Fix Haaretz Group (Semrushbot)

This commit is contained in:
magnolia1234 2024-01-02 19:44:11 +01:00
parent d68bbf5074
commit 73699adfa3
5 changed files with 22 additions and 5 deletions

View file

@ -60,7 +60,7 @@ var remove_cookies = [];
var remove_cookies_select_hold, remove_cookies_select_drop;
// Set User-Agent
var use_google_bot, use_bing_bot, use_facebook_bot;
var use_google_bot, use_bing_bot, use_facebook_bot, use_semrush_bot;
// Set Referer
var use_drudgereport_referer, use_facebook_referer, use_google_referer, use_twitter_referer;
// Set random IP-address
@ -107,6 +107,7 @@ function initSetRules() {
use_google_bot = [];
use_bing_bot = [];
use_facebook_bot = [];
use_semrush_bot = [];
use_drudgereport_referer = [];
use_facebook_referer = [];
use_google_referer = [];
@ -140,6 +141,8 @@ const userAgentMobileB = "Chrome/115.0.5790.171 Mobile Safari/537.36 (compatible
const userAgentDesktopF = 'facebookexternalhit/1.1 (+http://www.facebook.com/externalhit_uatext.php)';
const userAgentDesktopS = "Mozilla/5.0 (compatible; SemrushBot; +http://www.semrush.com/bot.html)";
var enabledSites = [];
var disabledSites = [];
var optionSites = {};
@ -260,6 +263,10 @@ function addRules(domain, rule) {
if (!use_facebook_bot.includes(domain))
use_facebook_bot.push(domain);
break;
case 'semrushbot':
if (!use_semrush_bot.includes(domain))
use_semrush_bot.push(domain);
break;
}
}
if (rule.referer) {
@ -414,7 +421,7 @@ function set_rules(sites, sites_updated, sites_custom) {
blockedJsInlineDomains = Object.keys(blockedJsInline);
disableJavascriptInline();
use_random_ip = Object.keys(random_ip);
change_headers = use_google_bot.concat(use_bing_bot, use_facebook_bot, use_drudgereport_referer, use_facebook_referer, use_google_referer, use_twitter_referer, use_random_ip);
change_headers = use_google_bot.concat(use_bing_bot, use_facebook_bot, use_semrush_bot, use_drudgereport_referer, use_facebook_referer, use_google_referer, use_twitter_referer, use_random_ip);
}
// add grouped sites to en/disabledSites (and exclude sites)
@ -1042,6 +1049,7 @@ if (matchUrlDomain(change_headers, details.url) && !ignore_types.includes(detail
!(matchUrlDomain('www.wsj.com', details.url));
var bingbotEnabled = matchUrlDomain(use_bing_bot, details.url);
var facebookbotEnabled = matchUrlDomain(use_facebook_bot, details.url);
var semrushbotEnabled = matchUrlDomain(use_semrush_bot, details.url);
// if referer exists, set it
requestHeaders = requestHeaders.map(function (requestHeader) {
@ -1116,6 +1124,14 @@ if (matchUrlDomain(change_headers, details.url) && !ignore_types.includes(detail
})
}
// override User-Agent to use Semrushbot
if (semrushbotEnabled) {
requestHeaders.push({
"name": "User-Agent",
"value": userAgentDesktopS
})
}
// random IP for sites in use_random_ip
let domain_random;
if (domain_random = matchUrlDomain(use_random_ip, details.url)) {

View file

@ -3,7 +3,7 @@ Changelog Bypass Paywalls Clean - Firefox
Updates (install signed xpi-file): https://gitlab.com/magnolia1234/bypass-paywalls-firefox-clean/-/releases
Post-release
Fix Haaretz.co.il (fetch from archive.is)
Fix Haaretz Group (Semrushbot)
Fix The Diplomat (magazine)
* v3.4.9.0 (2023-12-31)

View file

@ -51,5 +51,5 @@
"webRequestBlocking",
"*://*/*"
],
"version": "3.4.9.2"
"version": "3.4.9.3"
}

View file

@ -827,5 +827,5 @@
"*://archive.vn/*",
"*://webcache.googleusercontent.com/*"
],
"version": "3.4.9.2"
"version": "3.4.9.3"
}

View file

@ -1053,6 +1053,7 @@ var defaultSites = {
"themarker.com"
],
allow_cookies: 1,
useragent: "semrushbot",
cs_dompurify: 1
},
"Haas Mediengruppe (opt-in to custom sites)": {