Static site generator for university research lab website
CoffeeScript HTML CSS Shell
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
src
.gitignore
README.md
commit.sh
config.json
gulpfile.coffee
package.json
publish.sh
server.coffee

README.md

Magic Lab Website

Static site generator for university research lab website.

We use Gulp to assemble content into static pages that can be served by a CDN. We make use of standard plugins to transpile, concatenate, minify, and render higher-level languages into servable HTML and CSS. Care is taken to choose high-level languages that will age well, such as polyfills for backwards compatibility of recently established standards.

Versions of the generated site are stored in source control, which enables diff-based auditing and a push-to-deploy workflow. Push-to-deploy is supported by CDNs such as GitHub Pages, and easy to set up on any host with ssh access by writing a post-receive hook.

We store data in a normalized Google Spreadsheet and use a custom plugin to feed it into the gulp workflow. It generates one HTML page for each worksheet.

Installation

Install Node.js and run

npm install

from the project directory. If you don't want to install gulp globally, add the local install to your path:

export PATH="$PATH:./node_modules/.bin"

Configuration

Gulp encourages a "code over configuration" style, so in many cases it is best to edit the gulpfile directly.

Some situational behavior is controlled by environment variables:

  • MINIFY: Set to any non-empty value to enable minification of generated HTML and CSS.

  • PORT=8000: Port to serve site on when using gulp serve and gulp watch.

  • GSS_ID: ID of Google Spreadsheet to fetch data from.

  • GOOGLE_DRIVE_FOLDER_ID: Set this to the folder where assets are stored.

  • OAUTH_CLIENT_ID, OAUTH_CLIENT_SECRET, OAUTH_REFRESH_TOKEN: Set these to an app that has access to the assets folder.

Usage

To generate the website from data and templates, run

gulp build

then publish the public folder. (Use publish.sh to automatically commit that folder to the gh-pages branch.)

For development, run

gulp watch

and navigate to http://localhost:8000. Changes made to source files should be reflected immediately in your browser.

Content Management

Most common tasks can be accomplished by editing the spreadsheet and running gulp build.

  • To create a new page:

    Create a new worksheet with the title you want for the page title. It will be converted into kebab-case. Currently there is no automatic site-map generation, so you will have to link to the new page manually.

  • To create a new section on a page:

    Enter the section title you want in the "Section" column. Discontiguous sections will be grouped together, in the order of their first appearance.

  • To create a new data field:

    Create a column with the field name in the first row, and write a CSS rule in src/style/main.css to display it. For a field named xyz, the created element will have class .data-field-xyz

  • To write custom HTML in a field:

    Use markdown syntax.

  • To see what has changed between published revisions:

    You can view diffs on the published branch using any general-purpose Git history visualizer, such as SourceTree.

    We commit JSON data and rendered HTML to this branch, so a change on one worksheet would produce a change to the corresponding data file and HTML file.

Spreadsheet to DOM Transformation

On each worksheet, rows are grouped by their "section" column. Within the section a div is created for each row. Each cell is transformed with markdown and placed into a div with a class derived from its column title. This makes it easy to write custom CSS to lay out specific fields on specific pages.

Some fields are treated specially:

  • Title: GSS inserts a field by this name if there is not one already, so we hide it by default. To show it, set its display property to something other than none in the CSS.
  • Link: This is automatically converted into a hyperlink, and combined with the Name field if both are present. This is often more convenient than using markdown to get the same result.
  • Picture: This is automatically converted into an image in the "assets" folder.

For example, this data:

Section         Name                Link                      Description
Research Team   Mary-Anne Williams  http://themagiclab.org/   Director of the Magic Lab

transforms into this HTML:

<body class="data-page-People">
    <h2>People</h2>
    <section class="data-section-research-team">
        <h3>Research Team</h3>
        <div class="data-item">
            <span class="data-field-name">
                <a href="http://themagiclab.org/">Mary-Anne Williams</a>
            </span>
            <span class="data-field-description">
                Director of the Magic Lab
            </span>
        </div>
    </section>
</body>