll installation on your computer and then clone the web site source from Github.
By command line:
git clone https://github.com/neicnordic/neicweb.git
From the windows/mac GUI:
Cick the plus to add a repository, then click clone, then find and click neicnorduic/neicweb.
From the web interface:
Go to the repository at github. In the column on the right-hand side, there are two buttons for downloading the code: "Clone in Desktop" and "Download to ZIP".
The use Jekyll, you need to have the Ruby programming language installed on your computer, and associated libraries. The libraries are usually called "gems" in Ruby parlance. While it is possible to install a system-global Ruby installation and install libraries into that one, I recommend against it, due to complicated dependencies between Ruby versions and different gems. The canonical solution used in the Ruby community is instead to use separate Ruby installations for each site you are developing, much like virtualenv in Python.
So, start by installing rbenv, a tool to manage multiple Ruby installations:
git clone https://github.com/sstephenson/rbenv.git ~/.rbenv
Initialize rbenv with these commands:
export PATH="$HOME/.rbenv/bin:$PATH" eval "$(rbenv init -)"
To automatically do this at login you can for example add the same lines to your ~/.bash_profile.
Install the ruby-build plugin in order to build Ruby versions from source:
git clone https://github.com/sstephenson/ruby-build.git ~/.rbenv/plugins/ruby-build
Run eval "$(rbenv init -)" again to set everything up.
Check that it works by running:
rbenv install --list
That should show you a (long) list of available Ruby versions that can be compiled and installed. Install a Ruby (this was the latest 2.0.0 at the time). Package dependency note: on some systems, you need have the OpenSSL development package installed in order to compile Ruby. In Ubuntu distributions, this package is called "libssl-dev".
rbenv install 2.0.0-p353
Change into using the Ruby version:
rbenv global 2.0.0-p353
Install the Jekyll gem, same version as on the web server:
gem update gem install jekyll -v 2.5.3
We also need the external Pandoc document converter and its Ruby bindings pandoc-ruby to build the site. On a Linux machine, you will typically find Pandoc in the standard package repository. On a Mac or Windows computer, download the installer packages from the Pandoc home page instead.
yum install pandoc gem install pandoc-ruby
Finally, we need another gem, RedCloth, it adds Textile markup language support to Jekyll (although we currently do not use this feature on the web pages).
gem install RedCloth
We are now ready to work with Jekyll and the site. The commands to build the site are:
jekyll build jekyll serve (to launch the web server locally on port 4000)
There is a neat automatic rebuild feature of Jekyll, so that you don't need to rebuild and restart the server each time you edit a page.
jekyll serve --watch
Once you are satisfied with the changes, do not forget to commit them to git repository and push it to Github.