Plumage

Biomedical resource discovery for institutions, powered by eagle-i

About Plumage

Plumage is free software to make biomedical resources at large institutions more discoverable. It was developed at the University of California, San Francisco, to power UCSF Cores Search, the campus-wide search engine for core resources.

Why Plumage?

Designed to maximize resource discoverability:

Carefully optimized for search engine users:

Easy to deploy:

Designed for success:

Technical overview

Plumage is an application written in Perl 5.12, and tested on Linux and MacOS. It extracts data from an instance of eagle-i (or data marked up with the eagle-i ontology), and generates a new web site: a bundle of static HTML, JavaScript, and images that can be deployed on any server.

The software is bundled with a standard set of modern HTML5 web templates created with Bootstrap and Template Toolkit, and incorporating cores discoverability best practices originally implemented at UCSF. Generated website can be easily customized in two ways:

Who is Plumage?

The Plumage software was developed by Anirvan Chatterjee and the Virtual Home team at the Clinical & Translational Science Institute at the University of California, San Francisco, in collaboration with UCSF's Research Resources Program. This project was supported by the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through UCSF-CTSI Grant Numbers UL1 RR024131 and UL1 TR000004. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.

License

Plumage is Copyright (c) 2012-2013, The Regents of the University of California. All rights reserved.

This application is free software; you can redistribute it and/or modify it under the terms of the BSD license (revised, three clause). For more details, see the full text of the license in the file LICENSE. (TLDRLegal offers a non-binding human-readable description of this license.)

Bundled code

The default Plumage distribution is bundled with several other open source projects:

Plumage websites dynamically load jQuery and html5shiv JavaScript libraries (both released under the MIT license). The software uses a number of Perl CPAN modules during the build process, distributed under several open source licenses, typically under the same terms as Perl itself.

Support and hosting

UCSF is happy to help answer questions about Plumage, and may also be able to offer Plumage as a hosted service for external institutions. Email Anirvan Chatterjee at UCSF (anirvan.chatterjee at ucsf.edu) for more details.

Quick start guide

This quick start guide is intended to help technical users with an existing eagle-i installation get up and running with Plumage in minutes.

Ensure you have Perl 5.12 or higher installed on your server. Perl 5.12+ is installed on virtually all current Unix-like distributions, including MacOS 10.7+, RHEL/Centos 6+, and Ubuntu 11+. If you're unable to upgrade an older version of Perl bundled with your system, use Perlbrew to install a newer version of Perl. Plumage may work on Windows, but hasn't been tested.

Download a tarball or ZIP file of Plumage from Github. Unarchive it.

Install Plumage and its Perl dependencies. You may be prompted to configure CPAN settings; if so, just follow the instructions, and hit return to stick with the default options.

% perl Build.PL
% PERL_MM_USE_DEFAULT=1 ./Build installdeps
# if previous line doesn't work, use only "./Build installdeps"

Then try running the Plumage app.

./bin/plumage

If you see a help screen, everything worked.

Now we're going to configure the simplest possible configuration. Use a text editor to create a file called plumage.conf in your current directory with the following contents. (We're using Howard University as an example.)

site_name              = Howard Cores Search
institution_short_name = Howard
eagle_i_base_url       = http://howard.eagle-i.net/
template_path          = /home/webmaster/plumage/templates
output_path            = /var/www/html/plumage_example
url                    = http://localhost/plumage_example/

Here's how to set the configuration options:

Done? Let's build the website:

./bin/plumage --build

You should see messages showing the website being built.

Pay attention to error messages. Still having problems? Feel free to contact Anirvan (anirvan.chatterjee at ucsf.edu) with your questions.

If all goes well, your fancy new Plumage website will be written to output_path and you'll be able to view it in a web browser at url.

Real world usage

The quick start guide left you with a simple generic installation of Plumage. Now we're going to customize Plumage to meet your needs.

Using roles [IMPORTANT]

You can very easily configure multiple instances of Plumage using the same configuration file. We call each configured instance a role.

In your configuration file, put items common to every role at the top. Then add role blocks below that. Each role block begins with [rolename] on its own line. In this example configuration, the "development" role writes to an output_path corresponding to a development URL, while the "production" role specifies a different output path.

site_name              = Example University Cores
institution_short_name = Example
eagle_i_base_url       = http://example.eagle-i.net/
template_path          = /home/webmaster/plumage/templates

[production]
url = http://cores.example.edu/
google_analytics_id = UA-1234567-01
# build_deploy_command = <see documentation below>
output_path            = /var/www/site/cores.example.edu/

[development]
url                    = http://dev-cores.example.edu/
output_path            = /var/www/site/dev-cores.example.edu/

If you specify one or more roles, you can build that specific role by listing the role name as an argument to --build, like this:

./bin/plumage --build development

Customizing look and feel [IMPORTANT]

Almost every institution will want to customize the look and feel of Plumage to match your local branding needs. Plumage is designed to make it incredibly easy to make local look and feel changes without needing to tweak the default templates, so you can take advantage of upgraded default templates without losing your local changes.

The default templates are stored in the directory specified in template_path. We suggest that you don't edit these templates at all.

Instead, create a new directory for your custom local edits, and put the path in your configuration file as custom_template_path. For example:

template_path          = /home/webmaster/plumage/templates
custom_template_path   = /home/webmaster/plumage/custom_templates

When Plumage is looking for templates, it'll look first in custom_template_path, and only then in template_path. So if you wanted to override a default template, just copy it to custom_template_path, make some tweaks and it'll override the default. But most of the time, you don't even need to do that. Look at the HOWTOs below.

How templates work

All site templates are written using Template Toolkit, a popular and very well-documented templating system for Perl, sort of like PHP's Smarty or Ruby's ERB.

Pending further documentation, please see the bundled templates to see how to use loops and variables.

documentation to be written

Static and dynamic content

All static non-template content (e.g. images, CSS) should go inside the static subdirectory inside template_path and custom_template_path. Files in this directory will be copied as-is to the final website.

All dynamic *.tt Template Toolkit templates that should be run through a template engine should go inside the dynamic subdirectory inside template_path and custom_template_path. When Template Toolkit is running, all template includes will look first in custom_template_path/dynamic and then in template_path/dynamic.

When building the site, build will merge data from the two static directories, and then

For example, given the following files:

custom_templates/
    static/
        images/
            logo.gif
        css/
            custom.css
        dynamic/
            custom_navbar_top.html.tt
templates/
    static/
        images/
            xyz.png
        css/
            plumage.css
            bootstrap.css
            custom.css
        dynamic/
            _header.html.tt

...the build command will output:

output/
    index.html         <-- generated from templates
    some-resource.html <-- generated from templates
    images/
        logo.png
        xyz.png
    css/
        plumage.css
        bootstrap.css
        custom.css  <-- the custom_templates version

How Bootstrap works

All HTML and CSS on the site is written using Twitter Bootstrap 2, a popular responsive HTML5/CSS framework. Go read the Bootstrap documentation. You will be confused if you don't and then attempt to tweak the generated HTML.

HOWTO: Add custom CSS

  1. Configure a custom_template_path directory
  2. Create a new file at static/assets/css/custom.css inside the directory
  3. Put your CSS there
  4. The contents of this file will be read after the default Plumage CSS, which means it should override CSS rules of equal CSS specificity

HOWTO: Add a custom institutional navbar

  1. Configure a custom_template_path directory
  2. Create a new file called custom_navbar_top.html.tt inside the directory
  3. Put content there (preferably inside <div id="leaderboard" class="row"><div class="span12">)
  4. The contents of this file will be automatically included above the Plumage menu in the defaultpage header template (_header.html.tt)
  5. If needed, add custom CSS styles to #leaderboard as described above

HOWTO: Customize the footer

  1. Configure a custom_template_path directory
  2. Create a new file called custom_footer.html.tt inside the directory
  3. Put content there, inside one or more <div class="row"> blocks
  4. The content of this file will be automatically included in the default footer template (_footer.html.tt)
  5. If needed, add custom CSS styles to the contents of #footer as described above

HOWTO: Change the contents of the About page

  1. Configure a custom_template_path directory
  2. Create a new file called custom_about_page.html.tt inside the directory
  3. Put content there (e.g. <h2>s and <p>s)
  4. The content of this file will be automatically included in the default About page template (about.html.tt)

Upgrading the search engine via Swiftype

Plumage comes packaged with a minimal typeahead search, but we recommend plugging in a professional hosted search system.

Plumage works out of the box with Swiftype, a free search provider (like Google Custom Search, but more flexible, and free for most users). Swiftype is optional, but very highly recommended. Set up a new Swiftype account, and an engine for every website for which you want to use Swiftype search (free accounts only get 1 engine, so you'll want to save this for your deployed version).

For Swiftype to work, you need to configure both an overall swiftype_api_key and a swiftype_key for every website role.

Every time you do a new build, Plumage will contact Swiftype's servers, and use your API keys to kick off a reindex of your content. (Swiftype may not reindex as frequently as you'd like; check their documentation for details.)

Tracking usage via Google Analytics

Plumage comes with support for Google Analytics out of the box, just by adding one line to the configuration file. Start off by creating a new Google Analytics account for your Plumage instance.

Then in the configuration file, set google_analytics_id to your new site's Google Analytics account ID, e.g.:

google_analytics_id = UA-1234567-01

Make sure to create a new Google Analytics account for every production Plumage instance you create. For example, if you have a production cores.institution.edu and a development dev-cores.institution.edu, set a Google Analytics ID only for the production role, like this:

[production]
url = http://cores.yoursite.edu/
google_analytics_id = UA-1234567-01

[dev]
url = http://dev-cores.yoursite.edu/

Tracking Swiftype usage

documentation to be written

Ensuring search engine visibility of all pages

Plumage automatically creates a sitemap, and lists the location via robots.txt file. But search engines will automatically discover this robots.txt if and only if a Plumage site is installed at the top level of your site (e.g. http://cores.institution.edu/ works, but http://www.institution.edu/cores/ doesn't).

If you've put Plumage in a subdirectory (e.g. http://www.institution.edu/cores/) and want maximum search engine visibility, you have two choices:

  1. If you're using an automated process to build a sitemap for your whole site, make sure that automatic process picks up every .html generated by Plumage.

  2. Otherwise, make sure to add the sitemap URL to your site-wide robots.txt. For example, if you've deployed your site at http://www.institution.edu/cores/, add the line Sitemap: http://www.institution.edu/cores/assets/sitemap.xml inside http://www.institution.edu/robots.txt. (You can do this even if you have a preexisting sitemap listed there.)

Managing deployments

build_deploy_command is an optional command line that gets run after every Plumage site build. You can use this to create a deploy hook on specific roles.

For example, you might want to run Plumage on a staging server, and use rsync to copy the final production files to a live production server.

[production]
url = http://cores.yoursite.edu/
output_path = /var/www/html/cores-prod/
build_deploy_command = rsync -az -e ssh --delete /var/www/html/cores-prod/ user@cores.yoursite.edu:/var/www/html/

[dev]
url = http://dev-cores.yoursite.edu/
output_path = /var/www/html/cores-dev/