Example of Calibre news download recipe

README.md

Recipe for Calibre news download

An example of creating a "news recipe" for Calibre to download news stories and covert to mobi or epub.

calibre-recipe.py

import string, re
from calibre import strftime
from calibre.web.feeds.recipes import BasicNewsRecipe
from calibre.ebooks.BeautifulSoup import BeautifulSoup

class AdvancedUserRecipe1334031382(BasicNewsRecipe):
    title          = u'<name>'
    oldest_article = 400
    max_articles_per_feed = 150
    no_stylesheets = True
    remove_javascript = True
    
    remove_tags_before = dict(id='story')
    remove_tags_after = dict(id='story')

    remove_tags = [
        dict(name='div', attrs={'class':'share clearfix'}),
        dict(name='div', attrs={'class':'publish-meta clearfix'}),        
        dict(name='div', attrs={'class':'modal-header'}),
        dict(name='a', attrs={'class':'buttons prev'}),
        dict(name='div', attrs={'class':'viewport'}),
        dict(name='a', attrs={'class':'buttons next'}),
        dict(name='div', attrs={'class':'box related-images'}),
        dict(name='div', attrs={'class':'caption round4-bottom'}),
        dict(name='div', attrs={'class':'magnify'}),
        dict(name='div', attrs={'class':'box related-links'}),
        dict(name='div', attrs={'class':'tags clearfix'}),
        dict(name='div', attrs={'class':'view-comments'}),
        dict(name='div', attrs={'id':'comment-container'}),
        dict(name='div', attrs={'class':'related-block stories clearfix'}),
        dict(name='div', attrs={'class':'related-block clearfix'}),
        dict(name='div', attrs={'class':'OUTBRAIN'})
    ]
    
    feeds          = [(u'<name>', u'<feed>')]