Just recently (last night), Steven Loria updated TextBlob to v0.5.0. The module enabled a relatively easy way to do Natural Language Processing in Python. NLTK is a dependency so it’s familiar turfs with an easier getting started part. Based on this, I did also did an easy way to parse a set of words and documents to measure important keywords based on TF-IDF algorithm.
A few minutes ago I uploaded the module to PyPi and tagged it as v0.1.0. It’s still rough and what it does is just plain TF-IDF for now. The next version will incorporate the said graph building feature.
Basic use will be using it to create an Google-like autocomplete feature when you do a search but that’s a topic for another day. For now, I am using it to analyze my own tweets and to favorite other tweets relevant with my own tweets automatically.
Head on to http://tistaharahap.github.io/WordGraph/ for some action.