PyPLN

PyPLN is a distributed pipeline for natural language processing, made in Python. We use NLTK <http://nltk.org/>_ and Celery <http://www.celeryproject.org>_ as our foundations. The goal of the project is to create an easy way to use NLTK for processing big corpora, with a Web interface.

PyPLN is sponsored by Fundação Getulio Vargas <http://portal.fgv.br/>_.

License

PyPLN is free software, released under the GPLv3 <https://gnu.org/licenses/gpl-3.0.html>_.

Documentation

Our documentation is hosted using GitHub Pages <http://pages.github.com/>_:

• PyPLN Documentation <http://pypln.org/docs>_ (created using Sphinx <http://sphinx.pocoo.org/>_)
• Code reference <http://pypln.org/docs/reference/>_ (created using epydoc <http://epydoc.sourceforge.net/>_)

Requirements

You will need some Python packages, libmagic <http://www.darwinsys.com/file/>, poppler utils <http://poppler.freedesktop.org/>, libfreetype <http://www.freetype.org/>'s development headers and aspell dictionaries <ftp://ftp.gnu.org/gnu/aspell/dict/0index.html.>_ for english and portuguese.

To install dependencies (on a Debian-like GNU/Linux distribution)::

sudo apt-get install python-setuptools libmagic-dev poppler-utils libfreetype6-dev fonts-dejavu aspell-en aspell-pt
pip install virtualenv virtualenvwrapper
mkvirtualenv pypln.backend
# we need to install Cython first because of the way pip handles C extensions
pip install Cython
pip install -r requirements/production.txt

You will also need to download some NLTK data packages. You can do so executing::

python -m nltk.downloader genesis maxent_treebank_pos_tagger punkt stopwords averaged_perceptron_tagger

Developing

To run tests::

workon pypln.backend
pip install -r requirements/development.txt
make test

See our code guidelines <https://github.com/namd/pypln.backend/blob/develop/CONTRIBUTING.rst>_.

Creating a new Task

All analyses in PyPLN are performed by our workers. Every worker is a Celery
task that can be included in the canvas that will run when a document is
received in pypln.web.

New workers are very easy to create. All you need to do is write a subclass of `PyPLNTask <https://github.com/NAMD/pypln.backend/blob/develop/pypln/backend/celery_task.py#L36>`
that implements a "process" method. This method will receive the document as a
dictionary, and should return a dictionary that will be used to update the
existing document. As an example::


    from pypln.backend.celery_task import PyPLNTask

    class FreqDist(PyPLNTask):
        def process(self, document):
            value = document['value']
            square = value ** 2
            return {'squared_value': square}


This worker assumes that a previous worker has already included "value" in the
document and uses it to create a new one, called "squared_value".