miniflux/reader/scraper/rules.go

47 lines
2 KiB
Go
Raw Normal View History

2017-12-11 05:51:04 +01:00
// Copyright 2017 Frédéric Guillot. All rights reserved.
// Use of this source code is governed by the Apache 2.0
// license that can be found in the LICENSE file.
2018-08-25 06:51:50 +02:00
package scraper // import "miniflux.app/reader/scraper"
2017-12-11 05:51:04 +01:00
// List of predefined scraper rules (alphabetically sorted)
// domain => CSS selectors
var predefinedRules = map[string]string{
2018-12-14 05:25:30 +01:00
"bbc.co.uk": "div.vxp-column--single, div.story-body__inner, ul.gallery-images__list",
2018-01-05 04:32:24 +01:00
"cbc.ca": ".story-content",
2018-01-06 22:25:12 +01:00
"darkreading.com": "#article-main:not(header)",
2018-01-05 04:32:24 +01:00
"developpez.com": "div[itemprop=articleBody]",
"francetvinfo.fr": ".text",
"github.com": "article.entry-content",
2018-12-01 20:49:30 +01:00
"heise.de": "header .article-content__lead, header .article-image, div.article-layout__content.article-content",
2018-01-05 04:32:24 +01:00
"igen.fr": "section.corps",
"ing.dk": "section.body",
"lapresse.ca": ".amorce, .entry",
2019-12-20 03:35:29 +01:00
"lemonde.fr": "article",
2018-01-05 04:32:24 +01:00
"lepoint.fr": ".art-text",
"lesjoiesducode.fr": ".blog-post-content img",
"lesnumeriques.com": ".text",
"linux.com": "div.content, div[property]",
"medium.com": ".section-content",
"mac4ever.com": "div[itemprop=articleBody]",
"monwindows.com": ".blog-post-body",
"npr.org": "#storytext",
"oneindia.com": ".io-article-body",
"opensource.com": "div[property]",
"osnews.com": "div.newscontent1",
"phoronix.com": "div.content",
"pseudo-sciences.org": "#art_main",
"raywenderlich.com": "article",
2018-01-05 04:32:24 +01:00
"slate.fr": ".field-items",
"techcrunch.com": "div.article-entry",
"theoatmeal.com": "div#comic",
2018-01-05 04:32:24 +01:00
"theregister.co.uk": "#body",
"universfreebox.com": "#corps_corps",
"version2.dk": "section.body",
"wdwnt.com": "div.entry-content",
2018-01-05 04:32:24 +01:00
"wired.com": "main figure, article",
"zeit.de": ".summary, .article-body",
"zdnet.com": "div.storyBody",
"openingsource.org": "article.suxing-popup-gallery",
2017-12-11 05:51:04 +01:00
}