Top News Selectors (tns): Top news parsing from select websites
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
docs
tests
tns
.coveragerc
.gitignore
.travis.yml
LICENSE
README.md
requirements.txt
setup.py

README.md

Top-news-selectors (tns)

Build Status codecov

A static HTML site parser for parsing the top story titles and URIs for the following websites:

This parser is built on Beautifulsoup4 based on the CSS selectors found inside the respective HTML documents. These selectors were chosen based on the selectors that were present during the 30 days for the month of November 2016.

This package has not been tested beyond the months of May 2016 to May 2017 (05/01/2016 - 05/31/2017). Use at your own risk when going beyond that range. The coverage of the parser for a year time range is indicated by the figure below:

alt selectors_coverage_year

Install

This package is available via pip:

$ pip install tns

Usage

>>> from tns import SiteParser
>>> html = open('npr.html').read()
>>> parser = SiteParser(html)
>>> parser.npr()
{
'hero_text': "foo",
'hero_link': 'http://bar.com',
'headlines': [{'splash_title': 'baz', 'link': 'http://qux.com'}]
}
>>> parser.washingtonpost()
{'hero_text': '', 'hero_link': '', 'headlines': []}

The keys hero_text and hero_link refer to the top most identified news post, often identified by a enlarged picture or text across the top of the page. Headlines refer to the subsequent headlines after the hero article, where splash_title refers to the title found on the home page of the site not the actual title of the article.

You can see that the second function call with parser, parser.washingtonpost(), has zero entries because the document passed to SiteParser was not intended for that parser.

Each of the sites are paired to a function:

"http://abcnews.go.com/":          parser.abcnews()
"https://www.cbsnews.com/":        parser.cbsnews()
"https://www.nbcnews.com/":        parser.nbcnews()
"https://www.washingtonpost.com/": parser.washingtonpost()
"https://www.npr.org/":            parser.npr()
"http://www.latimes.com/":         parser.latimes()
"https://www.usatoday.com/":       parser.usatoday()
"https://www.wsj.com/":            parser.wsj()
"https://www.nytimes.com/":        parser.nytimes()
"http://www.foxnews.com":          parser.foxnews()
"http://www.chicagotribune.com/":  parser.chicagotribune()

Debugging

When the parser fails to find specific headlines or returns no headlines at all, this could be due to:

  • Iframes loading content dynamically
  • Headlines being injected via Javascript
  • The wrong parser is being used for the specified site