Skip to content
Go to file
This branch is 68 commits behind ecprice:master.

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time



A website and framework that tracks changes in online news articles over time.

Original installation at
A product of the Knight Mozilla MIT news hackathon in June 2012.
Authors: Eric Price (, Greg Price (,
 and Jennifer 8. Lee (

This is free software under the MIT/Expat license; see LICENSE.
The project's source code lives at .


You need to have installed on your local machine
* Git
* Python 2.6 or later
* Django and other Python libraries

On a Debian- or Ubuntu-based system, it may suffice (untested) to run
  $ sudo apt-get install git-core python-django python-django-south python-simplejson

On Mac OS, the easiest way may be to install pip:
and then
  $ pip install Django

Initial setup

  $ python website/ syncdb && python website/ migrate
  $ git init articles
  $ cd articles && touch x && git add x && git commit -m 'Initial commit'

Running NewsDiffs Locally

Do the initial setup above.  Then to start the webserver for testing:
  $ python website/ runserver

and visit http://localhost:8000/

Running the scraper

Do the initial setup above.  You will also need additional Python
libraries; on a Debian- or Ubuntu-based system, it may suffice
(untested) to run
  $ sudo apt-get install python-bs4 python-beautifulsoup

on a Mac, you will want something like

 $ pip install beautifulsoup4
 $ pip install beautifulsoup
 $ pip install html5lib

Note that we need two versions of BeautifulSoup, both 3.2 and 4.0;
some websites are parsed correctly in only one version.

Then run
  $ python website/ scraper

This will populate the articles repository with a list of current news
articles.  This is a snapshot at a single time, so the website will
not yet have any changes. To get changes, wait some time (say, 3
hours) and run 'python website/ scraper' again.  If any of
the articles have changed in the intervening time, the website should
display the associated changes.

The scraper will log progress to /tmp/newsdiffs_logging (which is
overwritten each run) and errors to /tmp/newsdiffs/logging_errs (which
is cumulative).

To run the scraper every hour, run something like:

 $ while true; do python website/ scraper; sleep 60m; done

or make a cron job.

Adding new sites to the scraper

The procedure for adding new sites to the scraper is outlined in
parsers/ .  You need to

  (1) Create a new parser module in parsers/ .  This should be a
      subclass of BaseParser (in parsers/  Model it off
      the other parsers in that directory.  You can test the parser
      with by running, e.g.,

$ python parsers/ bbc.BBCParser

      which will output a list of URLs to track, and

$ python parsers/ bbc.BBCParser

      which will output the text that NewsDiffs would store.

  (2) Add the parser to 'parsers' in parsers/

This should cause the scraper to start tracking the site.

To make the source display properly on the website, you will need
minor edits to two other files: website/frontend/ and
website/frontend/ (to define the display name and create a tab
for the source, respectively).  Search for 'bbc' to find the locations
to edit.


Automatic scraper that tracks changes in news articles over time.




No releases published
You can’t perform that action at this time.