Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/jplusplus/nobel

Nobel laureate top list for the Nobel Prize Foundation
https://github.com/jplusplus/nobel

Last synced: 12 months ago
JSON representation

Nobel laureate top list for the Nobel Prize Foundation

Awesome Lists containing this project

README 

          
Installation

============



 * Install Composer: `curl -sS https://getcomposer.org/installer | php`

 * Run `php composer.phar install`

 * `cp settings.default.php settings.php` and make any modifications needed to settings.php

 * Make sure you have php_curl enabled. For recent Ubuntu version run `sudo apt-get install php5-curl`, or uncomment `extension=php_curl.dll` in your `php.ini` file. Then restart Apache.

 * Open /demo/lists.php, /demo/ui.php, and demo/gallery.php, to confirm that everything works, and see how to include the provided widgets.



The scripts in the `$baseDir` directory must be accessible from the Internet, as they are called from the frontend scripts.



The PHP scripts have been tested on PHP 5.3 and 5.6. They use no deprecated functions as of December 2015, and should keep working for the foreseeable future. For PHP versions < 5.3.7, SSL support must be added by installing the php_openssl module (chances are you already have it installed).



All PHP code is contained in a namespace called `Toplist`.



The gallery widget depends on [Foundation.js](http://foundation.zurb.com/). It has been tested with Foundation 5 (the version used at nobelprize.org as of December 2015).



All list widget Javascript code is wrapped in an object called `TopList`. There is also one variable in the global scope, called `gToplistSettings`. Except for these objects, the code should be self contained, and not interfering with anything else.



End points

==========



 * list.php

     - See [demo/lists.php](demo/lists.php) and [demo/ui.php](demo/ui.php)

 * gallery.php

     - See [demo/gallery.php](demo/gallery.php)



Gallery

=======

_See [demo/gallery.php](demo/gallery.php) for how to include a gallery._



The gallery widget shows images from Wikimedia Commons, related to a specific laureate. The images are picked from article illustrations on selected Wikipedias. There will be irrelevant images showing up in some cases. These can be manually removed by adding them to the blacklist array `$gImageBlacklist` in [settings.php](settings.default.php).



Wikipedia editions can be added or removed in `$gImageSourceWPEditions` in [settings.php](settings.default.php). By default we include *large* Wikipedias with a *big number of contributors*. Size alone is not a good measure, as those figures are misleading for editions with high bot activity (e.g. [Swedish](https://sv.wikipedia.org/wiki/Portal:Huvudsida) or [Cebuano](https://ceb.wikipedia.org/wiki/Unang_Panid)). Missing from the list is Russian, despite its big size and high user activity. The Russian Wikipedia community is quite fond of using genre images for navigation templates, which will increase the number of irrelevant images shown, and thus require a large blacklist.

Not included by default, but worth contemplating, are French (`fr`), Swedish (`sv`), and Japanese (`ja`) Wikipedia, given the large number of laureates from France, Switzerland, Sweden and Japan.



Looking up images involves a number of steps: Finding the corresponding DbPedia uri from the nobelprize.org linked data API, Finding the corresponding English Wikipedia article from DbPedia, finding the corresponding articles on different Wikipedia edititions from the Wikidata API, and finally fetching all images from all articles, finding their Wikimedia Commons thumbnails, and filtering out obviously irrelevant, blacklisted and duplicated content. Caching is crucial, as these requests will block page rendering.



Image captions consist of an English image description, and an image credit/license text if (and only if) required. Very long descriptions are truncated. If descriptions look weird, or contain text in other languages, it is most likely because the images does not use the right templates at [Wikimedia Commons](https://commons.wikimedia.org). Such problems should be fixed there. The Wikimedia Commons community currently use the [`{{Information}}`](https://commons.wikimedia.org/wiki/Template:Information) template to make sure images have machine readable descriptions. While not strictly nescessary, using the Information template (or [some other template with similar functionality](https://commons.wikimedia.org/wiki/Commons:Machine-readable_data)) is the simplest way to ensure that image data is parsed correctly by our image galleries.



A quick note on the image selection criteria: Using Wikipedia article illustrations generally gave a better result for most laureates, compared to other methods we tried (using featured images from a Wikimedia Commons category, and using images from Wikimedia Commons pages), but it does return some irrelevant images for quite a few laureates, escpecially those who are more well known in another capacity than that as a Nobel prize laureate (e.g. Henry Kissinger). In such cases, simply keep adding images that feel out of place to `$gImageBlacklist` in [settings.php](settings.default.php).



List widget

===========

_See [demo/lists.php](demo/lists.php) for how to include a list widget._



The list widget shows a top list of the most popular laureates, optionally filtered by _gender_, _birth place_, or _award_, in any combination. Popularity can be based on either nobelprize.org page views (as returned by the [local toplist API](http://www.nobelprize.org/nobel_prizes/popular_api.php)), or by a weighted average of page views at a number of Wikipedia editions.



By default, the Wikipedia popularity is based on the editions of the world's largest languages. You can add or remove editions, or change their weights, in the `$gStatsWPEditions` array in [settings.php](settings.default.php).



Estimating the size of a language is notoriously difficult, and the situation is made even more complex when we start mapping languages to Wikipedia editions. Javanese, for instance, is generally regarded as one of the worlds 10 largest languages, but the Javanese Wikipedia is very, very small, with most Javanese speakers probably using the Indonesian Wikipedia. Some languages are divided among multiple Wikipedias (such as Arabic, with a separate Wikipedia in Egyptian Arabic). Some Wikipedias are more accessible than others to their main target audience, like the Chinese Wikipedia (occasionally blocked in mainland China) or Bengali Wikipedia (with Bangladesh having an extremely low Internet penetration).

By default, we use the following editions for page view statistics: Chinese (`zh`), including script varieties, but excluding Minnan, Yue (Cantonese), Mindong, Wu, Hakka, and Gan; English (`en`), not including simplified English or Scots; Spanish (`es`); Hindi (`hi`), not including Urdu; and Arabic (`ar`), excluding Egyptian Arabic. Portugese (`pt`) would probably be the next language to add, if the list were to be expanded, and perhaps Russian. (Bengali Wikipedia (`bn`) is too small to give any valuable contribution.)



Wikipedia statistics are fetched one editions at a time, making the first rendering of an uncached list _very_ slow. Caching is crucial, as these requests will block page rendering. Fetching statistics involves a number of steps: Finding the corresponding DbPedia uri from the nobelprize.org linked data API, Finding the corresponding English Wikipedia article from DbPedia, finding the corresponding articles on different Wikipedia edititions from the Wikidata API, and fetching their page views statistics from Wikimedia statistics API.



List ui

=======

_See [demo/ui.php](demo/ui.php) for how to include a list with a user interface._



The list UI widget allows the user to filter the list on gender, award, statistics type, and a subset of available regions.



*Only one ui widget can be used on the same page.*



Maintenance scripts

===================

The [maintenance](maintenance/) folder contains scripts to be run from the command line:



* `purgeCache.php` will purge _all_ caches. Depending on your cache settings and server setup, you might need to run this script as root: `sudo php purgeCache.php` (if you are using the default file cache you are very likely to need this).



* `populateCache.php` tries to populate the cache for lists and galleries. Running this script will take quite some time. You might want to put in in a cron job, to run during low trafic hours. Usage: `php populateCache.php`, `php populateCache.php --limit="10"` or `php populateCache.php -l10`