Skip to main content

jusText with language identification (fasttext) instead of stopwords

Project description

@jordiae Changes

Use Fasttext language identifier instead of stopwords.

Install:

pip install justext-lid

Usage (no stopwords required):

from justext_lid import justext
first_paragraph = justext(HTML, langs={'en', 'es'}, langid_th=0.3)[0]
first_paragraph_lang = first_paragraph.lang

EleutherAI Changes

Modified the classification function in order to remove elements that have quotes (in order to improve results on forums)

jusText

https://api.travis-ci.org/miso-belica/jusText.png?branch=master

Program jusText is a tool for removing boilerplate content, such as navigation links, headers, and footers from HTML pages. It is designed to preserve mainly text containing full sentences and it is therefore well suited for creating linguistic resources such as Web corpora. You can try it online.

This is a fork of original (currently unmaintained) code of jusText hosted on Google Code.

Adaptations of the algorithm to other languages:

Some libraries using jusText:

Some currently (Jan 2020) maintained alternatives:

Installation

Make sure you have Python 2.7+/3.4+ and pip (Windows, Linux) installed. Run simply:

$ [sudo] pip install justext

Dependencies

lxml (version depends on your Python version)

Usage

$ python -m justext -s Czech -o text.txt http://www.zdrojak.cz/clanky/automaticke-zabezpeceni/
$ python -m justext -s English -o plain_text.txt english_page.html
$ python -m justext --help # for more info

Python API

import requests
import justext

response = requests.get("http://planet.python.org/")
paragraphs = justext.justext(response.content, justext.get_stoplist("English"))
for paragraph in paragraphs:
  if not paragraph.is_boilerplate:
    print paragraph.text

Testing

Run tests via

$ py.test-2.7 && py.test-3.4 && py.test-3.5 && py.test-3.6 && py.test-3.7 && py.test-3.8

Acknowledgements

This software has been developed at the Natural Language Processing Centre of Masaryk University in Brno with a financial support from PRESEMT and Lexical Computing Ltd. It also relates to PhD research of Jan Pomikálek.

Changelog for jusText

2.2.0 (2016-03-06)

  • INCOMPATIBLE CHANGE: Stop words are case insensitive.

  • INCOMPATIBLE CHANGE: Dropped support for Python 3.2

  • BUG FIX: Preserve new lines from original text in paragraphs.

2.1.1 (2014-05-27)

  • BUG FIX: Function decode_html now respects parameter errors when falling to default_encoding #9.

2.1.0 (2014-01-25)

  • FEATURE: Added XPath selector to the paragrahs. XPath selector is also available in detailed output as xpath attribute of <p> tag #5.

2.0.0 (2013-08-26)

  • FEATURE: Added pluggable DOM preprocessor.

  • FEATURE: Added support for Python 3.2+.

  • INCOMPATIBLE CHANGE: Paragraphs are instances of justext.paragraph.Paragraph.

  • INCOMPATIBLE CHANGE: Script ‘justext’ removed in favour of command python -m justext.

  • FEATURE: It’s possible to enter an URI as input document in CLI.

  • FEATURE: It is possible to pass unicode string directly.

1.2.0 (2011-08-08)

  • FEATURE: Character counts used instead of word counts where possible in order to make the algorithm work well in the language independent mode (without a stoplist) for languages where counting words is not easy (Japanese, Chinese, Thai, etc).

  • BUG FIX: More robust parsing of meta tags containing the information about used charset.

  • BUG FIX: Corrected decoding of HTML entities &#128; to &#159;

1.1.0 (2011-03-09)

  • First public release.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

jusText-lid-2.2.8.tar.gz (824.3 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page