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{
|
2020-09-30 07:22:25 +02:00
|
|
|
"bbc.co.uk": "div.vxp-column--single, div.story-body__inner, ul.gallery-images__list",
|
|
|
|
|
"cbc.ca": ".story-content",
|
|
|
|
|
"darkreading.com": "#article-main:not(header)",
|
|
|
|
|
"developpez.com": "div[itemprop=articleBody]",
|
|
|
|
|
"dilbert.com": "span.comic-title-name, img.img-comic",
|
2022-07-31 03:45:57 +02:00
|
|
|
"explosm.net": "div#comic",
|
2020-05-24 17:51:47 +02:00
|
|
|
"financialsamurai.com": "article",
|
2020-09-30 07:22:25 +02:00
|
|
|
"francetvinfo.fr": ".text",
|
|
|
|
|
"github.com": "article.entry-content",
|
|
|
|
|
"heise.de": "header .article-content__lead, header .article-image, div.article-layout__content.article-content",
|
|
|
|
|
"igen.fr": "section.corps",
|
2021-12-27 10:57:23 +01:00
|
|
|
"ikiwiki.iki.fi": ".page.group",
|
2020-09-30 07:22:25 +02:00
|
|
|
"ing.dk": "section.body",
|
|
|
|
|
"lapresse.ca": ".amorce, .entry",
|
|
|
|
|
"lemonde.fr": "article",
|
|
|
|
|
"lepoint.fr": ".art-text",
|
|
|
|
|
"lesjoiesducode.fr": ".blog-post-content img",
|
|
|
|
|
"lesnumeriques.com": ".text",
|
|
|
|
|
"linux.com": "div.content, div[property]",
|
|
|
|
|
"mac4ever.com": "div[itemprop=articleBody]",
|
|
|
|
|
"monwindows.com": ".blog-post-body",
|
|
|
|
|
"npr.org": "#storytext",
|
|
|
|
|
"oneindia.com": ".io-article-body",
|
|
|
|
|
"opensource.com": "div[property]",
|
2022-08-16 18:28:26 +02:00
|
|
|
"openingsource.org": "article.suxing-popup-gallery",
|
2020-09-30 07:22:25 +02:00
|
|
|
"osnews.com": "div.newscontent1",
|
|
|
|
|
"phoronix.com": "div.content",
|
|
|
|
|
"pseudo-sciences.org": "#art_main",
|
2022-01-03 16:47:10 +01:00
|
|
|
"quantamagazine.org": ".outer--content, figure, script",
|
2020-09-30 07:22:25 +02:00
|
|
|
"raywenderlich.com": "article",
|
2022-08-17 14:31:31 +02:00
|
|
|
"royalroad.com": ".author-note-portlet,.chapter-content",
|
2020-09-30 07:22:25 +02:00
|
|
|
"slate.fr": ".field-items",
|
2022-08-16 18:30:54 +02:00
|
|
|
"smbc-comics.com": "div#cc-comicbody, div#aftercomic",
|
2022-08-16 18:31:44 +02:00
|
|
|
"swordscomic.com": "img#comic-image, div#info-frame.tab-content-area",
|
2020-09-30 07:22:25 +02:00
|
|
|
"techcrunch.com": "div.article-entry",
|
|
|
|
|
"theoatmeal.com": "div#comic",
|
2021-08-17 04:57:41 +02:00
|
|
|
"theregister.com": "#top-col-story h2, #body",
|
2022-10-15 00:27:05 +02:00
|
|
|
"theverge.com": "h2.inline:nth-child(2),h2.duet--article--dangerously-set-cms-markup,figure.w-full,div.duet--article--article-body-component",
|
2020-09-30 07:22:25 +02:00
|
|
|
"turnoff.us": "article.post-content",
|
|
|
|
|
"universfreebox.com": "#corps_corps",
|
|
|
|
|
"version2.dk": "section.body",
|
|
|
|
|
"wdwnt.com": "div.entry-content",
|
|
|
|
|
"wired.com": "main figure, article",
|
|
|
|
|
"zeit.de": ".summary, .article-body",
|
|
|
|
|
"zdnet.com": "div.storyBody",
|
2017-12-11 05:51:04 +01:00
|
|
|
}
|
