RAKEkeywords

RAKEkeywords implements RAKE (Rapid Automatic Keyword Extraction), an algorithm that finds keywords in a document. The algorithm comes from Rose, Engel, Cramer, and Cowley (2010), "Automatic keyword extraction from individual documents", in Text Mining: Applications and Theory (John Wiley and Sons).

The package includes stopword lists for many languages, from stopwords-iso. It also registers as a keyword-extractor plugin for OVOS Plugin Manager (OPM).

Install

pip install RAKEkeywords

Usage

from RAKEkeywords import Rake

rake = Rake()

keywords = rake.extract_keywords("Mycroft is a free and open-source voice assistant for Linux-based operating systems that uses a natural language user interface")

"""
[('natural language user interface', 16.0),
 ('open-source voice assistant', 9.0),
 ('linux-based operating systems', 9.0),
 ('mycroft', 1.0),
 ('free', 1.0)]
"""

Pass a lang code to use the stopword list for another language.

from RAKEkeywords import Rake

rake = Rake(lang="pt")
keywords = rake.extract_keywords("Portugal, oficialmente República Portuguesa, é um país cujo território se situa na zona ocidental da Península Ibérica.")

Pass stop_words_path to use a custom stopword file instead of the built-in list.

from RAKEkeywords import Rake
from os.path import join, dirname

my_list = join(dirname(__file__), "stopwords", "en", "FoxStoplist.txt")
rake = Rake(stop_words_path=my_list)

Related projects

• OpenVoiceOS/ovos-plugin-manager: the plugin framework this package registers with under the opm.keywords entry point.
• stopwords-iso/stopwords-iso: the multi-language stopword lists bundled with this package.

License

MIT