Skip to content

Dasudian/disco

 
 

Repository files navigation

Disco - Massive data, Minimal code

Disco Logo

Disco is a distributed map-reduce and big-data framework. Like the original framework, which was publicized by Google, Disco supports parallel computations over large data sets on an unreliable cluster of computers. This makes it a perfect tool for analyzing and processing large datasets without having to bother about difficult technical questions related to distributed computing, such as communication protocols, load balancing, locking, job scheduling or fault tolerance, all of which are taken care by Disco.

Writing a Disco job is very simple. For example, the following job counts the number of words in a document:

from disco.core import Job, result_iterator

def map(line, params):
    for word in line.split():
        yield word, 1

def reduce(iter, params):
    from disco.util import kvgroup
    for word, counts in kvgroup(sorted(iter)):
        yield word, sum(counts)

if __name__ == '__main__':
    input = ["http://discoproject.org/media/text/chekhov.txt"]
    job = Job().run(input=input, map=map, reduce=reduce)
    for word, count in result_iterator(job.wait()):
        print word, count

Note: For installing Disco, you cannot use the zip or tar.gz packages generated by github, instead you should clone this repository.

The develop branch contains the newest features and is not recommended for use in production. The master branch is the latest stable release and is tested in production. Important bug fixes will be first merged into the develop branch and then backported into the master branch.

Disco integrates with a lot of different tools. The following screenshot, for example, shows using ipython notebook to write a Disco job and using matplotlib to plot the results: ipython example

To learn more about the Disco Ecosystem see Disco Integrations. For some other resources, check out the Talks on Disco. Visit [discoproject.org] (http://discoproject.org) for more information.

Build Status: Travis-CI :: Travis-CI

About

a Map/Reduce framework for distributed computing

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Erlang 56.1%
  • Python 36.0%
  • JavaScript 4.3%
  • HTML 1.0%
  • Shell 1.0%
  • Makefile 0.6%
  • Other 1.0%