jevy146/pytextrank

Python implementation of TextRank for text document NLP parsing and summarization

Python Implementation for TextRank

This repository is based on pytextrank. Refer to original repository for more information.

Modifications:

• Removed unused argument Parse=True which causes error with spaCy spacy_nlp function. Refer to laxatives' Pull Request.
• Modified functions to avoid writing intermediate results to json files
• Added function top_keywords_sentences which return the top keywords and sentences to for easy use

Differences vs the initial Mihalcea paper

Python implementation of TextRank, based on the Mihalcea 2004 <http://web.eecs.umich.edu/~mihalcea/papers/mihalcea.emnlp04.pdf>_ paper.

Modifications to the original algorithm by Rada Mihalcea <https://web.eecs.umich.edu/~mihalcea/>_, et al. include:

• fixed bug; see Java impl, 2008 <https://github.com/ceteri/textrank>_
• use of lemmatization instead of stemming
• verbs included in the graph (but not in the resulting keyphrases)
• named entity recognition
• normalized keyphrase ranks used in summarization

The results produced by this implementation are intended more for use as feature vectors in machine learning, not as academic paper summaries.

Inspired by Williams 2016 <http://mike.place/2016/summarization/>_ talk on text summarization.

Example Usage

import xang1234_pytextrank as pyt

text='This is a test senctence for pytextrank'
sentence, keywords= pyt.top_keywords_sentences(text, phrase_limit=15, sent_word_limit=150)

For more details of usage, refer to this post

Dependencies and Installation

This code has dependencies on several other Python projects:

• spaCy <https://spacy.io/docs/usage/>_
• NetworkX <http://networkx.readthedocs.io/>_
• datasketch <https://github.com/ekzhu/datasketch>_
• graphviz <https://pypi.python.org/pypi/graphviz>_

Also, the runtime depends on a local file called stop.txt which contains a list of stopwords. You can override this in the normalize_key_phrases() call.