Confopy 0.1.4

Evaluates the linguistic and structural quality of scientific texts.

# Confopy

Asserting the linguistic and structural quality of scientific texts. Written in Python.
Name origin: Confopy := Conform + Python



# Installation

sudo apt-get install python-pdfminer
# lxml==2.3.2
sudo pip install -U lxml
sudo pip install -U numpy
sudo pip install -U pyyaml nltk
sudo pip install -U pyenchant # spell checking

sudo pip install -U pattern
#sudo pip install -U pyparsing # for nltk_contrib

# Install nltk_contrib:
#cd confopy/contrib/nltk_contrib
#python setup.py build
#sudo python setup.py install

# Getting a corpus

Confopy needs a corpus (collection of language data) to run.

For German (TIGER treebank):

1. Go to:
http://www.ims.uni-stuttgart.de/forschung/ressourcen/korpora/TIGERCorpus/license/htmlicense.html
2. Accept the license and download TIGER-XML Release 2.2:
http://www.ims.uni-stuttgart.de/forschung/ressourcen/korpora/TIGERCorpus/download/tigercorpus-2.2.xml.tar.gz
3. Unpack the archive into confopy/localization/de/corpus\_de/
4. Run the patch tiger\_release\_aug07.corrected.16012013\_patch.py in the same folder
5. Verify that the generated file is named exactly like in confopy/config.py

# Python 3

* The package python-pdfminer only works with python 2.4 or newer, but not with python 3


# Unicode errors

* Configure terminal to use unicode!
* For Python devs:
http://docs.python.org/2/howto/unicode.html#the-unicode-type
* Convert the TIGER Treebank Version 1 file
"tiger_release_july03.penn"
to utf-8 encoding before using Confopy!