Home | Documentation | Examples
nlpru is a library meant to simplify Natural Language Processing and analysis of Russian text. It is built with Russian language social media data in mind, specifically twitter.
The library makes it easy to use several methods commonly used with natural language processsing, such as categorizing tweets by topic, preprocessing (or cleaning text), etc. While several other libraries exist that can do many of these things, I found that general ones such as preprocessor don't seem to work well with Russian text. Russian specific libraries, such as pymorphy2 (which nlpru uses behind the scenes) offer useful NLP morphological analysis but not preprocessing or facilitation of different types of analysis.
Designed for Python 3.5
Author: @sergegoussev
WIP — while all methods work, the library will be expanded with time
You can install this library via git from github directly:
>>> pip install git+https://github.com/sergegoussev/nlpru.git
Once installed, you can call each method you require by importing it from nlpru. For example, let's say you have some tweets to preprocess:
#import method
from nlpru import Clean
#initiate the method
c = Clean()
tweet = """п***ц какое расследование,почему б не указать,что наш Самарский ио @D_Azaroff
,вот он не бот,можно обратиться напрямую,не откажет"""
#use the method and save its output
clean_tweet = c.Clean_document(tweet, remove_swears=True)If we print clean_tweet, it will return:
"какое расследование, почему б не указать, что наш Самарский ио @DAzaroff, вот он не бот, можно обратиться напрямую, не откажет"
The library also offers some useful and simple methods, such as categorizing topics and taking conversation threads into account
A full overview of the availible methods and examples is availible: