nlp-rake

A Python package that implements RAKE algorithm for keyword extraction.

Rapid Automatic Keyword Extraction (RAKE) algorithm is proposed by Stuart Rose, Dave Engel, Nick Cramer, Wendy Cowley.

Installation

python3 -m pip install --index-url https://test.pypi.org/simple/ --no-deps nlp-rake-eeeeeeeelias

Quick usage

    from nlp_rake import rake_text

    text = 'Compatibility of systems of...'
    ranked_result = rake_text(text)
    # [
    #     ('minimal generating sets', 8.67),
    #     ('linear diophantine equations', 8.5),
    #     ('minimal supporting set', 7.67),
    #     ...
    # ]

Advanced usage

    import nlp_rake

    text = 'Compatibility of systems of...'

    tokens = nlp_rake.split_to_tokens(text)
    phrases = nlp_rake.split_tokens_to_phrases(
        tokens, stoplist=nlp_rake.ENGLISH_WORDS_STOPLIST
    )
    cooccurrence = nlp_rake.get_cooccurrence_graph(phrases)
    degrees = nlp_rake.get_degrees(cooccurrence)
    frequencies = nlp_rake.get_frequencies(cooccurrence)
    ranked_result = nlp_rake.get_ranked_phrases(
        phrases, degrees=degrees, frequencies=frequencies
    )
    # [
    #     ('minimal generating sets', 8.67),
    #     ('linear diophantine equations', 8.5),
    #     ('minimal supporting set', 7.67),
    #     ...
    # ]

Other

Check out some exercises I did for an online computational linguistics course.

License

MIT