fukanchik / a41124835ed0

Course materials for O'Reilly Media "Get Started with NLP in Python" -- see private Slack room for more details

Set up for "Get Started with NLP in Python"

This repo includes the notebooks, source data, and other materials for: Get Started with Natural Language Processing in Python.

Virtual Environment

It's a good idea to use virtualenv to manage your Python 3 virtual environment:

virtualenv -p /usr/bin/python3 ~/venv

Then run:

source ~/venv/bin/activate

Installation

To install the required Python libraries and related data sets:

pip install -r requirements.txt
python -m nltk.downloader punkt
python -m nltk.downloader wordnet
python -m textblob.download_corpora
python -m spacy download en

TextBlob issues on Windows

The GitHub page for textblob-aptagger says the package is no longer needed as of TextBlob 0.11.0, which uses the NLTK perceptron tagger instead. Source code for nltk.tag.perceptron claims that it's a port of the TextBlob code. Not exactly -- however there may be issues on some versions of Windows.

If you run pip install textblob and DO NOT install textblob-aptagger that should work fine with only minor changes to Exercise 3 code and the pynlp.py module:

• change import textblob_aptagger as tag to import nltk
• change tag.PerceptronTagger() to nltk.tag.PerceptronTagger()
• change .tag(sent) to .tag(nltk.word_tokenize(sent)) (or the equivalent)

Results should look very much like the original results, although in general NTLK perceptron tagger has problems, e.g., it doesn't handle punctuation properly.

• kudos @blue_slacker

NB: these course materials will shift from TextBlob to spaCy, soon, although the latter still has a few rough edges

Docker support (needs update)

A Docker container -- courtesy of @montyz, @ashapochka -- was defined for an instance of this course a few months ago. May need updates?

https://github.com/montyz/nlp-12-14-2016