internetarchive / tvnews

Article 2 TV News

Uses python 3.4

Article2TVNews allows users to find TV news clips on subjects related to an article they are reading. It provides methods to extract text from an article and get tv news clips based on the article's subject. This package is written for the Internet Archive's WayBack Machine and its Context browser extension. Typical usage looks something like this::

#!/usr/bin/env python

from article2tvnews import ArticleExtraction, NewsRecommender

url = <news-article-url>
html = ArticleExtraction.getHTML(url)
article = ArticleExtraction.extract(html)
for clip in NewsRecommender.makeRecommendations(article):
    print(clip)

Entity Extraction

By default, NewsRecommender uses SpaCy for entity extraction. However, one can choose to use the OpenCalais API. First, the user must register to get an API key. Then pass it as calais_token when you call makeRecommendations()::

#!/usr/bin/env python

from article2tvnews import ArticleExtraction, NewsRecommender

url = <news-article-url>
html = ArticleExtraction.getHTML(url)
article = ArticleExtraction.extract(html)
for clip in NewsRecommender.makeRecommendations(article, calais_token = <api_key>):
    print(clip)

Possible Changes:

• Time period for videos
• extract embedded tweets (possible added weight)

Thanks

Thanks to the wonderful people at The Internet Archive for giving me this opportunity, and a special thanks to:

• Mark

• Dvd

• Vangelis

• Anish

• And of course, Brewster Kahle